Supplemental ArticleEvaluating MyPlate: An Expanded Framework UsingTraditional and Nontraditional Metrics for AssessingHealth Communication CampaignsElyse Levine, PhD, RD1; Jodie Abbatangelo-Gray, ScD2; Amy R. Mobley, PhD, RD3;Grant R. McLaughlin, MA1; Jill Herzog, BA1

1Booz Alle2Summit R3DepartmeSTATEMESUPPORTAddress foParkway,Levine_ely�2012 SOdoi:10.1016

S2

ABSTRACT

MyPlate, the icon and multimodal communication plan developed for the 2010 Dietary Guidelines forAmericans (DGA), provides an opportunity to consider new approaches to evaluating the effectivenessof communication initiatives. A review of indicators used in assessments for previous DGA communica-tion initiatives finds gaps in accounting for important intermediate and long-term outcomes. This evalu-ation framework for theMyPlate Communications Initiative builds onwell-known and underused modelsand theories to propose awide breadth of observations, outputs, and outcomes that can contribute to a fullerassessment of effectiveness. Two areas are suggested to focus evaluation efforts in order to advance under-standing of the effectiveness of theMyPlate Communications Initiative: understanding the extent to whichmessages and products from the initiative are associated with positive changes in social norms toward thedesired behaviors, and strategies to increase the effectiveness of communications about DGA in vulnerablepopulations.Key Words: MyPlate, dietary guidelines, health communication, program evaluation, nutrition policy(J Nutr Educ Behav. 2012;44:S2-S12.)

INTRODUCTION

In June, 2011, theMyPlate icon and itssupporting multicomponent commu-nications plan were unveiled by theUnited States Department of Agricul-ture (USDA) Center for NutritionPolicy andPromotion (CNPP) as aplat-form to support the translation of the2010DietaryGuidelines forAmericans(DGA).1 MyPlate replaces MyPyramidas USDA's healthy eating communica-tions initiative. Like its predecessors,the Food Guide Pyramid and MyPyra-mid, MyPlate illustrates the foodgroups and is supported by communi-cation tools and materials thatunderlie theDGA.However, accordingto USDA, MyPlate is a substantialcommunications departure from theprevious approach. The Pyramid was

n Hamilton, McLean, VAesearch Associates, New York, NYnt of Nutritional Sciences, UniversitNT OF POTENTIAL CONFLIC: See page S10.r correspondence: Elyse Levine, PhRockville, MD 20852; Phone: (240se@bah.comCIETY FOR NUTRITION EDUC/j.jneb.2012.05.011

designed to be a teaching tool to com-municate the DGA as a whole and rep-resented what and how much to eatover the course of a day. In contrast,theMyPlate icon is ‘‘a simple, yet pow-erful, visual cue to prompt consumersto think about their food choicesacross food groups and to builda healthy plate at meal times.’’2 Tothat end, MyPlate is part of a multi-modal communication strategy thatincludes the MyPlate Web site withthe SuperTracker tool to personalizefood plans, consumer educationalmaterials and e-tools, social media en-gagement, and a partnership initiativeto help coordinate and disseminateconsistent messages of the DGA.1,3

The MyPlate initiative has been de-signed for maximum visibility. Likethe USDA efforts that preceded it,

y of Connecticut, Storrs, CTT OF INTEREST AND FUNDING/

D, RD, Booz Allen Hamilton, 1 Preserve) 453-5387; Fax: (301) 838-3606; E-mail:

ATION AND BEHAVIOR

Journal of Nutrition Education and Beh

MyPlate will be incorporated intohealth curriculum resources createdfor nutrition education purposes forchildren and adults, translated intoseveral languages, and promoted bynutrition communicators, educators,and the food industry. The MyPlatecommunications initiative also shareswith its predecessors high expecta-tions for performance, and the evalua-tion of its effectiveness will bechallenging—an aspect shared by allhealth communication campaigns.Evaluation must be appropriate forits intended use and realistic basedon the stage of the communicationsinitiative. Yet with constant scrutinyof monies allocated for health educa-tion,4 communication researchersneed to demonstrate the worthinessof interventions, especially whencommunication plan componentshave indirect connections to long-term goals. Indeed, when it comestime to re-examine the DGA and thecommunication tactics, how will re-searchers know whether the MyPlateinitiative is effective? Howwill the ini-tiative's success be measured, and overwhat period of time?

The objective of this manuscript isto propose a framework for evaluatingthe DGA communications initiative,

avior � Volume 44, Number 4, 2012

Journal of Nutrition Education and Behavior � Volume 44, Number 4, 2012 Levine et al S3

including the MyPlate icon, the con-cepts of which can be applied to otherhealth communication efforts de-signed for the general population. Itis hoped that the framework will stim-ulate various approaches to evaluatingthe effectiveness of the DGA commu-nication and highlight the impor-tance of even small changes thatoccur before long-term goals arereached, for example, changes in per-ceptions of what a healthy amountof meat looks like for a meal. This pro-cess will, in turn, inform improvementof the initiative and provide a widertoolbox to share lessons learned. Tobuild the case for the framework, thenext section summarizes challengesof evaluating health communicationcampaigns and shows how communi-cation initiatives built around DGAserve as good case studies. This sectionis followed by the proposed frame-work for evaluating the MyPlate com-munication initiative and a discussionon considerations for acceptableevidence of effectiveness in healthcommunications. Finally, applica-tions for the framework are suggested,focusing on areas that would advanceunderstanding of the effectivenessof the 2010 DGA communicationsinitiative.

Health CommunicationEvaluation Challenges

Hornik notes ‘‘evaluations of publichealth communication programs willrarely produce the unequivocal evi-dence promised in randomized con-trolled trials of pills.’’5 This truismcan be explained by 3 main observa-tions. First, many consider behavioralchanges the most important outcomeof communication campaigns, butcommunication most frequentlyworks as an indirect contributor to be-havior change outcomes.6-8 Forexample, seat belt use was aided bymessages featuring car crashdummies, but real change cameabout through incentives frominsurance companies and ‘‘click it orticket’’ regulations, which were alsoinfluenced by communication toindustry and policy makers.9

Second, in interventions that findevidence of behavior change attrib-uted to communication, the magni-tude of the effects is much smaller

than those observed in medical trials.This finding was illustrated bySnyder's systematic analysis of healthcampaigns, which yielded an averageeffect size of about 5 percentagepoints.10 Although often seen as in-sufficient among critics, small effectsizes, she argues, can actually havea big impact at the population level,particularly in dietary health.

Finally, messages need substantiveexposure and sufficient time to work,yet many health communicationcampaigns cannot achieve the levelof saturation needed to produce mea-surable results.11 Real change, likethat seen in seat belt use, is typicallymeasured over decades, whereasmost behavioral researchers canhope to secure funding for only 3-5years.

