A Translational Science Model for HCILucas Colussocolussouwedu

Human Centered Design and EngineeringDUB Group University of Washington

Ridley Jonesrajoneuwedu

Human Centered Design and EngineeringDUB Group University of Washington

Sean A Munsonsmunsonuwedu

Human Centered Design and EngineeringDUB Group University of Washington

Gary Hsiehgaryhsuwedu

Human Centered Design and EngineeringDUB Group University of Washington

ABSTRACTUsing scientific discoveries to inform design practice is animportant but difficult objective in HCI In this paper weprovide an overview of Translational Science in HCI by trian-gulating literature related to the research-practice gap withinterview data from many parties engaged (or not) in trans-lating HCI knowledge We propose a model for TranslationalScience in HCI based on the concept of a continuum to de-scribe how knowledge progresses (or stalls) through multiplesteps and translations until it can influence design practiceThe model offers a conceptual framework that can be usedby researchers and practitioners to visualize and describethe progression of HCI knowledge through a sequence oftranslations Additionally the model may facilitate a preciseidentification of translational barriers which allows devis-ing more effective strategies to increase the use of scientificfindings in design practice

CCS CONCEPTSbullHuman-centered computingrarrHCI theory conceptsand models

KEYWORDSTranslational Science Translational Research Research-PracticeGap

Permission to make digital or hard copies of all or part of this work forpersonal or classroom use is granted without fee provided that copiesare not made or distributed for profit or commercial advantage and thatcopies bear this notice and the full citation on the first page Copyrightsfor components of this work owned by others than the author(s) mustbe honored Abstracting with credit is permitted To copy otherwise orrepublish to post on servers or to redistribute to lists requires prior specificpermission andor a fee Request permissions from permissionsacmorgCHI 2019 May 4ndash9 2019 Glasgow Scotland UKcopy 2019 Copyright held by the ownerauthor(s) Publication rights licensedto ACMACM ISBN 978-1-4503-5970-21905 $1500httpsdoiorg10114532906053300231

ACM Reference FormatLucas Colusso Ridley Jones Sean A Munson and Gary Hsieh2019 A Translational Science Model for HCI In Proceedings of CHIConference on Human Factors in Computing Systems Proceedings(CHI 2019) ACM New York NY USA 13 pages httpsdoiorg10114532906053300231

1 INTRODUCTIONTranslational Science (TS) is the study of scientific knowl-edge progression from academia to practice and back Takingfindings from a basic science discovery to real-world impactis a complex process that requires both research (eg lab-oratory and applied research) and non-research activities(eg design and engineering) [28] In this paper we workunder the premise that there is a general goal in appliedfields to translate scientific knowledge to inform the workof professionals In HCIrsquos case it means influencing designpractice as an endpoint [31 69]

However HCI papers offer limited support for practice Forexample only 7 of CHI 2011 papers were oriented towardssupporting design practice [56] At the same time thereis a growing expectation from industry practitioners gov-ernments and the general public that scientific knowledgeshould be useful to society [57 64] Practitioners specificallyhave voiced concerns with the applicability of HCI researchfindings [11 19 36 37] Consequently mapping the Trans-lational Science process is necessary to understand how toincrease the use of HCI discoveries in design practice

Past research in HCI has approached TS using the research-practice gap metaphor This metaphor implies a separationbetween two sides or communities academic researchersand design practitioners However the research-practicegap metaphor can oversimplify the translation work thatHCI scholars and design practitioners do For example HCIscholars often draw on other disciplines to inspire appliedresearch such as cognitive sciences psychology and anthro-pology theories (eg Hutchinsrsquos Distributed Cognition [34]from cognitive science Leontievrsquos Activity Theory [44] frompsychology Suchmanrsquos Situated Action [67] from anthro-pology) The research-practice gap narrative overlooks this

Table 1 Summary of participant information First self-reported percentage of participants with experience in dif-ferent areas within HCI Basic research experience con-tains research in Social Psychology Philosophy ChemistryConsequently experience in Basic research did not count to-wards Years of experience in HCI Second at the bottom weshow participant experience in the HCI field

Area PercentageDesign Practice 44Applied Research 42Teaching 40UX Research 26Basic Research in HCI 21Entrepreneurship 21Basic Research 16Media Outreach 14Policymaking 7

Measure Years in HCIMedian 13Max 47Min 1

type of translational work and if there are multiple gapsthey are likely more nuanced Focusing on a single set ofbarriers from one gap may limit how translational problemsare approached [7]If the research-practice gap metaphor is limited what

model might better depict Translational Science in HCI Inthis paper we present a literature review and interview studyof HCI community members representing various parties tothe Translational Science process Through our findings wecontribute descriptions of the translational research practicesinvolving HCI researchers other scholars design practition-ers and translators such as industry researchers and sciencecommunicators Ourmain contribution is a continuummodelfor Translational Science in HCI providing a foundation forfuture work to discuss Translational Science in HCI

2 METHODSFollowing Zimmerman et alrsquos process to create the model ofinteraction design research within HCI [75] we conducted aliterature review interviewed stakeholders in the HCI com-munity and constructed a model for Translational Sciencein HCI Our literature review consisted of publications onTS translational research and research-practice gaps in HCIand other applied fields such as public health managementscience communication and education Second we inter-viewed academic researchers design practitioners and stu-dents entrepreneurs and science communicators Literaturereview findings were compared to participant perspectives

For interviews we first recruited six HCI scholars experi-enced in research design community and media outreachentrepreneurship teaching and policymaking to obtain a

panorama of translation in HCI and to iterate on our researchprotocol and preliminary model In the second interviewstage we broadened recruiting criteria and interviewed 37participants engaged in HCI-related research and practicefields We highlight that our participants came from manydifferent traditions such as computer science design indus-trial engineering anthropology English literature socialpsychology interaction and user experience design designeducation user experience research and product manage-ment and worked in academia industry and governmentWe also interviewed science communicators and communi-cation managers We had participants with many years ofexperience in HCI-related fields as well as participants whoare newer members of this community and even some that atfirst did not consider themselves members (such as communi-cation managers and some design practitioners) See detailedparticipant information on Table 1 and in SupplementarymaterialsWe asked participants to define their perceptions of the

gap between research and practice where translational worktakes place in HCI and what they have done regarding trans-lation As in the first stage participants commented on thecurrent version of our model for TS in HCI

The development of the model the literature review andinterviews were interdependent The model provided key-words for literature searches and framing concepts for thecreation of the interview protocol At the same time we didnot limit our literature review to a predetermined set of key-words some concepts only emerged from our readings andinterviews We iterated on the model after each interviewAfter many months a detailed model emerged from formaldata analysis and in-depth critique sessions

3 TRANSLATIONAL SCIENCE IN HCI AND LINKEDFIELDS

We start by defining terms related to Translational Science(TS) such as research practice theory and the research-practice gap First TS studies the transformation of knowl-edge through successive fields of research from a basic sci-ence discovery to real-world impact a complex process thatrequires both theory-producing steps (basic and applied) andnon-research steps (dissemination implementation design)The study of each step is called translational research [28]

Human-Computer Interaction (HCI) is ldquoa discipline con-cerned with the design evaluation and implementation ofinteractive computing systems for human use and with thestudy of the major phenomenon surrounding themrdquo [69] Weacknowledge that HCI is a broad discipline that encompassesmany fields and that has evolved throughout the years [9] Inthe field of HCI TS is about translating rich understandingsof people and their interactions with technology with a goalof influencing the design of interactive systems

On one side of the HCI work are theoretical findings Inthis paper we use Yvonne Rogersrsquos and Ben Shneidermanrsquosdefinitions of theory knowledge enabling understandingsand generalizations about specific phenomena Shneidermaneven provides a list of instantiations of knowledge theories(rules frameworks models) descriptions (terminology tax-onomies ontologies) causal explanations and predictionsand more practical knowledge such as solutions (problemfixes improved processes) and guidelines (recommendationspatterns tutorials) [59 64] On the other end of the trans-lational work is design practice the process or practice ofdevising planning or constructing something focusing onthe design of effective user-friendly interactive computingsystems [26 31 70] The goal of the TS problem in HCI isthus to facilitate the adoption implementation and insti-tutionalization of theoretical findings into design practice(much like what is described in [57])

Despite acknowledgments of the value of theoretical in-sights about people in design practice studies have repeat-edly shown that these insights are rarely adopted [11 19 54]Prior work on HCI translations has primarily described theprocess using a gap metaphor [7] The research-practicegap metaphor is a common concept that has been used inHCI to describe an undesired space that allegedly exists be-tween the research and the practice of the discipline [7 31]Much work describes multiple translational barriers thatexplain why a single gap exists and the ldquogaprdquo problema-tization inspired many studies proposing ways to bridgeit [19 22 31 37 50 64]Building on the gap metaphor Gray et al [32] have sug-

gested a trickle-down and bubble-up model emphasizingthat the interaction between the research and practice com-munities is bidirectional Bubble-up describes the efforts ofthe HCI community to abstract situated knowledge and prac-tice of methods tools or concepts into refined theory Thetrickle-down describes the way adaptation of research andtheory takes place in design practice including the use ofmethods tools or concepts that originate in academia

However despite the importance of acknowledging a gapbetween research and practice [50] a gap metaphor mightbe simplistic and might limit how we think about TS inHCI Beck posited that the gap metaphor influences howHCI researchers set translational problems to focus on barri-ers which has implications for the kinds of solutions theyseek [7] Frequently HCI researchers attempt to bridge thegap by pushing HCI knowledge into practitioner use by pre-senting research papers in different formats [19 32 46 54]in an attempt to make the theoretical insights ldquomore perti-nent and easier to use for developersrdquo [12] The proposedsolutions tackle barriers that HCI researchers and practi-tioners have identified as reasons for the research-practicegap to exist For example recent work has described issues

with the terminology and applicability of research papershow practitioners may lack access to academic resourcesthe different cultures and skill sets of researchers and design-ers [19 50 53 56] and many more [36]Countering the focus on barriers Beck and Ekbia [7] en-

couraged the HCI community to focus on the connectionsand continuities between theory and practice in HCI In thiscontext Beck and Ekbia suggest that a new metaphor mdash thecontinuum metaphor mdash which draws attention to the mutualagreement harmony synergy and support between researchand practice However how one can think of the HCI TSproblem as a continuum is unclear

Expanding beyond HCI our literature review shows thatldquothe gaprdquo problem is also the dominant narrative in manyother fields such as Psychology Nursing Human ResourcesLibrary Sciences Management Education Social Work andmore [10 15 23 47 52 60 61 63] Similarly the gapmetaphoris used to describe a disconnect between the research com-munity and the practice community Common barriers in-clude practitioners being unaware of the latest findings [61]thinking that research is not applicable to real-world prob-lems [47 63] or if applicable too difficult to implement [10]and the insularity of academic work dissemination [15 23]In these fields the work of Everett Rogers on diffusion

of innovations [57] has been very influential For Rogersdiffusion is the process by which a central source communi-cates an innovation towards practitioners among whom theinnovation is adopted over time or rejected Rogersrsquos work isparticularly influential in its descriptions of the work of trans-lators to disseminate knowledge and measures of knowledgeadoption A major criticism of this work is that it representsa one-way model which is insufficient as knowledge flowsthrough multiple channels in parallel Also the centralityof research producers in the model disregards the needs ofpractitioners and what they have to offer researchers [55]The biomedical and health sciences have begun using a

continuum metaphor to describe the progression of researchfrom basic science discovery to proposed human applica-tion to clinical treatment [28] They have noted similarissues with using a single gap to describe the discontinu-ities between research and practice Using a single gap hadbecome a source of confusion as different research stepshad differing goals settings study designs investigatorsand outcomes [74] This confusion has hindered the properidentification of translational barriers and prevented properresource allocation by funding agencies [74]In their review article of the translational continuum in

Biomedicine Drolet and Lorenzi discuss three ldquotranslationalchasmsrdquo (Trsquos or gaps) as the main landmarks in knowledgeprogression in the continuum

T1 is the use of knowledge about disease mechanismsin the development of diagnosis methods therapy and pre-vention T1 research is performed by scientists trained inmolecular biology genetics and other basic sciences work-ing in laboratories In contrast ambulatory care settings arethe ldquolaboratoryrdquo for T2 research where applied interventionsfurther the understanding of T1 results by confronting themwith broader populations T2 is known as the clinical trialsstep where safety and efficacy research occur T3 requiresimplementing scientific knowledge applying and evaluatinginterventions in real-world settings and of the disciplinesthat inform the design of those interventions such as clini-cal epidemiology and evidence synthesis but also commu-nication theory behavioral science public policy financingorganizational theory design informatics and mixed meth-ods research Finally T3 is bridged when clinical practicesand guidelines are implemented to produce concrete publichealth changes For example if an effective clinical applica-tion is found (eg aspirin effectively decreases thrombosis inindividual patients) then it must be marketed and explainedto clinical practitioners who may sometimes be trained inthe proper ways to administer a drug before public healthgains can be achieved After T3 is bridged public health im-pact is studied via practice-based research and feedback forcontinual improvement follows (eg why are clinicians notadministering aspirin)

The most noticeable differences between health and HCIpertain to institutional and policy constraints that need to beaddressed to protect drugs and treatment users [48 74] Addi-tionally the continuum model shows the work of many par-ties to progress knowledge through successive steps ratherthan focusing on the work of researchers and clinicians suchas basic scientists applied researchers marketers design-ers clinical practitioners physicians social workers publichealth researchers policymakers [28 68 74]

There are drawbacks in using existing models to describeTS in HCI although we can learn from them The healthmodel is field-specific such as in how it deals with institu-tional constraints The diffusion model focuses on dissem-inating information and does not address the importanceof offering applicable resources to practitioners or learningwhat is useful for them On the other hand health showsmany stakeholders working on a continuum of knowledgeprogressions supporting Beck and Ekhbiarsquos call [7] Thediffusion of innovation shows how translators disseminateinformation resonating with Gray et alrsquos work [32]

Therefore how could we rethink the research-practice gapmetaphor to represent HCIrsquos unique process of knowledgeprogression incorporating learnings from other fields andpast work related to Translational Science

TBA

TBD

Bubble-up

TADAppliedResearch

DesignPractice

BasicResearch

TAA

Figure 1 The HCI Translational Science Model consists of 3main steps Basic and Applied Research (described only asldquoresearchrdquo in the gap metaphor) and Design Practice and 3main gaps TBA TAD and TBD The dotted line representsthe Bubble-up from practice to research TAA suggests a gapwithin Applied research Steps and gaps are defined on Ta-ble 2 and detailed in the remainder of this paper

4 MODEL OF TRANSLATIONAL SCIENCE IN HCIWe propose a model for Translational Science in HCI (Fig-ure 1) We ground the decisions that led to the design of theproposed model in our interview findings First we presenthow participants talked about the gap metaphor transla-tional science in the field and the main actors in translationalscience Second we describe how participants view and en-act relationships that point to a progression of knowledgebetween research and practiceWhile most participants first described a single gap in

Translational Science upon further reflection many notedissues with the dichotomy Initially participants defined TSin HCI as a ldquoBIG gaprdquo or a ldquobarrierrdquo between researchers anddesigners (P3 P11 P17 P22 P23 P26 P35 P39) Participantswould mention known research-practice barriers such asaccessing and understanding scientific work as well as howdissimilar incentive structures could be the biggest driver forthe gap between HCI researchers and Design practitionersMost participants regardless of role agreed that those

in the HCI field share the goal of improving products andservices for people even if not directly Therefore we set theend-goal of our proposed model as Design practice (Figure 1)As participants described their personal experiences re-

lated to translational science and ldquothe gaprdquo the roles of addi-tional parties in the progression of HCI knowledge surfacednaturally such as industry researchers and science commu-nicators as well as a nuanced and fluid variety of academicresearchers (eg P7 had worked in industry launched astartup in the past and now focuses on design research andteaching at her academic department)At the end of the interview after reflecting on their own

initial interpretations of a single research-practice gap itbecame clear to participants that the gap metaphor is insuf-ficient to describe how multiple parties conduct a series oftranslations Most participants especially HCI researchersnoted their work falls in the Applied research step they also

Table 2 Description of the steps and gaps of the HCI Translational Science model Definitions drawn from our data Shneider-man [64] and the NSFmdashNational Science Foundation [1] Of note similar to their academic counterparts industry researcherscan engage in both Basic and Applied research The model does not designate where academic or industry researchers residein the continuum

Steps DescriptionBasic Research Basic research is performed without thought of practical ends resulting in general knowledge to answer a large number of problems though

it may not give a complete specific answer to any one of them In HCI basic research results in a foundational understanding of peopletechnology and design It is achieved by researchers posing formative questions and investigating a topic through the scientific processoften in specific scenarioscontexts Participants mentioned 3 main types of basic research relative to HCI A) Research from other fieldssuch as anthropology and psychology that is drawn upon to conduct Applied research B) Ethnographies conducted within the context ofhow people use or are impacted by technologies and C) HCI research probing the technological boundaries demonstrating new potentialinteraction capabilities and new experiences made possible by technological advances

Applied Research Applied research provides complete answers to practical problems In HCI applied research results in practical goalsolution-orientedresearch It is achieved by HCI researchers posing questions through the use of an embodiment of knowledge study of interface efficacyand how people relate to it observation of tensions between people and technologies safety and ethics in realistic scenarios

Design Practice HCI practitioners using HCI and design knowledge to create something new in the acircAIJmade worldacircAI Involves the deep consideration ofspecific users related stakeholders technical and market requirements Integration of art science and engineering to make aestheticallyfunctional interfaces

Gaps Barriers Main translationsTBA Information over-

load skillsetTranslation of basic science to designs andinteractions through an exchange between Applied and Basic researchLearning and iteration of theories through scientific research processes and communication Translations of basicresearch findings into designs to be tested in Applied research often occurs through a collaboration between basicand applied researchers as well as HCI practitioners

TBD Understandingaccess

Translation of basic science into content that can be consumed more broadly For example pop psychology books orevents where researchers explain their findings to practitioners

TAD Applicability Translation and synthesis of knowledge into usable resources considering specific application domains Developmentof resources that can lead to new knowledge being adopted in Design practice

Bubble-up Personal interestincentives

Looking at practice to inform research observation of designs and practice breakdowns as opportunities to informfurther investigations HCI practitioners broadcasting learnings from their day-to-day problem-solving work on socialmediapublishing platforms

TAA Information over-load traditions

Dissemination and reuse of HCI research from different traditions

noticed that the single gap narrative does not distinguishBasic researchIn the HCI TS model we used the T-terminology to de-

scribe different gaps between steps to facilitate more pre-cise discussion about specific gaps (following Drolet andLorenzirsquos model [28]) Drawing from our interviews themostprominent gaps in the HCI Translational Science model seemto be between Basic and Applied research (TBA) between Ap-plied research and Design practice (TAD) and between Basicresearch and Design practice (TBD) We also heard about howBubble-up happens (supporting [32]) and finally about gapswithin applied research in HCI Table 2 shows descriptionsof the abovementioned steps and gaps with their uniqueknowledge exchanges translations and barriersAs can be seen on Figure 1 and Table 2 the model does

not designate where academic or industry researchers residein the continuum Basic and Applied research can occur bothin academic or industry settings and the same is true forthe design of interactive systems (eg university startupincubators) Additionally Applied research can be tightlylinked to basic research if conducted in an academic settingbut in the industry it gravitates towards design practice

TAD mdash Applied Research to Design PracticeThe gap between Applied research and Design practice iswhat people most often refer to as the research-practice gapParticipants said that basic research findings rarely influencethe work of design practitioners which is one of the reasonswhy translating basic into applied research is important

Participants described known barriers to the use of re-search findings by practitioners such as understanding is-sues due to the terminology in research contributions anddifficulty in accessing them However they highlighted thatapplicability might be the most significant barrier keepingapplied research from influencing design practiceP7 and P13 experienced in both academia and industry

said that expecting HCI practitioners to read papers is unre-alistic A practitioner noted designers will not mdash and cannotmdash spend the time to read papers ldquoI donrsquot have four hours toread one paper I just donrsquot I could get a lot done in those fourhours Frankly I do have 10 minutes to skim through a Mediumarticle Irsquove never understood why in academia they donrsquot setup an incentivization structure for academics to almost likepromote their research to industryrdquo (P35)

When researchers and designers meet and talk in personthe terminology issue is lessened as they can speak directlyto each other and ask for clarification (P11 P13 P14 P30P39 P40) However even at these rare occasions designers

do not get actionable information from researchers (P7 P11P29 P35) ldquoI have been to CHIplay but there is a big barrierbetween industry and academia Very little of what I saw therewas applicable and that event is the only direct contact I havewith academia During the year everyone is locked up in theircompanies or universities and donrsquot speakrdquo (P11 who runs avirtual reality gaming company)

HCI researchers translating between applied research anddesign practice described the applicability barrier as the mostimportant to address to successfully influence the work of de-sign practitioners P14 P30 and P32 all HCI professors saidthat designers do not care about using theories or detailedterminology only wanting to know what can be applied fortheir particular case

Many HCI researchers recognized that the long and diffi-cult applied research process does not result in actionableresources that can be offered to design practice which intheir view is a big problem for HCI (P4 P7 P14 P19 P27P34) An industry researcher who worked in a translationalmedical science project after her doctorate said that HCIldquodoes not have enough research study findings that have beenshown to work in practice or practice-based research Researchfindings are not real world enough to be meaningful and appli-cablerdquo (P16) This concern motivated P19 a long-time CHIcontributor to engage with industry partners tapping intowider populations to generalize her study results and tolearn how to translate her research findings into somethingldquousefulrdquo for practitioners

We then asked participants what has worked in terms oftranslations from applied research to design practice Re-searchers mentioned having success with a ldquomake it happenrdquoapproach They described embedding in organizations asconsultants or doing design work along with practitionersfirst learning what those practitioners care about before of-fering recommendations (P8 P19 P20 P21 P30 P32) andcreating actionable resources for practitioners to use such aseasy to use research and design methods (P7 P14 P19 P32)or open-source projects P8 used an open-source model ldquoitwas an open-source project so they literally put the code up onthe wall and like went through it and used that as the jumpingoff point to invent their startuprdquo and introduced the projectto practitioners in a hackathonP19 told the story of how the idea to create a design

method started ldquothis industry guy asked for resources andthen I asked myself what do I give to this guy Papers andacademic research are too hard for practitioners to get itrdquoHer approach was to study how to transform empirical datafrom published studies into a design method ldquoConnecting topractice was a study in itself a long onerdquo (P19)

Compatibility with existing workflows and demonstratedutility affect adoption of assets and methods by practice sup-porting [57] Scholars mentioned making assets compatible

with current industry practices (P8 P19) and the need forthese assets to prove their usefulness from the outset such asby connecting with metrics that practitioners value ldquoI havelearned one lesson Managers and bosses need to want it Theyneed to be convinced and make that a part of their processThey need to know what the practical outcome of the methodfor them is One manager told me lsquoI will fail if I donrsquot use thismethodrsquo and thatrsquos when I knew I had succeededrdquo (P19)

TBA mdash Basic to Applied ResearchParticipants specifically researchers pointed that apart fromTAD there is also an important gap in the knowledge progres-sion from basic to applied research with its unique barriersFirst translating basic knowledge into research products isa challenging process that happens through collaborationswith parties who possess supportive skills Second thereseems to be a disconnect between HCI and other disciplines

Participants noted that translating theory from the socialsciences ethnography or philosophy into tools for conduct-ing applied research is complicated and often depends onpartnerships as they lack the ldquoskillsetrdquo (P27 P28 P33 P43)P24 a psychology professor said ldquomy main goal is findingreproducible regular patterns in peoplersquos behavior that bothhave real-life potential applicability or consequences and atthe same time can be studied in a fairly controlled way But Iwant my studies to have an impact in real liferdquo

Regarding this challenge participant P27 posed a questionand a proposed path ldquoHow do we facilitate the interaction ofbasic and applied researcher in HCI long term sustainableThis is a big problem We need to do more work on understand-ing how this happens naturally which might help us foster howto facilitate this interactionrdquo P28 also present in the groupinterview agreed saying that he worked in projects withcollaborators from other fields in successful programs dedi-cated to foster basic and applied research exchange fundedby a Translational Medicine associationA few participants described how basic and applied re-

search interact through joint projects Basic researchers maywork with HCI researchers or practitioners to create designsthat can be used in detailed investigations However findingthese collaborations and establishing common-ground canbe challenging and costly (P7 P19 P24 P27 P28)

We hired a computer science PhD student whowas a really good programmer We met with himand he prototyped a version that we had in mindMy colleague and I may say rsquooh letrsquos do thisrsquonot because we know that we have to do it butbecause we happened to think of it but learningfrom the programmer that would be difficult toimplement we would think of something else(P27)

However P27 and P33 emphasized that collaborationswith students or designers are often temporary and expen-sive and that longer-term partnerships would be better InP27rsquos case the student soon graduated and left only the codeChanging and deploying the code required someone to havemdash or gain mdash familiarity with that particular technology stackP27 and his colleague spent a year with the project on a holdbecause of this issue P27 and P33 who are self-described ba-sic researchers even started learning design and engineeringskills to do applied work on their ownSecond as the first point already suggests participants

valued drawing on other domains of knowledge to creatework that is ldquonewrdquo and ldquofreshrdquo (P30 P26) Researchers thinkthat engaging with diverse scholarship transforms them intobetter scholars (P5 P17 P26 P28) Understanding is not abarrier in TBA while it can be difficult to quickly grasp otherresearch traditions there is institutional support throughcitations and a common understanding of papers as an outputto incentivize reading diverse workHowever HCI scholars mentioned barriers to engaging

with other scholarship Scholars do not feel incentivized toparticipate in communities of interest that are peripheral toHCI P28 P29 and P35 said this stifles research innovationMost of the knowledge circulation in HCI happens within theconfines of the CHI community and other ACM conferences

HCI researchers also said their work rarely feeds back intothe original bodies of theory from which they draw on ldquoIdraw on social psychology theories and behavior change andtranslate that into designs So thatrsquos a translation Usuallyyoursquore drawing on many different theories to help inform onecoherent thing Not the opposite I donrsquot feel like I contributeto basic sciencerdquo (P7) Again citation deficit was raised asevidence for a basic-applied research gap (P3 P14) Partici-pants did not mention efforts to facilitate engagement withother disciplines or to promote more mdashand more relevantmdashcitation exchanges

TAA mdash Gap within Applied researchParticipants described another disconnect within HCI inacademia one specifically pertaining to how HCI research isdisseminated and reused in the fieldWhile some participantswho frequently participate and publish at CHI boasted thatit is common to cross-pollinate work with their colleaguesrsquodiverse research (P1 P7 P19 P27 P32) others mentioned acritical barrier related to citations perhaps themain currencyin academia Within sub-communities research papers areoften not cited

I speak directly to those issues We use the samekeywords publish in the same venues It be-comes a personal grudge which is a problemwithin HCI You may not like my work but

it should be cited If you disagree with my ap-proach or findings cite it and explain why (P2)

The reasons given by participants for the citation failurewere information overload and academic traditions Firstparticipants mentioned that it is difficult to keep up with allpublications within an area of studies or across them (P1 P9P15 P17 P32 P43) which is curious as most of the knowl-edge circulation happens within a few well-known venuesRegarding academic traditions some researchers said thatauthors can devalue ethnographies or systems developmentif they do not engage with mdashor understandmdash it which cancause those researchers to not learn from or cite that work(P4 P5 P19 P27) As in TBA participants did not mention ef-forts to promote more and more relevant citation exchangesacross HCI conferences tracks

TBD mdash Basic Research to Design PracticeOur participants initially marked TBD as the rarest transla-tional path Basic researchers do not plan studies or writepapers with the objective of influencing practice and P19mentioned that ldquoHCI is so important because it sits in betweenbasic research and design practice to do the translationrdquo ForP6 P17 P24 and P33 TBD is where the understanding andaccess barriers are more salient However although rareparticipants acknowledged that basic research can influenceHCI practice and provided a few examplesOne approach is to translate basic science into content

to be consumed more broadly For example pop psychol-ogy books [6 29] are often written by practitioners withscientific training sharing scientific findings blended withpersonal industry experience or partnering with scholarsNone of our participants had written or helped to write suchkind of books instead they accomplished other types of out-reach such as participating in radio talks and contributingto news articles about science (P6 P17 P24) Second trans-lating content and disseminating it for a targeted group ofpractitioners in events (P24 P43) and in discussions withpractitioners that are actively searching for new knowledgeto inform product development (P4 P8 P17 P30 P33) Thirdparticipants agreed that training in basic research affordsthe abilities to partner with practitioners through researchmethod skills and deep scientific understanding (P6 P17 P24P28 P33 P40 P43)

Bubble-upBubble-up an important aspect of our model (Figure 2) hasbeen discussed in prior work [32] Our participants whoengaged in Bubble-up work described mostly researcher-ledstrategies to help practice knowledge to influence researchThey also voiced the impression that most of their colleaguesdo not do the same because they are either not personally

interested in design or for lack of incentive For P15 ldquoIt iscrucial for researchers to really engage with practitioners andknow how things are done in industry This is or should be theminimum requirement to contribute to designrdquoIn our interviews practitioners mentioned not engaging

with bubble-up UX researchers and designers are most oftennot incentivized to broadcast findings because of intellec-tual property restrictions (P12 P35) However disseminatingcarefully vetted case studies in platforms such as Mediumhas been more common among practitioners Although ourparticipants did not have experience with such forms of com-munication they mentioned an interest in doing so for per-sonal branding (P29 P39 P41 P42) Interestingly we heardfrom all HCI practitioners that instructors with professionalexperience as designers or engineers effectively made theldquoacademia-industry linkagerdquo (P35) in the classroom bringingknowledge from practice into the education of a new gener-ation of practitioners This view was consistent among HCIprofessors who had industry experience (P10 P11 P13 P15P34 P35 P38)

HCI researchers interested in learning insights from prac-tice put effort into a few strategies that they deem successfulsuch as organizing and attending practitioner-oriented con-ferences (P9 P13 P15 P24) keeping track of students whogo on to leading industry positions (P13 P29 P31) and usingonline design groups ldquoIrsquom part of Facebook and Slack groupsto talk to practitioners Most of the times we talk about newtools or books design trends but sometimes deeper issues andmethods Itrsquos very varied but focused on practice only ItacircĂŹsalmost like a crowdsourced design encyclopedia on currenttopicsrdquo (P15) The Bubble-up goals that our participants men-tioned were to learn what practitioners care about (P9 P13P15 P24) to gain insights about design field trends (P13 P29P31) to improve teaching (P15) and syllabi (P13)

Different Translators and Translational WorkThrough our interviews we uncovered different types oftranslations that address TS barriers the design of exampleproducts people transfer education use of research knowl-edge in industry research and science communication

Synthesizing Information into contextualized products Thistype of translation achieved through the collaborative workof basic researchers applied researchers and designers canentail both translating basic knowledge into research con-cepts for testing and applied research into prototypes andguidelines that can facilitate adoption in design practiceIn the interviews we heard specific instances of transla-

tional work altering the original research insight to create anew product to address more contextualized problems Bothin TBA and in TAD knowledge can get lost in translationbecause of the synthetic nature of design which is reflected

in the accounts of basic and applied researchers who haveengaged in translations (P8 P19 P24) P24 mentioned that aforay into applied research required making many conces-sions and changing their initial ideas because of technicaland design constraints P19 in turn mentioned the need tomake a design method compatible with industry practiceswhich required drastic adaptations

The challenge in translating research into products ormethods is described by Chilana et al through the creationof a new venture born out of HCI research focusing on adop-tion [17] and by the concept of appropriation used by Grayet al [32] As seen in the previous paragraph not only dodesign practitioners partially appropriate knowledge so doresearchers when attempting to bridge gaps

People transfer as knowledge transfer Chilana et alrsquos migra-tion from academia to starting a venture [17] represents asecond path that drives translations the movement of peo-ple People move across gaps and steps such as students orprofessors working in industry temporarily through sabbati-cals collaborative projects or consulting permanently aftergraduating or after a career change People carry knowledgeand skills with them Most participants mentioned that in-dustry researchers for example whether trained in appliedor basic research can help translate the terminology of aca-demic contributions (both in TBD and TAD) We interviewedindustry researchers with academic training who sharedhow they have used academic research to inform their workFirst industry researchers said that most commonly theirknowledge of research methods is employed on a day today basis Second they use scientific findings to inform thecreation of study protocols and initial product design direc-tions In these two latter cases industry researchers benefitfrom their academic experience in reading and conductingscientific studies to quickly evaluate and identify takeawaysin existing bodies of knowledge contextualizing that knowl-edge to the problems that their companies face for the designof products or services (P12 P16 P17 P40 P42)Connected to the points above most participants called

out one particular form of people transfer mdash education mdash asperhaps the most common path for HCI to influence designpractice Academia usually contributes to society with re-search findings but HCI researchers in academia also partici-pate in the training of practitioners Multiple HCI professorsmentioned that teaching HCI theory in classes is challengingand how to do so effectively is an open question ldquoHow dowe effectively train people to be sensitive and make them thinkabout these questions [referring to information privacy andpersuasive design] when they go into the workforcerdquo (P1)

