elaboration likelihood model: a missing intrinsic emotional … · 2017-08-23 · human history. it...

human history. It was approximately2,400 years ago that Aristotle clearlyidentified the three main aspects of the

INTRODUCTIONThe roots of persuasion are embeddeddeeply in the human psyche as well as in

� Palgrave Macmillan Ltd 1479-1862/05 $30.00 Vol. 14, 1, 79–98 Journal of Targeting, Measurement and Analysis for Marketing 79

Elaboration likelihood model:A missing intrinsicemotional implicationJon D. Morrisis a professor in the Department of Advertising, College of Journalism and Communications, University of Florida. He beganteaching at the university in 1984. Previous to his academic positions, he worked in the advertising agency business. Hebegan his advertising career in 1968, and has worked for several agencies, including Nicholson-Morris, Louisville, KY, andDoyle Dane Bernbach and Dancer Fitzgerald Sample, New York, NY. He earned his PhD from the University of Florida. Hisresearch has appeared in the Journal of Advertising Research, Educational Technology, the International Journal ofInstructional Media, and in the Proceedings of the American Academy of Advertising, among others. Over the past few years,he has been developing a model (AdSAM) for analysing emotional response to marketing communications.

ChongMoo Woois a doctoral student majoring in advertising in the Department of Advertising at the University of Florida. He received a BAand MA, both in advertising and public relations from the Chung-Ang University, Seoul, South Korea, and an MA inadvertising from the University of Florida. During the programme of study, he has completed course work toward an MA instatistics and has published in the Journal of Advertising Research, the Journal of Current Issues and Research in Advertising,Advances in Consumer Research, and others.

A. J. Singhcompleted his BA in Commerce in 1994 at the University of Mumbai and then obtained an MA in economics from the sameuniversity in 1997. He received his MA in mass communication with a specialisation in advertising in May 2003 from theUniversity of Florida. He is currently employed in Delhi, India, as Account Manager and Producer for the Online andInteractive Department of the Discovery Channel.

Abstract The Elaboration Likelihood Model (ELM), developed by Petty and Cacioppo,proposed two routes to attitude change: central and peripheral. The central routeemphasises a high relevance of the message to the individual. In the peripheral route,the individual concentrates on heuristic cues like attractive expert sources and numberrather than the content of arguments employed by the message to process themessage. If these cues produce an attitude change, this change is likely to be shorterlasting and unpredictable of that individual’s behaviour. Hence, the cognitive (central)aspect of the ELM overshadows its affective (peripheral) aspect, and the underlyingsuggestion of this model is that an attitude change is mostly reached through cognitionas opposed to emotion. This study attempts to show that the emotional aspect is asimportant as the cognitive aspect. The basis for this conclusion is that even as anindividual processes a message cognitively, that cognition has an emotional core. Inaddition, there is a possibility that content processing (elaboration) gives rise toemotions and that this leads to a longer-lasting change in attitudes.

Keywords: Elaboration Likelihood Model, cognition, emotion, structural equationmodelling, consumer behaviour, Self-Assessment Manikin, PAD(pleasure-arousal-dominance)

Jon D. MorrisUniversity of FloridaCollege of Journalism andCommunicationsPO Box 118400Gainesville, FL 32611-8400USA

Tel: �1 (352) 219 6000;�1 (352) 392 0443;E-mail: [email protected]

‘the most distinctive and indispensableconcept in contemporary socialpsychology’8 is that of a variable knownas attitude. The importance of theattitude concept stems from the fact thatit is believed to be a mediating variablefor knowledge acquisition as well asbehavioural change.9

Given the importance of attitude andattitude change in mediating behaviour,many social psychologists and sociologistshave tried to define attitude. As a result,there are a multitude of definitions ofthis term.10,11 Social psychologists,however, reached a sort of consensus thatthe term should be used to refer to ‘ageneral and enduring positive or negativefeeling about some person, object orissue’.12,13 The study of attitude changehas become a source of great debate andassumed primary focus for persuasionresearchers.

A change in affect (negative orpositive) as a result of a change incommercial stimuli has recently beenshown to be linked to behaviour andbehavioural changes,14 contradicting aprevious number of studies thatquestioned the ability of affect to predictbehaviour.15 In fact, the mid-1970s werefilled with disillusionment towards theaffective attitude concept. As a result, inthat period, there was a sharp decline inthe study of attitudes by socialpsychologists,16 even though Fishbein andAjzen17–20 have been able to conclude,with confidence, that attitude andbehaviours are strongly linked.

According to Fishbein and Ajzen,what distinguishes attitude from otherconcepts is its strongly affective natureand that ‘affect is the most essential partof the attitude concept’.21 A strong linkbetween attitudes and subsequentbehaviour had already been examined,albeit cognitively, by Petty and Cacioppoin their Elaboration Likelihood Model(ELM).22

persuasion situation: source, audience andmessage content.