The Food Guide Pyramid and itssuccessor MyPyramid serve as goodillustrations of these challenges. Bothgraphics benefitted from formative re-search, especially the communicationplan for the MyPyramid infographicsymbol, corresponding consumermes-sages, andWeb content.12,13 However,evaluation efforts were scarce, andmeasures that were reported relied ononly a few indicators, as will bediscussed in more detail below.Published findings indicate that manyAmericans report recognizing thePyramid and having knowledge of itsspecific recommendations.12,14 Infact, the criterion that most use tojudge the effectiveness of communi-cation about the DGA is dietaryquality among Americans, which didnot improve during the 2-decade reignof the Pyramid.8,15,16 Regardless ofwhether dietary intake is the bestmeasure of effectiveness, the lack ofinformation on other, intermediateoutcomes raises questions that can bebetter addressed within an evaluationframework. For example, whatexplains the apparent ineffectivenessof MyPyramid and its communi-cations? Was it overly complex? Didit send misleading messages, such as‘‘all fats are bad,’’ as was reportedduring news coverage of the MyPlatelaunch?17,18 Despite a lack ofinformation on why, how, and withwhom the MyPyramid communi-cations needed improvement, anexpert roundtable assembled after therelease of the 2010 DGA reachedconsensus that ‘‘consumer messages

around nutrition and especiallyweight loss need to be even simplerthan the past communications of theDGA.’’8

MyPlate CommunicationInitiative

In response to the perceived short-comings of theMyPyramid communi-cations, health professionals hopethat the MyPlate icon and corre-sponding communications will betterbridge the gap between knowledgeand behavior. The agency withprimary responsibility for setting nu-trition policy for USDA and its nutri-tion promotion, USDA's CNPP,developed a comprehensive commu-nication initiative to disseminate thekey messages of the DGA for optimaluse, and it set goals and objectivesfor the effort within the scope oftheory-based communication inter-ventions. The goal of the initiative is‘‘to support Americans in buildinghealthy diets.’’1 The USDA will try toaccomplish this goal by ‘‘providingan easy-to-understand icon that willhelp deliver a series of healthy eatingmessages that highlight key consumeractions based on the 2010 DGA’’ and‘‘empowering people with informa-tion they need to make healthy foodchoices.’’1

Toward these objectives, the USDACNPP designed the MyPlate icon to bea visual cue prompting consumers tobuild healthy plates at meal times.2

The plate icon with the 5 food groupsit symbolizes (fruit, vegetables, grains,protein food, and dairy food), alongwith a new Web site, were pretestedfor comprehension and overall ap-peal, as were the 7 key selected mes-sages, prioritized from 16 testedmessages.19 The icon and communi-cations initiative that supportsMyPlate includes a comprehensiveWeb site with interactive featuresand resources for educators, healthprofessionals, and consumers to putthe messages into practice, and 2other elements supported by commu-nication theory. The first element isa plan to release 1 key message ata time through multiple channels tohelp ensure maximum exposure, rais-ing awareness and building a founda-tion for subsequent messages.3 Thesecond feature is a 2-level partnership

S4 Levine et al Journal of Nutrition Education and Behavior � Volume 44, Number 4, 2012

plan to help disseminate the messagesthrough a wide range of like-mindedgroups and institutions with closeproximity to the public.

Although using partners to amplifyexposure and reach for public healthmessages can be beneficial,20-22 theMyPyramid initiative was criticizedfor its over-reliance on industrysources, which could weaken thevalidity of the message.23 The‘‘Community Partner’’ level is open toall who wish to help spread the wordthrough their communication vehicles;at the time of writing, the Web site tosign on as a Community Partnercounted over 6,000 individuals and or-ganizations. Organizations committedat the ‘‘National Partner’’ level, whichincludes over 85 representatives fromhealthassociations, researchandprofes-sional organizations, food service andrestaurant chains, the food industry,food retailers, and media with missionsconsistent with the USDA's, providein-kind resources to further the successof the communication initiative.24

Why Develop an EvaluationFramework for the MyPlateCommunications Initiative?

Grounded in evidence-based nutritionscience and pretested for optimal us-ability, the USDA CNPP used soundpractices to create the MyPlate initia-tive.19 However, current options avail-able for evaluating its effectiveness arelimited. Findings can be gleaned fromnational databases, including theNational Health and Nutrition Exami-nation Survey,25 the Healthy EatingIndex,26 the Centers for DiseaseControl and Prevention's (CDC)Behavioral Risk Factor SurveillanceSystem,27 as well as the InternationalFood Information Council Founda-tion's annual Food and HealthSurvey.14 Proprietary sources of data,such as Porter Novelli's HealthStyles,28

Gallup-Healthway's Well-Being In-dex,29 and panel data managed bythe NDP Group,30 may provide addi-tional information. Specific questionsthat have been used for previousDGA communication initiatives fromnonproprietary instruments are pre-sented in Table 1.

A quick survey of this compilationshows it taps a limited set of con-structs. Most items ask about aware-

ness of the DGA and/or MyPyramid,or they alternatively tap the antici-pated long-term outcomes—food con-sumption (eg, how many fruits andvegetables consumed) and measuresof obesity. Thesedata are undisputedlyimportant, but they are insufficient toascertain what works and what needsimprovement. As discussed earlier,food consumption and obesity are, atbest, uncertainmeasures of communi-cation effectiveness because changesin these outcomes cannot be attrib-uted exclusively to the communica-tions; rather, they result frommultiple determinants includingpolicy and environment. It is furthersuggested that careful tracking of con-tributing factors along amore compre-hensive causal pathway can helpidentify promising elements that canbest contribute to MyPlate communi-cations plan goals, which can also beleveraged in communication initia-tives or campaigns of related intent.

The limited measures availablefrom ongoing national surveys pro-vides an opportunity for partners ofthe MyPlate initiative and academicinstitutions to help evaluate the cam-paign through independent research.The MyPlate effort also offers a newopportunity for partnering organi-zations and nutrition and healtheducators to contribute to a pool ofknowledge that will help developand refine evaluation measures oflarge health communication initia-tives. To that end, a framework is pre-sented to help visualize the breadth ofinterventions and observations thatcan contribute to a fuller evaluationof campaign effectiveness.

A PROPOSEDEVALUATIONFRAMEWORK FOR THEMYPLATECOMMUNICATIONINITIATIVE

The primary goal of this comprehen-sive, use-focused evaluation frame-work, illustrated in the Figure, is toencourage consideration of a wideperspective of evidence when healthcommunicators seek to measure effec-tiveness of MyPlate and other healthcommunication initiatives. A multi-faceted evaluation framework is pro-

posed that invites qualitative andquantitative researchmethods tomea-sure both communications imple-mentation and outcomes. With thisapproach, health communicators canbetter identify the links between pro-gram inputs, activities, and outcomesthat can guide communications im-provement andhelp to assess the over-all effectiveness of the initiative.