Formal and informal science communication While papersare a primary communication artifact among researcherstime constraints hinder their use by industry researchers To

bridge TAD and TBD gaps generating curated resources is par-ticularly important (P6 P12 P16 P20 P34 P42) ldquoI have timeto read the abstract basically Irsquod love to read more academicpapers but everything is due yesterday in industryrdquo (P12) Par-ticipants valued Nielsen reports (P12 P20 P34) and meetingswith scholars (P21 P40 P43) for helping them learn aboutthe latest relevant readings in a particular areaWriting books blog posts presenting research in events

talks or informal conversations can also help bridge gapsScience communication can be done by writers researchersthemselves or collaboratively Participants saw it as an im-portant pathway to reach and influence scholars practition-ers and public opinion (P1 P6 P7 P22 P23 P25) supportingSmith et al [65] There is also a more informal approach toscience communication where industry researchers use aca-demic research opportunistically in discussions with theirproduct teams or when academic researchers engage incasual conversations with design practitioners about howacademic research can inform product developmentIn translational science communication a major barrier

is balancing scientific precision with writing content thatappeals to specific audiences As in previous studies HCIpractitioners reported difficulty reading research papers (P12P20 P21 P35 P37 P38) For the three science communicatorswe interviewed (P22 P23 P25) the most significant issue inworking with scholars is their difficulty in explaining thingssimply P25 said that academics dislike having their worktranslated ldquobecause they want the original language in thefinal piecerdquo HCI scholars acknowledged this barrier and men-tioned a struggle in balancing simple communication andscientific precision P1 has written blog posts and mentionedthat it is hard to break away from the academic writing styleldquoacademics tend to be very fact-driven and we have a certainstyle of writing and we get this feedback - you have to befriendlier you have to insert fun pieces itrsquos just a differentstyle of writing to make it an engaging piece So even if thedesire is there the ability sometimes can be difficult We aremore precise and nuanced in the way we describe thingsrdquo P7adds that she has hesitated to forego precision because studyresults are not entirely generalizable to all populations

5 DISCUSSIONBy triangulating information from multiple stakeholdersinvolved in the TS process in HCI and models from other ap-plied fields we develop a model for TS in HCI that presentsa more nuanced view with multiple gaps barriers withineach and corresponding translation efforts While this modelof multiple gaps has similarities to models developed inmedicine it also has attributes unique to HCI For exam-ple the HCI model is not linear This is likely because HCI-proposed innovations typically do not face the regulatory

constraints of medical research There is much more flexibil-ity for research insights to directly influence design practicewithout rigorous testing and evaluation

Below we discuss the implications of using our proposedmodel of TS in HCI We then consider how to coordinatetranslational efforts to move scientific HCI research intothe hands and minds of practitioners and the importanceof engaging mdash and studying how to engage mdash with othertranslators in Translational Science

The Value of a Translational Science Model for HCIThe Translational Science process in HCI can be describedas a ldquocontinuumrdquo since various resources and actions areinvolved in a progression of knowledge advancing discover-ies towards design practice There are gaps in this progres-sion and using models can make translational research morelikely to succeed [68] Woolf describes several benefits ofa TS model that distinguishes different steps translationsand disseminations [74] These include repositioning transla-tional barriers into a more specific arrangement helping tounderstand where translation occurs or has stalled Secondmodels of TS can support discussions about allocation ofresources that facilitate Translational Science

First our model can be used to reposition and study trans-lational barriers more precisely Previous work has generateda translational barriers taxonomy [36] resulting from discus-sions between HCI practitioners and researchers [11] thathave been framed by a view of a single gap between researchand practice In our model we specify nuanced barriers andoffer additional insights into how to address them For exam-ple publishing academic papers to broadcast findings withinthe scientific community is still important it is not howevermdash and should not be mdash a functional communication channelbetween research and practice [27]Also HCI scholars doing research on Translational Sci-

ence may describe the aim of their research more clearlywith our terms while helping to refine model constructs Forexample the TS HCI model may be used to describe whereresearch has progressed or stalled (eg use of design appli-cations found in HCI research is delayed in TAD) The modelhighlights explicitly that applicability is the most significantbarrier for HCI research to influence practice in TAD so tofacilitate the translation of work stalled at this stage the HCIcommunity might focus on translations that make it easierfor practitioners to apply theory-driven resources

Second our model can help understand areas where struc-tural support is needed Funding agencies and universitiesare essential pieces of Translational Science in HCI [17 38]as they help drive and support research with broad socialand economic implications [21] Our model can help organi-zations more precisely target steps and barriers where vitaladvances are stalled and catalyze work that can facilitate

translation For example with a more specific TS model thehealth field devised initiatives such as dedicated budgetsresearch centers scientific journals and conferences for dif-ferent Trsquos [18 24 74] While we do not argue that all of theinitiatives listed above are necessary for more robust TS inHCI they should at least be considered Foremost invest-ment in the translation of basic and applied research for usein design practice is vital for capitalizing on investments inproducing new insights An excellent example is a specificproject that emerged in the health domain context a plan toreduce cancer mortality by 2025 consisting of concentratingtranslational research in a few centers that can vet and testscientific findings clusters with the most potential [16]

Multiple Translations and TranslatorsHCI is uniquely positioned and capable of addressing trans-lational barriers as it sits at the intersection between socialsciences computer science and technology [14] FocusingHCIrsquos attention on each gap will help increase both the im-pact of research and the pace of advancement in the fieldThe constellation of parties involved in translating HCI

knowledge is broader than just the sub-communities of HCIresearchers and design practitioners We found that theseparties already cooperate and transition across steps andgaps Based on our findings and results of previous SIGs [36]rather than creating a new profession [50] we believe there isa need to leverage and coordinate the incredibly diverse andcapable parties already somehow involved in TranslationalScience efforts especially those who can mdash and want to mdashdo more translational work (while explaining the value ofdoing so to others who could become translators) Belowwe address efforts that in addition to the findings that wepresented hold promise as effective channels to strengthenTS in HCI and should be developed or further studied

The scientific publication cycle helps bridge TBA Whilethis gap suffers less with understanding barriers major ad-vances could be clustered and communicated to ensure thatknowledge is progressing throughout the TS model To ac-complish this existing infrastructure can be leveraged andexpanded upon such as crowdsourced sites to summarizepublished research [3] and perhaps new ways to aggregateand discover knowledge clusters on Google Scholar and otheracademic work search-engines [35]Our model also highlights an intra-community gap that

needs to be considered TAA gap raises concerns about frag-mentation within the HCI community As the HCI commu-nity grows it is essential to encourage learning about en-gaging with and citing research across application areas andways of knowing

In TBD and TAD translators such as science communica-tors and industry researchers help share curated academicknowledge with design practitioners (similar to what Everett

Rogers described as change agents [57]) HCI practitionersalready use social media sites such as Reddit and Mediumto publish and discuss prominent issues For example theruserexperience reddit community [2] provides a venuefor more than 32000 UX amateurs and professionals to com-municate [40] and 211000 users follow the UX CollectiveMedium publication [4] How can we better leverage theseplatforms for sharing HCI knowledge

One other dissemination pathway is the education of stu-dents and researchers on the complexities of translating sci-entific findings A key question for education is whethercurrent HCI education is sufficient or if the communityshould train new kinds of translators or develop translationresources for existing translators in other fields While Nor-man proposed the training of translational developers [50]our model suggests that different skills and training may beneeded to prepare people who will help bridge each distinctgap We could potentially train many translators and mustconsider this diversity of skills and roles in training for TSIn TAD specifically HCI has the opportunity to translate

research results into resources that practitioners can use HCIscientific findings are usually shared through papers whichintroduce and enforce the applicability barrier as productsof the research process are often not ready for immediate useby practitioners [8 71] and rarely consider the practicalitiesof everyday design practice [32] The literature describesmany different tools that can help communicate researchfindings such as training modules workshops technicalsupport and guides [30 68] In HCI specifically many ve-hicles for HCI knowledge have been proposed but shouldbe more systematically evaluated [7] such as strong con-cepts [33] different instantiations of design patterns [5]scenarios [13] personas [20] conceptual models [49] designconcepts [62] design heuristics [26] research objects [51] de-sign fiction [73] tutorials [25] methods [32] and assets [19]

Must we measure knowledge adoption Describing the pro-cess of adopting and adapting HCI research to practice Chi-lana et al [17] ask whether studying knowledge adoption isbeyond the scope of HCI We believe that there is a need toexplore what it means for an HCI contribution to be adoptedbeyond prototypes [31] mining research insights and trans-lating them to how designers want to be supported intopractical evidence-based resources [50 58] Designing eval-uations of knowledge adoption could bring empirical valueto the design field Our findings support this approach bysuggesting tackling the applicability barrier in TAD to facili-tate the adoption and application of knowledge is necessaryHere there are many obvious strands of work in the in-tersection of TAD with diffusion of innovation studies andknowledge adoption at individual and organizational lev-els [57] as well as many relatable user-centered methods to

Table 3 Additional Translators that can facilitate Translational Science in HCI and should be further investigated

Translators EvidencePolicymakers Mentioned by participants as a growing and effective role for regulating practice (P1 P3 P7 P26) Policymakers have a need for facts

based on the best knowledge currently available [72] Lazar [43] even affirms that for example for accessibility researchers to have anyreal impact outside of the research community they need to understand law and work with policymakers

Professional asso-ciations

UXPA or IxDA for example are organizations that frequently set up local events for the HCI community Our participants believe that astronger link with these associations can help establish interfaces between researchers and practitioners Researchers have partnered withmeetup groups and professional organizations to organize events combining academic and industry talks [19] but practitioners rarely meetwith researchers at these events [50 66] and it is unclear how effective they are at supporting adoption of HCI knowledge

Business stake-holders

Business-related stakeholders such as marketing financing and venture capital There is an understudied pathway in turning academicresearch into commercial products and services a path documented in [17] and mentioned by a few participants in our study (P1 P3 P7P8 P13 P15 P17)

Society end-users Some HCI researchers said that they are not interested in influencing design practice but in working directly with the populations that canbenefit from their work (P5 P8) Ladner points to working with end-users to tackle their problems through an open science approach asan alternative [41] Some participants also mentioned working with the media to influence public opinion This way users would demandchange from companies (P1 P7 P43)

Funding agencies as described in the discussion session top-down stimuli can promote structural change and reorganize infrastructures to facilitate Transla-tional Science More work about mdash and with mdash decision-makers of funding agencies is necessary

uncover stakeholder needs that help align user needs withbusiness needs [17] There is currently little incentive forHCI scholars to invest more time and resources in under-standing adoption [17] If a researcher aims at influencingdesign practice mdash to bridge from TBD or TAD mdash then evalu-ating knowledge adoption [45] may be necessary to knowwhether one has succeeded

Engaging mdash and studying engagement mdash with translators HCImust continue to engage with everyday successes and prob-lems practitioners face in their work An effective feedbackloop in Translational Science helps keep applied fields groundedin and relevant for practice For example asking questionssuch as ldquoIs this privacy recommendation found in a CHI pa-per effective or usedadapted in different ways in practicerdquoResearch with a focus on professional design practices allowsinsight into how practitioners refine and concretize abstractknowledge [32 42] Previous work provides guidance for thein-situ study of professional design practice [31] or usingsocial media traces to learn about design practice in [39]While the abovementioned examples focus on sharing

knowledge with design practitioners they are only one rolein TS in HCI Shneiderman advocates for building the ca-pacity to collaborate and coordinate with many stakehold-ers blending scientists engineers designers and end-usersto produce ldquohigher-impact research converging into a so-lutionrdquo [64] We were impressed by the many original andcreative translations that the HCI community representedthrough our 43 participants and referenced authors hasfound to bridge and study research and practice in HCI Fu-ture work should strive to understand the barriers to andeffective strategies for engaging the range of stakeholders inTS for HCI (See Table 3)

6 CONCLUSIONThe presence of barriers that hamper the progression ofknowledge into design practice is a significant issue within

HCI Therefore it is necessary to understand how knowledgeprogresses or fails to progress from research to practiceTo design the model for Translational Science in HCI wedrew on past work and interview data with researcherspractitioners andmultiple parties who are engaged ndash or notmdashin translating HCI knowledge In our continuum we describemultiple steps and gaps between basic and applied researchand design practice We also identify multiple translators andthe translational work they do This model offers insights onhow to bridge translational gaps and how to work with andtrain translators effectively It also acts as a foundation forfuture research on Translational Science in HCI

ACKNOWLEDGMENTSWe thank our study participants for their time and input intothis research Second we acknowledge reviewers for theirvaluable feedback that greatly improved the quality of thispaper Finally we thank Eclair Junchaya Hy Nguyen RickPaz as well as Sam Kolovson and Alison Kolberg for theirthoughtful feedback and support

REFERENCES[1] 2018 NSF Definitions of Research and Development An Anno-

tated Compilation of Official Sources httpswwwnsfgovstatisticsranddefrd-definitionspdf September 10 2018

[2] nd ruserexperience httpswwwredditcomruserexperienceRetrieved Sep 16 2018

[3] nd Stanford Scholar httpsscholarstanfordedu Retrieved Sep 142018

[4] nd UX Collective httpsuxdesigncc Retrieved Sep 16 2018[5] Christopher Alexander 1977 A pattern language towns buildings

construction Oxford university press[6] Dan Ariely 2008 Predictably Irrational HarperCollins New York NY

USA[7] Jordan Beck and Hamid R Ekbia 2018 The Theory-Practice Gap As

Generative Metaphor In Proceedings of the 2018 CHI Conference onHuman Factors in Computing Systems (CHI rsquo18) ACM New York NYUSA Article 620 11 pages httpsdoiorg10114531735743174194

On one side of the HCI work are theoretical findings Inthis paper we use Yvonne Rogersrsquos and Ben Shneidermanrsquosdefinitions of theory knowledge enabling understandingsand generalizations about specific phenomena Shneidermaneven provides a list of instantiations of knowledge theories(rules frameworks models) descriptions (terminology tax-onomies ontologies) causal explanations and predictionsand more practical knowledge such as solutions (problemfixes improved processes) and guidelines (recommendationspatterns tutorials) [59 64] On the other end of the trans-lational work is design practice the process or practice ofdevising planning or constructing something focusing onthe design of effective user-friendly interactive computingsystems [26 31 70] The goal of the TS problem in HCI isthus to facilitate the adoption implementation and insti-tutionalization of theoretical findings into design practice(much like what is described in [57])

Despite acknowledgments of the value of theoretical in-sights about people in design practice studies have repeat-edly shown that these insights are rarely adopted [11 19 54]Prior work on HCI translations has primarily described theprocess using a gap metaphor [7] The research-practicegap metaphor is a common concept that has been used inHCI to describe an undesired space that allegedly exists be-tween the research and the practice of the discipline [7 31]Much work describes multiple translational barriers thatexplain why a single gap exists and the ldquogaprdquo problema-tization inspired many studies proposing ways to bridgeit [19 22 31 37 50 64]Building on the gap metaphor Gray et al [32] have sug-

gested a trickle-down and bubble-up model emphasizingthat the interaction between the research and practice com-munities is bidirectional Bubble-up describes the efforts ofthe HCI community to abstract situated knowledge and prac-tice of methods tools or concepts into refined theory Thetrickle-down describes the way adaptation of research andtheory takes place in design practice including the use ofmethods tools or concepts that originate in academia

However despite the importance of acknowledging a gapbetween research and practice [50] a gap metaphor mightbe simplistic and might limit how we think about TS inHCI Beck posited that the gap metaphor influences howHCI researchers set translational problems to focus on barri-ers which has implications for the kinds of solutions theyseek [7] Frequently HCI researchers attempt to bridge thegap by pushing HCI knowledge into practitioner use by pre-senting research papers in different formats [19 32 46 54]in an attempt to make the theoretical insights ldquomore perti-nent and easier to use for developersrdquo [12] The proposedsolutions tackle barriers that HCI researchers and practi-tioners have identified as reasons for the research-practicegap to exist For example recent work has described issues

with the terminology and applicability of research papershow practitioners may lack access to academic resourcesthe different cultures and skill sets of researchers and design-ers [19 50 53 56] and many more [36]Countering the focus on barriers Beck and Ekbia [7] en-

couraged the HCI community to focus on the connectionsand continuities between theory and practice in HCI In thiscontext Beck and Ekbia suggest that a new metaphor mdash thecontinuum metaphor mdash which draws attention to the mutualagreement harmony synergy and support between researchand practice However how one can think of the HCI TSproblem as a continuum is unclear

Expanding beyond HCI our literature review shows thatldquothe gaprdquo problem is also the dominant narrative in manyother fields such as Psychology Nursing Human ResourcesLibrary Sciences Management Education Social Work andmore [10 15 23 47 52 60 61 63] Similarly the gapmetaphoris used to describe a disconnect between the research com-munity and the practice community Common barriers in-clude practitioners being unaware of the latest findings [61]thinking that research is not applicable to real-world prob-lems [47 63] or if applicable too difficult to implement [10]and the insularity of academic work dissemination [15 23]In these fields the work of Everett Rogers on diffusion

of innovations [57] has been very influential For Rogersdiffusion is the process by which a central source communi-cates an innovation towards practitioners among whom theinnovation is adopted over time or rejected Rogersrsquos work isparticularly influential in its descriptions of the work of trans-lators to disseminate knowledge and measures of knowledgeadoption A major criticism of this work is that it representsa one-way model which is insufficient as knowledge flowsthrough multiple channels in parallel Also the centralityof research producers in the model disregards the needs ofpractitioners and what they have to offer researchers [55]The biomedical and health sciences have begun using a

continuum metaphor to describe the progression of researchfrom basic science discovery to proposed human applica-tion to clinical treatment [28] They have noted similarissues with using a single gap to describe the discontinu-ities between research and practice Using a single gap hadbecome a source of confusion as different research stepshad differing goals settings study designs investigatorsand outcomes [74] This confusion has hindered the properidentification of translational barriers and prevented properresource allocation by funding agencies [74]In their review article of the translational continuum in

Biomedicine Drolet and Lorenzi discuss three ldquotranslationalchasmsrdquo (Trsquos or gaps) as the main landmarks in knowledgeprogression in the continuum

T1 is the use of knowledge about disease mechanismsin the development of diagnosis methods therapy and pre-vention T1 research is performed by scientists trained inmolecular biology genetics and other basic sciences work-ing in laboratories In contrast ambulatory care settings arethe ldquolaboratoryrdquo for T2 research where applied interventionsfurther the understanding of T1 results by confronting themwith broader populations T2 is known as the clinical trialsstep where safety and efficacy research occur T3 requiresimplementing scientific knowledge applying and evaluatinginterventions in real-world settings and of the disciplinesthat inform the design of those interventions such as clini-cal epidemiology and evidence synthesis but also commu-nication theory behavioral science public policy financingorganizational theory design informatics and mixed meth-ods research Finally T3 is bridged when clinical practicesand guidelines are implemented to produce concrete publichealth changes For example if an effective clinical applica-tion is found (eg aspirin effectively decreases thrombosis inindividual patients) then it must be marketed and explainedto clinical practitioners who may sometimes be trained inthe proper ways to administer a drug before public healthgains can be achieved After T3 is bridged public health im-pact is studied via practice-based research and feedback forcontinual improvement follows (eg why are clinicians notadministering aspirin)

The most noticeable differences between health and HCIpertain to institutional and policy constraints that need to beaddressed to protect drugs and treatment users [48 74] Addi-tionally the continuum model shows the work of many par-ties to progress knowledge through successive steps ratherthan focusing on the work of researchers and clinicians suchas basic scientists applied researchers marketers design-ers clinical practitioners physicians social workers publichealth researchers policymakers [28 68 74]

There are drawbacks in using existing models to describeTS in HCI although we can learn from them The healthmodel is field-specific such as in how it deals with institu-tional constraints The diffusion model focuses on dissem-inating information and does not address the importanceof offering applicable resources to practitioners or learningwhat is useful for them On the other hand health showsmany stakeholders working on a continuum of knowledgeprogressions supporting Beck and Ekhbiarsquos call [7] Thediffusion of innovation shows how translators disseminateinformation resonating with Gray et alrsquos work [32]

Therefore how could we rethink the research-practice gapmetaphor to represent HCIrsquos unique process of knowledgeprogression incorporating learnings from other fields andpast work related to Translational Science

TBA

TBD

Bubble-up

TADAppliedResearch

DesignPractice

BasicResearch

TAA

Figure 1 The HCI Translational Science Model consists of 3main steps Basic and Applied Research (described only asldquoresearchrdquo in the gap metaphor) and Design Practice and 3main gaps TBA TAD and TBD The dotted line representsthe Bubble-up from practice to research TAA suggests a gapwithin Applied research Steps and gaps are defined on Ta-ble 2 and detailed in the remainder of this paper

4 MODEL OF TRANSLATIONAL SCIENCE IN HCIWe propose a model for Translational Science in HCI (Fig-ure 1) We ground the decisions that led to the design of theproposed model in our interview findings First we presenthow participants talked about the gap metaphor transla-tional science in the field and the main actors in translationalscience Second we describe how participants view and en-act relationships that point to a progression of knowledgebetween research and practiceWhile most participants first described a single gap in

Translational Science upon further reflection many notedissues with the dichotomy Initially participants defined TSin HCI as a ldquoBIG gaprdquo or a ldquobarrierrdquo between researchers anddesigners (P3 P11 P17 P22 P23 P26 P35 P39) Participantswould mention known research-practice barriers such asaccessing and understanding scientific work as well as howdissimilar incentive structures could be the biggest driver forthe gap between HCI researchers and Design practitionersMost participants regardless of role agreed that those

in the HCI field share the goal of improving products andservices for people even if not directly Therefore we set theend-goal of our proposed model as Design practice (Figure 1)As participants described their personal experiences re-

lated to translational science and ldquothe gaprdquo the roles of addi-tional parties in the progression of HCI knowledge surfacednaturally such as industry researchers and science commu-nicators as well as a nuanced and fluid variety of academicresearchers (eg P7 had worked in industry launched astartup in the past and now focuses on design research andteaching at her academic department)At the end of the interview after reflecting on their own

initial interpretations of a single research-practice gap itbecame clear to participants that the gap metaphor is insuf-ficient to describe how multiple parties conduct a series oftranslations Most participants especially HCI researchersnoted their work falls in the Applied research step they also

Table 2 Description of the steps and gaps of the HCI Translational Science model Definitions drawn from our data Shneider-man [64] and the NSFmdashNational Science Foundation [1] Of note similar to their academic counterparts industry researcherscan engage in both Basic and Applied research The model does not designate where academic or industry researchers residein the continuum

Steps DescriptionBasic Research Basic research is performed without thought of practical ends resulting in general knowledge to answer a large number of problems though

it may not give a complete specific answer to any one of them In HCI basic research results in a foundational understanding of peopletechnology and design It is achieved by researchers posing formative questions and investigating a topic through the scientific processoften in specific scenarioscontexts Participants mentioned 3 main types of basic research relative to HCI A) Research from other fieldssuch as anthropology and psychology that is drawn upon to conduct Applied research B) Ethnographies conducted within the context ofhow people use or are impacted by technologies and C) HCI research probing the technological boundaries demonstrating new potentialinteraction capabilities and new experiences made possible by technological advances

Applied Research Applied research provides complete answers to practical problems In HCI applied research results in practical goalsolution-orientedresearch It is achieved by HCI researchers posing questions through the use of an embodiment of knowledge study of interface efficacyand how people relate to it observation of tensions between people and technologies safety and ethics in realistic scenarios

Design Practice HCI practitioners using HCI and design knowledge to create something new in the acircAIJmade worldacircAI Involves the deep consideration ofspecific users related stakeholders technical and market requirements Integration of art science and engineering to make aestheticallyfunctional interfaces

Gaps Barriers Main translationsTBA Information over-

load skillsetTranslation of basic science to designs andinteractions through an exchange between Applied and Basic researchLearning and iteration of theories through scientific research processes and communication Translations of basicresearch findings into designs to be tested in Applied research often occurs through a collaboration between basicand applied researchers as well as HCI practitioners

TBD Understandingaccess

Translation of basic science into content that can be consumed more broadly For example pop psychology books orevents where researchers explain their findings to practitioners

TAD Applicability Translation and synthesis of knowledge into usable resources considering specific application domains Developmentof resources that can lead to new knowledge being adopted in Design practice

Bubble-up Personal interestincentives

Looking at practice to inform research observation of designs and practice breakdowns as opportunities to informfurther investigations HCI practitioners broadcasting learnings from their day-to-day problem-solving work on socialmediapublishing platforms

TAA Information over-load traditions

Dissemination and reuse of HCI research from different traditions

noticed that the single gap narrative does not distinguishBasic researchIn the HCI TS model we used the T-terminology to de-

scribe different gaps between steps to facilitate more pre-cise discussion about specific gaps (following Drolet andLorenzirsquos model [28]) Drawing from our interviews themostprominent gaps in the HCI Translational Science model seemto be between Basic and Applied research (TBA) between Ap-plied research and Design practice (TAD) and between Basicresearch and Design practice (TBD) We also heard about howBubble-up happens (supporting [32]) and finally about gapswithin applied research in HCI Table 2 shows descriptionsof the abovementioned steps and gaps with their uniqueknowledge exchanges translations and barriersAs can be seen on Figure 1 and Table 2 the model does

not designate where academic or industry researchers residein the continuum Basic and Applied research can occur bothin academic or industry settings and the same is true forthe design of interactive systems (eg university startupincubators) Additionally Applied research can be tightlylinked to basic research if conducted in an academic settingbut in the industry it gravitates towards design practice

TAD mdash Applied Research to Design PracticeThe gap between Applied research and Design practice iswhat people most often refer to as the research-practice gapParticipants said that basic research findings rarely influencethe work of design practitioners which is one of the reasonswhy translating basic into applied research is important

Participants described known barriers to the use of re-search findings by practitioners such as understanding is-sues due to the terminology in research contributions anddifficulty in accessing them However they highlighted thatapplicability might be the most significant barrier keepingapplied research from influencing design practiceP7 and P13 experienced in both academia and industry

said that expecting HCI practitioners to read papers is unre-alistic A practitioner noted designers will not mdash and cannotmdash spend the time to read papers ldquoI donrsquot have four hours toread one paper I just donrsquot I could get a lot done in those fourhours Frankly I do have 10 minutes to skim through a Mediumarticle Irsquove never understood why in academia they donrsquot setup an incentivization structure for academics to almost likepromote their research to industryrdquo (P35)

When researchers and designers meet and talk in personthe terminology issue is lessened as they can speak directlyto each other and ask for clarification (P11 P13 P14 P30P39 P40) However even at these rare occasions designers

do not get actionable information from researchers (P7 P11P29 P35) ldquoI have been to CHIplay but there is a big barrierbetween industry and academia Very little of what I saw therewas applicable and that event is the only direct contact I havewith academia During the year everyone is locked up in theircompanies or universities and donrsquot speakrdquo (P11 who runs avirtual reality gaming company)

HCI researchers translating between applied research anddesign practice described the applicability barrier as the mostimportant to address to successfully influence the work of de-sign practitioners P14 P30 and P32 all HCI professors saidthat designers do not care about using theories or detailedterminology only wanting to know what can be applied fortheir particular case

Many HCI researchers recognized that the long and diffi-cult applied research process does not result in actionableresources that can be offered to design practice which intheir view is a big problem for HCI (P4 P7 P14 P19 P27P34) An industry researcher who worked in a translationalmedical science project after her doctorate said that HCIldquodoes not have enough research study findings that have beenshown to work in practice or practice-based research Researchfindings are not real world enough to be meaningful and appli-cablerdquo (P16) This concern motivated P19 a long-time CHIcontributor to engage with industry partners tapping intowider populations to generalize her study results and tolearn how to translate her research findings into somethingldquousefulrdquo for practitioners

We then asked participants what has worked in terms oftranslations from applied research to design practice Re-searchers mentioned having success with a ldquomake it happenrdquoapproach They described embedding in organizations asconsultants or doing design work along with practitionersfirst learning what those practitioners care about before of-fering recommendations (P8 P19 P20 P21 P30 P32) andcreating actionable resources for practitioners to use such aseasy to use research and design methods (P7 P14 P19 P32)or open-source projects P8 used an open-source model ldquoitwas an open-source project so they literally put the code up onthe wall and like went through it and used that as the jumpingoff point to invent their startuprdquo and introduced the projectto practitioners in a hackathonP19 told the story of how the idea to create a design

method started ldquothis industry guy asked for resources andthen I asked myself what do I give to this guy Papers andacademic research are too hard for practitioners to get itrdquoHer approach was to study how to transform empirical datafrom published studies into a design method ldquoConnecting topractice was a study in itself a long onerdquo (P19)

Compatibility with existing workflows and demonstratedutility affect adoption of assets and methods by practice sup-porting [57] Scholars mentioned making assets compatible

with current industry practices (P8 P19) and the need forthese assets to prove their usefulness from the outset such asby connecting with metrics that practitioners value ldquoI havelearned one lesson Managers and bosses need to want it Theyneed to be convinced and make that a part of their processThey need to know what the practical outcome of the methodfor them is One manager told me lsquoI will fail if I donrsquot use thismethodrsquo and thatrsquos when I knew I had succeededrdquo (P19)

TBA mdash Basic to Applied ResearchParticipants specifically researchers pointed that apart fromTAD there is also an important gap in the knowledge progres-sion from basic to applied research with its unique barriersFirst translating basic knowledge into research products isa challenging process that happens through collaborationswith parties who possess supportive skills Second thereseems to be a disconnect between HCI and other disciplines

Participants noted that translating theory from the socialsciences ethnography or philosophy into tools for conduct-ing applied research is complicated and often depends onpartnerships as they lack the ldquoskillsetrdquo (P27 P28 P33 P43)P24 a psychology professor said ldquomy main goal is findingreproducible regular patterns in peoplersquos behavior that bothhave real-life potential applicability or consequences and atthe same time can be studied in a fairly controlled way But Iwant my studies to have an impact in real liferdquo

Regarding this challenge participant P27 posed a questionand a proposed path ldquoHow do we facilitate the interaction ofbasic and applied researcher in HCI long term sustainableThis is a big problem We need to do more work on understand-ing how this happens naturally which might help us foster howto facilitate this interactionrdquo P28 also present in the groupinterview agreed saying that he worked in projects withcollaborators from other fields in successful programs dedi-cated to foster basic and applied research exchange fundedby a Translational Medicine associationA few participants described how basic and applied re-

search interact through joint projects Basic researchers maywork with HCI researchers or practitioners to create designsthat can be used in detailed investigations However findingthese collaborations and establishing common-ground canbe challenging and costly (P7 P19 P24 P27 P28)

We hired a computer science PhD student whowas a really good programmer We met with himand he prototyped a version that we had in mindMy colleague and I may say rsquooh letrsquos do thisrsquonot because we know that we have to do it butbecause we happened to think of it but learningfrom the programmer that would be difficult toimplement we would think of something else(P27)

However P27 and P33 emphasized that collaborationswith students or designers are often temporary and expen-sive and that longer-term partnerships would be better InP27rsquos case the student soon graduated and left only the codeChanging and deploying the code required someone to havemdash or gain mdash familiarity with that particular technology stackP27 and his colleague spent a year with the project on a holdbecause of this issue P27 and P33 who are self-described ba-sic researchers even started learning design and engineeringskills to do applied work on their ownSecond as the first point already suggests participants

valued drawing on other domains of knowledge to creatework that is ldquonewrdquo and ldquofreshrdquo (P30 P26) Researchers thinkthat engaging with diverse scholarship transforms them intobetter scholars (P5 P17 P26 P28) Understanding is not abarrier in TBA while it can be difficult to quickly grasp otherresearch traditions there is institutional support throughcitations and a common understanding of papers as an outputto incentivize reading diverse workHowever HCI scholars mentioned barriers to engaging

with other scholarship Scholars do not feel incentivized toparticipate in communities of interest that are peripheral toHCI P28 P29 and P35 said this stifles research innovationMost of the knowledge circulation in HCI happens within theconfines of the CHI community and other ACM conferences