Given this support for the long historyof persuasive processes, it is surprisingthat it was only 60 years ago that anysort of systematic research attention wasapplied to the study of persuasion. CarlHovland, a Yale psychologist, has beenattributed with the creating ‘the modernexperimental study of persuasion’.1

Systematic persuasion research, however,can be linked to media effects researchthat began with Walter Lippmann2 andHarold Lasswell.3

The problem with this research wasthat it was based on anecdotal evidenceand not empirical research.4 Then, in the1940s and 1950s, there was a shift inresearchers’ thinking, as evidenceincreasingly pointed to an indirect effectsmodel of persuasion. Hyman andSheatsely5 suggested that a mere increasein message flow could not achievepersuasion and that effective messagedissemination requires consideration ofspecific psychological barriers. This led tothe idea of a ‘two-step’ flow ofcommunication as suggested by Katz andLazarsfeld,6 who argued that the mediatend to influence opinion leaders, who inturn influence the public.

The persuasion debate was thenbrought to another level by the researchof Hovland et al.,7 who documented theimportance of different factors inpersuading people to accept aproposition. Subsequent research onpersuasion has focused on the moderatingvariables, such as knowledge and beliefs,to develop contemporary models ofpersuasion.

As with any important human process,the presence of both affect (or feelings)as well as information processing can befelt permeating the research onpersuasion. Although persuasionresearchers are split into the camps ofcognition and affect, they all agree that

80 Journal of Targeting, Measurement and Analysis for Marketing Vol. 14, 1, 79–98 � Palgrave Macmillan Ltd 1479-1862/05 $30.00

Morris, Woo and Singh

undertaken to examine the central routeto persuasion and the role of affect versuscognition. The basis for this conclusion isthat even as an individual processes amessage cognitively, the cognition has anemotional core.

CONCEPTUALISATION

Emotion: The affective neuroscience

The role of emotion was largely ignoreduntil the 1970s; indeed, there was aprevailing opinion that emotions, for themost part, ‘disrupt and disorganisebehaviour and are primarily a source ofhuman problems’.28 Naturally, there werescientists who opposed the prevailingview regarding emotions and supportedthe idea that emotions played a centralrole in behavioural changes that weresaid to represent learning.29 Despite thissupport, emotion remained an absententity in the classic works of cognitivescience and cognitive neuroscience.30

Recent research conducted in the fieldof affective neuroscience elevatedemotions from the status of merephysiological responses, by identifyingregions of the brain — specifically, theamygdala, ventromedial prefrontal,brain-stem nuclei, hypothalamus andbasal forebrain — as areas responsible forprocessing different emotions to varyingdegrees.31 Most of these regions aresubcortical, referring to their locationbelow the cerebral cortex, an area that isof primary importance to cognitiveneuroscientists. The proposition cited inthe ELM, stating that informationprocessing is a separate and more directroute to persuasion than affect, appears tobe contradicted by these physiologicalfindings.

The limbic, or primitive/subcorticalbrain system, was hypothesised to be theseat of emotion and the cortex, or thehigher brain, the seat of cognition;

This lack of support for affect as adominant figure in the creation ofpersuasion most likely led to the assertionin the ELM23 that there are two routesto attitude change: central and peripheral.The central route emphasises highrelevance of the message to theindividual. The greater the relevance andthe more interest that the individualshows in the subject of the message,usually a product, the higher the chancesthat they will think or elaborate on themessage. Another aspect of the centralroute is that it deals with the messagecontent — text, words, written materialused in the message — as opposed to theperipheral route that deals with themessage cues — colour use,people/lifestyle depiction, visuals, etc.24

This thinking seems to eliminate thepossibility that emotional processing maybe as important in the central route as inthe peripheral route. Why does cognitiveprocessing have to be void of emotionalinfluence?

The peripheral route is said to be thepath taken when the message had littleor no relevance to its receiver.25 Theindividual concentrates on heuristic cuessuch as attractive expert sources andnumbers rather than the content ofarguments employed by the message inorder to process it. According to theELM, if these cues produce an attitudechange, this change is likely to be shorterlasting and unpredictable of thatindividual’s behaviour.26 Hence, theinformation processing (central) aspect ofthe ELM overshadows the affective(peripheral) aspect, and the underlyingsuggestion of this model is that anattitude change is mostly throughinformation processing as opposed toemotion.