The terms ‘‘framework’’ and‘‘model’’ are sometimes used inter-changeably, but for the purpose ofthis article, a framework is defined asa rubric toprovideacommonsetofvar-iables to use in the design, collection,analysis, and application of findings.31

As described by Ostrom, ‘‘Withouta commonframework toorganizefind-ings, isolated knowledge does not cu-mulate.’’ Models for health educationresearch and practice are defined ascausal linkages among a set of con-cepts; models are often informed bymore than 1 theory, as well as by em-pirical findings.32 To use a familiarand fitting metaphor, frameworkshave been compared to maps thatshow the layof the land, theories repre-sent the highways and routes that leadto the destination, andmodels use 1 ormore theories as the preferred route.For the purposes of this illustration,the institutions that shape recommen-dations and the communities inwhichpeople interact can be compared to thecities and landmarks that mark prog-ress. Observations at each of theselevels will all contribute toward a col-lective pattern of ‘‘what works.’’

The framework is composedofwell-tested elements in the field of healthinterventions. The structure of a logicmodel is used to organize the majorcomponents into possible sources ofdata and the context from which in-formation will be derived (inputs), an-ticipated interventions (activities),and expected outcomes. The consen-sus of an expert panel convened atthe CDC posits that campaign plan-ners should use their logic model toidentify issues of relevance to specificaudiences (current beliefs, practices,group identification) before educationefforts commence.6 Messages shouldthen be tailored to meet the needs ofunique segments of the population.

Major classifications of research—formative, process, outcome research—are also represented in the model andinformeachkeycomponentof the logic

Table 1. National Data Collection Vehicles for Monitoring Previous Dietary Guidelines for Americans (DGA) and Related Initiatives

Data Collection Instrument DGA Indicators and Sample Measuresa,b Notes

Federal/state data collectionsNational Health and Nutrition

Examination Survey (NHANES)25Food consumption: How often did (you) drink any type of milk, including milk added to cereal?Overweight and obesity: Assessment using height and weight. Howmuch did you weigh 1 year ago?

Ten years ago? At 25 years of age?

Annual survey combining interviews and laboratorytests, n ¼ 5,000, administered annually since1999.

Flexible Consumer Behavior Survey72 Awareness of DGA icon/communication plan: Have you heard of MyPyramid?Use of DGAmessages: Have you looked up the MyPyramid plan for a (man/woman/person) your age

on the Internet? Have you tried to follow the (MyPyramid plan/Pyramid plan) recommended for you?Knowledge of DGA: Howmany ounces of meat and beans would you say a (man/woman) of your age

and physical activity should eat each day for good health?Food consumption: How often (does your family/do you) have any of these dark green vegetables

available at home? This includes fresh, dried, canned, and frozen vegetables.

Added to the National Health and NutritionExamination Survey 2007-2008, 2009-2010.

Behavioral Risk Factor SurveillanceSystem27

Awareness of DGA icon/communication plan: Have you heard about the US Department ofAgriculture’s Food Guide Pyramid?73

Fruit and vegetable consumption:c During the past month, howmany times per day, week, ormonthdid you eat orange-colored vegetables such as sweet potatoes, pumpkin, winter squash, or carrots?

Overweight and obesity: Assessment using self-reported height and weight.

State-based system of surveys; more than 350,000interviewed annually by telephone; separatemodules for states.

Youth Risk Behavior SurveillanceSystem74

Fruit and vegetable consumption: During the past 7 days, how many times did you eat fruit?Overweight and obesity: Assessment using self-reported height and weight.

Conducted every other year.Includes a national school-based survey conducted

by the CDC (n > 16,000 in 2009) in addition tostate, territorial, tribal, and local surveys.

Healthy Eating Index-2005d,26 Diet quality: Food intake or availability in conformance with Federal dietary guidancerecommendations.

US population scores use the most recent NHANESdata available when the HEI is analyzed; scoreshave been provided for subpopulations includingolder adults, children, and populations served byfood assistance programs.

California Health Interview Survey75 Food consumption: During the past month, how many times did you eat fruit? Do not count juices.Overweight and obesity: Assessment using self-reported height and weight.

Collected from > 50,000 individuals every other yearsince 2001.

Nongovernment data collectionsInternational Food Information

Council Food and Health Survey14Familiarity with DGA and MyPyramid: Which of the following best describes your familiarity with

the DGA, which are the US government-approved food and nutrition guidelines?Intention to act on specific guidelines: Which specific guideline are you most interested in

personally adopting?Exposure to DGA: Where, if at all, have you seen Dietary Guidelines information?

Web survey, n ¼ 1,000, conducted annually since2006.

Porter Novelli Styles28 Food and beverage consumption; fruit and vegetable consumption; healthy eating intent;nutrients respondents try to consume/avoid; familiarity with the DGA (2010); DGAimportance (2010); motivations to eat healthy.e

Conducted by Porter Novelli since 1995, fieldedthree times per year, n ¼ 6000 for Spring surveyand n = 4000 each for Summer and Fall surveys.

Well-Being Index29 Food consumption; overweight and obesity.e Daily survey n ¼ 1,000/d conducted by Gallup-Healthways since 2008.

NDP Group, Inc. National EatingTrends Panel Data30

Food consumption.e National Eating Trends panel data has been collectedsince 1980; monitors eating and drinking habits ofthousands of consumers.

CDC indicates Centers for Disease Control and Prevention; HEI, Healthy Eating Index; US, United States.aDGA indicators and measures reported are examples. Other measures were also used for assessment; bDGA indicators are bolded; cIn core questions on the Behavioral Risk Factor Sur-

veillance System every other year; dAt the time of writing the HEI was being revised to reflect the 2010 DGA; eProprietary data, individual questions are added by licensing organizations.

JournalofNutritio

nEducatio

nandBehavior�

Volume44,

Number

4,2012

Levineet

alS5

Figure. Evaluation framework for the MyPlate communication initiative. BMI indicates body mass index; DGA, Dietary Guidelines forAmericans; USDA, United States Department of Agriculture.

S6 Levine et al Journal of Nutrition Education and Behavior � Volume 44, Number 4, 2012

model. Examples of specific questionsrelated to the use-focused evaluationare presented in Table 2.

Formative research assesses the po-tential feasibility and desirability ofplanned program inputs within thecontext of the resources that theMyPlate communications initiativecan leverage to achieve its goals. For-mative research uses qualitative andquantitative research to obtain stake-holder feedback on knowledge,beliefs, attitudes and practices, moti-vators and barriers to desired out-comes, specific activities or outputsplanned for the campaign, and base-line information to measure changefor specific audiences. It is often an it-erative process of revising and solicit-ing further feedback within theconstraints of the research timeframeand resources. The USDA conductedformative research to develop the

icon and the messages that supportit from the DGA,19 however, it is ex-pected that more formative researchwill be forthcoming to tailor the useof MyPlate and other communicationinputs for special populations, as wellas to make the best use of emerging orchanging resources (eg, partnerships,new media applications).

Process evaluation is conceptualizedunder the ‘‘activities’’ header in theframework, and focuses on the imple-mentation of the initiative, assessingthe extent to which all program activ-ities (outputs) are being conducted asplanned. The systematic collection ofprocess information over time willhelp the MyPlate initiative andUSDA partners tomodify componentsand activities as needed. Research ef-forts should be designed to answerquestions such as: Is the initiative be-ing implemented as intended? Is it

reaching intended audiences? Haveexternal factors (policy and otherenvironmental factors that mightinfluence nutrition patterns) emergedduring the communication period?Which components are showingmost promise? Are there communica-tion plan elements that need to be al-tered in some way or even eliminatedaltogether? Are community-levelpartners receiving the materials andsupport they need to educate popula-tions of special interest?