HCI researchers also said their work rarely feeds back intothe original bodies of theory from which they draw on ldquoIdraw on social psychology theories and behavior change andtranslate that into designs So thatrsquos a translation Usuallyyoursquore drawing on many different theories to help inform onecoherent thing Not the opposite I donrsquot feel like I contributeto basic sciencerdquo (P7) Again citation deficit was raised asevidence for a basic-applied research gap (P3 P14) Partici-pants did not mention efforts to facilitate engagement withother disciplines or to promote more mdashand more relevantmdashcitation exchanges

TAA mdash Gap within Applied researchParticipants described another disconnect within HCI inacademia one specifically pertaining to how HCI research isdisseminated and reused in the fieldWhile some participantswho frequently participate and publish at CHI boasted thatit is common to cross-pollinate work with their colleaguesrsquodiverse research (P1 P7 P19 P27 P32) others mentioned acritical barrier related to citations perhaps themain currencyin academia Within sub-communities research papers areoften not cited

I speak directly to those issues We use the samekeywords publish in the same venues It be-comes a personal grudge which is a problemwithin HCI You may not like my work but

it should be cited If you disagree with my ap-proach or findings cite it and explain why (P2)

The reasons given by participants for the citation failurewere information overload and academic traditions Firstparticipants mentioned that it is difficult to keep up with allpublications within an area of studies or across them (P1 P9P15 P17 P32 P43) which is curious as most of the knowl-edge circulation happens within a few well-known venuesRegarding academic traditions some researchers said thatauthors can devalue ethnographies or systems developmentif they do not engage with mdashor understandmdash it which cancause those researchers to not learn from or cite that work(P4 P5 P19 P27) As in TBA participants did not mention ef-forts to promote more and more relevant citation exchangesacross HCI conferences tracks

TBD mdash Basic Research to Design PracticeOur participants initially marked TBD as the rarest transla-tional path Basic researchers do not plan studies or writepapers with the objective of influencing practice and P19mentioned that ldquoHCI is so important because it sits in betweenbasic research and design practice to do the translationrdquo ForP6 P17 P24 and P33 TBD is where the understanding andaccess barriers are more salient However although rareparticipants acknowledged that basic research can influenceHCI practice and provided a few examplesOne approach is to translate basic science into content

to be consumed more broadly For example pop psychol-ogy books [6 29] are often written by practitioners withscientific training sharing scientific findings blended withpersonal industry experience or partnering with scholarsNone of our participants had written or helped to write suchkind of books instead they accomplished other types of out-reach such as participating in radio talks and contributingto news articles about science (P6 P17 P24) Second trans-lating content and disseminating it for a targeted group ofpractitioners in events (P24 P43) and in discussions withpractitioners that are actively searching for new knowledgeto inform product development (P4 P8 P17 P30 P33) Thirdparticipants agreed that training in basic research affordsthe abilities to partner with practitioners through researchmethod skills and deep scientific understanding (P6 P17 P24P28 P33 P40 P43)

Bubble-upBubble-up an important aspect of our model (Figure 2) hasbeen discussed in prior work [32] Our participants whoengaged in Bubble-up work described mostly researcher-ledstrategies to help practice knowledge to influence researchThey also voiced the impression that most of their colleaguesdo not do the same because they are either not personally

interested in design or for lack of incentive For P15 ldquoIt iscrucial for researchers to really engage with practitioners andknow how things are done in industry This is or should be theminimum requirement to contribute to designrdquoIn our interviews practitioners mentioned not engaging

with bubble-up UX researchers and designers are most oftennot incentivized to broadcast findings because of intellec-tual property restrictions (P12 P35) However disseminatingcarefully vetted case studies in platforms such as Mediumhas been more common among practitioners Although ourparticipants did not have experience with such forms of com-munication they mentioned an interest in doing so for per-sonal branding (P29 P39 P41 P42) Interestingly we heardfrom all HCI practitioners that instructors with professionalexperience as designers or engineers effectively made theldquoacademia-industry linkagerdquo (P35) in the classroom bringingknowledge from practice into the education of a new gener-ation of practitioners This view was consistent among HCIprofessors who had industry experience (P10 P11 P13 P15P34 P35 P38)

HCI researchers interested in learning insights from prac-tice put effort into a few strategies that they deem successfulsuch as organizing and attending practitioner-oriented con-ferences (P9 P13 P15 P24) keeping track of students whogo on to leading industry positions (P13 P29 P31) and usingonline design groups ldquoIrsquom part of Facebook and Slack groupsto talk to practitioners Most of the times we talk about newtools or books design trends but sometimes deeper issues andmethods Itrsquos very varied but focused on practice only ItacircĂŹsalmost like a crowdsourced design encyclopedia on currenttopicsrdquo (P15) The Bubble-up goals that our participants men-tioned were to learn what practitioners care about (P9 P13P15 P24) to gain insights about design field trends (P13 P29P31) to improve teaching (P15) and syllabi (P13)

Different Translators and Translational WorkThrough our interviews we uncovered different types oftranslations that address TS barriers the design of exampleproducts people transfer education use of research knowl-edge in industry research and science communication

Synthesizing Information into contextualized products Thistype of translation achieved through the collaborative workof basic researchers applied researchers and designers canentail both translating basic knowledge into research con-cepts for testing and applied research into prototypes andguidelines that can facilitate adoption in design practiceIn the interviews we heard specific instances of transla-

tional work altering the original research insight to create anew product to address more contextualized problems Bothin TBA and in TAD knowledge can get lost in translationbecause of the synthetic nature of design which is reflected

in the accounts of basic and applied researchers who haveengaged in translations (P8 P19 P24) P24 mentioned that aforay into applied research required making many conces-sions and changing their initial ideas because of technicaland design constraints P19 in turn mentioned the need tomake a design method compatible with industry practiceswhich required drastic adaptations

The challenge in translating research into products ormethods is described by Chilana et al through the creationof a new venture born out of HCI research focusing on adop-tion [17] and by the concept of appropriation used by Grayet al [32] As seen in the previous paragraph not only dodesign practitioners partially appropriate knowledge so doresearchers when attempting to bridge gaps

People transfer as knowledge transfer Chilana et alrsquos migra-tion from academia to starting a venture [17] represents asecond path that drives translations the movement of peo-ple People move across gaps and steps such as students orprofessors working in industry temporarily through sabbati-cals collaborative projects or consulting permanently aftergraduating or after a career change People carry knowledgeand skills with them Most participants mentioned that in-dustry researchers for example whether trained in appliedor basic research can help translate the terminology of aca-demic contributions (both in TBD and TAD) We interviewedindustry researchers with academic training who sharedhow they have used academic research to inform their workFirst industry researchers said that most commonly theirknowledge of research methods is employed on a day today basis Second they use scientific findings to inform thecreation of study protocols and initial product design direc-tions In these two latter cases industry researchers benefitfrom their academic experience in reading and conductingscientific studies to quickly evaluate and identify takeawaysin existing bodies of knowledge contextualizing that knowl-edge to the problems that their companies face for the designof products or services (P12 P16 P17 P40 P42)Connected to the points above most participants called

out one particular form of people transfer mdash education mdash asperhaps the most common path for HCI to influence designpractice Academia usually contributes to society with re-search findings but HCI researchers in academia also partici-pate in the training of practitioners Multiple HCI professorsmentioned that teaching HCI theory in classes is challengingand how to do so effectively is an open question ldquoHow dowe effectively train people to be sensitive and make them thinkabout these questions [referring to information privacy andpersuasive design] when they go into the workforcerdquo (P1)

Formal and informal science communication While papersare a primary communication artifact among researcherstime constraints hinder their use by industry researchers To

bridge TAD and TBD gaps generating curated resources is par-ticularly important (P6 P12 P16 P20 P34 P42) ldquoI have timeto read the abstract basically Irsquod love to read more academicpapers but everything is due yesterday in industryrdquo (P12) Par-ticipants valued Nielsen reports (P12 P20 P34) and meetingswith scholars (P21 P40 P43) for helping them learn aboutthe latest relevant readings in a particular areaWriting books blog posts presenting research in events

talks or informal conversations can also help bridge gapsScience communication can be done by writers researchersthemselves or collaboratively Participants saw it as an im-portant pathway to reach and influence scholars practition-ers and public opinion (P1 P6 P7 P22 P23 P25) supportingSmith et al [65] There is also a more informal approach toscience communication where industry researchers use aca-demic research opportunistically in discussions with theirproduct teams or when academic researchers engage incasual conversations with design practitioners about howacademic research can inform product developmentIn translational science communication a major barrier

is balancing scientific precision with writing content thatappeals to specific audiences As in previous studies HCIpractitioners reported difficulty reading research papers (P12P20 P21 P35 P37 P38) For the three science communicatorswe interviewed (P22 P23 P25) the most significant issue inworking with scholars is their difficulty in explaining thingssimply P25 said that academics dislike having their worktranslated ldquobecause they want the original language in thefinal piecerdquo HCI scholars acknowledged this barrier and men-tioned a struggle in balancing simple communication andscientific precision P1 has written blog posts and mentionedthat it is hard to break away from the academic writing styleldquoacademics tend to be very fact-driven and we have a certainstyle of writing and we get this feedback - you have to befriendlier you have to insert fun pieces itrsquos just a differentstyle of writing to make it an engaging piece So even if thedesire is there the ability sometimes can be difficult We aremore precise and nuanced in the way we describe thingsrdquo P7adds that she has hesitated to forego precision because studyresults are not entirely generalizable to all populations

5 DISCUSSIONBy triangulating information from multiple stakeholdersinvolved in the TS process in HCI and models from other ap-plied fields we develop a model for TS in HCI that presentsa more nuanced view with multiple gaps barriers withineach and corresponding translation efforts While this modelof multiple gaps has similarities to models developed inmedicine it also has attributes unique to HCI For exam-ple the HCI model is not linear This is likely because HCI-proposed innovations typically do not face the regulatory

constraints of medical research There is much more flexibil-ity for research insights to directly influence design practicewithout rigorous testing and evaluation

Below we discuss the implications of using our proposedmodel of TS in HCI We then consider how to coordinatetranslational efforts to move scientific HCI research intothe hands and minds of practitioners and the importanceof engaging mdash and studying how to engage mdash with othertranslators in Translational Science

The Value of a Translational Science Model for HCIThe Translational Science process in HCI can be describedas a ldquocontinuumrdquo since various resources and actions areinvolved in a progression of knowledge advancing discover-ies towards design practice There are gaps in this progres-sion and using models can make translational research morelikely to succeed [68] Woolf describes several benefits ofa TS model that distinguishes different steps translationsand disseminations [74] These include repositioning transla-tional barriers into a more specific arrangement helping tounderstand where translation occurs or has stalled Secondmodels of TS can support discussions about allocation ofresources that facilitate Translational Science

First our model can be used to reposition and study trans-lational barriers more precisely Previous work has generateda translational barriers taxonomy [36] resulting from discus-sions between HCI practitioners and researchers [11] thathave been framed by a view of a single gap between researchand practice In our model we specify nuanced barriers andoffer additional insights into how to address them For exam-ple publishing academic papers to broadcast findings withinthe scientific community is still important it is not howevermdash and should not be mdash a functional communication channelbetween research and practice [27]Also HCI scholars doing research on Translational Sci-

ence may describe the aim of their research more clearlywith our terms while helping to refine model constructs Forexample the TS HCI model may be used to describe whereresearch has progressed or stalled (eg use of design appli-cations found in HCI research is delayed in TAD) The modelhighlights explicitly that applicability is the most significantbarrier for HCI research to influence practice in TAD so tofacilitate the translation of work stalled at this stage the HCIcommunity might focus on translations that make it easierfor practitioners to apply theory-driven resources

Second our model can help understand areas where struc-tural support is needed Funding agencies and universitiesare essential pieces of Translational Science in HCI [17 38]as they help drive and support research with broad socialand economic implications [21] Our model can help organi-zations more precisely target steps and barriers where vitaladvances are stalled and catalyze work that can facilitate

translation For example with a more specific TS model thehealth field devised initiatives such as dedicated budgetsresearch centers scientific journals and conferences for dif-ferent Trsquos [18 24 74] While we do not argue that all of theinitiatives listed above are necessary for more robust TS inHCI they should at least be considered Foremost invest-ment in the translation of basic and applied research for usein design practice is vital for capitalizing on investments inproducing new insights An excellent example is a specificproject that emerged in the health domain context a plan toreduce cancer mortality by 2025 consisting of concentratingtranslational research in a few centers that can vet and testscientific findings clusters with the most potential [16]

Multiple Translations and TranslatorsHCI is uniquely positioned and capable of addressing trans-lational barriers as it sits at the intersection between socialsciences computer science and technology [14] FocusingHCIrsquos attention on each gap will help increase both the im-pact of research and the pace of advancement in the fieldThe constellation of parties involved in translating HCI

knowledge is broader than just the sub-communities of HCIresearchers and design practitioners We found that theseparties already cooperate and transition across steps andgaps Based on our findings and results of previous SIGs [36]rather than creating a new profession [50] we believe there isa need to leverage and coordinate the incredibly diverse andcapable parties already somehow involved in TranslationalScience efforts especially those who can mdash and want to mdashdo more translational work (while explaining the value ofdoing so to others who could become translators) Belowwe address efforts that in addition to the findings that wepresented hold promise as effective channels to strengthenTS in HCI and should be developed or further studied

The scientific publication cycle helps bridge TBA Whilethis gap suffers less with understanding barriers major ad-vances could be clustered and communicated to ensure thatknowledge is progressing throughout the TS model To ac-complish this existing infrastructure can be leveraged andexpanded upon such as crowdsourced sites to summarizepublished research [3] and perhaps new ways to aggregateand discover knowledge clusters on Google Scholar and otheracademic work search-engines [35]Our model also highlights an intra-community gap that

needs to be considered TAA gap raises concerns about frag-mentation within the HCI community As the HCI commu-nity grows it is essential to encourage learning about en-gaging with and citing research across application areas andways of knowing

In TBD and TAD translators such as science communica-tors and industry researchers help share curated academicknowledge with design practitioners (similar to what Everett

Rogers described as change agents [57]) HCI practitionersalready use social media sites such as Reddit and Mediumto publish and discuss prominent issues For example theruserexperience reddit community [2] provides a venuefor more than 32000 UX amateurs and professionals to com-municate [40] and 211000 users follow the UX CollectiveMedium publication [4] How can we better leverage theseplatforms for sharing HCI knowledge

One other dissemination pathway is the education of stu-dents and researchers on the complexities of translating sci-entific findings A key question for education is whethercurrent HCI education is sufficient or if the communityshould train new kinds of translators or develop translationresources for existing translators in other fields While Nor-man proposed the training of translational developers [50]our model suggests that different skills and training may beneeded to prepare people who will help bridge each distinctgap We could potentially train many translators and mustconsider this diversity of skills and roles in training for TSIn TAD specifically HCI has the opportunity to translate

research results into resources that practitioners can use HCIscientific findings are usually shared through papers whichintroduce and enforce the applicability barrier as productsof the research process are often not ready for immediate useby practitioners [8 71] and rarely consider the practicalitiesof everyday design practice [32] The literature describesmany different tools that can help communicate researchfindings such as training modules workshops technicalsupport and guides [30 68] In HCI specifically many ve-hicles for HCI knowledge have been proposed but shouldbe more systematically evaluated [7] such as strong con-cepts [33] different instantiations of design patterns [5]scenarios [13] personas [20] conceptual models [49] designconcepts [62] design heuristics [26] research objects [51] de-sign fiction [73] tutorials [25] methods [32] and assets [19]

Must we measure knowledge adoption Describing the pro-cess of adopting and adapting HCI research to practice Chi-lana et al [17] ask whether studying knowledge adoption isbeyond the scope of HCI We believe that there is a need toexplore what it means for an HCI contribution to be adoptedbeyond prototypes [31] mining research insights and trans-lating them to how designers want to be supported intopractical evidence-based resources [50 58] Designing eval-uations of knowledge adoption could bring empirical valueto the design field Our findings support this approach bysuggesting tackling the applicability barrier in TAD to facili-tate the adoption and application of knowledge is necessaryHere there are many obvious strands of work in the in-tersection of TAD with diffusion of innovation studies andknowledge adoption at individual and organizational lev-els [57] as well as many relatable user-centered methods to

Table 3 Additional Translators that can facilitate Translational Science in HCI and should be further investigated

Translators EvidencePolicymakers Mentioned by participants as a growing and effective role for regulating practice (P1 P3 P7 P26) Policymakers have a need for facts

based on the best knowledge currently available [72] Lazar [43] even affirms that for example for accessibility researchers to have anyreal impact outside of the research community they need to understand law and work with policymakers

Professional asso-ciations

UXPA or IxDA for example are organizations that frequently set up local events for the HCI community Our participants believe that astronger link with these associations can help establish interfaces between researchers and practitioners Researchers have partnered withmeetup groups and professional organizations to organize events combining academic and industry talks [19] but practitioners rarely meetwith researchers at these events [50 66] and it is unclear how effective they are at supporting adoption of HCI knowledge

Business stake-holders

Business-related stakeholders such as marketing financing and venture capital There is an understudied pathway in turning academicresearch into commercial products and services a path documented in [17] and mentioned by a few participants in our study (P1 P3 P7P8 P13 P15 P17)

Society end-users Some HCI researchers said that they are not interested in influencing design practice but in working directly with the populations that canbenefit from their work (P5 P8) Ladner points to working with end-users to tackle their problems through an open science approach asan alternative [41] Some participants also mentioned working with the media to influence public opinion This way users would demandchange from companies (P1 P7 P43)

Funding agencies as described in the discussion session top-down stimuli can promote structural change and reorganize infrastructures to facilitate Transla-tional Science More work about mdash and with mdash decision-makers of funding agencies is necessary

uncover stakeholder needs that help align user needs withbusiness needs [17] There is currently little incentive forHCI scholars to invest more time and resources in under-standing adoption [17] If a researcher aims at influencingdesign practice mdash to bridge from TBD or TAD mdash then evalu-ating knowledge adoption [45] may be necessary to knowwhether one has succeeded

Engaging mdash and studying engagement mdash with translators HCImust continue to engage with everyday successes and prob-lems practitioners face in their work An effective feedbackloop in Translational Science helps keep applied fields groundedin and relevant for practice For example asking questionssuch as ldquoIs this privacy recommendation found in a CHI pa-per effective or usedadapted in different ways in practicerdquoResearch with a focus on professional design practices allowsinsight into how practitioners refine and concretize abstractknowledge [32 42] Previous work provides guidance for thein-situ study of professional design practice [31] or usingsocial media traces to learn about design practice in [39]While the abovementioned examples focus on sharing

knowledge with design practitioners they are only one rolein TS in HCI Shneiderman advocates for building the ca-pacity to collaborate and coordinate with many stakehold-ers blending scientists engineers designers and end-usersto produce ldquohigher-impact research converging into a so-lutionrdquo [64] We were impressed by the many original andcreative translations that the HCI community representedthrough our 43 participants and referenced authors hasfound to bridge and study research and practice in HCI Fu-ture work should strive to understand the barriers to andeffective strategies for engaging the range of stakeholders inTS for HCI (See Table 3)

6 CONCLUSIONThe presence of barriers that hamper the progression ofknowledge into design practice is a significant issue within

HCI Therefore it is necessary to understand how knowledgeprogresses or fails to progress from research to practiceTo design the model for Translational Science in HCI wedrew on past work and interview data with researcherspractitioners andmultiple parties who are engaged ndash or notmdashin translating HCI knowledge In our continuum we describemultiple steps and gaps between basic and applied researchand design practice We also identify multiple translators andthe translational work they do This model offers insights onhow to bridge translational gaps and how to work with andtrain translators effectively It also acts as a foundation forfuture research on Translational Science in HCI

ACKNOWLEDGMENTSWe thank our study participants for their time and input intothis research Second we acknowledge reviewers for theirvaluable feedback that greatly improved the quality of thispaper Finally we thank Eclair Junchaya Hy Nguyen RickPaz as well as Sam Kolovson and Alison Kolberg for theirthoughtful feedback and support

REFERENCES[1] 2018 NSF Definitions of Research and Development An Anno-

tated Compilation of Official Sources httpswwwnsfgovstatisticsranddefrd-definitionspdf September 10 2018

[2] nd ruserexperience httpswwwredditcomruserexperienceRetrieved Sep 16 2018

[3] nd Stanford Scholar httpsscholarstanfordedu Retrieved Sep 142018

[4] nd UX Collective httpsuxdesigncc Retrieved Sep 16 2018[5] Christopher Alexander 1977 A pattern language towns buildings

construction Oxford university press[6] Dan Ariely 2008 Predictably Irrational HarperCollins New York NY

USA[7] Jordan Beck and Hamid R Ekbia 2018 The Theory-Practice Gap As

Generative Metaphor In Proceedings of the 2018 CHI Conference onHuman Factors in Computing Systems (CHI rsquo18) ACM New York NYUSA Article 620 11 pages httpsdoiorg10114531735743174194

T1 is the use of knowledge about disease mechanismsin the development of diagnosis methods therapy and pre-vention T1 research is performed by scientists trained inmolecular biology genetics and other basic sciences work-ing in laboratories In contrast ambulatory care settings arethe ldquolaboratoryrdquo for T2 research where applied interventionsfurther the understanding of T1 results by confronting themwith broader populations T2 is known as the clinical trialsstep where safety and efficacy research occur T3 requiresimplementing scientific knowledge applying and evaluatinginterventions in real-world settings and of the disciplinesthat inform the design of those interventions such as clini-cal epidemiology and evidence synthesis but also commu-nication theory behavioral science public policy financingorganizational theory design informatics and mixed meth-ods research Finally T3 is bridged when clinical practicesand guidelines are implemented to produce concrete publichealth changes For example if an effective clinical applica-tion is found (eg aspirin effectively decreases thrombosis inindividual patients) then it must be marketed and explainedto clinical practitioners who may sometimes be trained inthe proper ways to administer a drug before public healthgains can be achieved After T3 is bridged public health im-pact is studied via practice-based research and feedback forcontinual improvement follows (eg why are clinicians notadministering aspirin)

The most noticeable differences between health and HCIpertain to institutional and policy constraints that need to beaddressed to protect drugs and treatment users [48 74] Addi-tionally the continuum model shows the work of many par-ties to progress knowledge through successive steps ratherthan focusing on the work of researchers and clinicians suchas basic scientists applied researchers marketers design-ers clinical practitioners physicians social workers publichealth researchers policymakers [28 68 74]

There are drawbacks in using existing models to describeTS in HCI although we can learn from them The healthmodel is field-specific such as in how it deals with institu-tional constraints The diffusion model focuses on dissem-inating information and does not address the importanceof offering applicable resources to practitioners or learningwhat is useful for them On the other hand health showsmany stakeholders working on a continuum of knowledgeprogressions supporting Beck and Ekhbiarsquos call [7] Thediffusion of innovation shows how translators disseminateinformation resonating with Gray et alrsquos work [32]

Therefore how could we rethink the research-practice gapmetaphor to represent HCIrsquos unique process of knowledgeprogression incorporating learnings from other fields andpast work related to Translational Science

TBA

TBD

Bubble-up

TADAppliedResearch

DesignPractice

BasicResearch

TAA

Figure 1 The HCI Translational Science Model consists of 3main steps Basic and Applied Research (described only asldquoresearchrdquo in the gap metaphor) and Design Practice and 3main gaps TBA TAD and TBD The dotted line representsthe Bubble-up from practice to research TAA suggests a gapwithin Applied research Steps and gaps are defined on Ta-ble 2 and detailed in the remainder of this paper

4 MODEL OF TRANSLATIONAL SCIENCE IN HCIWe propose a model for Translational Science in HCI (Fig-ure 1) We ground the decisions that led to the design of theproposed model in our interview findings First we presenthow participants talked about the gap metaphor transla-tional science in the field and the main actors in translationalscience Second we describe how participants view and en-act relationships that point to a progression of knowledgebetween research and practiceWhile most participants first described a single gap in

Translational Science upon further reflection many notedissues with the dichotomy Initially participants defined TSin HCI as a ldquoBIG gaprdquo or a ldquobarrierrdquo between researchers anddesigners (P3 P11 P17 P22 P23 P26 P35 P39) Participantswould mention known research-practice barriers such asaccessing and understanding scientific work as well as howdissimilar incentive structures could be the biggest driver forthe gap between HCI researchers and Design practitionersMost participants regardless of role agreed that those

in the HCI field share the goal of improving products andservices for people even if not directly Therefore we set theend-goal of our proposed model as Design practice (Figure 1)As participants described their personal experiences re-

lated to translational science and ldquothe gaprdquo the roles of addi-tional parties in the progression of HCI knowledge surfacednaturally such as industry researchers and science commu-nicators as well as a nuanced and fluid variety of academicresearchers (eg P7 had worked in industry launched astartup in the past and now focuses on design research andteaching at her academic department)At the end of the interview after reflecting on their own

initial interpretations of a single research-practice gap itbecame clear to participants that the gap metaphor is insuf-ficient to describe how multiple parties conduct a series oftranslations Most participants especially HCI researchersnoted their work falls in the Applied research step they also

Table 2 Description of the steps and gaps of the HCI Translational Science model Definitions drawn from our data Shneider-man [64] and the NSFmdashNational Science Foundation [1] Of note similar to their academic counterparts industry researcherscan engage in both Basic and Applied research The model does not designate where academic or industry researchers residein the continuum

Steps DescriptionBasic Research Basic research is performed without thought of practical ends resulting in general knowledge to answer a large number of problems though

it may not give a complete specific answer to any one of them In HCI basic research results in a foundational understanding of peopletechnology and design It is achieved by researchers posing formative questions and investigating a topic through the scientific processoften in specific scenarioscontexts Participants mentioned 3 main types of basic research relative to HCI A) Research from other fieldssuch as anthropology and psychology that is drawn upon to conduct Applied research B) Ethnographies conducted within the context ofhow people use or are impacted by technologies and C) HCI research probing the technological boundaries demonstrating new potentialinteraction capabilities and new experiences made possible by technological advances

Applied Research Applied research provides complete answers to practical problems In HCI applied research results in practical goalsolution-orientedresearch It is achieved by HCI researchers posing questions through the use of an embodiment of knowledge study of interface efficacyand how people relate to it observation of tensions between people and technologies safety and ethics in realistic scenarios

Design Practice HCI practitioners using HCI and design knowledge to create something new in the acircAIJmade worldacircAI Involves the deep consideration ofspecific users related stakeholders technical and market requirements Integration of art science and engineering to make aestheticallyfunctional interfaces

Gaps Barriers Main translationsTBA Information over-

load skillsetTranslation of basic science to designs andinteractions through an exchange between Applied and Basic researchLearning and iteration of theories through scientific research processes and communication Translations of basicresearch findings into designs to be tested in Applied research often occurs through a collaboration between basicand applied researchers as well as HCI practitioners

TBD Understandingaccess

Translation of basic science into content that can be consumed more broadly For example pop psychology books orevents where researchers explain their findings to practitioners

TAD Applicability Translation and synthesis of knowledge into usable resources considering specific application domains Developmentof resources that can lead to new knowledge being adopted in Design practice

Bubble-up Personal interestincentives

Looking at practice to inform research observation of designs and practice breakdowns as opportunities to informfurther investigations HCI practitioners broadcasting learnings from their day-to-day problem-solving work on socialmediapublishing platforms

TAA Information over-load traditions

Dissemination and reuse of HCI research from different traditions

noticed that the single gap narrative does not distinguishBasic researchIn the HCI TS model we used the T-terminology to de-

scribe different gaps between steps to facilitate more pre-cise discussion about specific gaps (following Drolet andLorenzirsquos model [28]) Drawing from our interviews themostprominent gaps in the HCI Translational Science model seemto be between Basic and Applied research (TBA) between Ap-plied research and Design practice (TAD) and between Basicresearch and Design practice (TBD) We also heard about howBubble-up happens (supporting [32]) and finally about gapswithin applied research in HCI Table 2 shows descriptionsof the abovementioned steps and gaps with their uniqueknowledge exchanges translations and barriersAs can be seen on Figure 1 and Table 2 the model does

not designate where academic or industry researchers residein the continuum Basic and Applied research can occur bothin academic or industry settings and the same is true forthe design of interactive systems (eg university startupincubators) Additionally Applied research can be tightlylinked to basic research if conducted in an academic settingbut in the industry it gravitates towards design practice

TAD mdash Applied Research to Design PracticeThe gap between Applied research and Design practice iswhat people most often refer to as the research-practice gapParticipants said that basic research findings rarely influencethe work of design practitioners which is one of the reasonswhy translating basic into applied research is important

Participants described known barriers to the use of re-search findings by practitioners such as understanding is-sues due to the terminology in research contributions anddifficulty in accessing them However they highlighted thatapplicability might be the most significant barrier keepingapplied research from influencing design practiceP7 and P13 experienced in both academia and industry

said that expecting HCI practitioners to read papers is unre-alistic A practitioner noted designers will not mdash and cannotmdash spend the time to read papers ldquoI donrsquot have four hours toread one paper I just donrsquot I could get a lot done in those fourhours Frankly I do have 10 minutes to skim through a Mediumarticle Irsquove never understood why in academia they donrsquot setup an incentivization structure for academics to almost likepromote their research to industryrdquo (P35)

When researchers and designers meet and talk in personthe terminology issue is lessened as they can speak directlyto each other and ask for clarification (P11 P13 P14 P30P39 P40) However even at these rare occasions designers

do not get actionable information from researchers (P7 P11P29 P35) ldquoI have been to CHIplay but there is a big barrierbetween industry and academia Very little of what I saw therewas applicable and that event is the only direct contact I havewith academia During the year everyone is locked up in theircompanies or universities and donrsquot speakrdquo (P11 who runs avirtual reality gaming company)

HCI researchers translating between applied research anddesign practice described the applicability barrier as the mostimportant to address to successfully influence the work of de-sign practitioners P14 P30 and P32 all HCI professors saidthat designers do not care about using theories or detailedterminology only wanting to know what can be applied fortheir particular case

Many HCI researchers recognized that the long and diffi-cult applied research process does not result in actionableresources that can be offered to design practice which intheir view is a big problem for HCI (P4 P7 P14 P19 P27P34) An industry researcher who worked in a translationalmedical science project after her doctorate said that HCIldquodoes not have enough research study findings that have beenshown to work in practice or practice-based research Researchfindings are not real world enough to be meaningful and appli-cablerdquo (P16) This concern motivated P19 a long-time CHIcontributor to engage with industry partners tapping intowider populations to generalize her study results and tolearn how to translate her research findings into somethingldquousefulrdquo for practitioners

We then asked participants what has worked in terms oftranslations from applied research to design practice Re-searchers mentioned having success with a ldquomake it happenrdquoapproach They described embedding in organizations asconsultants or doing design work along with practitionersfirst learning what those practitioners care about before of-fering recommendations (P8 P19 P20 P21 P30 P32) andcreating actionable resources for practitioners to use such aseasy to use research and design methods (P7 P14 P19 P32)or open-source projects P8 used an open-source model ldquoitwas an open-source project so they literally put the code up onthe wall and like went through it and used that as the jumpingoff point to invent their startuprdquo and introduced the projectto practitioners in a hackathonP19 told the story of how the idea to create a design

method started ldquothis industry guy asked for resources andthen I asked myself what do I give to this guy Papers andacademic research are too hard for practitioners to get itrdquoHer approach was to study how to transform empirical datafrom published studies into a design method ldquoConnecting topractice was a study in itself a long onerdquo (P19)

Compatibility with existing workflows and demonstratedutility affect adoption of assets and methods by practice sup-porting [57] Scholars mentioned making assets compatible

with current industry practices (P8 P19) and the need forthese assets to prove their usefulness from the outset such asby connecting with metrics that practitioners value ldquoI havelearned one lesson Managers and bosses need to want it Theyneed to be convinced and make that a part of their processThey need to know what the practical outcome of the methodfor them is One manager told me lsquoI will fail if I donrsquot use thismethodrsquo and thatrsquos when I knew I had succeededrdquo (P19)

TBA mdash Basic to Applied ResearchParticipants specifically researchers pointed that apart fromTAD there is also an important gap in the knowledge progres-sion from basic to applied research with its unique barriersFirst translating basic knowledge into research products isa challenging process that happens through collaborationswith parties who possess supportive skills Second thereseems to be a disconnect between HCI and other disciplines