Given recent findings, however,highlighting the importance of affectover information processing in predictingbehaviour,27 the current study was


Elaboration likelihood model

of emotions, they also lost the abilityto make rational decisions.39

Affective neuroscience, at the turn ofthe millennium, has firmly entrencheditself in the brain. Drawing from thefindings of forerunners such as Damasioand Ledoux, researchers have establishedthe fact that the brain circuitry ofemotion and cognition is not separate.40

Data have shown that no parts of thebrain are dedicated exclusively tocognition or affect. Emotion evolved tofacilitate an organism’s adaptation to thecomplex challenges it faced during itspast,41 and it is hard-wired in thebrain.42 It has been conclusively shownthat the architecture of the brain doesnot honour the age-old concept ofsegregation of cognition and affect.Most compellingly, cognition has beenshown to be rudderless withoutemotion, and studies in cognitiveneuroscience and behavioural sciencecannot be conducted without takingemotion into account.43

ELM: The cognitive response hierarchy

Since the 1950s, a great deal of researchhas been conducted to understand,describe and predict consumers’attitudinal responses to advertising.44

According to Petty and Cacioppo,45 mosttheories of attitude change that exist as aresult of research conducted in the 1950scan be grouped into seven majorapproaches, namely:

— conditioning and modellingapproaches that focus on the directeffects of rewards and punishments;

— the message-learning approach thatfocuses on an individual’s attention to,understanding and acceptance of apersuasive message;

— judgmental approaches that focus onpast experiences’ impacts onattitudinal judgments;

however, neuroscientific experimentsdisproved this theory.32 Certainsubcortical structures that were previouslythought only to regulate emotion havenow been shown to be intimatelyconnected to cognition as well (iehippocampus for memory). By contrast,certain cortical structures that wereexclusively the province of cognition (iethe prefrontal cortex) have been shownto be involved in emotionalprocessing.33–35

Neuroscientists have also beenattempting to solve one of the mostenduring controversies of social science— the primacy hypothesis. Thishypothesis attempts to identify theprimacy of affect over cognition or viceversa. Joseph Ledoux’s experimentssupport the primacy of affect hypothesisas expounded by Zajonc.36 Ledoux’sexperiments involving fear conditioningin rats conclusively showed that fear, oneof the basic emotions, is processed in thealmond-shaped, subcortical structure ofthe brain, the amygdala. The amygdalanot only processes the emotion of fear,but also does so without the aid ofconscious thought.37

Antonio Damasio is another pioneerin the field of affective neuroscience.His work with patients with disturbedemotions, as a result of frontal lobedamage shows that, although individualsmay have complete control over theirthought or logical processes, they donot have the same control over theiremotions, and in fact their verysurvival is endangered.38 In addition,Damasio experimentally proved thatreasoning and decision making cannotexist without the accompaniment ofemotion. In studies of patients whowere entirely rational until the timethat neurological damage affected areasof the brain involved in emotionalprocessing, it was seen that not onlydid these individuals lose a certain class



— favourable attitudes toward theproduct;

— preference over alternative products;and

— conviction that the purchase will bewise.

The formulation of these six steps thenlaid the basis for further models andresearch studies on how advertisingcommunication works (eg Krugman,51

Houston and Rothschild,52 Vaughn,53

Petty et al.,54 etc). These hierarchymodels have proved to be conceptuallyuseful in addition to being accepted bymany advertising academicians andpractitioners.55 Among the many modelsspawned by the response hierarchy, theELM received special attention becauseof its cognitive base and widespreadusage in a variety of disciplines.

Although a number of cognitivetheories of attitude change have beendeveloped over the years, the one thatprovides the foundation for the ELM isthe communication/persuasion matrixmodel.56,57 The McGuire model containsa few drawbacks58,59 that havepurportedly been addressed by thecognitive response theory developed byGreenwald60 and Petty et al.61 Thecognitive response theory, too, has had itsshare of drawbacks. The major criticismis that it focuses only on individuals whoprocess messages actively and failed totake into account individuals who do notthink actively about the information theyreceive.62 The ELM seems to fill this gapin the cognitive response theory bysuggesting that there are two routes topersuasion: (1) the route taken byindividuals who think actively aboutinformation in a message — that is, thecentral route — and (2) the route takenby individuals who do not think activelyabout the informational content of amessage — that is, the peripheral route.

The central route emphasises

— motivational approaches that examinethe relationship between humanmotives and attitudinal changes;

— attributional approaches that focus onthe relationship between inferencesabout self or others and attitudinalchanges;

— combinatory approaches that focus onthe integration of information into anoverall attitude; and

— self-persuasion approaches that focuson self-generated information in thepresence or absence of a persuasivemessage.

Although these seven approaches utilisedifferent variables and different processes,‘they really seem to indicate that thereare only two fundamentally different‘‘routes’’ to changing a person’sattitudes’.46 These two routes, central andperipheral, form the basis of the ELM,which, according to its authors, is anattempt to outline a framework that‘takes one step toward a general theoryof attitude change’.47

The ELM belongs to a family ofmodels described as the responsehierarchy models. These models attemptto depict the changes that a consumerundergoes when they move from a stateof relative unawareness about a certainbrand or product to relative awarenessand then, finally, to purchaseintentions/behaviour.48 One of the first(and perhaps the best known) responsehierarchy models was the ‘hierarchy ofeffects’ (HOE) model.49 According toLavidge and Steiner,50 there are six stepsfollowed by consumers before actualpurchasing, namely:

— complete unawareness of the existenceof a product or service;

— mere awareness of the existence ofthe product or service;

— knowledge about what the productoffers;



attitude formation and change bykeeping it on the periphery of thatprocess. The model was developed at atime when the role of emotions incommunication was not clearlyunderstood by social scientists. Thisthinking seems to eliminate thepossibility that emotional processing maybe as important in the central route as inthe peripheral route. Why does cognitiveprocessing have to be void of emotionalinfluence?