Outcome research allows identifica-tion of links between communicationcomponents and desired outcomes inboth audience knowledge and dietarybehaviors. This research will enableresearchers to answer questions suchas: What are the short-term, interme-diate, and long-term outcomes of theinitiative? Are desired outcomes beingreached and, and to what extent?

Table 2. Sample Research Questions for a Use-Based Approach to Evaluation Research

Levels of Influence

Research Type

Formative Process OutcomePersonal/ interpersonal What social media tools

designed to communicateMyPlate initiative-relatedmessages are desired andwould be used by targetedpopulations?

How many MyPlate initiativesocial media tools have beendeveloped?

How often are MyPlate initiativesocial media tools used bytargeted populations?

Have social media toolsinfluenced knowledge/awareness of MyPlateinitiative-related messages?

Environmental How can MyPlate initiative-related messages best beincorporated into the schoolsystem (eg, school cafeteria)?

What proportion of institutions—schools, pre-school and after-school programs—haveadopted MyPlate initiative-related materials/messages?

To what extent have schoolsadopted DGArecommendations? (eg, hasdietary quality of menus forschools improved?)

Is successful implementation andadoption of DGA in schoolsassociated with improvementsin children’s dietary intake?

Systems and policy What are the costs and benefitsof agribusiness to alignsystems and policies tosupport wide adoption andmaintenance of DGA?

Towhat extent have agribusinessand manufacturers changedfood ingredients to reflectDGAs?

To what extent are changes inagribusiness and policy tosupport adoption of DGAassociated with improveddietary quality, decreasedprevalence of overweight andobesity, and decreased healthcare costs?

DGA indicates Dietary Guidelines for Americans.Note: Selected questions are presented as examples; this is not an exhaustive list.

Journal of Nutrition Education and Behavior � Volume 44, Number 4, 2012 Levine et al S7

As discussed earlier, a goal of theframework is to expand knowledgeof important intermediate outcomes,such as changes in self-efficacy forchoosing healthy food, or the numberand quality of organizational partner-ships that promote the messages thatsupport MyPlate. The intermediateoutcomes not only contribute to de-sired long-term outcomes, but theycan be identified as short-term cam-paign successes in and of themselves.The framework illustrates that com-bining process and outcome measureswidens the range of factors contribut-ing to program successes and failures.

Another overarching structure in-corporated into the framework is thesocio-ecological (SE) model, which isreferenced widely with respect toinfluences on food choices and wasnotably used in the development ofthe DGA. The SE model provided or-ganizing frameworks for systematicreviews conducted by Contento andby Thomson.7,33 Story et al basedapproaches to healthy foodenvironments on the SE model,34

and Medeiros et al used the SE modelto develop a logic model for commu-nity nutrition education.35 The inclu-sion of the SE model in the proposedevaluation framework for the MyPlatecommunications initiative acknow-ledges many levels of influenceaffecting food choices, including indi-vidual, household, and communitydynamics; work and school environ-ments; access to healthy food; poli-cies; and market forces. The SEmodel populates each of the compo-nents of the framework with exam-ples specific to the MyPlate initiative.

In arguing for the importance ofprogram theory, Contento presentsevidence gleaned from over 300nutrition-related communicationstudies, which suggest that ‘‘nutritioneducation is more likely to be effectivewhen it . systematically links rele-vant theory, research and practice.’’7

Thus, the major constructs withineach level of the framework are de-rived from theories of behaviorchange, including the Health BeliefModel,36 the Theory of Reasoned

Action,37 and the Theory of PlannedBehavior38 for studies at the personaland interpersonal level. Theories thatare useful at the levels of environmen-tal settings include Social CognitiveTheory and Diffusion of Innova-tions.39,40 Reciprocal determinism,a concept from Social CognitiveTheory that acknowledges continu-ous interactions between personalfactors and the environment, isillustrated in the framework (Figure)with arrows depicting multiple inter-actions that can be realized for favor-able behavior change to occur. Thisconcept was also noted by Contento,who emphasized that interactions be-tween biological, behavioral, and en-vironmental factors are at play indiet-related health.7 An expert panelassembled by the CDC also notedthat some mass media communica-tion efforts ‘‘reverberate at multiplelevels of the sociopolitical environ-ment’’ and thus may change the con-text within which individuals receiveand process campaign messages.6 Infact, as discussed below, altering social

S8 Levine et al Journal of Nutrition Education and Behavior � Volume 44, Number 4, 2012

norms and popular culture around di-etary attitudes, beliefs, and ultimatelybehaviors is a desired consequence ofthe MyPlate initiative.

At the systems and policy level,theories can be used to assess leverag-ing of public and political will. Thesetheories include Framing Theory, forwhich framing is defined as ‘‘the pro-cess by which people develop a partic-ular conceptualization of an issue orreorient their thinking about an is-sue.’’41 Framing theory can be usedwith other communication theoriesto examine how changes in the pre-sentation of an issue or an event canproduce (sometimes large) changesof opinion. Logic models, sometimescalled ‘‘theories of change models’’are also used at this level to help visu-alize how resources and strategies canbe used to achieve change.42

In order to help describe the rangeof possible outcomes that can resultfrom a communications initiativeas far-reaching as MyPlate, the RE-AIM dimensions of reach, efficacy,adoption, implementation, andmain-tenance developed by Glasgow andcolleagues could be used.43 The reachand efficacy dimensions capture de-sired impact at the personal/interper-sonal level. Reach refers to theproportion of the target audiencethat was exposed to DGA andMyPlate-related communications,and efficacy refers to the rate of ‘‘suc-cess,’’ which can include behaviorchange, among those who wereexposed if the guidelines were imple-mented as intended. The dimensionsof adoption and implementation cap-ture impact at the environmental andsystem/policy levels. Adoption refersto the proportion of settings, such asworksites, health departments, orcommunities that adopt theMyPlate-related messages. Implemen-tation refers to the extent to whichthe intervention is executed as in-tended in the real world. Maintenancerefers to the extent to which MyPlate-related messages and guidelines aresustained over time. This dimensionacts at the personal/interpersonallevel and at the systems and policylevel. All are crucial constructs forevaluating programs intended forwide-scale dissemination. Althoughthe evaluation framework is not de-signed to accommodate scoring orquantifying the outcomes of the

MyPlate communication initiative tocompare with other programs, as hasbeen suggested by Glasgow et al,43

RE-AIM dimensions, or similarmodelsthat focus on overall population-based impact, are useful in their fullassessment of strengths and limita-tions of public health interventions.