Participants noted that translating theory from the socialsciences ethnography or philosophy into tools for conduct-ing applied research is complicated and often depends onpartnerships as they lack the ldquoskillsetrdquo (P27 P28 P33 P43)P24 a psychology professor said ldquomy main goal is findingreproducible regular patterns in peoplersquos behavior that bothhave real-life potential applicability or consequences and atthe same time can be studied in a fairly controlled way But Iwant my studies to have an impact in real liferdquo

Regarding this challenge participant P27 posed a questionand a proposed path ldquoHow do we facilitate the interaction ofbasic and applied researcher in HCI long term sustainableThis is a big problem We need to do more work on understand-ing how this happens naturally which might help us foster howto facilitate this interactionrdquo P28 also present in the groupinterview agreed saying that he worked in projects withcollaborators from other fields in successful programs dedi-cated to foster basic and applied research exchange fundedby a Translational Medicine associationA few participants described how basic and applied re-

search interact through joint projects Basic researchers maywork with HCI researchers or practitioners to create designsthat can be used in detailed investigations However findingthese collaborations and establishing common-ground canbe challenging and costly (P7 P19 P24 P27 P28)

We hired a computer science PhD student whowas a really good programmer We met with himand he prototyped a version that we had in mindMy colleague and I may say rsquooh letrsquos do thisrsquonot because we know that we have to do it butbecause we happened to think of it but learningfrom the programmer that would be difficult toimplement we would think of something else(P27)

However P27 and P33 emphasized that collaborationswith students or designers are often temporary and expen-sive and that longer-term partnerships would be better InP27rsquos case the student soon graduated and left only the codeChanging and deploying the code required someone to havemdash or gain mdash familiarity with that particular technology stackP27 and his colleague spent a year with the project on a holdbecause of this issue P27 and P33 who are self-described ba-sic researchers even started learning design and engineeringskills to do applied work on their ownSecond as the first point already suggests participants

valued drawing on other domains of knowledge to creatework that is ldquonewrdquo and ldquofreshrdquo (P30 P26) Researchers thinkthat engaging with diverse scholarship transforms them intobetter scholars (P5 P17 P26 P28) Understanding is not abarrier in TBA while it can be difficult to quickly grasp otherresearch traditions there is institutional support throughcitations and a common understanding of papers as an outputto incentivize reading diverse workHowever HCI scholars mentioned barriers to engaging

with other scholarship Scholars do not feel incentivized toparticipate in communities of interest that are peripheral toHCI P28 P29 and P35 said this stifles research innovationMost of the knowledge circulation in HCI happens within theconfines of the CHI community and other ACM conferences

HCI researchers also said their work rarely feeds back intothe original bodies of theory from which they draw on ldquoIdraw on social psychology theories and behavior change andtranslate that into designs So thatrsquos a translation Usuallyyoursquore drawing on many different theories to help inform onecoherent thing Not the opposite I donrsquot feel like I contributeto basic sciencerdquo (P7) Again citation deficit was raised asevidence for a basic-applied research gap (P3 P14) Partici-pants did not mention efforts to facilitate engagement withother disciplines or to promote more mdashand more relevantmdashcitation exchanges

TAA mdash Gap within Applied researchParticipants described another disconnect within HCI inacademia one specifically pertaining to how HCI research isdisseminated and reused in the fieldWhile some participantswho frequently participate and publish at CHI boasted thatit is common to cross-pollinate work with their colleaguesrsquodiverse research (P1 P7 P19 P27 P32) others mentioned acritical barrier related to citations perhaps themain currencyin academia Within sub-communities research papers areoften not cited

I speak directly to those issues We use the samekeywords publish in the same venues It be-comes a personal grudge which is a problemwithin HCI You may not like my work but

it should be cited If you disagree with my ap-proach or findings cite it and explain why (P2)

The reasons given by participants for the citation failurewere information overload and academic traditions Firstparticipants mentioned that it is difficult to keep up with allpublications within an area of studies or across them (P1 P9P15 P17 P32 P43) which is curious as most of the knowl-edge circulation happens within a few well-known venuesRegarding academic traditions some researchers said thatauthors can devalue ethnographies or systems developmentif they do not engage with mdashor understandmdash it which cancause those researchers to not learn from or cite that work(P4 P5 P19 P27) As in TBA participants did not mention ef-forts to promote more and more relevant citation exchangesacross HCI conferences tracks

TBD mdash Basic Research to Design PracticeOur participants initially marked TBD as the rarest transla-tional path Basic researchers do not plan studies or writepapers with the objective of influencing practice and P19mentioned that ldquoHCI is so important because it sits in betweenbasic research and design practice to do the translationrdquo ForP6 P17 P24 and P33 TBD is where the understanding andaccess barriers are more salient However although rareparticipants acknowledged that basic research can influenceHCI practice and provided a few examplesOne approach is to translate basic science into content

to be consumed more broadly For example pop psychol-ogy books [6 29] are often written by practitioners withscientific training sharing scientific findings blended withpersonal industry experience or partnering with scholarsNone of our participants had written or helped to write suchkind of books instead they accomplished other types of out-reach such as participating in radio talks and contributingto news articles about science (P6 P17 P24) Second trans-lating content and disseminating it for a targeted group ofpractitioners in events (P24 P43) and in discussions withpractitioners that are actively searching for new knowledgeto inform product development (P4 P8 P17 P30 P33) Thirdparticipants agreed that training in basic research affordsthe abilities to partner with practitioners through researchmethod skills and deep scientific understanding (P6 P17 P24P28 P33 P40 P43)

Bubble-upBubble-up an important aspect of our model (Figure 2) hasbeen discussed in prior work [32] Our participants whoengaged in Bubble-up work described mostly researcher-ledstrategies to help practice knowledge to influence researchThey also voiced the impression that most of their colleaguesdo not do the same because they are either not personally

interested in design or for lack of incentive For P15 ldquoIt iscrucial for researchers to really engage with practitioners andknow how things are done in industry This is or should be theminimum requirement to contribute to designrdquoIn our interviews practitioners mentioned not engaging

with bubble-up UX researchers and designers are most oftennot incentivized to broadcast findings because of intellec-tual property restrictions (P12 P35) However disseminatingcarefully vetted case studies in platforms such as Mediumhas been more common among practitioners Although ourparticipants did not have experience with such forms of com-munication they mentioned an interest in doing so for per-sonal branding (P29 P39 P41 P42) Interestingly we heardfrom all HCI practitioners that instructors with professionalexperience as designers or engineers effectively made theldquoacademia-industry linkagerdquo (P35) in the classroom bringingknowledge from practice into the education of a new gener-ation of practitioners This view was consistent among HCIprofessors who had industry experience (P10 P11 P13 P15P34 P35 P38)

HCI researchers interested in learning insights from prac-tice put effort into a few strategies that they deem successfulsuch as organizing and attending practitioner-oriented con-ferences (P9 P13 P15 P24) keeping track of students whogo on to leading industry positions (P13 P29 P31) and usingonline design groups ldquoIrsquom part of Facebook and Slack groupsto talk to practitioners Most of the times we talk about newtools or books design trends but sometimes deeper issues andmethods Itrsquos very varied but focused on practice only ItacircĂŹsalmost like a crowdsourced design encyclopedia on currenttopicsrdquo (P15) The Bubble-up goals that our participants men-tioned were to learn what practitioners care about (P9 P13P15 P24) to gain insights about design field trends (P13 P29P31) to improve teaching (P15) and syllabi (P13)

Different Translators and Translational WorkThrough our interviews we uncovered different types oftranslations that address TS barriers the design of exampleproducts people transfer education use of research knowl-edge in industry research and science communication

Synthesizing Information into contextualized products Thistype of translation achieved through the collaborative workof basic researchers applied researchers and designers canentail both translating basic knowledge into research con-cepts for testing and applied research into prototypes andguidelines that can facilitate adoption in design practiceIn the interviews we heard specific instances of transla-

tional work altering the original research insight to create anew product to address more contextualized problems Bothin TBA and in TAD knowledge can get lost in translationbecause of the synthetic nature of design which is reflected

in the accounts of basic and applied researchers who haveengaged in translations (P8 P19 P24) P24 mentioned that aforay into applied research required making many conces-sions and changing their initial ideas because of technicaland design constraints P19 in turn mentioned the need tomake a design method compatible with industry practiceswhich required drastic adaptations

The challenge in translating research into products ormethods is described by Chilana et al through the creationof a new venture born out of HCI research focusing on adop-tion [17] and by the concept of appropriation used by Grayet al [32] As seen in the previous paragraph not only dodesign practitioners partially appropriate knowledge so doresearchers when attempting to bridge gaps

People transfer as knowledge transfer Chilana et alrsquos migra-tion from academia to starting a venture [17] represents asecond path that drives translations the movement of peo-ple People move across gaps and steps such as students orprofessors working in industry temporarily through sabbati-cals collaborative projects or consulting permanently aftergraduating or after a career change People carry knowledgeand skills with them Most participants mentioned that in-dustry researchers for example whether trained in appliedor basic research can help translate the terminology of aca-demic contributions (both in TBD and TAD) We interviewedindustry researchers with academic training who sharedhow they have used academic research to inform their workFirst industry researchers said that most commonly theirknowledge of research methods is employed on a day today basis Second they use scientific findings to inform thecreation of study protocols and initial product design direc-tions In these two latter cases industry researchers benefitfrom their academic experience in reading and conductingscientific studies to quickly evaluate and identify takeawaysin existing bodies of knowledge contextualizing that knowl-edge to the problems that their companies face for the designof products or services (P12 P16 P17 P40 P42)Connected to the points above most participants called

out one particular form of people transfer mdash education mdash asperhaps the most common path for HCI to influence designpractice Academia usually contributes to society with re-search findings but HCI researchers in academia also partici-pate in the training of practitioners Multiple HCI professorsmentioned that teaching HCI theory in classes is challengingand how to do so effectively is an open question ldquoHow dowe effectively train people to be sensitive and make them thinkabout these questions [referring to information privacy andpersuasive design] when they go into the workforcerdquo (P1)

Formal and informal science communication While papersare a primary communication artifact among researcherstime constraints hinder their use by industry researchers To

bridge TAD and TBD gaps generating curated resources is par-ticularly important (P6 P12 P16 P20 P34 P42) ldquoI have timeto read the abstract basically Irsquod love to read more academicpapers but everything is due yesterday in industryrdquo (P12) Par-ticipants valued Nielsen reports (P12 P20 P34) and meetingswith scholars (P21 P40 P43) for helping them learn aboutthe latest relevant readings in a particular areaWriting books blog posts presenting research in events

talks or informal conversations can also help bridge gapsScience communication can be done by writers researchersthemselves or collaboratively Participants saw it as an im-portant pathway to reach and influence scholars practition-ers and public opinion (P1 P6 P7 P22 P23 P25) supportingSmith et al [65] There is also a more informal approach toscience communication where industry researchers use aca-demic research opportunistically in discussions with theirproduct teams or when academic researchers engage incasual conversations with design practitioners about howacademic research can inform product developmentIn translational science communication a major barrier

is balancing scientific precision with writing content thatappeals to specific audiences As in previous studies HCIpractitioners reported difficulty reading research papers (P12P20 P21 P35 P37 P38) For the three science communicatorswe interviewed (P22 P23 P25) the most significant issue inworking with scholars is their difficulty in explaining thingssimply P25 said that academics dislike having their worktranslated ldquobecause they want the original language in thefinal piecerdquo HCI scholars acknowledged this barrier and men-tioned a struggle in balancing simple communication andscientific precision P1 has written blog posts and mentionedthat it is hard to break away from the academic writing styleldquoacademics tend to be very fact-driven and we have a certainstyle of writing and we get this feedback - you have to befriendlier you have to insert fun pieces itrsquos just a differentstyle of writing to make it an engaging piece So even if thedesire is there the ability sometimes can be difficult We aremore precise and nuanced in the way we describe thingsrdquo P7adds that she has hesitated to forego precision because studyresults are not entirely generalizable to all populations

5 DISCUSSIONBy triangulating information from multiple stakeholdersinvolved in the TS process in HCI and models from other ap-plied fields we develop a model for TS in HCI that presentsa more nuanced view with multiple gaps barriers withineach and corresponding translation efforts While this modelof multiple gaps has similarities to models developed inmedicine it also has attributes unique to HCI For exam-ple the HCI model is not linear This is likely because HCI-proposed innovations typically do not face the regulatory

constraints of medical research There is much more flexibil-ity for research insights to directly influence design practicewithout rigorous testing and evaluation

Below we discuss the implications of using our proposedmodel of TS in HCI We then consider how to coordinatetranslational efforts to move scientific HCI research intothe hands and minds of practitioners and the importanceof engaging mdash and studying how to engage mdash with othertranslators in Translational Science

The Value of a Translational Science Model for HCIThe Translational Science process in HCI can be describedas a ldquocontinuumrdquo since various resources and actions areinvolved in a progression of knowledge advancing discover-ies towards design practice There are gaps in this progres-sion and using models can make translational research morelikely to succeed [68] Woolf describes several benefits ofa TS model that distinguishes different steps translationsand disseminations [74] These include repositioning transla-tional barriers into a more specific arrangement helping tounderstand where translation occurs or has stalled Secondmodels of TS can support discussions about allocation ofresources that facilitate Translational Science

First our model can be used to reposition and study trans-lational barriers more precisely Previous work has generateda translational barriers taxonomy [36] resulting from discus-sions between HCI practitioners and researchers [11] thathave been framed by a view of a single gap between researchand practice In our model we specify nuanced barriers andoffer additional insights into how to address them For exam-ple publishing academic papers to broadcast findings withinthe scientific community is still important it is not howevermdash and should not be mdash a functional communication channelbetween research and practice [27]Also HCI scholars doing research on Translational Sci-

ence may describe the aim of their research more clearlywith our terms while helping to refine model constructs Forexample the TS HCI model may be used to describe whereresearch has progressed or stalled (eg use of design appli-cations found in HCI research is delayed in TAD) The modelhighlights explicitly that applicability is the most significantbarrier for HCI research to influence practice in TAD so tofacilitate the translation of work stalled at this stage the HCIcommunity might focus on translations that make it easierfor practitioners to apply theory-driven resources

Second our model can help understand areas where struc-tural support is needed Funding agencies and universitiesare essential pieces of Translational Science in HCI [17 38]as they help drive and support research with broad socialand economic implications [21] Our model can help organi-zations more precisely target steps and barriers where vitaladvances are stalled and catalyze work that can facilitate

translation For example with a more specific TS model thehealth field devised initiatives such as dedicated budgetsresearch centers scientific journals and conferences for dif-ferent Trsquos [18 24 74] While we do not argue that all of theinitiatives listed above are necessary for more robust TS inHCI they should at least be considered Foremost invest-ment in the translation of basic and applied research for usein design practice is vital for capitalizing on investments inproducing new insights An excellent example is a specificproject that emerged in the health domain context a plan toreduce cancer mortality by 2025 consisting of concentratingtranslational research in a few centers that can vet and testscientific findings clusters with the most potential [16]

Multiple Translations and TranslatorsHCI is uniquely positioned and capable of addressing trans-lational barriers as it sits at the intersection between socialsciences computer science and technology [14] FocusingHCIrsquos attention on each gap will help increase both the im-pact of research and the pace of advancement in the fieldThe constellation of parties involved in translating HCI

knowledge is broader than just the sub-communities of HCIresearchers and design practitioners We found that theseparties already cooperate and transition across steps andgaps Based on our findings and results of previous SIGs [36]rather than creating a new profession [50] we believe there isa need to leverage and coordinate the incredibly diverse andcapable parties already somehow involved in TranslationalScience efforts especially those who can mdash and want to mdashdo more translational work (while explaining the value ofdoing so to others who could become translators) Belowwe address efforts that in addition to the findings that wepresented hold promise as effective channels to strengthenTS in HCI and should be developed or further studied

The scientific publication cycle helps bridge TBA Whilethis gap suffers less with understanding barriers major ad-vances could be clustered and communicated to ensure thatknowledge is progressing throughout the TS model To ac-complish this existing infrastructure can be leveraged andexpanded upon such as crowdsourced sites to summarizepublished research [3] and perhaps new ways to aggregateand discover knowledge clusters on Google Scholar and otheracademic work search-engines [35]Our model also highlights an intra-community gap that

needs to be considered TAA gap raises concerns about frag-mentation within the HCI community As the HCI commu-nity grows it is essential to encourage learning about en-gaging with and citing research across application areas andways of knowing

In TBD and TAD translators such as science communica-tors and industry researchers help share curated academicknowledge with design practitioners (similar to what Everett

Rogers described as change agents [57]) HCI practitionersalready use social media sites such as Reddit and Mediumto publish and discuss prominent issues For example theruserexperience reddit community [2] provides a venuefor more than 32000 UX amateurs and professionals to com-municate [40] and 211000 users follow the UX CollectiveMedium publication [4] How can we better leverage theseplatforms for sharing HCI knowledge

One other dissemination pathway is the education of stu-dents and researchers on the complexities of translating sci-entific findings A key question for education is whethercurrent HCI education is sufficient or if the communityshould train new kinds of translators or develop translationresources for existing translators in other fields While Nor-man proposed the training of translational developers [50]our model suggests that different skills and training may beneeded to prepare people who will help bridge each distinctgap We could potentially train many translators and mustconsider this diversity of skills and roles in training for TSIn TAD specifically HCI has the opportunity to translate

research results into resources that practitioners can use HCIscientific findings are usually shared through papers whichintroduce and enforce the applicability barrier as productsof the research process are often not ready for immediate useby practitioners [8 71] and rarely consider the practicalitiesof everyday design practice [32] The literature describesmany different tools that can help communicate researchfindings such as training modules workshops technicalsupport and guides [30 68] In HCI specifically many ve-hicles for HCI knowledge have been proposed but shouldbe more systematically evaluated [7] such as strong con-cepts [33] different instantiations of design patterns [5]scenarios [13] personas [20] conceptual models [49] designconcepts [62] design heuristics [26] research objects [51] de-sign fiction [73] tutorials [25] methods [32] and assets [19]

Must we measure knowledge adoption Describing the pro-cess of adopting and adapting HCI research to practice Chi-lana et al [17] ask whether studying knowledge adoption isbeyond the scope of HCI We believe that there is a need toexplore what it means for an HCI contribution to be adoptedbeyond prototypes [31] mining research insights and trans-lating them to how designers want to be supported intopractical evidence-based resources [50 58] Designing eval-uations of knowledge adoption could bring empirical valueto the design field Our findings support this approach bysuggesting tackling the applicability barrier in TAD to facili-tate the adoption and application of knowledge is necessaryHere there are many obvious strands of work in the in-tersection of TAD with diffusion of innovation studies andknowledge adoption at individual and organizational lev-els [57] as well as many relatable user-centered methods to

Table 3 Additional Translators that can facilitate Translational Science in HCI and should be further investigated

Translators EvidencePolicymakers Mentioned by participants as a growing and effective role for regulating practice (P1 P3 P7 P26) Policymakers have a need for facts

based on the best knowledge currently available [72] Lazar [43] even affirms that for example for accessibility researchers to have anyreal impact outside of the research community they need to understand law and work with policymakers

Professional asso-ciations

UXPA or IxDA for example are organizations that frequently set up local events for the HCI community Our participants believe that astronger link with these associations can help establish interfaces between researchers and practitioners Researchers have partnered withmeetup groups and professional organizations to organize events combining academic and industry talks [19] but practitioners rarely meetwith researchers at these events [50 66] and it is unclear how effective they are at supporting adoption of HCI knowledge

Business stake-holders

Business-related stakeholders such as marketing financing and venture capital There is an understudied pathway in turning academicresearch into commercial products and services a path documented in [17] and mentioned by a few participants in our study (P1 P3 P7P8 P13 P15 P17)

Society end-users Some HCI researchers said that they are not interested in influencing design practice but in working directly with the populations that canbenefit from their work (P5 P8) Ladner points to working with end-users to tackle their problems through an open science approach asan alternative [41] Some participants also mentioned working with the media to influence public opinion This way users would demandchange from companies (P1 P7 P43)

Funding agencies as described in the discussion session top-down stimuli can promote structural change and reorganize infrastructures to facilitate Transla-tional Science More work about mdash and with mdash decision-makers of funding agencies is necessary

uncover stakeholder needs that help align user needs withbusiness needs [17] There is currently little incentive forHCI scholars to invest more time and resources in under-standing adoption [17] If a researcher aims at influencingdesign practice mdash to bridge from TBD or TAD mdash then evalu-ating knowledge adoption [45] may be necessary to knowwhether one has succeeded

Engaging mdash and studying engagement mdash with translators HCImust continue to engage with everyday successes and prob-lems practitioners face in their work An effective feedbackloop in Translational Science helps keep applied fields groundedin and relevant for practice For example asking questionssuch as ldquoIs this privacy recommendation found in a CHI pa-per effective or usedadapted in different ways in practicerdquoResearch with a focus on professional design practices allowsinsight into how practitioners refine and concretize abstractknowledge [32 42] Previous work provides guidance for thein-situ study of professional design practice [31] or usingsocial media traces to learn about design practice in [39]While the abovementioned examples focus on sharing

knowledge with design practitioners they are only one rolein TS in HCI Shneiderman advocates for building the ca-pacity to collaborate and coordinate with many stakehold-ers blending scientists engineers designers and end-usersto produce ldquohigher-impact research converging into a so-lutionrdquo [64] We were impressed by the many original andcreative translations that the HCI community representedthrough our 43 participants and referenced authors hasfound to bridge and study research and practice in HCI Fu-ture work should strive to understand the barriers to andeffective strategies for engaging the range of stakeholders inTS for HCI (See Table 3)

6 CONCLUSIONThe presence of barriers that hamper the progression ofknowledge into design practice is a significant issue within

HCI Therefore it is necessary to understand how knowledgeprogresses or fails to progress from research to practiceTo design the model for Translational Science in HCI wedrew on past work and interview data with researcherspractitioners andmultiple parties who are engaged ndash or notmdashin translating HCI knowledge In our continuum we describemultiple steps and gaps between basic and applied researchand design practice We also identify multiple translators andthe translational work they do This model offers insights onhow to bridge translational gaps and how to work with andtrain translators effectively It also acts as a foundation forfuture research on Translational Science in HCI

ACKNOWLEDGMENTSWe thank our study participants for their time and input intothis research Second we acknowledge reviewers for theirvaluable feedback that greatly improved the quality of thispaper Finally we thank Eclair Junchaya Hy Nguyen RickPaz as well as Sam Kolovson and Alison Kolberg for theirthoughtful feedback and support

REFERENCES[1] 2018 NSF Definitions of Research and Development An Anno-

tated Compilation of Official Sources httpswwwnsfgovstatisticsranddefrd-definitionspdf September 10 2018

[2] nd ruserexperience httpswwwredditcomruserexperienceRetrieved Sep 16 2018

[3] nd Stanford Scholar httpsscholarstanfordedu Retrieved Sep 142018

[4] nd UX Collective httpsuxdesigncc Retrieved Sep 16 2018[5] Christopher Alexander 1977 A pattern language towns buildings

construction Oxford university press[6] Dan Ariely 2008 Predictably Irrational HarperCollins New York NY

USA[7] Jordan Beck and Hamid R Ekbia 2018 The Theory-Practice Gap As

Generative Metaphor In Proceedings of the 2018 CHI Conference onHuman Factors in Computing Systems (CHI rsquo18) ACM New York NYUSA Article 620 11 pages httpsdoiorg10114531735743174194

Table 2 Description of the steps and gaps of the HCI Translational Science model Definitions drawn from our data Shneider-man [64] and the NSFmdashNational Science Foundation [1] Of note similar to their academic counterparts industry researcherscan engage in both Basic and Applied research The model does not designate where academic or industry researchers residein the continuum

Steps DescriptionBasic Research Basic research is performed without thought of practical ends resulting in general knowledge to answer a large number of problems though

it may not give a complete specific answer to any one of them In HCI basic research results in a foundational understanding of peopletechnology and design It is achieved by researchers posing formative questions and investigating a topic through the scientific processoften in specific scenarioscontexts Participants mentioned 3 main types of basic research relative to HCI A) Research from other fieldssuch as anthropology and psychology that is drawn upon to conduct Applied research B) Ethnographies conducted within the context ofhow people use or are impacted by technologies and C) HCI research probing the technological boundaries demonstrating new potentialinteraction capabilities and new experiences made possible by technological advances

Applied Research Applied research provides complete answers to practical problems In HCI applied research results in practical goalsolution-orientedresearch It is achieved by HCI researchers posing questions through the use of an embodiment of knowledge study of interface efficacyand how people relate to it observation of tensions between people and technologies safety and ethics in realistic scenarios

Design Practice HCI practitioners using HCI and design knowledge to create something new in the acircAIJmade worldacircAI Involves the deep consideration ofspecific users related stakeholders technical and market requirements Integration of art science and engineering to make aestheticallyfunctional interfaces

Gaps Barriers Main translationsTBA Information over-

load skillsetTranslation of basic science to designs andinteractions through an exchange between Applied and Basic researchLearning and iteration of theories through scientific research processes and communication Translations of basicresearch findings into designs to be tested in Applied research often occurs through a collaboration between basicand applied researchers as well as HCI practitioners

TBD Understandingaccess

Translation of basic science into content that can be consumed more broadly For example pop psychology books orevents where researchers explain their findings to practitioners

TAD Applicability Translation and synthesis of knowledge into usable resources considering specific application domains Developmentof resources that can lead to new knowledge being adopted in Design practice

Bubble-up Personal interestincentives

Looking at practice to inform research observation of designs and practice breakdowns as opportunities to informfurther investigations HCI practitioners broadcasting learnings from their day-to-day problem-solving work on socialmediapublishing platforms

TAA Information over-load traditions

Dissemination and reuse of HCI research from different traditions

noticed that the single gap narrative does not distinguishBasic researchIn the HCI TS model we used the T-terminology to de-

scribe different gaps between steps to facilitate more pre-cise discussion about specific gaps (following Drolet andLorenzirsquos model [28]) Drawing from our interviews themostprominent gaps in the HCI Translational Science model seemto be between Basic and Applied research (TBA) between Ap-plied research and Design practice (TAD) and between Basicresearch and Design practice (TBD) We also heard about howBubble-up happens (supporting [32]) and finally about gapswithin applied research in HCI Table 2 shows descriptionsof the abovementioned steps and gaps with their uniqueknowledge exchanges translations and barriersAs can be seen on Figure 1 and Table 2 the model does

not designate where academic or industry researchers residein the continuum Basic and Applied research can occur bothin academic or industry settings and the same is true forthe design of interactive systems (eg university startupincubators) Additionally Applied research can be tightlylinked to basic research if conducted in an academic settingbut in the industry it gravitates towards design practice

TAD mdash Applied Research to Design PracticeThe gap between Applied research and Design practice iswhat people most often refer to as the research-practice gapParticipants said that basic research findings rarely influencethe work of design practitioners which is one of the reasonswhy translating basic into applied research is important

Participants described known barriers to the use of re-search findings by practitioners such as understanding is-sues due to the terminology in research contributions anddifficulty in accessing them However they highlighted thatapplicability might be the most significant barrier keepingapplied research from influencing design practiceP7 and P13 experienced in both academia and industry

said that expecting HCI practitioners to read papers is unre-alistic A practitioner noted designers will not mdash and cannotmdash spend the time to read papers ldquoI donrsquot have four hours toread one paper I just donrsquot I could get a lot done in those fourhours Frankly I do have 10 minutes to skim through a Mediumarticle Irsquove never understood why in academia they donrsquot setup an incentivization structure for academics to almost likepromote their research to industryrdquo (P35)

When researchers and designers meet and talk in personthe terminology issue is lessened as they can speak directlyto each other and ask for clarification (P11 P13 P14 P30P39 P40) However even at these rare occasions designers

do not get actionable information from researchers (P7 P11P29 P35) ldquoI have been to CHIplay but there is a big barrierbetween industry and academia Very little of what I saw therewas applicable and that event is the only direct contact I havewith academia During the year everyone is locked up in theircompanies or universities and donrsquot speakrdquo (P11 who runs avirtual reality gaming company)

HCI researchers translating between applied research anddesign practice described the applicability barrier as the mostimportant to address to successfully influence the work of de-sign practitioners P14 P30 and P32 all HCI professors saidthat designers do not care about using theories or detailedterminology only wanting to know what can be applied fortheir particular case

Many HCI researchers recognized that the long and diffi-cult applied research process does not result in actionableresources that can be offered to design practice which intheir view is a big problem for HCI (P4 P7 P14 P19 P27P34) An industry researcher who worked in a translationalmedical science project after her doctorate said that HCIldquodoes not have enough research study findings that have beenshown to work in practice or practice-based research Researchfindings are not real world enough to be meaningful and appli-cablerdquo (P16) This concern motivated P19 a long-time CHIcontributor to engage with industry partners tapping intowider populations to generalize her study results and tolearn how to translate her research findings into somethingldquousefulrdquo for practitioners

We then asked participants what has worked in terms oftranslations from applied research to design practice Re-searchers mentioned having success with a ldquomake it happenrdquoapproach They described embedding in organizations asconsultants or doing design work along with practitionersfirst learning what those practitioners care about before of-fering recommendations (P8 P19 P20 P21 P30 P32) andcreating actionable resources for practitioners to use such aseasy to use research and design methods (P7 P14 P19 P32)or open-source projects P8 used an open-source model ldquoitwas an open-source project so they literally put the code up onthe wall and like went through it and used that as the jumpingoff point to invent their startuprdquo and introduced the projectto practitioners in a hackathonP19 told the story of how the idea to create a design

method started ldquothis industry guy asked for resources andthen I asked myself what do I give to this guy Papers andacademic research are too hard for practitioners to get itrdquoHer approach was to study how to transform empirical datafrom published studies into a design method ldquoConnecting topractice was a study in itself a long onerdquo (P19)

Compatibility with existing workflows and demonstratedutility affect adoption of assets and methods by practice sup-porting [57] Scholars mentioned making assets compatible

with current industry practices (P8 P19) and the need forthese assets to prove their usefulness from the outset such asby connecting with metrics that practitioners value ldquoI havelearned one lesson Managers and bosses need to want it Theyneed to be convinced and make that a part of their processThey need to know what the practical outcome of the methodfor them is One manager told me lsquoI will fail if I donrsquot use thismethodrsquo and thatrsquos when I knew I had succeededrdquo (P19)

TBA mdash Basic to Applied ResearchParticipants specifically researchers pointed that apart fromTAD there is also an important gap in the knowledge progres-sion from basic to applied research with its unique barriersFirst translating basic knowledge into research products isa challenging process that happens through collaborationswith parties who possess supportive skills Second thereseems to be a disconnect between HCI and other disciplines

Participants noted that translating theory from the socialsciences ethnography or philosophy into tools for conduct-ing applied research is complicated and often depends onpartnerships as they lack the ldquoskillsetrdquo (P27 P28 P33 P43)P24 a psychology professor said ldquomy main goal is findingreproducible regular patterns in peoplersquos behavior that bothhave real-life potential applicability or consequences and atthe same time can be studied in a fairly controlled way But Iwant my studies to have an impact in real liferdquo

Regarding this challenge participant P27 posed a questionand a proposed path ldquoHow do we facilitate the interaction ofbasic and applied researcher in HCI long term sustainableThis is a big problem We need to do more work on understand-ing how this happens naturally which might help us foster howto facilitate this interactionrdquo P28 also present in the groupinterview agreed saying that he worked in projects withcollaborators from other fields in successful programs dedi-cated to foster basic and applied research exchange fundedby a Translational Medicine associationA few participants described how basic and applied re-

search interact through joint projects Basic researchers maywork with HCI researchers or practitioners to create designsthat can be used in detailed investigations However findingthese collaborations and establishing common-ground canbe challenging and costly (P7 P19 P24 P27 P28)

We hired a computer science PhD student whowas a really good programmer We met with himand he prototyped a version that we had in mindMy colleague and I may say rsquooh letrsquos do thisrsquonot because we know that we have to do it butbecause we happened to think of it but learningfrom the programmer that would be difficult toimplement we would think of something else(P27)