The resurgence of interest inemotions66–69 has given rise to newresearch and provided eye-opening resultsof the role that emotions play in anindividual’s life. Hence, it becomesnecessary to re-examine the ELM in thelight of these findings and wonderwhether affect should be kept at theperiphery of such a widely used modelof attitude formation and change ormoved to centre.

The peripheral route was said to bethe path taken when the message hadlittle or no relevance to its receiver.70

The individual concentrates on heuristiccues such as attractive expert sources andnumbers rather than the content ofarguments employed by the message inorder to process it.

Current research indicates thatemotional reactions influence bothsystematic and heuristic processing. Itshows, for example, that positive emotionsinfluence peripheral informationprocessing while negative emotionsinfluence central processing.71–74

Researchers have then gone a step further,claiming that there may be no central orperipheral cues and that the two routes topersuasion can interact with each otherwith one route dominating in the face ofcontradictory information and both routesoccurring concurrently when theprovided persuasive information does notcontradict the attitude structure of anindividual.75

cognitive/effortful thinking and minimalemotional involvement in processingmessages. This path to persuasioninvolves ‘paying careful attention to therelevant information in the message andrelating that information to previousknowledge stored in memory andgenerating new implications for theinformation’.63 This process of cognitiveresponses to a message is considered thelikelihood that an individual willelaborate on a particular message. Thisroute, however, demands that a messagehas high relevance to a particularindividual, and this individual shouldhave the motivation and ability toprocess the arguments present in themessage. Individuals who travel this routemay be referred to as ‘cognitiveelaborators’.

The peripheral route is almost theexact opposite of the central route topersuasion. Individuals who travel thisroute are referred to as ‘cognitivemisers’64 and process messages based onsimple affective cues such as sourceattractiveness and message length. Theindividual takes this route in a lowrelevance/low motivation situation andprocesses the message at a very heuristiclevel by observing the audio-visual cuespresent in the message and withoutthinking too much of the argumentspresent therein. Hence, the peripheralroute is characterised by ‘an absence ofeffortful message elaboration’ and theattitude formations and changesengendered by this route are ‘lessaccessible, persistent, resistant andpredictive of behaviour’ as comparedwith attitudes developed by the centralroute.65

Emotion and ELM

As a model of cognitive responses, theELM attempts to account for thepresence of affect in the process of



purchase intention of aconsumer in the attitude changeprocess outlined by the ELM:cognition or affect?

RQ3: Does emotion play a morecentral role in the attitudechange process outlined by theELM?

METHOD

Research design

Following the process outlined in theELM, where the direct route topersuasion was defined as a focus ofattention on the product’s features andbenefits and the peripheral route isdefined as a focus on the executionalvariables of the communications, therespondents in this study were split intotwo groups, labelled cognitive elaboratorsor cognitive misers. The two groupswere formed on the basis of theparticipants’ spontaneous reactions to twocar advertisements. If the spontaneousfirst reaction contained a reference to theexecutional variables — that is theelements of the commercial such as thesetting or characters — their reactionclassified the respondents as cognitivemisers. Likewise, if the first reactioncontained reference to the product andits features, then the respondents wereclassified as cognitive elaborators. Bydefinition, the cognitive elaboratorswould focus on the information andhave minimal emotional involvementwhile the cognitive misers would focuson the execution with an absence ofeffortful elaboration. The data setsobtained from these two groups wereanalysed to determine whether cognitionor affect dictated the intended futurepurchasing behaviour of the respondents.This, in turn, helped to determinewhether it was justified to keep emotionat the periphery of the attitude-change

Subsequent research supported this lineof thought, and while investigating theinfluence of print advertisements on theaffective and cognitive responses ofadolescents, Edens and McCormick76

discovered that ‘many adolescents wereunable to detect the explicit claim of anadvertisement yet maintained that the ad‘‘made sense’’, which suggests thatperipheral visual (or emotional)information becomes the centralmessage’. Such research, therefore,provides a strong case for placing affectat the centre of the ELM.