So how does one evaluate successfor an iconic communication initia-tive? A multifaceted framework ex-pands the paradigm for determiningcommunications ‘‘success’’ beyonda singular focus on positive behaviorchanges. In retrospect, MyPyramidcould have been said to be successfulin dimensions of reach to individualsand adaptation by institutions, butnot in other dimensions that areneeded to achieve the anticipatedoutcomes from increased dietaryquality. Although communicationcampaigns can contribute meaning-fully to long-term outcomes,10 thisframework illustrates that successmay also be found in small, positivechanges at each level of the socio-ecological model and within inter-mediate outcomes. Also, it is hopedthat the proposed evaluation frame-work will enhance evidence of suc-cess for the key purposes of healthcommunication initiatives—to raiseawareness, increase knowledge ofbenefits and risks, shape attitudes,heighten self-efficacy toward desiredactions, and motivate healthfulbehaviors.

The framework also helps identifyhow success for MyPlate communica-tions (or lack of success) depends onthe context of ‘‘upstream’’ factorsand the ‘‘downstream’’ effects. Up-stream factors are those that contrib-ute to adverse health practices, suchas low health literacy or poor socio-economic conditions, whereas down-stream refers to interventions thataim to change adverse behaviors.44

As discussed below in more detail, itis equally important to understandthe mediating effects of upstream fac-tors on downstream effects of MyPlatecommunications.

This explanation does not meanthat health communications havebeen without frameworks. Rather,frameworks can and must evolveover time, in the same way that be-havior change theories such as SocialCognitive Theory39 and Theory ofReasoned Action37 can be considered

evolutionary refinements of the basicknowledge-attitude-behavior model.

DEFINING EVIDENCE INEVIDENCE-BASEDHEALTHCOMMUNICATIONEVALUATION

Not surprisingly, there are more struc-tures and processes in place for ad-vancing nutrition science than thereare for the art and science of nutritioncommunication. Frameworks formedical science have evolved to opti-mize evidence-based findings. Severalorganizations for nutrition profes-sionals, such as the Academy of Nutri-tion and Dietetics, American Societyfor Parenteral and Enteral Nutrition,and health and medical researchagencies including the United StatesAgency for HealthCare Research andQuality, the Cochrane Collaboration,and the Natural Standard ResearchCollaboration, have publishedevidence-based practice guidelines.45-49

Recommendations from the DietaryGuidelines for Americans AdvisoryCommittee are, appropriately, evi-dence based. However, no such‘‘evidence-based’’ or even ‘‘practice-based’’ report exists as it relates to whatworks for communicating the DGAand/or MyPlate to change consumerdietary behavior. According to theroundtable discussion summarized byRowe et al, USDA and other govern-ment agencies ‘‘apply [the] equivalentof clinical judgment’’ to develop con-sumer guidance and implementationtools.8

Swinburn and colleagues recog-nized the importance of evidence-based/practice-based processes andcampaigns, but they also acknowl-edged the limitations of what is con-sidered acceptable evidence whenthey developed an evaluation frame-work for obesity prevention.50 CitingRychetnik et al and Kroke et al,51,52

they observed:

The term ‘evidence-based’—a termnow quite familiar to health pro-fessionals and policy makers—has become somewhat problematicbecause (i) it tends to be under-stood as referring only to frame-works used in ‘evidence-basedmedicine’ (EBM), which heavily

Journal of Nutrition Education and Behavior � Volume 44, Number 4, 2012 Levine et al S9

weights internal validity as the de-fining characteristic of evidence,and (ii) it largely ignores, andtherefore devalues, the importanceof external validity as well asa host of additional social, politi-cal, and commercial consider-ations that actually drive decisionmaking on policies and programs.

Similar calls for more comprehen-sive approaches and dimensions tobroaden ‘‘admissible’’ or ‘‘appropri-ate’’ evidence of effectiveness, includ-ing contextual and organizationalinformation, as well as informed opin-ion,43,50 have been noted outside ofhealth initiatives. An example of thistype of thinking to advance mathand science teaching was articulatedin the National Science Foundationpublication Footprints: Strategies forNon-Traditional Evaluation.53 The titleof this compilation refers to the elu-sive nature of evidence of a program'simpact. The introduction suggests thefollowing questions as indicators forknowing (p. 5):

How do we know we have collectedall the evidence? Where are thelikely places to look for missing ev-idence? For example: What are theuntouched areas of research?What is not being done or is beingdone ineffectively? Are there keytarget groups that are not beingserved or are being served inade-quately? What rival hypothesescan we formulate, e.g., wherewould we have been if this programdid not exist?

These questions are especially rele-vant for public health communica-tion initiatives like MyPlate, andthey support the need for an inclusiveframework that encourages applica-tion of a greater diversity of modelsand theories to health communica-tion and its evaluation. The CDC's ex-pert panel recommended exploringa wide range of theories, which pur-portedly are currently underused, in-cluding communication theories thataddress attention, attitudes, and sub-jective norms;54 theories of languagecomprehension and information pro-cessing;55 and theories examining theeffects of emotion on communica-tion.56 Social network theory57-59

and network analysis60 also hold un-realized promise for illuminating the

social dynamics within which com-munication operates. The ideationmodel,61 which takes into accountcognitive, emotional, and social fac-tors and is frequently used abroad,also has great potential use withinUS-based studies. Further, Davis sug-gests that the outcomes derived fromusing the Memorable Message Frame-work62 have the potential to enhancecampaign evaluation efforts and evenredefine ‘‘how we measure campaignsuccess.’’63 These theories and modelsare amenable for a use-focused evalua-tion approach, and their applicationscan fit within the proposed frame-work for evaluating the MyPlate com-munications initiative illustrated inthe Figure. Used collectively, theysupport the basic premise of healthcommunication theory as detailedearlier: that it is important to collectevidence of small changes occurringalong the pathway toward long-termgoals.

What More Should beMeasured, and in WhatContext?

Beyond recall of initiative messagesand components and reportedchanges in diet-related behaviors, theauthors suggest more attention tooverlooked measurement opportuni-ties that can provide evidence ofsuccess particular to the MyPlatecommunication effort.

Focus on social norms. Having animpact at the level of social normsshould be considered an important in-termediary outcome for the MyPlateinitiative, and its evidence shouldthus be identified as markers ofcampaign success. The icon status ofMyPlate is a strength of the initiativethat is challenging to measure, as the‘‘footprints’’ are often woven intothe very fabric of society (vis �a vispop culture and secular trends) ratherthan easily identified in traditionaldata sources. However, evidenceabounds of social normalization ofDGA. For example, Sesame Street'sCookie Monster singing that ‘‘ACookie is a Sometime Food’’ echoedthe ‘‘anytime and sometime food’’concepts associated with the Pyramidmessages and responded to a growingawareness that a popular children's

character eating only cookies was nolonger acceptable in a child obesityepidemic.64 In charting the pervasive-ness of the MyPlate icon resultingfrom efforts, researchers may considerindicators of cultural adjustment thatcan be used cumulatively for trackingnational sentiment: general refer-ences toMyPlate in pop culture (socialmedia and traditional media venuementions, including both intendedand unintended product placementwithin programming, etc); smart-phone and other electronic mediaapplications; active ‘‘counter-movement’’ activities and rhetoric(for example, Harvard School of Pub-lic Health's alternative to the USDA'sMyPlate, ‘‘Healthy Eating Plate’’);65

environmental scans to track in-creased depictions of meals that looklike the DGA; specific policy outcomes(eg, subsidization of healthy food andtaxation of unhealthy food); changesin sentiment toward agriculturalsubsidies and nutrition program regu-lations that do not reflect DGA;changes in food manufacturing prac-tices; and/or changes in fast-food res-taurant menus to be more inclusiveof DGA recommendations.