However P27 and P33 emphasized that collaborationswith students or designers are often temporary and expen-sive and that longer-term partnerships would be better InP27rsquos case the student soon graduated and left only the codeChanging and deploying the code required someone to havemdash or gain mdash familiarity with that particular technology stackP27 and his colleague spent a year with the project on a holdbecause of this issue P27 and P33 who are self-described ba-sic researchers even started learning design and engineeringskills to do applied work on their ownSecond as the first point already suggests participants

valued drawing on other domains of knowledge to creatework that is ldquonewrdquo and ldquofreshrdquo (P30 P26) Researchers thinkthat engaging with diverse scholarship transforms them intobetter scholars (P5 P17 P26 P28) Understanding is not abarrier in TBA while it can be difficult to quickly grasp otherresearch traditions there is institutional support throughcitations and a common understanding of papers as an outputto incentivize reading diverse workHowever HCI scholars mentioned barriers to engaging

with other scholarship Scholars do not feel incentivized toparticipate in communities of interest that are peripheral toHCI P28 P29 and P35 said this stifles research innovationMost of the knowledge circulation in HCI happens within theconfines of the CHI community and other ACM conferences

HCI researchers also said their work rarely feeds back intothe original bodies of theory from which they draw on ldquoIdraw on social psychology theories and behavior change andtranslate that into designs So thatrsquos a translation Usuallyyoursquore drawing on many different theories to help inform onecoherent thing Not the opposite I donrsquot feel like I contributeto basic sciencerdquo (P7) Again citation deficit was raised asevidence for a basic-applied research gap (P3 P14) Partici-pants did not mention efforts to facilitate engagement withother disciplines or to promote more mdashand more relevantmdashcitation exchanges

TAA mdash Gap within Applied researchParticipants described another disconnect within HCI inacademia one specifically pertaining to how HCI research isdisseminated and reused in the fieldWhile some participantswho frequently participate and publish at CHI boasted thatit is common to cross-pollinate work with their colleaguesrsquodiverse research (P1 P7 P19 P27 P32) others mentioned acritical barrier related to citations perhaps themain currencyin academia Within sub-communities research papers areoften not cited

I speak directly to those issues We use the samekeywords publish in the same venues It be-comes a personal grudge which is a problemwithin HCI You may not like my work but

it should be cited If you disagree with my ap-proach or findings cite it and explain why (P2)

The reasons given by participants for the citation failurewere information overload and academic traditions Firstparticipants mentioned that it is difficult to keep up with allpublications within an area of studies or across them (P1 P9P15 P17 P32 P43) which is curious as most of the knowl-edge circulation happens within a few well-known venuesRegarding academic traditions some researchers said thatauthors can devalue ethnographies or systems developmentif they do not engage with mdashor understandmdash it which cancause those researchers to not learn from or cite that work(P4 P5 P19 P27) As in TBA participants did not mention ef-forts to promote more and more relevant citation exchangesacross HCI conferences tracks

TBD mdash Basic Research to Design PracticeOur participants initially marked TBD as the rarest transla-tional path Basic researchers do not plan studies or writepapers with the objective of influencing practice and P19mentioned that ldquoHCI is so important because it sits in betweenbasic research and design practice to do the translationrdquo ForP6 P17 P24 and P33 TBD is where the understanding andaccess barriers are more salient However although rareparticipants acknowledged that basic research can influenceHCI practice and provided a few examplesOne approach is to translate basic science into content

to be consumed more broadly For example pop psychol-ogy books [6 29] are often written by practitioners withscientific training sharing scientific findings blended withpersonal industry experience or partnering with scholarsNone of our participants had written or helped to write suchkind of books instead they accomplished other types of out-reach such as participating in radio talks and contributingto news articles about science (P6 P17 P24) Second trans-lating content and disseminating it for a targeted group ofpractitioners in events (P24 P43) and in discussions withpractitioners that are actively searching for new knowledgeto inform product development (P4 P8 P17 P30 P33) Thirdparticipants agreed that training in basic research affordsthe abilities to partner with practitioners through researchmethod skills and deep scientific understanding (P6 P17 P24P28 P33 P40 P43)

Bubble-upBubble-up an important aspect of our model (Figure 2) hasbeen discussed in prior work [32] Our participants whoengaged in Bubble-up work described mostly researcher-ledstrategies to help practice knowledge to influence researchThey also voiced the impression that most of their colleaguesdo not do the same because they are either not personally

interested in design or for lack of incentive For P15 ldquoIt iscrucial for researchers to really engage with practitioners andknow how things are done in industry This is or should be theminimum requirement to contribute to designrdquoIn our interviews practitioners mentioned not engaging

with bubble-up UX researchers and designers are most oftennot incentivized to broadcast findings because of intellec-tual property restrictions (P12 P35) However disseminatingcarefully vetted case studies in platforms such as Mediumhas been more common among practitioners Although ourparticipants did not have experience with such forms of com-munication they mentioned an interest in doing so for per-sonal branding (P29 P39 P41 P42) Interestingly we heardfrom all HCI practitioners that instructors with professionalexperience as designers or engineers effectively made theldquoacademia-industry linkagerdquo (P35) in the classroom bringingknowledge from practice into the education of a new gener-ation of practitioners This view was consistent among HCIprofessors who had industry experience (P10 P11 P13 P15P34 P35 P38)

HCI researchers interested in learning insights from prac-tice put effort into a few strategies that they deem successfulsuch as organizing and attending practitioner-oriented con-ferences (P9 P13 P15 P24) keeping track of students whogo on to leading industry positions (P13 P29 P31) and usingonline design groups ldquoIrsquom part of Facebook and Slack groupsto talk to practitioners Most of the times we talk about newtools or books design trends but sometimes deeper issues andmethods Itrsquos very varied but focused on practice only ItacircĂŹsalmost like a crowdsourced design encyclopedia on currenttopicsrdquo (P15) The Bubble-up goals that our participants men-tioned were to learn what practitioners care about (P9 P13P15 P24) to gain insights about design field trends (P13 P29P31) to improve teaching (P15) and syllabi (P13)

Different Translators and Translational WorkThrough our interviews we uncovered different types oftranslations that address TS barriers the design of exampleproducts people transfer education use of research knowl-edge in industry research and science communication

Synthesizing Information into contextualized products Thistype of translation achieved through the collaborative workof basic researchers applied researchers and designers canentail both translating basic knowledge into research con-cepts for testing and applied research into prototypes andguidelines that can facilitate adoption in design practiceIn the interviews we heard specific instances of transla-

tional work altering the original research insight to create anew product to address more contextualized problems Bothin TBA and in TAD knowledge can get lost in translationbecause of the synthetic nature of design which is reflected

in the accounts of basic and applied researchers who haveengaged in translations (P8 P19 P24) P24 mentioned that aforay into applied research required making many conces-sions and changing their initial ideas because of technicaland design constraints P19 in turn mentioned the need tomake a design method compatible with industry practiceswhich required drastic adaptations

The challenge in translating research into products ormethods is described by Chilana et al through the creationof a new venture born out of HCI research focusing on adop-tion [17] and by the concept of appropriation used by Grayet al [32] As seen in the previous paragraph not only dodesign practitioners partially appropriate knowledge so doresearchers when attempting to bridge gaps

People transfer as knowledge transfer Chilana et alrsquos migra-tion from academia to starting a venture [17] represents asecond path that drives translations the movement of peo-ple People move across gaps and steps such as students orprofessors working in industry temporarily through sabbati-cals collaborative projects or consulting permanently aftergraduating or after a career change People carry knowledgeand skills with them Most participants mentioned that in-dustry researchers for example whether trained in appliedor basic research can help translate the terminology of aca-demic contributions (both in TBD and TAD) We interviewedindustry researchers with academic training who sharedhow they have used academic research to inform their workFirst industry researchers said that most commonly theirknowledge of research methods is employed on a day today basis Second they use scientific findings to inform thecreation of study protocols and initial product design direc-tions In these two latter cases industry researchers benefitfrom their academic experience in reading and conductingscientific studies to quickly evaluate and identify takeawaysin existing bodies of knowledge contextualizing that knowl-edge to the problems that their companies face for the designof products or services (P12 P16 P17 P40 P42)Connected to the points above most participants called

out one particular form of people transfer mdash education mdash asperhaps the most common path for HCI to influence designpractice Academia usually contributes to society with re-search findings but HCI researchers in academia also partici-pate in the training of practitioners Multiple HCI professorsmentioned that teaching HCI theory in classes is challengingand how to do so effectively is an open question ldquoHow dowe effectively train people to be sensitive and make them thinkabout these questions [referring to information privacy andpersuasive design] when they go into the workforcerdquo (P1)

Formal and informal science communication While papersare a primary communication artifact among researcherstime constraints hinder their use by industry researchers To

bridge TAD and TBD gaps generating curated resources is par-ticularly important (P6 P12 P16 P20 P34 P42) ldquoI have timeto read the abstract basically Irsquod love to read more academicpapers but everything is due yesterday in industryrdquo (P12) Par-ticipants valued Nielsen reports (P12 P20 P34) and meetingswith scholars (P21 P40 P43) for helping them learn aboutthe latest relevant readings in a particular areaWriting books blog posts presenting research in events

talks or informal conversations can also help bridge gapsScience communication can be done by writers researchersthemselves or collaboratively Participants saw it as an im-portant pathway to reach and influence scholars practition-ers and public opinion (P1 P6 P7 P22 P23 P25) supportingSmith et al [65] There is also a more informal approach toscience communication where industry researchers use aca-demic research opportunistically in discussions with theirproduct teams or when academic researchers engage incasual conversations with design practitioners about howacademic research can inform product developmentIn translational science communication a major barrier

is balancing scientific precision with writing content thatappeals to specific audiences As in previous studies HCIpractitioners reported difficulty reading research papers (P12P20 P21 P35 P37 P38) For the three science communicatorswe interviewed (P22 P23 P25) the most significant issue inworking with scholars is their difficulty in explaining thingssimply P25 said that academics dislike having their worktranslated ldquobecause they want the original language in thefinal piecerdquo HCI scholars acknowledged this barrier and men-tioned a struggle in balancing simple communication andscientific precision P1 has written blog posts and mentionedthat it is hard to break away from the academic writing styleldquoacademics tend to be very fact-driven and we have a certainstyle of writing and we get this feedback - you have to befriendlier you have to insert fun pieces itrsquos just a differentstyle of writing to make it an engaging piece So even if thedesire is there the ability sometimes can be difficult We aremore precise and nuanced in the way we describe thingsrdquo P7adds that she has hesitated to forego precision because studyresults are not entirely generalizable to all populations

5 DISCUSSIONBy triangulating information from multiple stakeholdersinvolved in the TS process in HCI and models from other ap-plied fields we develop a model for TS in HCI that presentsa more nuanced view with multiple gaps barriers withineach and corresponding translation efforts While this modelof multiple gaps has similarities to models developed inmedicine it also has attributes unique to HCI For exam-ple the HCI model is not linear This is likely because HCI-proposed innovations typically do not face the regulatory

constraints of medical research There is much more flexibil-ity for research insights to directly influence design practicewithout rigorous testing and evaluation

Below we discuss the implications of using our proposedmodel of TS in HCI We then consider how to coordinatetranslational efforts to move scientific HCI research intothe hands and minds of practitioners and the importanceof engaging mdash and studying how to engage mdash with othertranslators in Translational Science

The Value of a Translational Science Model for HCIThe Translational Science process in HCI can be describedas a ldquocontinuumrdquo since various resources and actions areinvolved in a progression of knowledge advancing discover-ies towards design practice There are gaps in this progres-sion and using models can make translational research morelikely to succeed [68] Woolf describes several benefits ofa TS model that distinguishes different steps translationsand disseminations [74] These include repositioning transla-tional barriers into a more specific arrangement helping tounderstand where translation occurs or has stalled Secondmodels of TS can support discussions about allocation ofresources that facilitate Translational Science

First our model can be used to reposition and study trans-lational barriers more precisely Previous work has generateda translational barriers taxonomy [36] resulting from discus-sions between HCI practitioners and researchers [11] thathave been framed by a view of a single gap between researchand practice In our model we specify nuanced barriers andoffer additional insights into how to address them For exam-ple publishing academic papers to broadcast findings withinthe scientific community is still important it is not howevermdash and should not be mdash a functional communication channelbetween research and practice [27]Also HCI scholars doing research on Translational Sci-

ence may describe the aim of their research more clearlywith our terms while helping to refine model constructs Forexample the TS HCI model may be used to describe whereresearch has progressed or stalled (eg use of design appli-cations found in HCI research is delayed in TAD) The modelhighlights explicitly that applicability is the most significantbarrier for HCI research to influence practice in TAD so tofacilitate the translation of work stalled at this stage the HCIcommunity might focus on translations that make it easierfor practitioners to apply theory-driven resources

Second our model can help understand areas where struc-tural support is needed Funding agencies and universitiesare essential pieces of Translational Science in HCI [17 38]as they help drive and support research with broad socialand economic implications [21] Our model can help organi-zations more precisely target steps and barriers where vitaladvances are stalled and catalyze work that can facilitate

translation For example with a more specific TS model thehealth field devised initiatives such as dedicated budgetsresearch centers scientific journals and conferences for dif-ferent Trsquos [18 24 74] While we do not argue that all of theinitiatives listed above are necessary for more robust TS inHCI they should at least be considered Foremost invest-ment in the translation of basic and applied research for usein design practice is vital for capitalizing on investments inproducing new insights An excellent example is a specificproject that emerged in the health domain context a plan toreduce cancer mortality by 2025 consisting of concentratingtranslational research in a few centers that can vet and testscientific findings clusters with the most potential [16]

Multiple Translations and TranslatorsHCI is uniquely positioned and capable of addressing trans-lational barriers as it sits at the intersection between socialsciences computer science and technology [14] FocusingHCIrsquos attention on each gap will help increase both the im-pact of research and the pace of advancement in the fieldThe constellation of parties involved in translating HCI

knowledge is broader than just the sub-communities of HCIresearchers and design practitioners We found that theseparties already cooperate and transition across steps andgaps Based on our findings and results of previous SIGs [36]rather than creating a new profession [50] we believe there isa need to leverage and coordinate the incredibly diverse andcapable parties already somehow involved in TranslationalScience efforts especially those who can mdash and want to mdashdo more translational work (while explaining the value ofdoing so to others who could become translators) Belowwe address efforts that in addition to the findings that wepresented hold promise as effective channels to strengthenTS in HCI and should be developed or further studied

The scientific publication cycle helps bridge TBA Whilethis gap suffers less with understanding barriers major ad-vances could be clustered and communicated to ensure thatknowledge is progressing throughout the TS model To ac-complish this existing infrastructure can be leveraged andexpanded upon such as crowdsourced sites to summarizepublished research [3] and perhaps new ways to aggregateand discover knowledge clusters on Google Scholar and otheracademic work search-engines [35]Our model also highlights an intra-community gap that

needs to be considered TAA gap raises concerns about frag-mentation within the HCI community As the HCI commu-nity grows it is essential to encourage learning about en-gaging with and citing research across application areas andways of knowing

In TBD and TAD translators such as science communica-tors and industry researchers help share curated academicknowledge with design practitioners (similar to what Everett

Rogers described as change agents [57]) HCI practitionersalready use social media sites such as Reddit and Mediumto publish and discuss prominent issues For example theruserexperience reddit community [2] provides a venuefor more than 32000 UX amateurs and professionals to com-municate [40] and 211000 users follow the UX CollectiveMedium publication [4] How can we better leverage theseplatforms for sharing HCI knowledge

One other dissemination pathway is the education of stu-dents and researchers on the complexities of translating sci-entific findings A key question for education is whethercurrent HCI education is sufficient or if the communityshould train new kinds of translators or develop translationresources for existing translators in other fields While Nor-man proposed the training of translational developers [50]our model suggests that different skills and training may beneeded to prepare people who will help bridge each distinctgap We could potentially train many translators and mustconsider this diversity of skills and roles in training for TSIn TAD specifically HCI has the opportunity to translate

research results into resources that practitioners can use HCIscientific findings are usually shared through papers whichintroduce and enforce the applicability barrier as productsof the research process are often not ready for immediate useby practitioners [8 71] and rarely consider the practicalitiesof everyday design practice [32] The literature describesmany different tools that can help communicate researchfindings such as training modules workshops technicalsupport and guides [30 68] In HCI specifically many ve-hicles for HCI knowledge have been proposed but shouldbe more systematically evaluated [7] such as strong con-cepts [33] different instantiations of design patterns [5]scenarios [13] personas [20] conceptual models [49] designconcepts [62] design heuristics [26] research objects [51] de-sign fiction [73] tutorials [25] methods [32] and assets [19]

Must we measure knowledge adoption Describing the pro-cess of adopting and adapting HCI research to practice Chi-lana et al [17] ask whether studying knowledge adoption isbeyond the scope of HCI We believe that there is a need toexplore what it means for an HCI contribution to be adoptedbeyond prototypes [31] mining research insights and trans-lating them to how designers want to be supported intopractical evidence-based resources [50 58] Designing eval-uations of knowledge adoption could bring empirical valueto the design field Our findings support this approach bysuggesting tackling the applicability barrier in TAD to facili-tate the adoption and application of knowledge is necessaryHere there are many obvious strands of work in the in-tersection of TAD with diffusion of innovation studies andknowledge adoption at individual and organizational lev-els [57] as well as many relatable user-centered methods to

Table 3 Additional Translators that can facilitate Translational Science in HCI and should be further investigated

Translators EvidencePolicymakers Mentioned by participants as a growing and effective role for regulating practice (P1 P3 P7 P26) Policymakers have a need for facts

based on the best knowledge currently available [72] Lazar [43] even affirms that for example for accessibility researchers to have anyreal impact outside of the research community they need to understand law and work with policymakers

Professional asso-ciations

UXPA or IxDA for example are organizations that frequently set up local events for the HCI community Our participants believe that astronger link with these associations can help establish interfaces between researchers and practitioners Researchers have partnered withmeetup groups and professional organizations to organize events combining academic and industry talks [19] but practitioners rarely meetwith researchers at these events [50 66] and it is unclear how effective they are at supporting adoption of HCI knowledge

Business stake-holders

Business-related stakeholders such as marketing financing and venture capital There is an understudied pathway in turning academicresearch into commercial products and services a path documented in [17] and mentioned by a few participants in our study (P1 P3 P7P8 P13 P15 P17)

Society end-users Some HCI researchers said that they are not interested in influencing design practice but in working directly with the populations that canbenefit from their work (P5 P8) Ladner points to working with end-users to tackle their problems through an open science approach asan alternative [41] Some participants also mentioned working with the media to influence public opinion This way users would demandchange from companies (P1 P7 P43)

Funding agencies as described in the discussion session top-down stimuli can promote structural change and reorganize infrastructures to facilitate Transla-tional Science More work about mdash and with mdash decision-makers of funding agencies is necessary

uncover stakeholder needs that help align user needs withbusiness needs [17] There is currently little incentive forHCI scholars to invest more time and resources in under-standing adoption [17] If a researcher aims at influencingdesign practice mdash to bridge from TBD or TAD mdash then evalu-ating knowledge adoption [45] may be necessary to knowwhether one has succeeded

Engaging mdash and studying engagement mdash with translators HCImust continue to engage with everyday successes and prob-lems practitioners face in their work An effective feedbackloop in Translational Science helps keep applied fields groundedin and relevant for practice For example asking questionssuch as ldquoIs this privacy recommendation found in a CHI pa-per effective or usedadapted in different ways in practicerdquoResearch with a focus on professional design practices allowsinsight into how practitioners refine and concretize abstractknowledge [32 42] Previous work provides guidance for thein-situ study of professional design practice [31] or usingsocial media traces to learn about design practice in [39]While the abovementioned examples focus on sharing

knowledge with design practitioners they are only one rolein TS in HCI Shneiderman advocates for building the ca-pacity to collaborate and coordinate with many stakehold-ers blending scientists engineers designers and end-usersto produce ldquohigher-impact research converging into a so-lutionrdquo [64] We were impressed by the many original andcreative translations that the HCI community representedthrough our 43 participants and referenced authors hasfound to bridge and study research and practice in HCI Fu-ture work should strive to understand the barriers to andeffective strategies for engaging the range of stakeholders inTS for HCI (See Table 3)

6 CONCLUSIONThe presence of barriers that hamper the progression ofknowledge into design practice is a significant issue within

HCI Therefore it is necessary to understand how knowledgeprogresses or fails to progress from research to practiceTo design the model for Translational Science in HCI wedrew on past work and interview data with researcherspractitioners andmultiple parties who are engaged ndash or notmdashin translating HCI knowledge In our continuum we describemultiple steps and gaps between basic and applied researchand design practice We also identify multiple translators andthe translational work they do This model offers insights onhow to bridge translational gaps and how to work with andtrain translators effectively It also acts as a foundation forfuture research on Translational Science in HCI

ACKNOWLEDGMENTSWe thank our study participants for their time and input intothis research Second we acknowledge reviewers for theirvaluable feedback that greatly improved the quality of thispaper Finally we thank Eclair Junchaya Hy Nguyen RickPaz as well as Sam Kolovson and Alison Kolberg for theirthoughtful feedback and support

REFERENCES[1] 2018 NSF Definitions of Research and Development An Anno-

tated Compilation of Official Sources httpswwwnsfgovstatisticsranddefrd-definitionspdf September 10 2018

[2] nd ruserexperience httpswwwredditcomruserexperienceRetrieved Sep 16 2018

[3] nd Stanford Scholar httpsscholarstanfordedu Retrieved Sep 142018

[4] nd UX Collective httpsuxdesigncc Retrieved Sep 16 2018[5] Christopher Alexander 1977 A pattern language towns buildings

construction Oxford university press[6] Dan Ariely 2008 Predictably Irrational HarperCollins New York NY

USA[7] Jordan Beck and Hamid R Ekbia 2018 The Theory-Practice Gap As

Generative Metaphor In Proceedings of the 2018 CHI Conference onHuman Factors in Computing Systems (CHI rsquo18) ACM New York NYUSA Article 620 11 pages httpsdoiorg10114531735743174194

do not get actionable information from researchers (P7 P11P29 P35) ldquoI have been to CHIplay but there is a big barrierbetween industry and academia Very little of what I saw therewas applicable and that event is the only direct contact I havewith academia During the year everyone is locked up in theircompanies or universities and donrsquot speakrdquo (P11 who runs avirtual reality gaming company)

HCI researchers translating between applied research anddesign practice described the applicability barrier as the mostimportant to address to successfully influence the work of de-sign practitioners P14 P30 and P32 all HCI professors saidthat designers do not care about using theories or detailedterminology only wanting to know what can be applied fortheir particular case

Many HCI researchers recognized that the long and diffi-cult applied research process does not result in actionableresources that can be offered to design practice which intheir view is a big problem for HCI (P4 P7 P14 P19 P27P34) An industry researcher who worked in a translationalmedical science project after her doctorate said that HCIldquodoes not have enough research study findings that have beenshown to work in practice or practice-based research Researchfindings are not real world enough to be meaningful and appli-cablerdquo (P16) This concern motivated P19 a long-time CHIcontributor to engage with industry partners tapping intowider populations to generalize her study results and tolearn how to translate her research findings into somethingldquousefulrdquo for practitioners

We then asked participants what has worked in terms oftranslations from applied research to design practice Re-searchers mentioned having success with a ldquomake it happenrdquoapproach They described embedding in organizations asconsultants or doing design work along with practitionersfirst learning what those practitioners care about before of-fering recommendations (P8 P19 P20 P21 P30 P32) andcreating actionable resources for practitioners to use such aseasy to use research and design methods (P7 P14 P19 P32)or open-source projects P8 used an open-source model ldquoitwas an open-source project so they literally put the code up onthe wall and like went through it and used that as the jumpingoff point to invent their startuprdquo and introduced the projectto practitioners in a hackathonP19 told the story of how the idea to create a design

method started ldquothis industry guy asked for resources andthen I asked myself what do I give to this guy Papers andacademic research are too hard for practitioners to get itrdquoHer approach was to study how to transform empirical datafrom published studies into a design method ldquoConnecting topractice was a study in itself a long onerdquo (P19)

Compatibility with existing workflows and demonstratedutility affect adoption of assets and methods by practice sup-porting [57] Scholars mentioned making assets compatible

with current industry practices (P8 P19) and the need forthese assets to prove their usefulness from the outset such asby connecting with metrics that practitioners value ldquoI havelearned one lesson Managers and bosses need to want it Theyneed to be convinced and make that a part of their processThey need to know what the practical outcome of the methodfor them is One manager told me lsquoI will fail if I donrsquot use thismethodrsquo and thatrsquos when I knew I had succeededrdquo (P19)

TBA mdash Basic to Applied ResearchParticipants specifically researchers pointed that apart fromTAD there is also an important gap in the knowledge progres-sion from basic to applied research with its unique barriersFirst translating basic knowledge into research products isa challenging process that happens through collaborationswith parties who possess supportive skills Second thereseems to be a disconnect between HCI and other disciplines

Participants noted that translating theory from the socialsciences ethnography or philosophy into tools for conduct-ing applied research is complicated and often depends onpartnerships as they lack the ldquoskillsetrdquo (P27 P28 P33 P43)P24 a psychology professor said ldquomy main goal is findingreproducible regular patterns in peoplersquos behavior that bothhave real-life potential applicability or consequences and atthe same time can be studied in a fairly controlled way But Iwant my studies to have an impact in real liferdquo

Regarding this challenge participant P27 posed a questionand a proposed path ldquoHow do we facilitate the interaction ofbasic and applied researcher in HCI long term sustainableThis is a big problem We need to do more work on understand-ing how this happens naturally which might help us foster howto facilitate this interactionrdquo P28 also present in the groupinterview agreed saying that he worked in projects withcollaborators from other fields in successful programs dedi-cated to foster basic and applied research exchange fundedby a Translational Medicine associationA few participants described how basic and applied re-

search interact through joint projects Basic researchers maywork with HCI researchers or practitioners to create designsthat can be used in detailed investigations However findingthese collaborations and establishing common-ground canbe challenging and costly (P7 P19 P24 P27 P28)

We hired a computer science PhD student whowas a really good programmer We met with himand he prototyped a version that we had in mindMy colleague and I may say rsquooh letrsquos do thisrsquonot because we know that we have to do it butbecause we happened to think of it but learningfrom the programmer that would be difficult toimplement we would think of something else(P27)

However P27 and P33 emphasized that collaborationswith students or designers are often temporary and expen-sive and that longer-term partnerships would be better InP27rsquos case the student soon graduated and left only the codeChanging and deploying the code required someone to havemdash or gain mdash familiarity with that particular technology stackP27 and his colleague spent a year with the project on a holdbecause of this issue P27 and P33 who are self-described ba-sic researchers even started learning design and engineeringskills to do applied work on their ownSecond as the first point already suggests participants

valued drawing on other domains of knowledge to creatework that is ldquonewrdquo and ldquofreshrdquo (P30 P26) Researchers thinkthat engaging with diverse scholarship transforms them intobetter scholars (P5 P17 P26 P28) Understanding is not abarrier in TBA while it can be difficult to quickly grasp otherresearch traditions there is institutional support throughcitations and a common understanding of papers as an outputto incentivize reading diverse workHowever HCI scholars mentioned barriers to engaging

with other scholarship Scholars do not feel incentivized toparticipate in communities of interest that are peripheral toHCI P28 P29 and P35 said this stifles research innovationMost of the knowledge circulation in HCI happens within theconfines of the CHI community and other ACM conferences

HCI researchers also said their work rarely feeds back intothe original bodies of theory from which they draw on ldquoIdraw on social psychology theories and behavior change andtranslate that into designs So thatrsquos a translation Usuallyyoursquore drawing on many different theories to help inform onecoherent thing Not the opposite I donrsquot feel like I contributeto basic sciencerdquo (P7) Again citation deficit was raised asevidence for a basic-applied research gap (P3 P14) Partici-pants did not mention efforts to facilitate engagement withother disciplines or to promote more mdashand more relevantmdashcitation exchanges

TAA mdash Gap within Applied researchParticipants described another disconnect within HCI inacademia one specifically pertaining to how HCI research isdisseminated and reused in the fieldWhile some participantswho frequently participate and publish at CHI boasted thatit is common to cross-pollinate work with their colleaguesrsquodiverse research (P1 P7 P19 P27 P32) others mentioned acritical barrier related to citations perhaps themain currencyin academia Within sub-communities research papers areoften not cited

I speak directly to those issues We use the samekeywords publish in the same venues It be-comes a personal grudge which is a problemwithin HCI You may not like my work but

it should be cited If you disagree with my ap-proach or findings cite it and explain why (P2)

The reasons given by participants for the citation failurewere information overload and academic traditions Firstparticipants mentioned that it is difficult to keep up with allpublications within an area of studies or across them (P1 P9P15 P17 P32 P43) which is curious as most of the knowl-edge circulation happens within a few well-known venuesRegarding academic traditions some researchers said thatauthors can devalue ethnographies or systems developmentif they do not engage with mdashor understandmdash it which cancause those researchers to not learn from or cite that work(P4 P5 P19 P27) As in TBA participants did not mention ef-forts to promote more and more relevant citation exchangesacross HCI conferences tracks

TBD mdash Basic Research to Design PracticeOur participants initially marked TBD as the rarest transla-tional path Basic researchers do not plan studies or writepapers with the objective of influencing practice and P19mentioned that ldquoHCI is so important because it sits in betweenbasic research and design practice to do the translationrdquo ForP6 P17 P24 and P33 TBD is where the understanding andaccess barriers are more salient However although rareparticipants acknowledged that basic research can influenceHCI practice and provided a few examplesOne approach is to translate basic science into content

to be consumed more broadly For example pop psychol-ogy books [6 29] are often written by practitioners withscientific training sharing scientific findings blended withpersonal industry experience or partnering with scholarsNone of our participants had written or helped to write suchkind of books instead they accomplished other types of out-reach such as participating in radio talks and contributingto news articles about science (P6 P17 P24) Second trans-lating content and disseminating it for a targeted group ofpractitioners in events (P24 P43) and in discussions withpractitioners that are actively searching for new knowledgeto inform product development (P4 P8 P17 P30 P33) Thirdparticipants agreed that training in basic research affordsthe abilities to partner with practitioners through researchmethod skills and deep scientific understanding (P6 P17 P24P28 P33 P40 P43)

Bubble-upBubble-up an important aspect of our model (Figure 2) hasbeen discussed in prior work [32] Our participants whoengaged in Bubble-up work described mostly researcher-ledstrategies to help practice knowledge to influence researchThey also voiced the impression that most of their colleaguesdo not do the same because they are either not personally

interested in design or for lack of incentive For P15 ldquoIt iscrucial for researchers to really engage with practitioners andknow how things are done in industry This is or should be theminimum requirement to contribute to designrdquoIn our interviews practitioners mentioned not engaging

with bubble-up UX researchers and designers are most oftennot incentivized to broadcast findings because of intellec-tual property restrictions (P12 P35) However disseminatingcarefully vetted case studies in platforms such as Mediumhas been more common among practitioners Although ourparticipants did not have experience with such forms of com-munication they mentioned an interest in doing so for per-sonal branding (P29 P39 P41 P42) Interestingly we heardfrom all HCI practitioners that instructors with professionalexperience as designers or engineers effectively made theldquoacademia-industry linkagerdquo (P35) in the classroom bringingknowledge from practice into the education of a new gener-ation of practitioners This view was consistent among HCIprofessors who had industry experience (P10 P11 P13 P15P34 P35 P38)

HCI researchers interested in learning insights from prac-tice put effort into a few strategies that they deem successfulsuch as organizing and attending practitioner-oriented con-ferences (P9 P13 P15 P24) keeping track of students whogo on to leading industry positions (P13 P29 P31) and usingonline design groups ldquoIrsquom part of Facebook and Slack groupsto talk to practitioners Most of the times we talk about newtools or books design trends but sometimes deeper issues andmethods Itrsquos very varied but focused on practice only ItacircĂŹsalmost like a crowdsourced design encyclopedia on currenttopicsrdquo (P15) The Bubble-up goals that our participants men-tioned were to learn what practitioners care about (P9 P13P15 P24) to gain insights about design field trends (P13 P29P31) to improve teaching (P15) and syllabi (P13)

Different Translators and Translational WorkThrough our interviews we uncovered different types oftranslations that address TS barriers the design of exampleproducts people transfer education use of research knowl-edge in industry research and science communication