Research questions

Advertising rests on the fundamentalbelief that a consumer can be persuadedto purchase a given product.77 The ELMis a theory of persuasion that is widelyused in a disparate array of disciplines toexplain persuasion as a result ofcognition. It is clear from a review ofthe model, that the ELM has a cognitivebase.78 What is also clear is that themodel dispenses of affect as ‘a simpleperipheral cue’ that causes a lesspersistent change in attitude andconcentrates on promoting cognition asthe main force behind a longer-lastingchange in attitude.79

The existing theory of affect gives thisstudy a twofold purpose: to examine theELM from an emotional standpoint andto determine whether emotions are acomponent in the central route topersuasion, thus showing that a change inattitude and the resultant behaviour is dueto affective as well as cognitive drives.

RQ1: Would cognitive elaboratorsdisplay positive emotionalresponse to advertising, and, ifso, would this response besignificantly different fromcognitive misers?

RQ2: What has a greater effect on the



The data sets that were generated as aresult of this testing were then examinedwith special emphasis on the variable ofspontaneous comments. The questionsthat isolate this variable for analysis were:What went through your mind as you watchedthis commercial? What were your thoughts asyou watched it? In the given data sets, thereactions to this question were evaluatedand recoded as cognitive elaborators andcognitive misers. AdSAM�, a non-verbalmeasure that assesses emotional response,was used in the questionnaire in order togauge the emotional reactions of theparticipants on three dimensions: pleasure,arousal and dominance. The emotionalresponses to the car commercial(s)generated by AdSAM� were then used inthe analysis of the recoded data with thepurchase intention. Based on the internaldata of the motor company that sponsoredthe research, the time frame of thepurchase intention was set to be twoyears. The participants were asked toreport their intention to purchase a newvehicle within the next two years.Additional qualifications includeddemographics such as age, annualhousehold income and education, andbehaviours such as driving frequency (seesample).

Coding procedure

Three coders (two males and one female)independently coded the spontaneouscomments variable in accordance withthe following set of rules. First, if thecomments of a participant were relatedto the product and its features, thatreaction would be considered a cognitiveresponse and the participant would beplaced in a group labelled cognitiveelaborators. Secondly, if the spontaneouscomments of a participant were relatedto the execution/setting of thecommercial, the response would beconsidered non-cognitive and the

process as defined by the ELM or give ita more central role.

Therefore, the hypotheses addressed bythe study were:

H1: The emotional response (ER) ofthe cognitive elaborators asmeasured by the AdSAM�emotional response pleasure,arousal and dominance (PAD)scale will be significantly greaterthan that of the cognitive misers.Thus, ElabER > MiserER.

H2: Emotional response and purchaseintent are directly related to eachother. Thus, the higher the PADscore, the higher the purchaseintent of that group.

To evaluate the effectiveness of amini-campaign of two automobileadvertisements that depicted a mid-sizeluxury sedan, mall intercepts wereconducted in the cities of Atlanta,Chicago, Cleveland, Dallas, Los Angeles,Melbourne, Minneapolis, New York,Orlando, Philadelphia, Phoenix and SanFrancisco. The participants were paid towatch a series of advertisements and thento respond to questions about them.

The copy testing of the carcommercials intended to measure theperformance of each commercial on thebasis of four key dimensions:impact/brand name registration,comprehension, evaluation and intendedresponse. The participants were exposedto a clutter of seven commercials, and thecar commercial was placed in the centreof the clutter. The idea was to test: recallthrough clutter and main ideas throughclutter. In order to test a number ofvariables, including spontaneouscomments (thoughts and feelings),emotional response (AdSAM) andpurchase intention, an isolatedre-exposure to the car commercial(s) wasthen conducted.



qualified participants towards whom thecar advertisements were targeted. Theywere not measures of involvement.

According to the ELM, it ismotivation that essentially determines anindividual’s choice of the central or theperipheral route to persuasion.82

Although there is a strong link betweenmotivation and involvement, there isevidence that involvement’s impact onmotivation is limited. It has been foundthat ‘[i]n most instances, consumersexperience situational (temporary)involvement with an offering or activity’rather than enduring involvement.83

Additionally, Cacioppo et al. argued that‘[t]he fact that the situational factors thathave been used to manipulate the extentto which attitudes are based onissue-relevant thinking sometimesaccount for only a small portion ofvariance is theoretically due, in part, tosystematic individual differences amongpeople in their desire to engage inissue-relevant thinking when theyformulate their attitudes’.84 Further, theyargued that one of the systematicindividual differences is need forcognition and individuals low in need forcognition are cognitive misers.

85Without

measuring or controlling involvement,these researchers86 examined thepersuasion routers travelled by individualslow in need for cognition (cognitivemisers) and high in need for cognition(cognitive elaborators) and reported amoderating effect of need-for-cognitionon argument quality’s influence onattitude.

Instrumentation

The non-verbal measure of emotionalresponse, the Self-Assessment Manikin(SAM)87 and AdSAM�,88 which visuallymeasures the dimensions of PAD with agraphic character arrayed along acontinuous 9-point scale, was used in

participant would be placed in a grouplabelled cognitive misers.