This list is by no means exhaustivefor possible footprints left by MyPlatein the culture, but hopefully theseideas can inspire creative means of ac-counting for the initiative's potentialto affect social norms.

Focus on vulnerable populations.Another area that is often overlookedwhen evaluating health campaignsfor the general public is the effective-ness with vulnerable populations. Itis hoped that research on the effec-tiveness of the MyPlate communica-tion initiative will be conductedamong populations in greatest needof healthier food consumption, in-cluding those with low health liter-acy, low socioeconomic status, and/or with language barriers. Few exam-ples of this focus on ‘‘upstream fac-tors’’ that mediate food choices canbe found in the literature; a notablestudy conducted by Zoellner et al inthe Lower Mississippi Delta illustratesthe need for greater attention tosocioeconomic status and culturalbackground, which mediate the effec-tiveness of campaigns like MyPlate.66

A survey of 177 adults from the region

S10 Levine et al Journal of Nutrition Education and Behavior � Volume 44, Number 4, 2012

revealed that only 12% of adults sur-veyed could identify the MyPyramidgraphic about 2 years after it was re-leased. The authors noted that the In-ternet was the least used and leasttrusted source of nutrition informa-tion among the respondents, yetMyPyramid was the cornerstone ofthe DGA communication campaign.However, more recent work indicatesthat some segments of the low-income population are using theInternet to search for nutrition infor-mation.67 Thus, several factors needto be taken into account to tailor com-munication strategies to the needsand preferences of vulnerable audi-ences. In terms of identifying the the-oretical underpinnings of a logicmodel for a campaign, one must alsoconsider whether factors of interestwill operate as predicted among eth-nic minority and/or limited resourcepopulations as they will among gen-eral audiences. If not, logic modelsand measures should be adjusted,and supplemental theories incorpo-rated (such as theories of culturalassimilation).6

Expand the range of strategic analy-ses. Improved understanding of theeffectiveness of MyPlate across theframework can be realized with analy-ses that account for latent or underly-ing relationships between factors.These methods include analysis ofupstream and downstream factors asmediators, moderators, covariates, orconfounders. All of these terms referto variables that help explain nuancesin the relationship between an inde-pendent variable and the dependentor outcome variable. A good reviewof these analyses as they pertain tonutrition research is presented byLockwood et al.68 Conjoint analysisis frequently used in marketing,69

but it may also be applicable to studiesof individual decision making aroundfood choices. Conjoint analysis as-sesses how people make tradeoffs insituations in which one has to chooseamong options with criteria of impor-tance, for example, nutrition, taste, orconvenience. Multilevel modeling,hierarchal linear analysis, or mixedmethods70 are analytical methodsthat account for group or nestingeffects, which is important in studiesfocusing on environmental settings.

At the systems and policy level,analytical methods such as socialnetworking, which uses a structuralapproach to illustrate patterns of link-ages between actors,58 can help assessthe interplay of institutions and pol-icy makers working for (and against)policy change amenable to the DGA.

IMPLICATIONS FORRESEARCH ANDPRACTICE

Assessing the effectiveness of nutri-tion communication initiatives is dif-ficult, even for narrowly targetedpopulations in controlled environ-ments like school classrooms. Initia-tives such as MyPlate, which arebroadcast to the general population,present myriad challenges to ade-quately capture measures of success.The USDA's CNPP has appealed to itspartners and academia to conductstudies and share findings that assesseffectiveness of this carefully plannedinitiative. The framework presentedhere takes this approach a step furtherby suggesting a version of professional‘‘crowd sourcing,’’71 inviting not onlywell-executed studies, but also docu-mentation of MyPlate's effectivenessin practice-based evidence and itsfootprint in policies, markets, andthe culture. By widening the net, it ishoped that health communicatorscan capture more evidence of effec-tiveness. The accumulated evidencecan provide more guidance than iscurrently available in communicatingDGA. Two important areas warrantingmore research and observation areMyPlate's message use in vulnerablepopulations, and signs of its impacton social norms.

The framework is untested—a ma-jor limitation—but it is composed ofvalidated social science characteris-tics, most notably the overallstructure of a logic model, the socio-ecological model, and constructsfrom theories of behavior change. Itis presented as a starting point to facil-itate hypothesis generation, study de-sign, indicator definition, and datacollection from disparate sources ofevidence for the purpose of docu-menting the effectiveness of MyPlate.It is expected that health communica-tion frameworks will continue toevolve with the progress of social sci-

ence, and the framework will also berefined as information on the perfor-mance of MyPlate communicationsbecomes available.

STATEMENT OFPOTENTIAL CONFLICT OFINTEREST

This supplemental article was fundedin part by Booz Allen Hamilton(authors EL, GRM, and JH) for themanuscript preparation. JAG works inconsultation with Booz Allen Hamil-ton. The opinions expressed are thoseof the authors and do not necessarilyrepresent the views or recommenda-tions of their respective institutions.

FUNDING/SUPPORT

This work was funded in part by BoozAllen Hamilton. The Society for Nutri-tion Education and Behavior, whichpublishes the Journal of Nutrition Edu-cation and Behavior, is a National Part-ner of the MyPlate CommunicationInitiative.

ACKNOWLEDGMENTS

Our thanks also to Dr. Timothy Tinkerand Jim Bender of Booz AllenHamilton, whose review of earlierdrafts provided helpful edits. Finally,the authors thank staff from theUSDA, Center for Nutrition Policyand Promotion, who reviewed anearlier version of the manuscript foraccuracy in describing the MyPlateCommunications Initiative.

REFERENCES

1. United StatesDepartment ofAgriculture:Center for Nutrition Policy and Promo-tion. MyPlate Background. http://www.cnpp.usda.gov/Publications/MyPlate/Backgrounder.pdf. Accessed January6, 2012.

2. Post R. A new approach to DietaryGuidelines communications: makeMyPlate, your plate. Child Obes. 2011;7:349-351.

3. United States Department of Agricul-ture, Center for Nutrition Policy andPromotion. USDA 2010 DietaryGuidelines Communication MessageCalendar. http://www.cnpp.usda.gov/Publications/MyPlate/Communications

Journal of Nutrition Education and Behavior � Volume 44, Number 4, 2012 Levine et al S11

MessageCalendar.pdf. Accessed March30, 2012.

4. Mendoza M. Nutrition education inef-fective. USA Today July 4, 2007; http://www.usatoday.com/news/health/2007-07-04-fightingfat_N.htm. AccessedJanuary 4, 2012.

5. Hornik RC. Public health communica-tion: making sense of contradictory ev-idence. In: Hornik RC, ed. PublicHealth Communication: Evidence for Be-havior Change. Mahwah, NJ: LawrenceErlbaum Associates Inc; 2002:1-22.