Synthesizing Information into contextualized products Thistype of translation achieved through the collaborative workof basic researchers applied researchers and designers canentail both translating basic knowledge into research con-cepts for testing and applied research into prototypes andguidelines that can facilitate adoption in design practiceIn the interviews we heard specific instances of transla-

tional work altering the original research insight to create anew product to address more contextualized problems Bothin TBA and in TAD knowledge can get lost in translationbecause of the synthetic nature of design which is reflected

in the accounts of basic and applied researchers who haveengaged in translations (P8 P19 P24) P24 mentioned that aforay into applied research required making many conces-sions and changing their initial ideas because of technicaland design constraints P19 in turn mentioned the need tomake a design method compatible with industry practiceswhich required drastic adaptations

The challenge in translating research into products ormethods is described by Chilana et al through the creationof a new venture born out of HCI research focusing on adop-tion [17] and by the concept of appropriation used by Grayet al [32] As seen in the previous paragraph not only dodesign practitioners partially appropriate knowledge so doresearchers when attempting to bridge gaps

People transfer as knowledge transfer Chilana et alrsquos migra-tion from academia to starting a venture [17] represents asecond path that drives translations the movement of peo-ple People move across gaps and steps such as students orprofessors working in industry temporarily through sabbati-cals collaborative projects or consulting permanently aftergraduating or after a career change People carry knowledgeand skills with them Most participants mentioned that in-dustry researchers for example whether trained in appliedor basic research can help translate the terminology of aca-demic contributions (both in TBD and TAD) We interviewedindustry researchers with academic training who sharedhow they have used academic research to inform their workFirst industry researchers said that most commonly theirknowledge of research methods is employed on a day today basis Second they use scientific findings to inform thecreation of study protocols and initial product design direc-tions In these two latter cases industry researchers benefitfrom their academic experience in reading and conductingscientific studies to quickly evaluate and identify takeawaysin existing bodies of knowledge contextualizing that knowl-edge to the problems that their companies face for the designof products or services (P12 P16 P17 P40 P42)Connected to the points above most participants called

out one particular form of people transfer mdash education mdash asperhaps the most common path for HCI to influence designpractice Academia usually contributes to society with re-search findings but HCI researchers in academia also partici-pate in the training of practitioners Multiple HCI professorsmentioned that teaching HCI theory in classes is challengingand how to do so effectively is an open question ldquoHow dowe effectively train people to be sensitive and make them thinkabout these questions [referring to information privacy andpersuasive design] when they go into the workforcerdquo (P1)

Formal and informal science communication While papersare a primary communication artifact among researcherstime constraints hinder their use by industry researchers To

bridge TAD and TBD gaps generating curated resources is par-ticularly important (P6 P12 P16 P20 P34 P42) ldquoI have timeto read the abstract basically Irsquod love to read more academicpapers but everything is due yesterday in industryrdquo (P12) Par-ticipants valued Nielsen reports (P12 P20 P34) and meetingswith scholars (P21 P40 P43) for helping them learn aboutthe latest relevant readings in a particular areaWriting books blog posts presenting research in events

talks or informal conversations can also help bridge gapsScience communication can be done by writers researchersthemselves or collaboratively Participants saw it as an im-portant pathway to reach and influence scholars practition-ers and public opinion (P1 P6 P7 P22 P23 P25) supportingSmith et al [65] There is also a more informal approach toscience communication where industry researchers use aca-demic research opportunistically in discussions with theirproduct teams or when academic researchers engage incasual conversations with design practitioners about howacademic research can inform product developmentIn translational science communication a major barrier

is balancing scientific precision with writing content thatappeals to specific audiences As in previous studies HCIpractitioners reported difficulty reading research papers (P12P20 P21 P35 P37 P38) For the three science communicatorswe interviewed (P22 P23 P25) the most significant issue inworking with scholars is their difficulty in explaining thingssimply P25 said that academics dislike having their worktranslated ldquobecause they want the original language in thefinal piecerdquo HCI scholars acknowledged this barrier and men-tioned a struggle in balancing simple communication andscientific precision P1 has written blog posts and mentionedthat it is hard to break away from the academic writing styleldquoacademics tend to be very fact-driven and we have a certainstyle of writing and we get this feedback - you have to befriendlier you have to insert fun pieces itrsquos just a differentstyle of writing to make it an engaging piece So even if thedesire is there the ability sometimes can be difficult We aremore precise and nuanced in the way we describe thingsrdquo P7adds that she has hesitated to forego precision because studyresults are not entirely generalizable to all populations

5 DISCUSSIONBy triangulating information from multiple stakeholdersinvolved in the TS process in HCI and models from other ap-plied fields we develop a model for TS in HCI that presentsa more nuanced view with multiple gaps barriers withineach and corresponding translation efforts While this modelof multiple gaps has similarities to models developed inmedicine it also has attributes unique to HCI For exam-ple the HCI model is not linear This is likely because HCI-proposed innovations typically do not face the regulatory

constraints of medical research There is much more flexibil-ity for research insights to directly influence design practicewithout rigorous testing and evaluation

Below we discuss the implications of using our proposedmodel of TS in HCI We then consider how to coordinatetranslational efforts to move scientific HCI research intothe hands and minds of practitioners and the importanceof engaging mdash and studying how to engage mdash with othertranslators in Translational Science

The Value of a Translational Science Model for HCIThe Translational Science process in HCI can be describedas a ldquocontinuumrdquo since various resources and actions areinvolved in a progression of knowledge advancing discover-ies towards design practice There are gaps in this progres-sion and using models can make translational research morelikely to succeed [68] Woolf describes several benefits ofa TS model that distinguishes different steps translationsand disseminations [74] These include repositioning transla-tional barriers into a more specific arrangement helping tounderstand where translation occurs or has stalled Secondmodels of TS can support discussions about allocation ofresources that facilitate Translational Science

First our model can be used to reposition and study trans-lational barriers more precisely Previous work has generateda translational barriers taxonomy [36] resulting from discus-sions between HCI practitioners and researchers [11] thathave been framed by a view of a single gap between researchand practice In our model we specify nuanced barriers andoffer additional insights into how to address them For exam-ple publishing academic papers to broadcast findings withinthe scientific community is still important it is not howevermdash and should not be mdash a functional communication channelbetween research and practice [27]Also HCI scholars doing research on Translational Sci-

ence may describe the aim of their research more clearlywith our terms while helping to refine model constructs Forexample the TS HCI model may be used to describe whereresearch has progressed or stalled (eg use of design appli-cations found in HCI research is delayed in TAD) The modelhighlights explicitly that applicability is the most significantbarrier for HCI research to influence practice in TAD so tofacilitate the translation of work stalled at this stage the HCIcommunity might focus on translations that make it easierfor practitioners to apply theory-driven resources

Second our model can help understand areas where struc-tural support is needed Funding agencies and universitiesare essential pieces of Translational Science in HCI [17 38]as they help drive and support research with broad socialand economic implications [21] Our model can help organi-zations more precisely target steps and barriers where vitaladvances are stalled and catalyze work that can facilitate

translation For example with a more specific TS model thehealth field devised initiatives such as dedicated budgetsresearch centers scientific journals and conferences for dif-ferent Trsquos [18 24 74] While we do not argue that all of theinitiatives listed above are necessary for more robust TS inHCI they should at least be considered Foremost invest-ment in the translation of basic and applied research for usein design practice is vital for capitalizing on investments inproducing new insights An excellent example is a specificproject that emerged in the health domain context a plan toreduce cancer mortality by 2025 consisting of concentratingtranslational research in a few centers that can vet and testscientific findings clusters with the most potential [16]

Multiple Translations and TranslatorsHCI is uniquely positioned and capable of addressing trans-lational barriers as it sits at the intersection between socialsciences computer science and technology [14] FocusingHCIrsquos attention on each gap will help increase both the im-pact of research and the pace of advancement in the fieldThe constellation of parties involved in translating HCI

knowledge is broader than just the sub-communities of HCIresearchers and design practitioners We found that theseparties already cooperate and transition across steps andgaps Based on our findings and results of previous SIGs [36]rather than creating a new profession [50] we believe there isa need to leverage and coordinate the incredibly diverse andcapable parties already somehow involved in TranslationalScience efforts especially those who can mdash and want to mdashdo more translational work (while explaining the value ofdoing so to others who could become translators) Belowwe address efforts that in addition to the findings that wepresented hold promise as effective channels to strengthenTS in HCI and should be developed or further studied

The scientific publication cycle helps bridge TBA Whilethis gap suffers less with understanding barriers major ad-vances could be clustered and communicated to ensure thatknowledge is progressing throughout the TS model To ac-complish this existing infrastructure can be leveraged andexpanded upon such as crowdsourced sites to summarizepublished research [3] and perhaps new ways to aggregateand discover knowledge clusters on Google Scholar and otheracademic work search-engines [35]Our model also highlights an intra-community gap that

needs to be considered TAA gap raises concerns about frag-mentation within the HCI community As the HCI commu-nity grows it is essential to encourage learning about en-gaging with and citing research across application areas andways of knowing

In TBD and TAD translators such as science communica-tors and industry researchers help share curated academicknowledge with design practitioners (similar to what Everett

Rogers described as change agents [57]) HCI practitionersalready use social media sites such as Reddit and Mediumto publish and discuss prominent issues For example theruserexperience reddit community [2] provides a venuefor more than 32000 UX amateurs and professionals to com-municate [40] and 211000 users follow the UX CollectiveMedium publication [4] How can we better leverage theseplatforms for sharing HCI knowledge

One other dissemination pathway is the education of stu-dents and researchers on the complexities of translating sci-entific findings A key question for education is whethercurrent HCI education is sufficient or if the communityshould train new kinds of translators or develop translationresources for existing translators in other fields While Nor-man proposed the training of translational developers [50]our model suggests that different skills and training may beneeded to prepare people who will help bridge each distinctgap We could potentially train many translators and mustconsider this diversity of skills and roles in training for TSIn TAD specifically HCI has the opportunity to translate

research results into resources that practitioners can use HCIscientific findings are usually shared through papers whichintroduce and enforce the applicability barrier as productsof the research process are often not ready for immediate useby practitioners [8 71] and rarely consider the practicalitiesof everyday design practice [32] The literature describesmany different tools that can help communicate researchfindings such as training modules workshops technicalsupport and guides [30 68] In HCI specifically many ve-hicles for HCI knowledge have been proposed but shouldbe more systematically evaluated [7] such as strong con-cepts [33] different instantiations of design patterns [5]scenarios [13] personas [20] conceptual models [49] designconcepts [62] design heuristics [26] research objects [51] de-sign fiction [73] tutorials [25] methods [32] and assets [19]

Must we measure knowledge adoption Describing the pro-cess of adopting and adapting HCI research to practice Chi-lana et al [17] ask whether studying knowledge adoption isbeyond the scope of HCI We believe that there is a need toexplore what it means for an HCI contribution to be adoptedbeyond prototypes [31] mining research insights and trans-lating them to how designers want to be supported intopractical evidence-based resources [50 58] Designing eval-uations of knowledge adoption could bring empirical valueto the design field Our findings support this approach bysuggesting tackling the applicability barrier in TAD to facili-tate the adoption and application of knowledge is necessaryHere there are many obvious strands of work in the in-tersection of TAD with diffusion of innovation studies andknowledge adoption at individual and organizational lev-els [57] as well as many relatable user-centered methods to

Table 3 Additional Translators that can facilitate Translational Science in HCI and should be further investigated

Translators EvidencePolicymakers Mentioned by participants as a growing and effective role for regulating practice (P1 P3 P7 P26) Policymakers have a need for facts

based on the best knowledge currently available [72] Lazar [43] even affirms that for example for accessibility researchers to have anyreal impact outside of the research community they need to understand law and work with policymakers

Professional asso-ciations

UXPA or IxDA for example are organizations that frequently set up local events for the HCI community Our participants believe that astronger link with these associations can help establish interfaces between researchers and practitioners Researchers have partnered withmeetup groups and professional organizations to organize events combining academic and industry talks [19] but practitioners rarely meetwith researchers at these events [50 66] and it is unclear how effective they are at supporting adoption of HCI knowledge

Business stake-holders

Business-related stakeholders such as marketing financing and venture capital There is an understudied pathway in turning academicresearch into commercial products and services a path documented in [17] and mentioned by a few participants in our study (P1 P3 P7P8 P13 P15 P17)

Society end-users Some HCI researchers said that they are not interested in influencing design practice but in working directly with the populations that canbenefit from their work (P5 P8) Ladner points to working with end-users to tackle their problems through an open science approach asan alternative [41] Some participants also mentioned working with the media to influence public opinion This way users would demandchange from companies (P1 P7 P43)

Funding agencies as described in the discussion session top-down stimuli can promote structural change and reorganize infrastructures to facilitate Transla-tional Science More work about mdash and with mdash decision-makers of funding agencies is necessary

uncover stakeholder needs that help align user needs withbusiness needs [17] There is currently little incentive forHCI scholars to invest more time and resources in under-standing adoption [17] If a researcher aims at influencingdesign practice mdash to bridge from TBD or TAD mdash then evalu-ating knowledge adoption [45] may be necessary to knowwhether one has succeeded

Engaging mdash and studying engagement mdash with translators HCImust continue to engage with everyday successes and prob-lems practitioners face in their work An effective feedbackloop in Translational Science helps keep applied fields groundedin and relevant for practice For example asking questionssuch as ldquoIs this privacy recommendation found in a CHI pa-per effective or usedadapted in different ways in practicerdquoResearch with a focus on professional design practices allowsinsight into how practitioners refine and concretize abstractknowledge [32 42] Previous work provides guidance for thein-situ study of professional design practice [31] or usingsocial media traces to learn about design practice in [39]While the abovementioned examples focus on sharing

knowledge with design practitioners they are only one rolein TS in HCI Shneiderman advocates for building the ca-pacity to collaborate and coordinate with many stakehold-ers blending scientists engineers designers and end-usersto produce ldquohigher-impact research converging into a so-lutionrdquo [64] We were impressed by the many original andcreative translations that the HCI community representedthrough our 43 participants and referenced authors hasfound to bridge and study research and practice in HCI Fu-ture work should strive to understand the barriers to andeffective strategies for engaging the range of stakeholders inTS for HCI (See Table 3)

6 CONCLUSIONThe presence of barriers that hamper the progression ofknowledge into design practice is a significant issue within

HCI Therefore it is necessary to understand how knowledgeprogresses or fails to progress from research to practiceTo design the model for Translational Science in HCI wedrew on past work and interview data with researcherspractitioners andmultiple parties who are engaged ndash or notmdashin translating HCI knowledge In our continuum we describemultiple steps and gaps between basic and applied researchand design practice We also identify multiple translators andthe translational work they do This model offers insights onhow to bridge translational gaps and how to work with andtrain translators effectively It also acts as a foundation forfuture research on Translational Science in HCI

ACKNOWLEDGMENTSWe thank our study participants for their time and input intothis research Second we acknowledge reviewers for theirvaluable feedback that greatly improved the quality of thispaper Finally we thank Eclair Junchaya Hy Nguyen RickPaz as well as Sam Kolovson and Alison Kolberg for theirthoughtful feedback and support

REFERENCES[1] 2018 NSF Definitions of Research and Development An Anno-

tated Compilation of Official Sources httpswwwnsfgovstatisticsranddefrd-definitionspdf September 10 2018

[2] nd ruserexperience httpswwwredditcomruserexperienceRetrieved Sep 16 2018

[3] nd Stanford Scholar httpsscholarstanfordedu Retrieved Sep 142018

[4] nd UX Collective httpsuxdesigncc Retrieved Sep 16 2018[5] Christopher Alexander 1977 A pattern language towns buildings

construction Oxford university press[6] Dan Ariely 2008 Predictably Irrational HarperCollins New York NY

USA[7] Jordan Beck and Hamid R Ekbia 2018 The Theory-Practice Gap As

Generative Metaphor In Proceedings of the 2018 CHI Conference onHuman Factors in Computing Systems (CHI rsquo18) ACM New York NYUSA Article 620 11 pages httpsdoiorg10114531735743174194

However P27 and P33 emphasized that collaborationswith students or designers are often temporary and expen-sive and that longer-term partnerships would be better InP27rsquos case the student soon graduated and left only the codeChanging and deploying the code required someone to havemdash or gain mdash familiarity with that particular technology stackP27 and his colleague spent a year with the project on a holdbecause of this issue P27 and P33 who are self-described ba-sic researchers even started learning design and engineeringskills to do applied work on their ownSecond as the first point already suggests participants

valued drawing on other domains of knowledge to creatework that is ldquonewrdquo and ldquofreshrdquo (P30 P26) Researchers thinkthat engaging with diverse scholarship transforms them intobetter scholars (P5 P17 P26 P28) Understanding is not abarrier in TBA while it can be difficult to quickly grasp otherresearch traditions there is institutional support throughcitations and a common understanding of papers as an outputto incentivize reading diverse workHowever HCI scholars mentioned barriers to engaging

with other scholarship Scholars do not feel incentivized toparticipate in communities of interest that are peripheral toHCI P28 P29 and P35 said this stifles research innovationMost of the knowledge circulation in HCI happens within theconfines of the CHI community and other ACM conferences

HCI researchers also said their work rarely feeds back intothe original bodies of theory from which they draw on ldquoIdraw on social psychology theories and behavior change andtranslate that into designs So thatrsquos a translation Usuallyyoursquore drawing on many different theories to help inform onecoherent thing Not the opposite I donrsquot feel like I contributeto basic sciencerdquo (P7) Again citation deficit was raised asevidence for a basic-applied research gap (P3 P14) Partici-pants did not mention efforts to facilitate engagement withother disciplines or to promote more mdashand more relevantmdashcitation exchanges

TAA mdash Gap within Applied researchParticipants described another disconnect within HCI inacademia one specifically pertaining to how HCI research isdisseminated and reused in the fieldWhile some participantswho frequently participate and publish at CHI boasted thatit is common to cross-pollinate work with their colleaguesrsquodiverse research (P1 P7 P19 P27 P32) others mentioned acritical barrier related to citations perhaps themain currencyin academia Within sub-communities research papers areoften not cited

I speak directly to those issues We use the samekeywords publish in the same venues It be-comes a personal grudge which is a problemwithin HCI You may not like my work but

it should be cited If you disagree with my ap-proach or findings cite it and explain why (P2)

The reasons given by participants for the citation failurewere information overload and academic traditions Firstparticipants mentioned that it is difficult to keep up with allpublications within an area of studies or across them (P1 P9P15 P17 P32 P43) which is curious as most of the knowl-edge circulation happens within a few well-known venuesRegarding academic traditions some researchers said thatauthors can devalue ethnographies or systems developmentif they do not engage with mdashor understandmdash it which cancause those researchers to not learn from or cite that work(P4 P5 P19 P27) As in TBA participants did not mention ef-forts to promote more and more relevant citation exchangesacross HCI conferences tracks

TBD mdash Basic Research to Design PracticeOur participants initially marked TBD as the rarest transla-tional path Basic researchers do not plan studies or writepapers with the objective of influencing practice and P19mentioned that ldquoHCI is so important because it sits in betweenbasic research and design practice to do the translationrdquo ForP6 P17 P24 and P33 TBD is where the understanding andaccess barriers are more salient However although rareparticipants acknowledged that basic research can influenceHCI practice and provided a few examplesOne approach is to translate basic science into content

to be consumed more broadly For example pop psychol-ogy books [6 29] are often written by practitioners withscientific training sharing scientific findings blended withpersonal industry experience or partnering with scholarsNone of our participants had written or helped to write suchkind of books instead they accomplished other types of out-reach such as participating in radio talks and contributingto news articles about science (P6 P17 P24) Second trans-lating content and disseminating it for a targeted group ofpractitioners in events (P24 P43) and in discussions withpractitioners that are actively searching for new knowledgeto inform product development (P4 P8 P17 P30 P33) Thirdparticipants agreed that training in basic research affordsthe abilities to partner with practitioners through researchmethod skills and deep scientific understanding (P6 P17 P24P28 P33 P40 P43)

Bubble-upBubble-up an important aspect of our model (Figure 2) hasbeen discussed in prior work [32] Our participants whoengaged in Bubble-up work described mostly researcher-ledstrategies to help practice knowledge to influence researchThey also voiced the impression that most of their colleaguesdo not do the same because they are either not personally

interested in design or for lack of incentive For P15 ldquoIt iscrucial for researchers to really engage with practitioners andknow how things are done in industry This is or should be theminimum requirement to contribute to designrdquoIn our interviews practitioners mentioned not engaging

with bubble-up UX researchers and designers are most oftennot incentivized to broadcast findings because of intellec-tual property restrictions (P12 P35) However disseminatingcarefully vetted case studies in platforms such as Mediumhas been more common among practitioners Although ourparticipants did not have experience with such forms of com-munication they mentioned an interest in doing so for per-sonal branding (P29 P39 P41 P42) Interestingly we heardfrom all HCI practitioners that instructors with professionalexperience as designers or engineers effectively made theldquoacademia-industry linkagerdquo (P35) in the classroom bringingknowledge from practice into the education of a new gener-ation of practitioners This view was consistent among HCIprofessors who had industry experience (P10 P11 P13 P15P34 P35 P38)

HCI researchers interested in learning insights from prac-tice put effort into a few strategies that they deem successfulsuch as organizing and attending practitioner-oriented con-ferences (P9 P13 P15 P24) keeping track of students whogo on to leading industry positions (P13 P29 P31) and usingonline design groups ldquoIrsquom part of Facebook and Slack groupsto talk to practitioners Most of the times we talk about newtools or books design trends but sometimes deeper issues andmethods Itrsquos very varied but focused on practice only ItacircĂŹsalmost like a crowdsourced design encyclopedia on currenttopicsrdquo (P15) The Bubble-up goals that our participants men-tioned were to learn what practitioners care about (P9 P13P15 P24) to gain insights about design field trends (P13 P29P31) to improve teaching (P15) and syllabi (P13)

Different Translators and Translational WorkThrough our interviews we uncovered different types oftranslations that address TS barriers the design of exampleproducts people transfer education use of research knowl-edge in industry research and science communication

Synthesizing Information into contextualized products Thistype of translation achieved through the collaborative workof basic researchers applied researchers and designers canentail both translating basic knowledge into research con-cepts for testing and applied research into prototypes andguidelines that can facilitate adoption in design practiceIn the interviews we heard specific instances of transla-

tional work altering the original research insight to create anew product to address more contextualized problems Bothin TBA and in TAD knowledge can get lost in translationbecause of the synthetic nature of design which is reflected

in the accounts of basic and applied researchers who haveengaged in translations (P8 P19 P24) P24 mentioned that aforay into applied research required making many conces-sions and changing their initial ideas because of technicaland design constraints P19 in turn mentioned the need tomake a design method compatible with industry practiceswhich required drastic adaptations

The challenge in translating research into products ormethods is described by Chilana et al through the creationof a new venture born out of HCI research focusing on adop-tion [17] and by the concept of appropriation used by Grayet al [32] As seen in the previous paragraph not only dodesign practitioners partially appropriate knowledge so doresearchers when attempting to bridge gaps

People transfer as knowledge transfer Chilana et alrsquos migra-tion from academia to starting a venture [17] represents asecond path that drives translations the movement of peo-ple People move across gaps and steps such as students orprofessors working in industry temporarily through sabbati-cals collaborative projects or consulting permanently aftergraduating or after a career change People carry knowledgeand skills with them Most participants mentioned that in-dustry researchers for example whether trained in appliedor basic research can help translate the terminology of aca-demic contributions (both in TBD and TAD) We interviewedindustry researchers with academic training who sharedhow they have used academic research to inform their workFirst industry researchers said that most commonly theirknowledge of research methods is employed on a day today basis Second they use scientific findings to inform thecreation of study protocols and initial product design direc-tions In these two latter cases industry researchers benefitfrom their academic experience in reading and conductingscientific studies to quickly evaluate and identify takeawaysin existing bodies of knowledge contextualizing that knowl-edge to the problems that their companies face for the designof products or services (P12 P16 P17 P40 P42)Connected to the points above most participants called

out one particular form of people transfer mdash education mdash asperhaps the most common path for HCI to influence designpractice Academia usually contributes to society with re-search findings but HCI researchers in academia also partici-pate in the training of practitioners Multiple HCI professorsmentioned that teaching HCI theory in classes is challengingand how to do so effectively is an open question ldquoHow dowe effectively train people to be sensitive and make them thinkabout these questions [referring to information privacy andpersuasive design] when they go into the workforcerdquo (P1)

Formal and informal science communication While papersare a primary communication artifact among researcherstime constraints hinder their use by industry researchers To

bridge TAD and TBD gaps generating curated resources is par-ticularly important (P6 P12 P16 P20 P34 P42) ldquoI have timeto read the abstract basically Irsquod love to read more academicpapers but everything is due yesterday in industryrdquo (P12) Par-ticipants valued Nielsen reports (P12 P20 P34) and meetingswith scholars (P21 P40 P43) for helping them learn aboutthe latest relevant readings in a particular areaWriting books blog posts presenting research in events

talks or informal conversations can also help bridge gapsScience communication can be done by writers researchersthemselves or collaboratively Participants saw it as an im-portant pathway to reach and influence scholars practition-ers and public opinion (P1 P6 P7 P22 P23 P25) supportingSmith et al [65] There is also a more informal approach toscience communication where industry researchers use aca-demic research opportunistically in discussions with theirproduct teams or when academic researchers engage incasual conversations with design practitioners about howacademic research can inform product developmentIn translational science communication a major barrier

is balancing scientific precision with writing content thatappeals to specific audiences As in previous studies HCIpractitioners reported difficulty reading research papers (P12P20 P21 P35 P37 P38) For the three science communicatorswe interviewed (P22 P23 P25) the most significant issue inworking with scholars is their difficulty in explaining thingssimply P25 said that academics dislike having their worktranslated ldquobecause they want the original language in thefinal piecerdquo HCI scholars acknowledged this barrier and men-tioned a struggle in balancing simple communication andscientific precision P1 has written blog posts and mentionedthat it is hard to break away from the academic writing styleldquoacademics tend to be very fact-driven and we have a certainstyle of writing and we get this feedback - you have to befriendlier you have to insert fun pieces itrsquos just a differentstyle of writing to make it an engaging piece So even if thedesire is there the ability sometimes can be difficult We aremore precise and nuanced in the way we describe thingsrdquo P7adds that she has hesitated to forego precision because studyresults are not entirely generalizable to all populations

5 DISCUSSIONBy triangulating information from multiple stakeholdersinvolved in the TS process in HCI and models from other ap-plied fields we develop a model for TS in HCI that presentsa more nuanced view with multiple gaps barriers withineach and corresponding translation efforts While this modelof multiple gaps has similarities to models developed inmedicine it also has attributes unique to HCI For exam-ple the HCI model is not linear This is likely because HCI-proposed innovations typically do not face the regulatory

constraints of medical research There is much more flexibil-ity for research insights to directly influence design practicewithout rigorous testing and evaluation

Below we discuss the implications of using our proposedmodel of TS in HCI We then consider how to coordinatetranslational efforts to move scientific HCI research intothe hands and minds of practitioners and the importanceof engaging mdash and studying how to engage mdash with othertranslators in Translational Science

The Value of a Translational Science Model for HCIThe Translational Science process in HCI can be describedas a ldquocontinuumrdquo since various resources and actions areinvolved in a progression of knowledge advancing discover-ies towards design practice There are gaps in this progres-sion and using models can make translational research morelikely to succeed [68] Woolf describes several benefits ofa TS model that distinguishes different steps translationsand disseminations [74] These include repositioning transla-tional barriers into a more specific arrangement helping tounderstand where translation occurs or has stalled Secondmodels of TS can support discussions about allocation ofresources that facilitate Translational Science

First our model can be used to reposition and study trans-lational barriers more precisely Previous work has generateda translational barriers taxonomy [36] resulting from discus-sions between HCI practitioners and researchers [11] thathave been framed by a view of a single gap between researchand practice In our model we specify nuanced barriers andoffer additional insights into how to address them For exam-ple publishing academic papers to broadcast findings withinthe scientific community is still important it is not howevermdash and should not be mdash a functional communication channelbetween research and practice [27]Also HCI scholars doing research on Translational Sci-

ence may describe the aim of their research more clearlywith our terms while helping to refine model constructs Forexample the TS HCI model may be used to describe whereresearch has progressed or stalled (eg use of design appli-cations found in HCI research is delayed in TAD) The modelhighlights explicitly that applicability is the most significantbarrier for HCI research to influence practice in TAD so tofacilitate the translation of work stalled at this stage the HCIcommunity might focus on translations that make it easierfor practitioners to apply theory-driven resources

Second our model can help understand areas where struc-tural support is needed Funding agencies and universitiesare essential pieces of Translational Science in HCI [17 38]as they help drive and support research with broad socialand economic implications [21] Our model can help organi-zations more precisely target steps and barriers where vitaladvances are stalled and catalyze work that can facilitate

translation For example with a more specific TS model thehealth field devised initiatives such as dedicated budgetsresearch centers scientific journals and conferences for dif-ferent Trsquos [18 24 74] While we do not argue that all of theinitiatives listed above are necessary for more robust TS inHCI they should at least be considered Foremost invest-ment in the translation of basic and applied research for usein design practice is vital for capitalizing on investments inproducing new insights An excellent example is a specificproject that emerged in the health domain context a plan toreduce cancer mortality by 2025 consisting of concentratingtranslational research in a few centers that can vet and testscientific findings clusters with the most potential [16]

Multiple Translations and TranslatorsHCI is uniquely positioned and capable of addressing trans-lational barriers as it sits at the intersection between socialsciences computer science and technology [14] FocusingHCIrsquos attention on each gap will help increase both the im-pact of research and the pace of advancement in the fieldThe constellation of parties involved in translating HCI

knowledge is broader than just the sub-communities of HCIresearchers and design practitioners We found that theseparties already cooperate and transition across steps andgaps Based on our findings and results of previous SIGs [36]rather than creating a new profession [50] we believe there isa need to leverage and coordinate the incredibly diverse andcapable parties already somehow involved in TranslationalScience efforts especially those who can mdash and want to mdashdo more translational work (while explaining the value ofdoing so to others who could become translators) Belowwe address efforts that in addition to the findings that wepresented hold promise as effective channels to strengthenTS in HCI and should be developed or further studied

The scientific publication cycle helps bridge TBA Whilethis gap suffers less with understanding barriers major ad-vances could be clustered and communicated to ensure thatknowledge is progressing throughout the TS model To ac-complish this existing infrastructure can be leveraged andexpanded upon such as crowdsourced sites to summarizepublished research [3] and perhaps new ways to aggregateand discover knowledge clusters on Google Scholar and otheracademic work search-engines [35]Our model also highlights an intra-community gap that

needs to be considered TAA gap raises concerns about frag-mentation within the HCI community As the HCI commu-nity grows it is essential to encourage learning about en-gaging with and citing research across application areas andways of knowing

In TBD and TAD translators such as science communica-tors and industry researchers help share curated academicknowledge with design practitioners (similar to what Everett

Rogers described as change agents [57]) HCI practitionersalready use social media sites such as Reddit and Mediumto publish and discuss prominent issues For example theruserexperience reddit community [2] provides a venuefor more than 32000 UX amateurs and professionals to com-municate [40] and 211000 users follow the UX CollectiveMedium publication [4] How can we better leverage theseplatforms for sharing HCI knowledge

One other dissemination pathway is the education of stu-dents and researchers on the complexities of translating sci-entific findings A key question for education is whethercurrent HCI education is sufficient or if the communityshould train new kinds of translators or develop translationresources for existing translators in other fields While Nor-man proposed the training of translational developers [50]our model suggests that different skills and training may beneeded to prepare people who will help bridge each distinctgap We could potentially train many translators and mustconsider this diversity of skills and roles in training for TSIn TAD specifically HCI has the opportunity to translate

research results into resources that practitioners can use HCIscientific findings are usually shared through papers whichintroduce and enforce the applicability barrier as productsof the research process are often not ready for immediate useby practitioners [8 71] and rarely consider the practicalitiesof everyday design practice [32] The literature describesmany different tools that can help communicate researchfindings such as training modules workshops technicalsupport and guides [30 68] In HCI specifically many ve-hicles for HCI knowledge have been proposed but shouldbe more systematically evaluated [7] such as strong con-cepts [33] different instantiations of design patterns [5]scenarios [13] personas [20] conceptual models [49] designconcepts [62] design heuristics [26] research objects [51] de-sign fiction [73] tutorials [25] methods [32] and assets [19]