Intercoder reliability was calculatedusing Holsti’s formula80 for reliabilityamong coders:

Reliability � 3M/N1 � N2 � N3.

The M denotes the number of codingdecisions the coders agree on, whereasN1, N2 and N3 represent the totalnumber of coding decisions made byeach of the three coders. The threecoders compared the results of theircoding for the sample. The inter-coderreliability was found to be 99 per cent.

Sample

The research sample was obtained from alarge marketing communications firm andwas part of a comprehensive copy test. Atotal of 254 participants (50 per centmale and 50 per cent female) from majorUS cities were extracted for the test ofthe car advertisements.

Participants in the study were screenedto ensure that the sample was involvedin the car buying and had motivation aswell as ability to process the givenpersuasive communication message ofhigh involvement/thinking product.81

These qualifications were: age 35–54years (50 per cent 35–44 years, 50 percent 45–54 years), with a householdincome of $50,000+ and a minimum ofsome college education. They needed tobe a principal driver and primary/shareddecision-maker of a 1992 or newervehicle (not a van or sport utility),bought new, who intended to purchase anew, not used, vehicle from a qualifyingvehicle set within the next two years.They had to be willing to consider bothdomestics and imports and qualify on akey psychographic battery.

As noted earlier, thesocio-demographics were used to screen



series of three recall questions, they werethen asked to identify their feelingsassociated with the test commercial usingAdSAM�, Using this non-verbal process,respondents were asked to indicate ‘howthe commercial made them feel’. Onaverage, they took from 15 to 30seconds to answer.

RESULTS

Descriptive statistics

The results from the comparison of thetwo groups, cognitive elaborators andcognitive misers, were obtained byexamining for differences among fourmeasured dependent variables. Thedescriptive statistics are shown in Table 1.

The descriptive statistics in Table 1provide an interesting insight into thevariation of spontaneous reaction andshow that the emotional reactions asmeasured by AdSAM’s PAD score areconsistently higher for the cognitiveelaborators than for the cognitive misers.On all three major emotional dimensions(PAD) the cognitive elaborators havehigher mean scores than the cognitivemisers. The mean scores were convergedon a 9-unit bipolar scale.

Tests of hypotheses: Emotional coreof cognition

The first hypothesis (H1) stated that theemotional response of the cognitive

this study. The PAD bipolar dimensionsconsist of pleasure/displeasure,arousal/non-arousal anddominance/submissiveness. Thesedimensions adequately describe the fullspectrum of human emotions. They havebeen reliably measured and are sufficientto define all emotional states.89 PAD hasgained popularity due to its simplicityand the ability to cover the full range ofhuman emotions.90

The use of AdSAM� as aninstrument helped to overcomeproblems associated with cumbersomeand time-consuming verbal self-reportmeasures. SAM and AdSAM� weredeveloped to adequately representRussell and Mehrabian’s PAD. It uses agraphic character that eliminates muchof the biases associated with verbal andother non-verbal measures.91 Thegraphic character is easily understoodand identifiable by both adults andchildren.92 In addition, SAM reducesrespondent wear-out and is not limitedby age, gender, culture or languagedifferences.93 AdSAM� has also provento be an effective and efficientinstrument to determine emotionalresponses to advertisements94 and, mostimportantly, it accurately measures therespondents’ feelings to anadvertisement.95

In these copy tests of a mini-campaignof two car advertisements, AdSAM� wasplaced near the beginning of theinterview. After respondents were asked a



Table 1: Descriptive statistics of measurement items: mean, standard deviation and responses

Group Variable Mean SD N

Cognitive elaborators Pleasure 7.20 1.61 113Arousal 6.22 2.10 113Dominance 6.19 1.93 113Purchase intent 3.52 1.02 113

Cognitive misers Pleasure 6.43 1.95 141Arousal 5.80 2.48 141Dominance 5.69 2.31 141Purchase intent 3.08 1.07 254

significantly higher emotional loadingthan the cognitive misers.97

For the MANOVA result to be valid,several underlying assumptions werechecked prior to interpretation. Theseassumptions included multivariatenormality, linearity among dependentvariables and equal variance forindependent group.98 The normality waschecked on each six marginal distributionusing both Shapiro-Wilk test statistics(N < 2,000) for normality and visualnormal QQ plot inspection. The sampledata generally trended the straight

elaborators would be significantly greaterthan that of the cognitive misers. To testthis hypothesis, a multi-group LISRELmeasurement model and a multivariateanalysis of variance (MANOVA) wereconducted.