6. Abbatangelo-Gray J, Cole GE,Kennedy MG. Guidance for evaluatingmass communication health initiatives:summary of an expert panel discussionsponsored by the Centers for DiseaseControl and Prevention. Eval HealthProf. 2007;30:229-253.

7. Contento IR. Nutrition education: link-ing research, theory, and practice. AsiaPac JClinNutr. 2008;17(suppl 1):176-179.

8. Rowe S, Alexander N, Almeida NG,et al. Translating the Dietary Guidelinesfor Americans 2010 to bring about realbehavior change. J AmDiet Assoc. 2011;111:28-39.

9. Williams AF, Wells JK. The role of en-forcement programs in increasing seatbelt use. J Safety Res. 2004;35:175-180.

10. Snyder LB. Health communicationcampaigns and their impact on behav-ior. J Nutr Educ Behav. 2007;39(2suppl):S32-S40.

11. Hornik R, Kelly B. Communicationand diet: an overview of experienceand principles. J Nutr Educ Behav.2007;39(2 suppl):S5-S12.

12. Britten P, Haven J, Davis C. Consumerresearch for development of educa-tional messages for the MyPyramidFood Guidance System. J Nutr Educ Be-hav. 2006;38(6 suppl):S108-S123.

13. Haven J, Burns A, Britten P, Davis C.Developing the consumer interface forthe MyPyramid Food Guidance Sys-tem. J Nutr Educ Behav. 2006;38(6suppl):S124-S135.

14. International Food Information CouncilFoundation. 2011 Food & Health Sur-vey. Consumer Attitudes Toward Nu-trition, Food Safety & Health. http://www.foodinsight.org/Content/3840/2011%20IFIC%20FDTN%20Food%20and%20Health%20Survey.pdf. Ac-cessed January 2, 2012.

15. Krebs-Smith S. Diet Quality of Ameri-cans in 1994-96 and 2001-02 asMeasuredby the Healthy Eating Index-2005.Nutrition Insight 37. http://www.cnpp.usda.gov/Publications/Nutrition

Insights/Insight37.pdf.Accessed January5, 2012.

16. Krebs-Smith SM, Guenther PM,Subar AF, Kirkpatrick SI, Dodd KW.Americans do not meet federal dietaryrecommendations. J Nutr. 2010;140:1832-1838.

17. Carollo K, ABC News Medical Unit.No more Pyramid: nutritional icon isnow a plate. http://abcnews.go.com/Health/food-pyramid-food-plate/story?id¼13743137. Accessed January 5, 2012.

18. JalonickMC,NBCnews. Foodpyramiddumped for ‘‘MyPlate.’’ http://www.msnbc.msn.com/id/43253092/ns/health-diet_and_nutrition/t/food-pyramid-dumped-my-plate/. Accessed January5, 2012.

19. United States Department of Agricul-ture: Center for Nutrition Policy andPromotion. Development of 2010Dietary Guidelines for Americans Con-sumer Messages and New Food Icon.Executive Summary of Formative Re-search. June 2011; http://www.choosemyplate.gov/food-groups/downloads/MyPlate/ExecutiveSummaryOfFormativeResearch.pdf. Accessed January4, 2012.

20. Donato KA. National health educationprograms to promote healthy eatingand physical activity. Nutr Rev. 2006;64(2 Pt 2):S65-S70.

21. Pivonka E, Foerster SB, Di Sogra L,Chapelsky Massimilla D. IndustryInitiatives 5 A Day for Better Health Pro-gram. Bethesda,MD:National Institutesof Health, National Cancer Institute;2001.

22. Reich MR. Public-private partnershipsfor public health. Nat Med. 2000;6:617-620.

23. Brownell KD, Ludwig DS. FightingObesity and the Food Lobby. TheWashington Post June 9, 2002. http://www.washingtonpost.com/ac2/wp-dyn/A15232-2002Jun7?language¼printer.Ac-cessed January 6, 2012.

24. UnitedStatesDepartmentofAgriculture,Center for Nutrition Policy and Promo-tion. ChooseMyPlate.gov PartneringProgram. http://www.choosemyplate.gov/partnering-program.html. AccessedFebruary 11, 2012.

25. Department of Health and Human Ser-vices, Centers for Disease Control andPrevention, National Center for HealthStatistics, National Health and Nutri-tion Examination Survey. NationalHealth andNutrition Examination Sur-vey: questionnaires, datasets, and relateddocumentation. http://www.cdc.gov/

nchs/nhanes/nhanes_questionnaires.htm.Accessed March 30, 2012.

26. United States Department of Agricul-ture, Center for Nutrition Policy andPromotion. Healthy Eating Index.http://www.cnpp.usda.gov/healthyeatingindex.htm. Accessed March 30, 2012.

27. United States Department of Agricul-ture, Centers for Disease Control andPrevention, Office of Surveillance, Ep-idemiology, and Laboratory Services,Public Health Surveillance ProgramOffice. Behavioral Risk Factor Surveil-lance System. http://www.cdc.gov/brfss/. Accessed March 30, 2012.

28. Centers for Disease Control andPrevention, Porter Novelli. Healthstyles survey. http://www.cdc.gov/healthcommunication/ToolsTemplates/EntertainmentEd/HealthstylesSurvey.html. Accessed March 30, 2012.

29. Gallup Inc. Gallup Healthways Well-Being Index. https://www.npd.com/wps/portal/npd/us/industryexpertise/foodandbeverage/. Accessed March30, 2012.

30. NPD Group. Food and beverage mar-ket research. http://www.npd.com.Accessed March 30, 2012.

31. Ostrom E. A general framework for an-alyzing sustainability of social-ecologicalsystems. Science. 2009;325:419-422.

32. Earp JA, Ennett ST. Conceptual modelsfor health education research and prac-tice. Health Educ Res. 1991;6:163-171.

33. Thomson C. Fruits, vegetables and be-havior change: a scientific overview.Produce for Better Health Foundation.http://www.pbhfoundation.org/about/res/pbh_res/. Accessed May 23, 2012.

34. Story M, Kaphingst KM, Robinson-O’Brien R, Glanz K. Creating healthyfood and eating environments: policyand environmental approaches. AnnuRev Public Health. 2008;29:253-272.

35. Medeiros LC, Butkus SN, Chipman H,Cox RH, Jones L, Little D. A logicmodel framework for community nu-trition education. J Nutr Educ Behav.2005;37:197-202.

36. Becker MH. The Health Belief Modeland Personal Health Behavior. San Fran-cisco, CA: Society for Public HealthEducation; 1974.

37. Fishbein M. Readings in Attitude Theoryand Measurement. New York, NY: Wi-ley; 1967.

38. Ajzen I. The Theory of Planned Behav-ior. Organ Behav Hum Decis Process.1991;50:179-211.

39. Bandura A. Social Foundations ofThought and Action: A Social Cognitive

S12 Levine et al Journal of Nutrition Education and Behavior � Volume 44, Number 4, 2012

Theory. Englewood Cliffs, NJ: PrenticeHall; 1986.