Must we measure knowledge adoption Describing the pro-cess of adopting and adapting HCI research to practice Chi-lana et al [17] ask whether studying knowledge adoption isbeyond the scope of HCI We believe that there is a need toexplore what it means for an HCI contribution to be adoptedbeyond prototypes [31] mining research insights and trans-lating them to how designers want to be supported intopractical evidence-based resources [50 58] Designing eval-uations of knowledge adoption could bring empirical valueto the design field Our findings support this approach bysuggesting tackling the applicability barrier in TAD to facili-tate the adoption and application of knowledge is necessaryHere there are many obvious strands of work in the in-tersection of TAD with diffusion of innovation studies andknowledge adoption at individual and organizational lev-els [57] as well as many relatable user-centered methods to

Table 3 Additional Translators that can facilitate Translational Science in HCI and should be further investigated

Translators EvidencePolicymakers Mentioned by participants as a growing and effective role for regulating practice (P1 P3 P7 P26) Policymakers have a need for facts

based on the best knowledge currently available [72] Lazar [43] even affirms that for example for accessibility researchers to have anyreal impact outside of the research community they need to understand law and work with policymakers

Professional asso-ciations

UXPA or IxDA for example are organizations that frequently set up local events for the HCI community Our participants believe that astronger link with these associations can help establish interfaces between researchers and practitioners Researchers have partnered withmeetup groups and professional organizations to organize events combining academic and industry talks [19] but practitioners rarely meetwith researchers at these events [50 66] and it is unclear how effective they are at supporting adoption of HCI knowledge

Business stake-holders

Business-related stakeholders such as marketing financing and venture capital There is an understudied pathway in turning academicresearch into commercial products and services a path documented in [17] and mentioned by a few participants in our study (P1 P3 P7P8 P13 P15 P17)

Society end-users Some HCI researchers said that they are not interested in influencing design practice but in working directly with the populations that canbenefit from their work (P5 P8) Ladner points to working with end-users to tackle their problems through an open science approach asan alternative [41] Some participants also mentioned working with the media to influence public opinion This way users would demandchange from companies (P1 P7 P43)

Funding agencies as described in the discussion session top-down stimuli can promote structural change and reorganize infrastructures to facilitate Transla-tional Science More work about mdash and with mdash decision-makers of funding agencies is necessary

uncover stakeholder needs that help align user needs withbusiness needs [17] There is currently little incentive forHCI scholars to invest more time and resources in under-standing adoption [17] If a researcher aims at influencingdesign practice mdash to bridge from TBD or TAD mdash then evalu-ating knowledge adoption [45] may be necessary to knowwhether one has succeeded

Engaging mdash and studying engagement mdash with translators HCImust continue to engage with everyday successes and prob-lems practitioners face in their work An effective feedbackloop in Translational Science helps keep applied fields groundedin and relevant for practice For example asking questionssuch as ldquoIs this privacy recommendation found in a CHI pa-per effective or usedadapted in different ways in practicerdquoResearch with a focus on professional design practices allowsinsight into how practitioners refine and concretize abstractknowledge [32 42] Previous work provides guidance for thein-situ study of professional design practice [31] or usingsocial media traces to learn about design practice in [39]While the abovementioned examples focus on sharing

knowledge with design practitioners they are only one rolein TS in HCI Shneiderman advocates for building the ca-pacity to collaborate and coordinate with many stakehold-ers blending scientists engineers designers and end-usersto produce ldquohigher-impact research converging into a so-lutionrdquo [64] We were impressed by the many original andcreative translations that the HCI community representedthrough our 43 participants and referenced authors hasfound to bridge and study research and practice in HCI Fu-ture work should strive to understand the barriers to andeffective strategies for engaging the range of stakeholders inTS for HCI (See Table 3)

6 CONCLUSIONThe presence of barriers that hamper the progression ofknowledge into design practice is a significant issue within

HCI Therefore it is necessary to understand how knowledgeprogresses or fails to progress from research to practiceTo design the model for Translational Science in HCI wedrew on past work and interview data with researcherspractitioners andmultiple parties who are engaged ndash or notmdashin translating HCI knowledge In our continuum we describemultiple steps and gaps between basic and applied researchand design practice We also identify multiple translators andthe translational work they do This model offers insights onhow to bridge translational gaps and how to work with andtrain translators effectively It also acts as a foundation forfuture research on Translational Science in HCI

ACKNOWLEDGMENTSWe thank our study participants for their time and input intothis research Second we acknowledge reviewers for theirvaluable feedback that greatly improved the quality of thispaper Finally we thank Eclair Junchaya Hy Nguyen RickPaz as well as Sam Kolovson and Alison Kolberg for theirthoughtful feedback and support

REFERENCES[1] 2018 NSF Definitions of Research and Development An Anno-

tated Compilation of Official Sources httpswwwnsfgovstatisticsranddefrd-definitionspdf September 10 2018

[2] nd ruserexperience httpswwwredditcomruserexperienceRetrieved Sep 16 2018

[3] nd Stanford Scholar httpsscholarstanfordedu Retrieved Sep 142018

[4] nd UX Collective httpsuxdesigncc Retrieved Sep 16 2018[5] Christopher Alexander 1977 A pattern language towns buildings

construction Oxford university press[6] Dan Ariely 2008 Predictably Irrational HarperCollins New York NY

USA[7] Jordan Beck and Hamid R Ekbia 2018 The Theory-Practice Gap As

Generative Metaphor In Proceedings of the 2018 CHI Conference onHuman Factors in Computing Systems (CHI rsquo18) ACM New York NYUSA Article 620 11 pages httpsdoiorg10114531735743174194

interested in design or for lack of incentive For P15 ldquoIt iscrucial for researchers to really engage with practitioners andknow how things are done in industry This is or should be theminimum requirement to contribute to designrdquoIn our interviews practitioners mentioned not engaging

with bubble-up UX researchers and designers are most oftennot incentivized to broadcast findings because of intellec-tual property restrictions (P12 P35) However disseminatingcarefully vetted case studies in platforms such as Mediumhas been more common among practitioners Although ourparticipants did not have experience with such forms of com-munication they mentioned an interest in doing so for per-sonal branding (P29 P39 P41 P42) Interestingly we heardfrom all HCI practitioners that instructors with professionalexperience as designers or engineers effectively made theldquoacademia-industry linkagerdquo (P35) in the classroom bringingknowledge from practice into the education of a new gener-ation of practitioners This view was consistent among HCIprofessors who had industry experience (P10 P11 P13 P15P34 P35 P38)

HCI researchers interested in learning insights from prac-tice put effort into a few strategies that they deem successfulsuch as organizing and attending practitioner-oriented con-ferences (P9 P13 P15 P24) keeping track of students whogo on to leading industry positions (P13 P29 P31) and usingonline design groups ldquoIrsquom part of Facebook and Slack groupsto talk to practitioners Most of the times we talk about newtools or books design trends but sometimes deeper issues andmethods Itrsquos very varied but focused on practice only ItacircĂŹsalmost like a crowdsourced design encyclopedia on currenttopicsrdquo (P15) The Bubble-up goals that our participants men-tioned were to learn what practitioners care about (P9 P13P15 P24) to gain insights about design field trends (P13 P29P31) to improve teaching (P15) and syllabi (P13)

Different Translators and Translational WorkThrough our interviews we uncovered different types oftranslations that address TS barriers the design of exampleproducts people transfer education use of research knowl-edge in industry research and science communication

Synthesizing Information into contextualized products Thistype of translation achieved through the collaborative workof basic researchers applied researchers and designers canentail both translating basic knowledge into research con-cepts for testing and applied research into prototypes andguidelines that can facilitate adoption in design practiceIn the interviews we heard specific instances of transla-

tional work altering the original research insight to create anew product to address more contextualized problems Bothin TBA and in TAD knowledge can get lost in translationbecause of the synthetic nature of design which is reflected

in the accounts of basic and applied researchers who haveengaged in translations (P8 P19 P24) P24 mentioned that aforay into applied research required making many conces-sions and changing their initial ideas because of technicaland design constraints P19 in turn mentioned the need tomake a design method compatible with industry practiceswhich required drastic adaptations

The challenge in translating research into products ormethods is described by Chilana et al through the creationof a new venture born out of HCI research focusing on adop-tion [17] and by the concept of appropriation used by Grayet al [32] As seen in the previous paragraph not only dodesign practitioners partially appropriate knowledge so doresearchers when attempting to bridge gaps

People transfer as knowledge transfer Chilana et alrsquos migra-tion from academia to starting a venture [17] represents asecond path that drives translations the movement of peo-ple People move across gaps and steps such as students orprofessors working in industry temporarily through sabbati-cals collaborative projects or consulting permanently aftergraduating or after a career change People carry knowledgeand skills with them Most participants mentioned that in-dustry researchers for example whether trained in appliedor basic research can help translate the terminology of aca-demic contributions (both in TBD and TAD) We interviewedindustry researchers with academic training who sharedhow they have used academic research to inform their workFirst industry researchers said that most commonly theirknowledge of research methods is employed on a day today basis Second they use scientific findings to inform thecreation of study protocols and initial product design direc-tions In these two latter cases industry researchers benefitfrom their academic experience in reading and conductingscientific studies to quickly evaluate and identify takeawaysin existing bodies of knowledge contextualizing that knowl-edge to the problems that their companies face for the designof products or services (P12 P16 P17 P40 P42)Connected to the points above most participants called

out one particular form of people transfer mdash education mdash asperhaps the most common path for HCI to influence designpractice Academia usually contributes to society with re-search findings but HCI researchers in academia also partici-pate in the training of practitioners Multiple HCI professorsmentioned that teaching HCI theory in classes is challengingand how to do so effectively is an open question ldquoHow dowe effectively train people to be sensitive and make them thinkabout these questions [referring to information privacy andpersuasive design] when they go into the workforcerdquo (P1)

Formal and informal science communication While papersare a primary communication artifact among researcherstime constraints hinder their use by industry researchers To

bridge TAD and TBD gaps generating curated resources is par-ticularly important (P6 P12 P16 P20 P34 P42) ldquoI have timeto read the abstract basically Irsquod love to read more academicpapers but everything is due yesterday in industryrdquo (P12) Par-ticipants valued Nielsen reports (P12 P20 P34) and meetingswith scholars (P21 P40 P43) for helping them learn aboutthe latest relevant readings in a particular areaWriting books blog posts presenting research in events

talks or informal conversations can also help bridge gapsScience communication can be done by writers researchersthemselves or collaboratively Participants saw it as an im-portant pathway to reach and influence scholars practition-ers and public opinion (P1 P6 P7 P22 P23 P25) supportingSmith et al [65] There is also a more informal approach toscience communication where industry researchers use aca-demic research opportunistically in discussions with theirproduct teams or when academic researchers engage incasual conversations with design practitioners about howacademic research can inform product developmentIn translational science communication a major barrier

is balancing scientific precision with writing content thatappeals to specific audiences As in previous studies HCIpractitioners reported difficulty reading research papers (P12P20 P21 P35 P37 P38) For the three science communicatorswe interviewed (P22 P23 P25) the most significant issue inworking with scholars is their difficulty in explaining thingssimply P25 said that academics dislike having their worktranslated ldquobecause they want the original language in thefinal piecerdquo HCI scholars acknowledged this barrier and men-tioned a struggle in balancing simple communication andscientific precision P1 has written blog posts and mentionedthat it is hard to break away from the academic writing styleldquoacademics tend to be very fact-driven and we have a certainstyle of writing and we get this feedback - you have to befriendlier you have to insert fun pieces itrsquos just a differentstyle of writing to make it an engaging piece So even if thedesire is there the ability sometimes can be difficult We aremore precise and nuanced in the way we describe thingsrdquo P7adds that she has hesitated to forego precision because studyresults are not entirely generalizable to all populations

5 DISCUSSIONBy triangulating information from multiple stakeholdersinvolved in the TS process in HCI and models from other ap-plied fields we develop a model for TS in HCI that presentsa more nuanced view with multiple gaps barriers withineach and corresponding translation efforts While this modelof multiple gaps has similarities to models developed inmedicine it also has attributes unique to HCI For exam-ple the HCI model is not linear This is likely because HCI-proposed innovations typically do not face the regulatory

constraints of medical research There is much more flexibil-ity for research insights to directly influence design practicewithout rigorous testing and evaluation

Below we discuss the implications of using our proposedmodel of TS in HCI We then consider how to coordinatetranslational efforts to move scientific HCI research intothe hands and minds of practitioners and the importanceof engaging mdash and studying how to engage mdash with othertranslators in Translational Science

The Value of a Translational Science Model for HCIThe Translational Science process in HCI can be describedas a ldquocontinuumrdquo since various resources and actions areinvolved in a progression of knowledge advancing discover-ies towards design practice There are gaps in this progres-sion and using models can make translational research morelikely to succeed [68] Woolf describes several benefits ofa TS model that distinguishes different steps translationsand disseminations [74] These include repositioning transla-tional barriers into a more specific arrangement helping tounderstand where translation occurs or has stalled Secondmodels of TS can support discussions about allocation ofresources that facilitate Translational Science

First our model can be used to reposition and study trans-lational barriers more precisely Previous work has generateda translational barriers taxonomy [36] resulting from discus-sions between HCI practitioners and researchers [11] thathave been framed by a view of a single gap between researchand practice In our model we specify nuanced barriers andoffer additional insights into how to address them For exam-ple publishing academic papers to broadcast findings withinthe scientific community is still important it is not howevermdash and should not be mdash a functional communication channelbetween research and practice [27]Also HCI scholars doing research on Translational Sci-

ence may describe the aim of their research more clearlywith our terms while helping to refine model constructs Forexample the TS HCI model may be used to describe whereresearch has progressed or stalled (eg use of design appli-cations found in HCI research is delayed in TAD) The modelhighlights explicitly that applicability is the most significantbarrier for HCI research to influence practice in TAD so tofacilitate the translation of work stalled at this stage the HCIcommunity might focus on translations that make it easierfor practitioners to apply theory-driven resources

Second our model can help understand areas where struc-tural support is needed Funding agencies and universitiesare essential pieces of Translational Science in HCI [17 38]as they help drive and support research with broad socialand economic implications [21] Our model can help organi-zations more precisely target steps and barriers where vitaladvances are stalled and catalyze work that can facilitate

translation For example with a more specific TS model thehealth field devised initiatives such as dedicated budgetsresearch centers scientific journals and conferences for dif-ferent Trsquos [18 24 74] While we do not argue that all of theinitiatives listed above are necessary for more robust TS inHCI they should at least be considered Foremost invest-ment in the translation of basic and applied research for usein design practice is vital for capitalizing on investments inproducing new insights An excellent example is a specificproject that emerged in the health domain context a plan toreduce cancer mortality by 2025 consisting of concentratingtranslational research in a few centers that can vet and testscientific findings clusters with the most potential [16]

Multiple Translations and TranslatorsHCI is uniquely positioned and capable of addressing trans-lational barriers as it sits at the intersection between socialsciences computer science and technology [14] FocusingHCIrsquos attention on each gap will help increase both the im-pact of research and the pace of advancement in the fieldThe constellation of parties involved in translating HCI

knowledge is broader than just the sub-communities of HCIresearchers and design practitioners We found that theseparties already cooperate and transition across steps andgaps Based on our findings and results of previous SIGs [36]rather than creating a new profession [50] we believe there isa need to leverage and coordinate the incredibly diverse andcapable parties already somehow involved in TranslationalScience efforts especially those who can mdash and want to mdashdo more translational work (while explaining the value ofdoing so to others who could become translators) Belowwe address efforts that in addition to the findings that wepresented hold promise as effective channels to strengthenTS in HCI and should be developed or further studied

The scientific publication cycle helps bridge TBA Whilethis gap suffers less with understanding barriers major ad-vances could be clustered and communicated to ensure thatknowledge is progressing throughout the TS model To ac-complish this existing infrastructure can be leveraged andexpanded upon such as crowdsourced sites to summarizepublished research [3] and perhaps new ways to aggregateand discover knowledge clusters on Google Scholar and otheracademic work search-engines [35]Our model also highlights an intra-community gap that

needs to be considered TAA gap raises concerns about frag-mentation within the HCI community As the HCI commu-nity grows it is essential to encourage learning about en-gaging with and citing research across application areas andways of knowing

In TBD and TAD translators such as science communica-tors and industry researchers help share curated academicknowledge with design practitioners (similar to what Everett

Rogers described as change agents [57]) HCI practitionersalready use social media sites such as Reddit and Mediumto publish and discuss prominent issues For example theruserexperience reddit community [2] provides a venuefor more than 32000 UX amateurs and professionals to com-municate [40] and 211000 users follow the UX CollectiveMedium publication [4] How can we better leverage theseplatforms for sharing HCI knowledge

One other dissemination pathway is the education of stu-dents and researchers on the complexities of translating sci-entific findings A key question for education is whethercurrent HCI education is sufficient or if the communityshould train new kinds of translators or develop translationresources for existing translators in other fields While Nor-man proposed the training of translational developers [50]our model suggests that different skills and training may beneeded to prepare people who will help bridge each distinctgap We could potentially train many translators and mustconsider this diversity of skills and roles in training for TSIn TAD specifically HCI has the opportunity to translate

research results into resources that practitioners can use HCIscientific findings are usually shared through papers whichintroduce and enforce the applicability barrier as productsof the research process are often not ready for immediate useby practitioners [8 71] and rarely consider the practicalitiesof everyday design practice [32] The literature describesmany different tools that can help communicate researchfindings such as training modules workshops technicalsupport and guides [30 68] In HCI specifically many ve-hicles for HCI knowledge have been proposed but shouldbe more systematically evaluated [7] such as strong con-cepts [33] different instantiations of design patterns [5]scenarios [13] personas [20] conceptual models [49] designconcepts [62] design heuristics [26] research objects [51] de-sign fiction [73] tutorials [25] methods [32] and assets [19]

Must we measure knowledge adoption Describing the pro-cess of adopting and adapting HCI research to practice Chi-lana et al [17] ask whether studying knowledge adoption isbeyond the scope of HCI We believe that there is a need toexplore what it means for an HCI contribution to be adoptedbeyond prototypes [31] mining research insights and trans-lating them to how designers want to be supported intopractical evidence-based resources [50 58] Designing eval-uations of knowledge adoption could bring empirical valueto the design field Our findings support this approach bysuggesting tackling the applicability barrier in TAD to facili-tate the adoption and application of knowledge is necessaryHere there are many obvious strands of work in the in-tersection of TAD with diffusion of innovation studies andknowledge adoption at individual and organizational lev-els [57] as well as many relatable user-centered methods to

Table 3 Additional Translators that can facilitate Translational Science in HCI and should be further investigated

Translators EvidencePolicymakers Mentioned by participants as a growing and effective role for regulating practice (P1 P3 P7 P26) Policymakers have a need for facts

based on the best knowledge currently available [72] Lazar [43] even affirms that for example for accessibility researchers to have anyreal impact outside of the research community they need to understand law and work with policymakers

Professional asso-ciations

UXPA or IxDA for example are organizations that frequently set up local events for the HCI community Our participants believe that astronger link with these associations can help establish interfaces between researchers and practitioners Researchers have partnered withmeetup groups and professional organizations to organize events combining academic and industry talks [19] but practitioners rarely meetwith researchers at these events [50 66] and it is unclear how effective they are at supporting adoption of HCI knowledge

Business stake-holders

Business-related stakeholders such as marketing financing and venture capital There is an understudied pathway in turning academicresearch into commercial products and services a path documented in [17] and mentioned by a few participants in our study (P1 P3 P7P8 P13 P15 P17)

Society end-users Some HCI researchers said that they are not interested in influencing design practice but in working directly with the populations that canbenefit from their work (P5 P8) Ladner points to working with end-users to tackle their problems through an open science approach asan alternative [41] Some participants also mentioned working with the media to influence public opinion This way users would demandchange from companies (P1 P7 P43)

Funding agencies as described in the discussion session top-down stimuli can promote structural change and reorganize infrastructures to facilitate Transla-tional Science More work about mdash and with mdash decision-makers of funding agencies is necessary

uncover stakeholder needs that help align user needs withbusiness needs [17] There is currently little incentive forHCI scholars to invest more time and resources in under-standing adoption [17] If a researcher aims at influencingdesign practice mdash to bridge from TBD or TAD mdash then evalu-ating knowledge adoption [45] may be necessary to knowwhether one has succeeded

Engaging mdash and studying engagement mdash with translators HCImust continue to engage with everyday successes and prob-lems practitioners face in their work An effective feedbackloop in Translational Science helps keep applied fields groundedin and relevant for practice For example asking questionssuch as ldquoIs this privacy recommendation found in a CHI pa-per effective or usedadapted in different ways in practicerdquoResearch with a focus on professional design practices allowsinsight into how practitioners refine and concretize abstractknowledge [32 42] Previous work provides guidance for thein-situ study of professional design practice [31] or usingsocial media traces to learn about design practice in [39]While the abovementioned examples focus on sharing

knowledge with design practitioners they are only one rolein TS in HCI Shneiderman advocates for building the ca-pacity to collaborate and coordinate with many stakehold-ers blending scientists engineers designers and end-usersto produce ldquohigher-impact research converging into a so-lutionrdquo [64] We were impressed by the many original andcreative translations that the HCI community representedthrough our 43 participants and referenced authors hasfound to bridge and study research and practice in HCI Fu-ture work should strive to understand the barriers to andeffective strategies for engaging the range of stakeholders inTS for HCI (See Table 3)

6 CONCLUSIONThe presence of barriers that hamper the progression ofknowledge into design practice is a significant issue within

HCI Therefore it is necessary to understand how knowledgeprogresses or fails to progress from research to practiceTo design the model for Translational Science in HCI wedrew on past work and interview data with researcherspractitioners andmultiple parties who are engaged ndash or notmdashin translating HCI knowledge In our continuum we describemultiple steps and gaps between basic and applied researchand design practice We also identify multiple translators andthe translational work they do This model offers insights onhow to bridge translational gaps and how to work with andtrain translators effectively It also acts as a foundation forfuture research on Translational Science in HCI

ACKNOWLEDGMENTSWe thank our study participants for their time and input intothis research Second we acknowledge reviewers for theirvaluable feedback that greatly improved the quality of thispaper Finally we thank Eclair Junchaya Hy Nguyen RickPaz as well as Sam Kolovson and Alison Kolberg for theirthoughtful feedback and support

REFERENCES[1] 2018 NSF Definitions of Research and Development An Anno-

tated Compilation of Official Sources httpswwwnsfgovstatisticsranddefrd-definitionspdf September 10 2018

[2] nd ruserexperience httpswwwredditcomruserexperienceRetrieved Sep 16 2018

[3] nd Stanford Scholar httpsscholarstanfordedu Retrieved Sep 142018

[4] nd UX Collective httpsuxdesigncc Retrieved Sep 16 2018[5] Christopher Alexander 1977 A pattern language towns buildings

construction Oxford university press[6] Dan Ariely 2008 Predictably Irrational HarperCollins New York NY

USA[7] Jordan Beck and Hamid R Ekbia 2018 The Theory-Practice Gap As

Generative Metaphor In Proceedings of the 2018 CHI Conference onHuman Factors in Computing Systems (CHI rsquo18) ACM New York NYUSA Article 620 11 pages httpsdoiorg10114531735743174194

bridge TAD and TBD gaps generating curated resources is par-ticularly important (P6 P12 P16 P20 P34 P42) ldquoI have timeto read the abstract basically Irsquod love to read more academicpapers but everything is due yesterday in industryrdquo (P12) Par-ticipants valued Nielsen reports (P12 P20 P34) and meetingswith scholars (P21 P40 P43) for helping them learn aboutthe latest relevant readings in a particular areaWriting books blog posts presenting research in events

talks or informal conversations can also help bridge gapsScience communication can be done by writers researchersthemselves or collaboratively Participants saw it as an im-portant pathway to reach and influence scholars practition-ers and public opinion (P1 P6 P7 P22 P23 P25) supportingSmith et al [65] There is also a more informal approach toscience communication where industry researchers use aca-demic research opportunistically in discussions with theirproduct teams or when academic researchers engage incasual conversations with design practitioners about howacademic research can inform product developmentIn translational science communication a major barrier

is balancing scientific precision with writing content thatappeals to specific audiences As in previous studies HCIpractitioners reported difficulty reading research papers (P12P20 P21 P35 P37 P38) For the three science communicatorswe interviewed (P22 P23 P25) the most significant issue inworking with scholars is their difficulty in explaining thingssimply P25 said that academics dislike having their worktranslated ldquobecause they want the original language in thefinal piecerdquo HCI scholars acknowledged this barrier and men-tioned a struggle in balancing simple communication andscientific precision P1 has written blog posts and mentionedthat it is hard to break away from the academic writing styleldquoacademics tend to be very fact-driven and we have a certainstyle of writing and we get this feedback - you have to befriendlier you have to insert fun pieces itrsquos just a differentstyle of writing to make it an engaging piece So even if thedesire is there the ability sometimes can be difficult We aremore precise and nuanced in the way we describe thingsrdquo P7adds that she has hesitated to forego precision because studyresults are not entirely generalizable to all populations

5 DISCUSSIONBy triangulating information from multiple stakeholdersinvolved in the TS process in HCI and models from other ap-plied fields we develop a model for TS in HCI that presentsa more nuanced view with multiple gaps barriers withineach and corresponding translation efforts While this modelof multiple gaps has similarities to models developed inmedicine it also has attributes unique to HCI For exam-ple the HCI model is not linear This is likely because HCI-proposed innovations typically do not face the regulatory

constraints of medical research There is much more flexibil-ity for research insights to directly influence design practicewithout rigorous testing and evaluation

Below we discuss the implications of using our proposedmodel of TS in HCI We then consider how to coordinatetranslational efforts to move scientific HCI research intothe hands and minds of practitioners and the importanceof engaging mdash and studying how to engage mdash with othertranslators in Translational Science

The Value of a Translational Science Model for HCIThe Translational Science process in HCI can be describedas a ldquocontinuumrdquo since various resources and actions areinvolved in a progression of knowledge advancing discover-ies towards design practice There are gaps in this progres-sion and using models can make translational research morelikely to succeed [68] Woolf describes several benefits ofa TS model that distinguishes different steps translationsand disseminations [74] These include repositioning transla-tional barriers into a more specific arrangement helping tounderstand where translation occurs or has stalled Secondmodels of TS can support discussions about allocation ofresources that facilitate Translational Science

First our model can be used to reposition and study trans-lational barriers more precisely Previous work has generateda translational barriers taxonomy [36] resulting from discus-sions between HCI practitioners and researchers [11] thathave been framed by a view of a single gap between researchand practice In our model we specify nuanced barriers andoffer additional insights into how to address them For exam-ple publishing academic papers to broadcast findings withinthe scientific community is still important it is not howevermdash and should not be mdash a functional communication channelbetween research and practice [27]Also HCI scholars doing research on Translational Sci-

ence may describe the aim of their research more clearlywith our terms while helping to refine model constructs Forexample the TS HCI model may be used to describe whereresearch has progressed or stalled (eg use of design appli-cations found in HCI research is delayed in TAD) The modelhighlights explicitly that applicability is the most significantbarrier for HCI research to influence practice in TAD so tofacilitate the translation of work stalled at this stage the HCIcommunity might focus on translations that make it easierfor practitioners to apply theory-driven resources

Second our model can help understand areas where struc-tural support is needed Funding agencies and universitiesare essential pieces of Translational Science in HCI [17 38]as they help drive and support research with broad socialand economic implications [21] Our model can help organi-zations more precisely target steps and barriers where vitaladvances are stalled and catalyze work that can facilitate

translation For example with a more specific TS model thehealth field devised initiatives such as dedicated budgetsresearch centers scientific journals and conferences for dif-ferent Trsquos [18 24 74] While we do not argue that all of theinitiatives listed above are necessary for more robust TS inHCI they should at least be considered Foremost invest-ment in the translation of basic and applied research for usein design practice is vital for capitalizing on investments inproducing new insights An excellent example is a specificproject that emerged in the health domain context a plan toreduce cancer mortality by 2025 consisting of concentratingtranslational research in a few centers that can vet and testscientific findings clusters with the most potential [16]

Multiple Translations and TranslatorsHCI is uniquely positioned and capable of addressing trans-lational barriers as it sits at the intersection between socialsciences computer science and technology [14] FocusingHCIrsquos attention on each gap will help increase both the im-pact of research and the pace of advancement in the fieldThe constellation of parties involved in translating HCI

knowledge is broader than just the sub-communities of HCIresearchers and design practitioners We found that theseparties already cooperate and transition across steps andgaps Based on our findings and results of previous SIGs [36]rather than creating a new profession [50] we believe there isa need to leverage and coordinate the incredibly diverse andcapable parties already somehow involved in TranslationalScience efforts especially those who can mdash and want to mdashdo more translational work (while explaining the value ofdoing so to others who could become translators) Belowwe address efforts that in addition to the findings that wepresented hold promise as effective channels to strengthenTS in HCI and should be developed or further studied

The scientific publication cycle helps bridge TBA Whilethis gap suffers less with understanding barriers major ad-vances could be clustered and communicated to ensure thatknowledge is progressing throughout the TS model To ac-complish this existing infrastructure can be leveraged andexpanded upon such as crowdsourced sites to summarizepublished research [3] and perhaps new ways to aggregateand discover knowledge clusters on Google Scholar and otheracademic work search-engines [35]Our model also highlights an intra-community gap that

needs to be considered TAA gap raises concerns about frag-mentation within the HCI community As the HCI commu-nity grows it is essential to encourage learning about en-gaging with and citing research across application areas andways of knowing

In TBD and TAD translators such as science communica-tors and industry researchers help share curated academicknowledge with design practitioners (similar to what Everett

Rogers described as change agents [57]) HCI practitionersalready use social media sites such as Reddit and Mediumto publish and discuss prominent issues For example theruserexperience reddit community [2] provides a venuefor more than 32000 UX amateurs and professionals to com-municate [40] and 211000 users follow the UX CollectiveMedium publication [4] How can we better leverage theseplatforms for sharing HCI knowledge

One other dissemination pathway is the education of stu-dents and researchers on the complexities of translating sci-entific findings A key question for education is whethercurrent HCI education is sufficient or if the communityshould train new kinds of translators or develop translationresources for existing translators in other fields While Nor-man proposed the training of translational developers [50]our model suggests that different skills and training may beneeded to prepare people who will help bridge each distinctgap We could potentially train many translators and mustconsider this diversity of skills and roles in training for TSIn TAD specifically HCI has the opportunity to translate

research results into resources that practitioners can use HCIscientific findings are usually shared through papers whichintroduce and enforce the applicability barrier as productsof the research process are often not ready for immediate useby practitioners [8 71] and rarely consider the practicalitiesof everyday design practice [32] The literature describesmany different tools that can help communicate researchfindings such as training modules workshops technicalsupport and guides [30 68] In HCI specifically many ve-hicles for HCI knowledge have been proposed but shouldbe more systematically evaluated [7] such as strong con-cepts [33] different instantiations of design patterns [5]scenarios [13] personas [20] conceptual models [49] designconcepts [62] design heuristics [26] research objects [51] de-sign fiction [73] tutorials [25] methods [32] and assets [19]

Must we measure knowledge adoption Describing the pro-cess of adopting and adapting HCI research to practice Chi-lana et al [17] ask whether studying knowledge adoption isbeyond the scope of HCI We believe that there is a need toexplore what it means for an HCI contribution to be adoptedbeyond prototypes [31] mining research insights and trans-lating them to how designers want to be supported intopractical evidence-based resources [50 58] Designing eval-uations of knowledge adoption could bring empirical valueto the design field Our findings support this approach bysuggesting tackling the applicability barrier in TAD to facili-tate the adoption and application of knowledge is necessaryHere there are many obvious strands of work in the in-tersection of TAD with diffusion of innovation studies andknowledge adoption at individual and organizational lev-els [57] as well as many relatable user-centered methods to

Table 3 Additional Translators that can facilitate Translational Science in HCI and should be further investigated

Translators EvidencePolicymakers Mentioned by participants as a growing and effective role for regulating practice (P1 P3 P7 P26) Policymakers have a need for facts

based on the best knowledge currently available [72] Lazar [43] even affirms that for example for accessibility researchers to have anyreal impact outside of the research community they need to understand law and work with policymakers

Professional asso-ciations

UXPA or IxDA for example are organizations that frequently set up local events for the HCI community Our participants believe that astronger link with these associations can help establish interfaces between researchers and practitioners Researchers have partnered withmeetup groups and professional organizations to organize events combining academic and industry talks [19] but practitioners rarely meetwith researchers at these events [50 66] and it is unclear how effective they are at supporting adoption of HCI knowledge