As shown in Figure 1 and Table 2,the emotional responses are highly loadedon cognitive elaborators and cognitivemisers and meet thegoodness-of-model-fit criteria.96 Inaddition, the likelihood ratio test (LRT),using chi-square statistics, endorses theidea that the cognitive elaborators have



Figure 1 Multi-group LISREL measurement model with standardised factor loadings

Table 2: LISREL goodness-of-model-fit test statistics

SRMR-CFI LISREL Model fit criteriaGroup SRMR (<0.09) CFI (>0.95)

Cognitive elaborators 0.07 0.91Cognitive misers 0.00 1.00

SRMR-CFI, standardised root mean square residual and comparative fit index; CFI, comparative fit index

Cognitive misers

Cognitive elaborators

Arousal DominancePleasure

0.84 0.55 0.49

Emotion

Arousal DominancePleasure

Emotion

0.39 0.28 0.24

were higher for the cognitive elaboratorsthan for the cognitive misers. The resultsof the MANOVA and the LISRELmodel confirm the hypothesis that thecognitive elaborators do indeed show asignificant level of response over thecognitive misers.

Tests of hypotheses: Explanatory powerof affect

RQ2 and H2 suggest that affect may havea great influence on the purchaseintention of a consumer in the attitudechange process. As shown in Figure 2,the path coefficient of the cognitivemisers is greater than that of cognitiveelaborators.

To locate the origin of explanatorypower of affect in LISREL, asupplementary simple regression analysiswas performed100 for each factor ofemotion, namely PAD (Table 4). Theresults show that pleasure and dominanceare the significant indicator variables forthe explanatory power of affect(cognitive misers) as well as cognition(cognitive elaborators). Previous researchresults have indicated that emotional

diagonal line of QQ plot showingnormality. The multivariate andunivariate homoscedasticity calculatedthrough Box’s M (M � 22.86, p � 0.00)and Levene’s test statistics(F � 2.82–6.77, p � 0.04–0.10) assertedunequal variance in some dependentvariables; however, any erratic shape ortrends was not found in residual plotinspection and, moreover, group cellsizes were sufficiently balanced.99

Bartlett’s test of sphericity (�2 � 74.6,p � 0.00) was significant at the 0.05significance level and implies that there isa linear relationship among thePAD-dependent variables.

Table 3 shows the MANOVA andsubsequent analysis of variance (ANOVA)analysis results. After linearlydecomposing the effects, a significantmain effect of spontaneous responsegrouping was found at the 0.05significance level (Wilk’s � � F � 3.98,p < 0.05). As per the analysis (see Table2), the cognitive elaborators show asignificantly higher level of pleasure(p < 0.05) than the cognitive misers, and,although not significantly different, themeans scores on arousal and dominance



Table 3: The effects of SR group (elaborators vs misers) on the linear combination of pleasure, arousal anddominance

Analysis Dependent variable F p-value

MANOVA Pleasure-arousal-dominance 3.98 (df � 3, Wilk’s � � 0.95) 0.01ANOVA Pleasure 11.49 (df � 1) 0.00

Arousal 1.89 (df � 1) 0.17Dominance 2.96 (df � 1) 0.09

ANOVA Purchase intention 23.39 (df � 1) 0.00

MANOVA, multivariate analysis of variance; ANOVA, analysis of variance.

Table 4: Pleasure, arousal and dominance simple unstructured regression on the purchase intention

Group Indicator B t-statistics p-value

Cognitive SR Pleasure 0.04 0.64 0.53Arousal 0.02 0.49 0.63Dominance 0.15 2.80 0.01

Affective SR Pleasure 0.18 3.62 0.00Arousal 0.00 0.01 0.99Dominance 0.00 0.04 0.97

SR, simple unstructured regression.

to each other and thus, that the higherthe PAD scores, the higher thepurchase intent of that group, anadditional ANOVA analysis on thepurchase intention was annexed to theprevious MANOVA and confirmatoryfactor analysis results. According to theANOVA analysis result (see Table 3),the cognitive elaborators (3.52 meanpurchase intention) would be morelikely to contact a dealer than thecognitive misers (3.08 mean purchaseintention). This result was shown to beof statistical significance (see Table 2,p < 0.05). Hence, H2 was supportedbecause the cognitive elaborators, whichshowed a higher emotional response tothe commercial, also showed a higherPI.

Figure 3 illustrates that the cognitiveelaborators’ responses have a moreemotional core compared with thecognitive misers. The cognitive

reactions influenced both central(systematic) and peripheral (heuristic)processing, specifically, positive emotionsinfluenced affective peripheralinformation processing and negativeemotions influenced cognitive centralprocessing.101–104 In this case, the moresystematic group, the cognitiveelaborators, emoted more than theheuristic group, the cognitive misers.This supports the contention thatcognition and affect are an importantpart of evaluation and decision making,and even those individuals who focus onthe more cognitive information might bethe same group that also reacts viscerallyto the stimulus.