40. Rogers EM.Diffusion of Innovations. 4thed. New York, NY: Free Press; 1995.

41. Austin L, Mitchko J, Freeman C,Kirby S, Milne J. Using framing theoryto unite the field of injury and violenceprevention and response: ‘‘AddingPower to Our Voices.’’ Soc Mar Q.2009;15(S1):35-54.

42. Harvard Graduate School of Education.Harvard Family Research Project. Auser’s guide to advocacy planning. http://www.hfrp.org/evaluation/publications-resources. Accessed April 6, 2012.

43. Glasgow RE, Vogt TM, Boles SM.Evaluating the public health impact ofhealth promotion interventions: theRE-AIM framework. Am J PublicHealth. 1999;89:1322-1327.

44. KellyMP,McDaid D, Ludbrook A, Po-well J. Economic appraisal of publichealth interventions, Briefing paper.http://www.nice.org.uk/aboutnice/whoweare/aboutthehda/hdapublications/hda_publications.jsp?o¼707. AccessedJanuary 22, 2012.

45. Academy of Nutrition and Dietetics.Evidence Analysis Library. http://www.adaevidencelibrary.com/default.cfm?auth¼1. Accessed January 6, 2012.

46. Agency for Healthcare Research andQuality. Evidence-based practice.http://www.ahrq.gov/clinic/epcix.htm.Accessed January 6, 2012.

47. American Society for Enteral andParenteral Nutrition. Clinical Guide-lines. http://www.nutritioncare.org/Professional_Resources/Guidelines_and_Standards/Guidelines_and_Standards_Library_Features/. Accessed January6, 2012.

48. Natural Standard Research Collabora-tion. Natural Standard database.http://www.naturalstandard.com/. Ac-cessed January 6, 2012.

49. The Cochrane Collaboration. CochraneReviews. http://www.cochrane.org/cochrane-reviews. Accessed January6, 2012.

50. Swinburn B, Gill T, Kumanyika S.Obesity prevention: a proposed frame-work for translating evidence into ac-tion. Obes Rev. 2005;6:23-33.

51. Rychetnik L, Hawe P, Waters E,Barratt A, Frommer M. A glossary forevidence based public health. J EpidemiolCommunity Health. 2004;58:538-545.

52. Kroke A, Boeing H, Rossnagel K,Willich SN. History of the concept of‘‘levels of evidence’’ and their currentstatus in relation to primary preventionthrough lifestyle interventions. PublicHealth Nutr. 2004;7:279-284.

53. Frechtling JA: National Science Foun-dation (US). Footprints: Strategies forNon-Traditional Program Evaluation.Arlington, VA: National Science Foun-dation; 1995.

54. Petty RE, Cacioppo JT. Issue involve-ment can increase or decrease persua-sion by enhancing message-relevantcognitive responses. J Pers Soc Psychol.1979;37:1915-1926.

55. McGuire W. Theoretical foundationsof campaigns. In: Rice R, Paisley W,eds. Public Communication Campaigns.Thousand Oaks, CA: Sage Publica-tions; 1981:41-70.

56. Brehm JW. A Theory of PsychologicalReactance. New York, NY: Holt, Rine-hart & Winston; 1966.

57. Boulay M, Valente TW. The selection offamily planning discussion partners inNepal. JHealthCommun. 2005;10:519-536.

58. Luke DA, Harris JK. Network analysisin public health: history, methods, andapplications. Annu Rev Public Health.2007;28:69-93.

59. Valente TW, SabaWP.Massmedia andinterpersonal influence in the BoliviaNational Reproductive Health Cam-paign. Communic Res. 1998;25:96-124.

60. Kincaid DL. From innovation to socialnorm: bounded normative influence. JHealth Commun. 2004;9(suppl 1):37-57.

61. Kincaid DL. Mass media, ideation, andbehavior: a longituduinal analysis ofcontraceptive change in the Philippines.Communic Res. 2000;27:723-763.

62. Knapp ML, Stohl C, Reardon KK.‘‘Memorable’’ messages. Journal of Com-munication. 1981;31:27-41.

63. Davis LA. Beyond ‘5-a-Day’: An Exami-nation of Memorable Messages in a Nutri-tion Education Program for Low-IncomeAfrican Americans [dissertation]. WestLafayette, IN: PurdueUniversity; 2011.

64. Newsweek. Sesame Street: the showthat counts. http://www.thedailybeast.com/newsweek/2009/05/22/sesame-street.html. Accessed March 30, 2012.

65. Harvard Health Publications. Compari-son of the Healthy Eating Plate andthe USDA’s MyPlate. http://www.health.harvard.edu/plate/comparison-of-

healthy-eating-plate-and-usda-myplate.Accessed May 6, 2012.

66. Zoellner J, Connell C, Bounds W,Crook L, Yadrick K. Nutrition literacystatus and preferred nutrition commu-nication channels among adults in theLower Mississippi Delta. Prev ChronDis. 2009;6:1-11.

67. Neuenschwander LM, Abbott A,Mobley AR. Assessment of low-income adults’ access to technology:implications for nutrition education.J Nutri Educ Behav. 2012;44:60-65.

68. Lockwood CM, DeFrancesco CA,Elliot DL, Beresford SA, Toobert DJ.Mediation analyses: applications in nu-trition research and reading the litera-ture. J AmDiet Assoc. 2010;110:753-762.

69. Green PE, Srinivasan V. Conjoint anal-ysis in consumer research: issues andoutlook. J Consum Res. 1990;5:103-123.

70. Fisher JO, Mennella JA, Hughes SO,LiuY,MendozaPM,PatrickH.Offering‘‘dip’’ promotes intake of a moderatelyliked raw vegetables among preschoolerswithgenetic sensitivity tobitterness. JAmDiet Assoc. 2012;112:235-245.

71. Howe J, WIRED. The rise of crowd-sourcing. http://www.wired.com/wired/archive/14.06/crowds.html?pg¼1&topic¼crowds&topic_set¼. Ac-cessed January 15, 2012.

72. Department of Health and Human Ser-vices, Centers for Disease Control andPrevention, National Center for HealthStatistics, National Health andNutritionExamination Survey. Flexible Con-sumer Behavior Module. http://www.cdc.gov/nchs/data/nhanes/nhanes_09_10/FCBS_f.pdf. Accessed March30, 2012.

73. Centers for Disease Control and Pre-vention. State-Added Question Data-base, 1995-1997. https://www.ark.org/adh_brfss_questions/results.aspx. Ac-cessed January 4, 2012.

74. United States Department of Healthand Human Services, Centers for Dis-ease Control and Prevention. YouthRisk Factor Surveillance System Sur-vey Questionnaire, 2009. http://www.cdc.gov/healthyyouth/yrbs/index.htm.Accessed January 4, 2012.

75. UCLA Center for Health Policy Re-search. California Health InterviewSurvey CHIS, 2012. http://www.chis.ucla.edu/main/PUF/default.asp. Ac-cessed January 2, 2012.