Business stake-holders

Business-related stakeholders such as marketing financing and venture capital There is an understudied pathway in turning academicresearch into commercial products and services a path documented in [17] and mentioned by a few participants in our study (P1 P3 P7P8 P13 P15 P17)

Society end-users Some HCI researchers said that they are not interested in influencing design practice but in working directly with the populations that canbenefit from their work (P5 P8) Ladner points to working with end-users to tackle their problems through an open science approach asan alternative [41] Some participants also mentioned working with the media to influence public opinion This way users would demandchange from companies (P1 P7 P43)

Funding agencies as described in the discussion session top-down stimuli can promote structural change and reorganize infrastructures to facilitate Transla-tional Science More work about mdash and with mdash decision-makers of funding agencies is necessary

uncover stakeholder needs that help align user needs withbusiness needs [17] There is currently little incentive forHCI scholars to invest more time and resources in under-standing adoption [17] If a researcher aims at influencingdesign practice mdash to bridge from TBD or TAD mdash then evalu-ating knowledge adoption [45] may be necessary to knowwhether one has succeeded

Engaging mdash and studying engagement mdash with translators HCImust continue to engage with everyday successes and prob-lems practitioners face in their work An effective feedbackloop in Translational Science helps keep applied fields groundedin and relevant for practice For example asking questionssuch as ldquoIs this privacy recommendation found in a CHI pa-per effective or usedadapted in different ways in practicerdquoResearch with a focus on professional design practices allowsinsight into how practitioners refine and concretize abstractknowledge [32 42] Previous work provides guidance for thein-situ study of professional design practice [31] or usingsocial media traces to learn about design practice in [39]While the abovementioned examples focus on sharing

knowledge with design practitioners they are only one rolein TS in HCI Shneiderman advocates for building the ca-pacity to collaborate and coordinate with many stakehold-ers blending scientists engineers designers and end-usersto produce ldquohigher-impact research converging into a so-lutionrdquo [64] We were impressed by the many original andcreative translations that the HCI community representedthrough our 43 participants and referenced authors hasfound to bridge and study research and practice in HCI Fu-ture work should strive to understand the barriers to andeffective strategies for engaging the range of stakeholders inTS for HCI (See Table 3)

6 CONCLUSIONThe presence of barriers that hamper the progression ofknowledge into design practice is a significant issue within

HCI Therefore it is necessary to understand how knowledgeprogresses or fails to progress from research to practiceTo design the model for Translational Science in HCI wedrew on past work and interview data with researcherspractitioners andmultiple parties who are engaged ndash or notmdashin translating HCI knowledge In our continuum we describemultiple steps and gaps between basic and applied researchand design practice We also identify multiple translators andthe translational work they do This model offers insights onhow to bridge translational gaps and how to work with andtrain translators effectively It also acts as a foundation forfuture research on Translational Science in HCI

ACKNOWLEDGMENTSWe thank our study participants for their time and input intothis research Second we acknowledge reviewers for theirvaluable feedback that greatly improved the quality of thispaper Finally we thank Eclair Junchaya Hy Nguyen RickPaz as well as Sam Kolovson and Alison Kolberg for theirthoughtful feedback and support

REFERENCES[1] 2018 NSF Definitions of Research and Development An Anno-

tated Compilation of Official Sources httpswwwnsfgovstatisticsranddefrd-definitionspdf September 10 2018

[2] nd ruserexperience httpswwwredditcomruserexperienceRetrieved Sep 16 2018

[3] nd Stanford Scholar httpsscholarstanfordedu Retrieved Sep 142018

[4] nd UX Collective httpsuxdesigncc Retrieved Sep 16 2018[5] Christopher Alexander 1977 A pattern language towns buildings

construction Oxford university press[6] Dan Ariely 2008 Predictably Irrational HarperCollins New York NY

USA[7] Jordan Beck and Hamid R Ekbia 2018 The Theory-Practice Gap As

Generative Metaphor In Proceedings of the 2018 CHI Conference onHuman Factors in Computing Systems (CHI rsquo18) ACM New York NYUSA Article 620 11 pages httpsdoiorg10114531735743174194

translation For example with a more specific TS model thehealth field devised initiatives such as dedicated budgetsresearch centers scientific journals and conferences for dif-ferent Trsquos [18 24 74] While we do not argue that all of theinitiatives listed above are necessary for more robust TS inHCI they should at least be considered Foremost invest-ment in the translation of basic and applied research for usein design practice is vital for capitalizing on investments inproducing new insights An excellent example is a specificproject that emerged in the health domain context a plan toreduce cancer mortality by 2025 consisting of concentratingtranslational research in a few centers that can vet and testscientific findings clusters with the most potential [16]

Multiple Translations and TranslatorsHCI is uniquely positioned and capable of addressing trans-lational barriers as it sits at the intersection between socialsciences computer science and technology [14] FocusingHCIrsquos attention on each gap will help increase both the im-pact of research and the pace of advancement in the fieldThe constellation of parties involved in translating HCI

knowledge is broader than just the sub-communities of HCIresearchers and design practitioners We found that theseparties already cooperate and transition across steps andgaps Based on our findings and results of previous SIGs [36]rather than creating a new profession [50] we believe there isa need to leverage and coordinate the incredibly diverse andcapable parties already somehow involved in TranslationalScience efforts especially those who can mdash and want to mdashdo more translational work (while explaining the value ofdoing so to others who could become translators) Belowwe address efforts that in addition to the findings that wepresented hold promise as effective channels to strengthenTS in HCI and should be developed or further studied

The scientific publication cycle helps bridge TBA Whilethis gap suffers less with understanding barriers major ad-vances could be clustered and communicated to ensure thatknowledge is progressing throughout the TS model To ac-complish this existing infrastructure can be leveraged andexpanded upon such as crowdsourced sites to summarizepublished research [3] and perhaps new ways to aggregateand discover knowledge clusters on Google Scholar and otheracademic work search-engines [35]Our model also highlights an intra-community gap that

needs to be considered TAA gap raises concerns about frag-mentation within the HCI community As the HCI commu-nity grows it is essential to encourage learning about en-gaging with and citing research across application areas andways of knowing

In TBD and TAD translators such as science communica-tors and industry researchers help share curated academicknowledge with design practitioners (similar to what Everett

Rogers described as change agents [57]) HCI practitionersalready use social media sites such as Reddit and Mediumto publish and discuss prominent issues For example theruserexperience reddit community [2] provides a venuefor more than 32000 UX amateurs and professionals to com-municate [40] and 211000 users follow the UX CollectiveMedium publication [4] How can we better leverage theseplatforms for sharing HCI knowledge

One other dissemination pathway is the education of stu-dents and researchers on the complexities of translating sci-entific findings A key question for education is whethercurrent HCI education is sufficient or if the communityshould train new kinds of translators or develop translationresources for existing translators in other fields While Nor-man proposed the training of translational developers [50]our model suggests that different skills and training may beneeded to prepare people who will help bridge each distinctgap We could potentially train many translators and mustconsider this diversity of skills and roles in training for TSIn TAD specifically HCI has the opportunity to translate

research results into resources that practitioners can use HCIscientific findings are usually shared through papers whichintroduce and enforce the applicability barrier as productsof the research process are often not ready for immediate useby practitioners [8 71] and rarely consider the practicalitiesof everyday design practice [32] The literature describesmany different tools that can help communicate researchfindings such as training modules workshops technicalsupport and guides [30 68] In HCI specifically many ve-hicles for HCI knowledge have been proposed but shouldbe more systematically evaluated [7] such as strong con-cepts [33] different instantiations of design patterns [5]scenarios [13] personas [20] conceptual models [49] designconcepts [62] design heuristics [26] research objects [51] de-sign fiction [73] tutorials [25] methods [32] and assets [19]

Must we measure knowledge adoption Describing the pro-cess of adopting and adapting HCI research to practice Chi-lana et al [17] ask whether studying knowledge adoption isbeyond the scope of HCI We believe that there is a need toexplore what it means for an HCI contribution to be adoptedbeyond prototypes [31] mining research insights and trans-lating them to how designers want to be supported intopractical evidence-based resources [50 58] Designing eval-uations of knowledge adoption could bring empirical valueto the design field Our findings support this approach bysuggesting tackling the applicability barrier in TAD to facili-tate the adoption and application of knowledge is necessaryHere there are many obvious strands of work in the in-tersection of TAD with diffusion of innovation studies andknowledge adoption at individual and organizational lev-els [57] as well as many relatable user-centered methods to

Table 3 Additional Translators that can facilitate Translational Science in HCI and should be further investigated

Translators EvidencePolicymakers Mentioned by participants as a growing and effective role for regulating practice (P1 P3 P7 P26) Policymakers have a need for facts

based on the best knowledge currently available [72] Lazar [43] even affirms that for example for accessibility researchers to have anyreal impact outside of the research community they need to understand law and work with policymakers

Professional asso-ciations

UXPA or IxDA for example are organizations that frequently set up local events for the HCI community Our participants believe that astronger link with these associations can help establish interfaces between researchers and practitioners Researchers have partnered withmeetup groups and professional organizations to organize events combining academic and industry talks [19] but practitioners rarely meetwith researchers at these events [50 66] and it is unclear how effective they are at supporting adoption of HCI knowledge

Business stake-holders

Business-related stakeholders such as marketing financing and venture capital There is an understudied pathway in turning academicresearch into commercial products and services a path documented in [17] and mentioned by a few participants in our study (P1 P3 P7P8 P13 P15 P17)

Society end-users Some HCI researchers said that they are not interested in influencing design practice but in working directly with the populations that canbenefit from their work (P5 P8) Ladner points to working with end-users to tackle their problems through an open science approach asan alternative [41] Some participants also mentioned working with the media to influence public opinion This way users would demandchange from companies (P1 P7 P43)

Funding agencies as described in the discussion session top-down stimuli can promote structural change and reorganize infrastructures to facilitate Transla-tional Science More work about mdash and with mdash decision-makers of funding agencies is necessary

uncover stakeholder needs that help align user needs withbusiness needs [17] There is currently little incentive forHCI scholars to invest more time and resources in under-standing adoption [17] If a researcher aims at influencingdesign practice mdash to bridge from TBD or TAD mdash then evalu-ating knowledge adoption [45] may be necessary to knowwhether one has succeeded

Engaging mdash and studying engagement mdash with translators HCImust continue to engage with everyday successes and prob-lems practitioners face in their work An effective feedbackloop in Translational Science helps keep applied fields groundedin and relevant for practice For example asking questionssuch as ldquoIs this privacy recommendation found in a CHI pa-per effective or usedadapted in different ways in practicerdquoResearch with a focus on professional design practices allowsinsight into how practitioners refine and concretize abstractknowledge [32 42] Previous work provides guidance for thein-situ study of professional design practice [31] or usingsocial media traces to learn about design practice in [39]While the abovementioned examples focus on sharing

knowledge with design practitioners they are only one rolein TS in HCI Shneiderman advocates for building the ca-pacity to collaborate and coordinate with many stakehold-ers blending scientists engineers designers and end-usersto produce ldquohigher-impact research converging into a so-lutionrdquo [64] We were impressed by the many original andcreative translations that the HCI community representedthrough our 43 participants and referenced authors hasfound to bridge and study research and practice in HCI Fu-ture work should strive to understand the barriers to andeffective strategies for engaging the range of stakeholders inTS for HCI (See Table 3)

6 CONCLUSIONThe presence of barriers that hamper the progression ofknowledge into design practice is a significant issue within

HCI Therefore it is necessary to understand how knowledgeprogresses or fails to progress from research to practiceTo design the model for Translational Science in HCI wedrew on past work and interview data with researcherspractitioners andmultiple parties who are engaged ndash or notmdashin translating HCI knowledge In our continuum we describemultiple steps and gaps between basic and applied researchand design practice We also identify multiple translators andthe translational work they do This model offers insights onhow to bridge translational gaps and how to work with andtrain translators effectively It also acts as a foundation forfuture research on Translational Science in HCI

ACKNOWLEDGMENTSWe thank our study participants for their time and input intothis research Second we acknowledge reviewers for theirvaluable feedback that greatly improved the quality of thispaper Finally we thank Eclair Junchaya Hy Nguyen RickPaz as well as Sam Kolovson and Alison Kolberg for theirthoughtful feedback and support

REFERENCES[1] 2018 NSF Definitions of Research and Development An Anno-

tated Compilation of Official Sources httpswwwnsfgovstatisticsranddefrd-definitionspdf September 10 2018

[2] nd ruserexperience httpswwwredditcomruserexperienceRetrieved Sep 16 2018

[3] nd Stanford Scholar httpsscholarstanfordedu Retrieved Sep 142018

[4] nd UX Collective httpsuxdesigncc Retrieved Sep 16 2018[5] Christopher Alexander 1977 A pattern language towns buildings

construction Oxford university press[6] Dan Ariely 2008 Predictably Irrational HarperCollins New York NY

USA[7] Jordan Beck and Hamid R Ekbia 2018 The Theory-Practice Gap As

Generative Metaphor In Proceedings of the 2018 CHI Conference onHuman Factors in Computing Systems (CHI rsquo18) ACM New York NYUSA Article 620 11 pages httpsdoiorg10114531735743174194

Table 3 Additional Translators that can facilitate Translational Science in HCI and should be further investigated

Translators EvidencePolicymakers Mentioned by participants as a growing and effective role for regulating practice (P1 P3 P7 P26) Policymakers have a need for facts

based on the best knowledge currently available [72] Lazar [43] even affirms that for example for accessibility researchers to have anyreal impact outside of the research community they need to understand law and work with policymakers

Professional asso-ciations

UXPA or IxDA for example are organizations that frequently set up local events for the HCI community Our participants believe that astronger link with these associations can help establish interfaces between researchers and practitioners Researchers have partnered withmeetup groups and professional organizations to organize events combining academic and industry talks [19] but practitioners rarely meetwith researchers at these events [50 66] and it is unclear how effective they are at supporting adoption of HCI knowledge

Business stake-holders

Business-related stakeholders such as marketing financing and venture capital There is an understudied pathway in turning academicresearch into commercial products and services a path documented in [17] and mentioned by a few participants in our study (P1 P3 P7P8 P13 P15 P17)

Society end-users Some HCI researchers said that they are not interested in influencing design practice but in working directly with the populations that canbenefit from their work (P5 P8) Ladner points to working with end-users to tackle their problems through an open science approach asan alternative [41] Some participants also mentioned working with the media to influence public opinion This way users would demandchange from companies (P1 P7 P43)

Funding agencies as described in the discussion session top-down stimuli can promote structural change and reorganize infrastructures to facilitate Transla-tional Science More work about mdash and with mdash decision-makers of funding agencies is necessary

uncover stakeholder needs that help align user needs withbusiness needs [17] There is currently little incentive forHCI scholars to invest more time and resources in under-standing adoption [17] If a researcher aims at influencingdesign practice mdash to bridge from TBD or TAD mdash then evalu-ating knowledge adoption [45] may be necessary to knowwhether one has succeeded

Engaging mdash and studying engagement mdash with translators HCImust continue to engage with everyday successes and prob-lems practitioners face in their work An effective feedbackloop in Translational Science helps keep applied fields groundedin and relevant for practice For example asking questionssuch as ldquoIs this privacy recommendation found in a CHI pa-per effective or usedadapted in different ways in practicerdquoResearch with a focus on professional design practices allowsinsight into how practitioners refine and concretize abstractknowledge [32 42] Previous work provides guidance for thein-situ study of professional design practice [31] or usingsocial media traces to learn about design practice in [39]While the abovementioned examples focus on sharing

knowledge with design practitioners they are only one rolein TS in HCI Shneiderman advocates for building the ca-pacity to collaborate and coordinate with many stakehold-ers blending scientists engineers designers and end-usersto produce ldquohigher-impact research converging into a so-lutionrdquo [64] We were impressed by the many original andcreative translations that the HCI community representedthrough our 43 participants and referenced authors hasfound to bridge and study research and practice in HCI Fu-ture work should strive to understand the barriers to andeffective strategies for engaging the range of stakeholders inTS for HCI (See Table 3)

6 CONCLUSIONThe presence of barriers that hamper the progression ofknowledge into design practice is a significant issue within

HCI Therefore it is necessary to understand how knowledgeprogresses or fails to progress from research to practiceTo design the model for Translational Science in HCI wedrew on past work and interview data with researcherspractitioners andmultiple parties who are engaged ndash or notmdashin translating HCI knowledge In our continuum we describemultiple steps and gaps between basic and applied researchand design practice We also identify multiple translators andthe translational work they do This model offers insights onhow to bridge translational gaps and how to work with andtrain translators effectively It also acts as a foundation forfuture research on Translational Science in HCI

ACKNOWLEDGMENTSWe thank our study participants for their time and input intothis research Second we acknowledge reviewers for theirvaluable feedback that greatly improved the quality of thispaper Finally we thank Eclair Junchaya Hy Nguyen RickPaz as well as Sam Kolovson and Alison Kolberg for theirthoughtful feedback and support

REFERENCES[1] 2018 NSF Definitions of Research and Development An Anno-

tated Compilation of Official Sources httpswwwnsfgovstatisticsranddefrd-definitionspdf September 10 2018

[2] nd ruserexperience httpswwwredditcomruserexperienceRetrieved Sep 16 2018

[3] nd Stanford Scholar httpsscholarstanfordedu Retrieved Sep 142018

[4] nd UX Collective httpsuxdesigncc Retrieved Sep 16 2018[5] Christopher Alexander 1977 A pattern language towns buildings

construction Oxford university press[6] Dan Ariely 2008 Predictably Irrational HarperCollins New York NY

USA[7] Jordan Beck and Hamid R Ekbia 2018 The Theory-Practice Gap As

Generative Metaphor In Proceedings of the 2018 CHI Conference onHuman Factors in Computing Systems (CHI rsquo18) ACM New York NYUSA Article 620 11 pages httpsdoiorg10114531735743174194

[8] Lisa A Bero Roberto Grilli Jeremy M Grimshaw Emma Harvey An-drew D Oxman and Mary Ann Thomson 1998 Closing the gapbetween research and practice an overview of systematic reviews ofinterventions to promote the implementation of research findings TheCochrane Effective Practice and Organization of Care Review GroupBMJ (Clinical research ed) 317 7156 (1998) 465ndash468

[9] Susanne Boslashdker 2015 Third-wave HCI 10 Years Latermdashparticipationand Sharing interactions 22 5 (Aug 2015) 24ndash31 httpsdoiorg1011452804405

[10] Andrew Booth 2003 Bridging the research-practice gap The role ofevidence based librarianship New Review of Information and LibraryResearch 9 1 (2003) 3ndash23

[11] Elizabeth A Buie Susan M Dray Keith E Instone Jhilmil Jain GitteLindgaard and Arnold M Lund 2010 Researcher-practitioner In-teraction In CHI rsquo10 Extended Abstracts on Human Factors in Com-puting Systems (CHI EA rsquo10) ACM New York NY USA 4469ndash4472httpsdoiorg10114517538461754176

[12] Keith A Butler 1985 Connecting Theory and Practice A Case Studyof Achieving Usability Goals In Proceedings of the SIGCHI Conferenceon Human Factors in Computing Systems (CHI rsquo85) ACM New YorkNY USA 85ndash88 httpsdoiorg101145317456317472

[13] John M Carroll 1995 Scenario-based design envisioning work andtechnology in system development (1995)

[14] John M Carroll 1997 Human-computer interaction psychology as ascience of design Annual review of psychology 48 1 (1997) 61ndash83

[15] Craig R Carter 2008 Knowledge production and knowledge transferclosing the researchndashpractice gap Journal of Supply Chain Manage-ment 44 2 (2008) 78ndash82

[16] Martin A Cheever James P Allison Andrea S Ferris Olivera J FinnBenjamin M Hastings Toby T Hecht Ira Mellman Sheila A PrindivilleJaye L Viner Louis M Weiner et al 2009 The prioritization of cancerantigens a national cancer institute pilot project for the accelerationof translational research Clinical cancer research 15 17 (2009) 5323ndash5337

[17] Parmit K Chilana Andrew J Ko and Jacob Wobbrock 2015 FromUser-Centered to Adoption-Centered Design A Case Study of an HCIResearch Innovation Becoming a Product In Proceedings of the 33rdAnnual ACM Conference on Human Factors in Computing Systems (CHIrsquo15) ACM New York NY USA 1749ndash1758 httpsdoiorg10114527021232702412

[18] Francis S Collins 2011 Reengineering translational science the timeis right Science translational medicine 3 90 (2011) 90cm17ndash90cm17

[19] Lucas Colusso Cynthia L Bennett Gary Hsieh and Sean A Munson2017 Translational Resources Reducing the Gap Between AcademicResearch and HCI Practice In Proceedings of the 2017 Conference onDesigning Interactive Systems (DIS rsquo17) ACM New York NY USA957ndash968 httpsdoiorg10114530646633064667

[20] Alan Cooper Robert Reimann and David Cronin 2007 About face 3the essentials of interaction design John Wiley amp Sons

[21] Juri Dachtera Dave Randall and Volker Wulf 2014 Research onResearch Design Research at the Margins Academia Industry andEnd-users In Proceedings of the SIGCHI Conference on Human Factorsin Computing Systems (CHI rsquo14) ACM New York NY USA 713ndash722httpsdoiorg10114525562882557261

[22] Peter Dalsgaard and Christian Dindler 2014 Between Theory andPractice Bridging Concepts in HCI Research In Proceedings of theSIGCHI Conference on Human Factors in Computing Systems (CHI rsquo14)ACM NewYork NY USA 1635ndash1644 httpsdoiorg10114525562882557342

[23] Diana L Deadrick and Pamela A Gibson 2007 An examination ofthe researchndashpractice gap in HR Comparing topics of interest to HRacademics and HR professionals Human Resource Management Review

17 2 (2007) 131ndash139[24] James W Dearing and Kerk F Kee 2012 Historical roots of dissemina-

tion and implementation science Dissemination and implementationresearch in health Translating science to practice 55 (2012) 71

[25] Audrey Desjardins Ron Wakkary Will Odom Henry Lin andMarkus Lorenz Schilling 2017 Exploring DIY Tutorials As a Wayto Disseminate Research Through Design Interactions 24 4 (June2017) 78ndash82 httpsdoiorg1011453098319

[26] Alan Dix Janet E Finlay Gregory D Abowd and Russell Beale 2003Human-Computer Interaction (3rd Edition) Prentice-Hall Inc UpperSaddle River NJ USA

[27] Paul Dourish 2006 Implications for Design In Proceedings of theSIGCHI Conference on Human Factors in Computing Systems (CHI rsquo06)ACM New York NY USA 541ndash550 httpsdoiorg10114511247721124855

[28] Brian C Drolet and Nancy M Lorenzi 2011 Translational researchunderstanding the continuum from bench to bedside TranslationalResearch 157 1 (2011) 1ndash5

[29] Nir Eyal 2014 Hooked How to build habit-forming products Penguin[30] Dean L Fixsen Sandra FNaoom KarenABlase and RobertM Friedman

2005 Implementation research a synthesis of the literature (2005)[31] Elizabeth Goodman Erik Stolterman and Ron Wakkary 2011 Under-

standing Interaction Design Practices In Proceedings of the SIGCHI Con-ference on Human Factors in Computing Systems (CHI rsquo11) ACM NewYork NY USA 1061ndash1070 httpsdoiorg10114519789421979100

[32] Colin M Gray Erik Stolterman and Martin A Siegel 2014 Repriori-tizing the Relationship Between HCI Research and Practice Bubble-upand Trickle-down Effects In Proceedings of the 2014 Conference onDesigning Interactive Systems (DIS rsquo14) ACM New York NY USA725ndash734 httpsdoiorg10114525985102598595

[33] Kristina Houmloumlk and Jonas Loumlwgren 2012 Strong ConceptsIntermediate-level Knowledge in Interaction Design Research ACMTrans Comput-Hum Interact 19 3 Article 23 (Oct 2012) 18 pageshttpsdoiorg10114523623642362371

[34] Edwin Hutchins 1995 Cognition in the Wild MIT press[35] Google Inc nd Google Scholar httpsscholargooglecom[36] Keith Instone 2010 Researcher-practitioner interaction update

(UXRPI) httpinstoneorguxrpi-blogtopic Retrieved August 92018

[37] Jon Kolko 2010 On Academic Knowledge Production Interactions 175 (Sept 2010) 80ndash80 httpsdoiorg10114518362161836237

[38] Ilpo Koskinen John Zimmerman Thomas Binder Johan Redstromand Stephan Wensveen 2011 Design research through practice Fromthe lab field and showroom Elsevier

[39] Yubo Kou and Colin M Gray 2018 Distinctions between the Commu-nication of Experiential and Academic Design Knowledge A LinguisticAnalysis Design Research Society (DRS) (2018) httpsdoiorgDOI1021606dma2018532

[40] Yubo Kou and Colin M Gray 2018 What Do You Recommend aComplete Beginner Like Me to Practice Professional Self-Disclosurein an Online Community Proc ACM Hum-Comput Interact 2 CSCWArticle 94 (Nov 2018) 24 pages httpsdoiorg1011453274363

[41] Richard E Ladner 2014 My Path to Becoming an Accessibility Re-searcher SIGACCESS Access Comput 110 (Sept 2014) 5ndash16 httpsdoiorg10114526709622670964

[42] Carine Lallemand 2015 Towards consolidated methods for the designand evaluation of user experience PhD Dissertation University ofLuxembourgacircĂŃacircĂŃ Luxembourg

[43] Jonathan Lazar 2017 Letrsquos Strengthen the HCI Community by Takinga Gap Year Interactions 25 1 (Dec 2017) 20ndash21 httpsdoiorg1011453155054

[44] Aleksei Leontiev [n d] Activity consciousness and personality ([nd])

[45] Joseph Lindley Paul Coulton and Miriam Sturdee 2017 Implicationsfor Adoption In Proceedings of the 2017 CHI Conference on HumanFactors in Computing Systems (CHI rsquo17) ACM New York NY USA265ndash277 httpsdoiorg10114530254533025742

[46] Daniel Lockton 2013 Design with intent a design pattern toolkit forenvironmental and social behaviour change PhD Dissertation BrunelUniversity School of Engineering and Design PhD Theses

[47] Matthew B Miles 1964 Innovation in education Bureau of PublicationTeachers College Columbia University

[48] Zoeuml Slote Morris Steven Wooding and Jonathan Grant 2011 Theanswer is 17 years what is the question understanding time lags intranslational research Journal of the Royal Society of Medicine 104 12(2011) 510ndash520

[49] Donald Norman 1986 A amp Draper SW User Centered System DesignNew Perspectives on Human-Computer Interaction L Erlbaum AssocInc Hillsdale NJ USA (1986)

[50] Donald A Norman 2010 The Research-practice Gap The Need forTranslational Developers Interactions 17 4 (July 2010) 9ndash12 httpsdoiorg10114518064911806494

[51] William Odom Tom Jenkins Kristina Andersen Bill Gaver JamesPierce Anna Vallgaringrda Andy Boucher David Chatting Janne vanKollenburg and Kevin Lefeuvre 2017 Crafting a Place for Attendingto the Things of Design at CHI Interactions 25 1 (Dec 2017) 52ndash57httpsdoiorg1011453161605

[52] Carrie J Petrucci and Kathleen M Quinlan 2007 Bridging theresearch-practice gap Concept mapping as a mixed-methods strategyin practice-based research and evaluation Journal of Social ServiceResearch 34 2 (2007) 25ndash42

[53] Stuart Reeves nd What Is the Relationship Between HCI Researchand UX Practice httpswwwuxmatterscommtarchives201408what-is-the-relationship-between-hci-research-and-ux-practicephp Retrieved Sep 4 2018

[54] Christian Remy Silke Gegenbauer and Elaine M Huang 2015 Bridg-ing the Theory-Practice Gap Lessons and Challenges of Applyingthe Attachment Framework for Sustainable HCI Design In Proceed-ings of the 33rd Annual ACM Conference on Human Factors in Com-puting Systems (CHI rsquo15) ACM New York NY USA 1305ndash1314httpsdoiorg10114527021232702567

[55] Maxine Robertson Jacky Swan and Sue Newell 1996 The role ofnetworks in the diffusion of technological innovation Journal ofManagement Studies 33 3 (1996) 333ndash359

[56] David J Roedl and Erik Stolterman 2013 Design Research at CHI andIts Applicability to Design Practice In Proceedings of the SIGCHI Con-ference on Human Factors in Computing Systems (CHI rsquo13) ACM NewYork NY USA 1951ndash1954 httpsdoiorg10114524706542466257

[57] Everett M Rogers 2010 Diffusion of innovations Simon and Schuster[58] Yvonne Rogers 2004 New theoretical approaches for human-computer

interaction Annual review of information science and technology 38 1(2004) 87ndash143

[59] Yvonne Rogers 2012 HCI theory classical modern and contemporarySynthesis Lectures on Human-Centered Informatics 5 2 (2012) 1ndash129

[60] Doris McGartland Rubio Ellie E Schoenbaum Linda S Lee David ESchteingart Paul R Marantz Karl E Anderson Lauren Dewey PlattAdriana Baez and Karin Esposito 2010 Defining translational researchimplications for training Academic medicine journal of the Associationof American Medical Colleges 85 3 (2010) 470

[61] Sara L Rynes 2012 The research-practice gap in IO psychology andrelated fields Challenges and potential solutions The Oxford handbookof organizational psychology 1 (2012) 409ndash452

[62] Corina Sas Steve Whittaker Steven Dow Jodi Forlizzi and John Zim-merman 2014 Generating Implications for Design Through DesignResearch In Proceedings of the SIGCHI Conference on Human Factors inComputing Systems (CHI rsquo14) ACM New York NY USA 1971ndash1980httpsdoiorg10114525562882557357

[63] Beth Seymour Sue Kinn and Norrie Sutherland 2003 Valuingboth critical and creative thinking in clinical practice narrowing theresearchndashpractice gap Journal of Advanced Nursing 42 3 (2003)288ndash296

[64] Ben Shneiderman 2016 The new ABCs of research Achieving break-through collaborations Oxford University Press

[65] C Estelle Smith Xinyi Wang Raghav Pavan Karumur and HaiyiZhu 2018 [Un]Breaking News Design Opportunities for EnhancingCollaboration in Scientific Media Production In Proceedings of the2018 CHI Conference on Human Factors in Computing Systems (CHI rsquo18)ACM New York NY USA Article 381 13 pages httpsdoiorg10114531735743173955

[66] Erik Stolterman 2008 The nature of design practice and implicationsfor interaction design research International Journal of Design 2 1(2008)

[67] Lucy A Suchman 1987 Plans and situated actions The problem ofhuman-machine communication Cambridge university press

[68] Rachel G Tabak Elaine C Khoong David A Chambers and Ross CBrownson 2012 Bridging research and practice models for dissemi-nation and implementation research American journal of preventivemedicine 43 3 (2012) 337ndash350

[69] Stuart Card Tom Carey Jean Gasen Marilyn Mantei Gary PerlmanGary Strong Thomas T Hewett Ronald Baecker andWilliam Verplank1996 ACM SIGCHI Curricula for Human-Computer Interaction httpsigchiorgcdgcdg2html2_1 Retrieved Sept 10 2018

[70] Raphael Velt 2018 Putting trajectories to work translating a HCIframework into design practice PhD Dissertation University of Not-tingham

[71] Abraham Wandersman Jennifer Duffy Paul Flaspohler Rita NoonanKeri Lubell Lindsey Stillman Morris Blachman Richard Dunville andJanet Saul 2008 Bridging the gap between prevention research andpractice The interactive systems framework for dissemination andimplementation American journal of community psychology 41 3-4(2008) 171ndash181

[72] Carol H Weiss 1995 The haphazard connection social science andpublic policy International Journal of Educational Research 23 2 (1995)137ndash150

[73] Richmond Y Wong Ellen Van Wyk and James Pierce 2017 Real-Fictional Entanglements Using Science Fiction and Design Fiction toInterrogate Sensing Technologies In Proceedings of the 2017 Conferenceon Designing Interactive Systems (DIS rsquo17) ACM New York NY USA567ndash579 httpsdoiorg10114530646633064682

[74] Steven H Woolf 2008 The meaning of translational research and whyit matters Jama 299 2 (2008) 211ndash213

[75] John Zimmerman Jodi Forlizzi and Shelley Evenson 2007 ResearchThrough Design As a Method for Interaction Design Research in HCIIn Proceedings of the SIGCHI Conference on Human Factors in ComputingSystems (CHI rsquo07) ACM New York NY USA 493ndash502 httpsdoiorg10114512406241240704