Tests of hypotheses: Emotion andpurchase intention

To test H2, that emotional responseand purchase intent are directly related



Figure 2 Multi-group LISREL hybrid model with standardised factor loadings and path coefficients

Pleasure

0.84 0.55 0.49

DominanceArousal

Central route

---- Peripheral route

0.22

Purchase intent


Emotion

DominancePleasure Arousal

0.39 0.28

0.31

Cognitive misers

0.24

Emotion

Purchase intent

DISCUSSIONThis study has found that affect is morethan just a peripheral occurrence in theprocess of persuasion as suggested in theELM. Indeed, the results contradict thenotion that affect is the product ofperipheral cues in the advertisementexecution more than the factualinformation related to the product.According to the ELM those respondentswho focus on the style of the commercial,the cognitive misers, would have thestronger emotional responses (but lowerpurchase intent) than the cognitiveelaborators. The cognitive misers whofocused on the commercial style ratherthan the product attributes did have lowerpurchase intent scores but they also hadlower PAD scores upon seeing thecommercials. It was the cognitiveelaborators who had higher PAD scores aswell as higher purchase intent scores.

elaborators also have a higher purchaseintention score than cognitive misers.Consequently, the elaborators, or thosewho are said to have selected the moredirect route to persuasion, have a higheremotion loading and purchase intentionscore than the misers or those who aresaid to have chosen the peripheral routeto persuasion.

The enlightened consumer’s purchaseintention fluctuated with the degree ofemotion, and the PAD scores werehigher for those citing product featuresrather than those citing executionalvariables. This is in contrast to thetheory of direct and peripheral route topersuasion as outlined by the ELM. Afterthe respondents were exposed to thecommercials, affective reaction showed astrong link to purchase intent, but wasstronger in the cognitive elaboratorsgroup.



Figure 3 Significant relative loading and path patterns of structural equated ELM

High (�)

Low (�)

(High �)

Pleasure DominanceArousal

Purchase intent


Emotion

Central route

---- Peripheral route

Low (�)

Cognitive misers

Pleasure DominanceArousal

Purchase intent

Emotion

(Low �)

(High �)

Purchase intent is significantly higher forthe cognitive elaborators than for thecognitive misers. Due to the cognitiveelaborators showing a higher level ofemotional response when compared withcognitive misers, the cognitive elaboratorsemote more and can be considered thehigher emotional group. This would thensupport the hypotheses: cognitiveelaborators may generate a higheremotional response than cognitive misersand purchase intent is related to emotionalresponse; the higher emotional response,the stronger the purchase intention.

These hypotheses were based onresearch that gave evidence regarding theimportance of affect, and as a result, itwas extrapolated that this concept couldnot be kept on the periphery of theELM. The data analysis seems to supportthis theory by providing significant proofof the existence of emotion in thecognitive sphere. In addition, by linkingpurchase intent with affect, support isalso provided for an attitude change viaan affective route that could be at thecentre of the ELM. This is furthersupport that emotional response is astrong driver of the intent

It is important to determine theemotional response as accurately aspossible. AdSAM� is an excellent toolfor measuring viewer’s responses.AdSAM� is a quick, valid and reliablemethod of testing,107–110 and it does soacross cultures and ethnic groups,without the problem of subjectivemeaning or other biases. It is anexcellent tool for dissecting emotionalresponses, and it provides invaluableinsights that help to highlight the role ofemotion in a persuasive setting such astelevision advertising.

Limitations

Although the limitations do notovershadow the findings, the sample used

The results of the data analysis givefurther insight105 into the role of emotionsin the process of persuasion. In fact, thereis some suggestion in the discussion of theELM that a higher affective responsewould interfere with the cognitiveprocessing and therefore produce lowerpersuasion scores.106 The acceptance ofthe two hypotheses in this studydemonstrates that cognitive processing hasan emotional core. Of interest is the factthat the more cognitive group, thecognitive elaborators, shows significantlyhigher pleasure responses than the lesscognitive group, the cognitive misers. Ahigher purchase intent accompanies thisincrease in pleasure. These results give aninsight into the importance of affect andsuggest that the ELM should be refined toinclude the role of emotions in the directroute to persuasion and the relationship ofaffect to behavioural intent.

The measurement model conductedbetween the two groups reveals asignificant presence of affect in thecognitive elaborators. This demonstratesthat even though information is beingprocessed cognitively, the procedure is notbeing conducted in an emotional vacuum.These traces of emotion in cognitiveprocessing provide support for thenecessity of re-examining the role ofemotion in the ELM and not simplyassigning it to a peripheral role.

Although significant differences werefound for one emotional variable —pleasure — the other two emotionalvariables, arousal and dominance, showhigher mean scores in the cognitiveelaborators group. This finding isinteresting as it suggests that even incognitive elaboration, there is a certainamount of emotional reaction and thatthis reaction is significantly different fromthe reactions of the supposedly moreaffective cognitive miser group.

The behavioural intentionsmeasurement supports these findings.



identifying whether emotion or cognitionwas the main driving force behind thepurchase.

The ELM is widely used in a numberof disciplines and has increased inpopularity ever since its inceptionbecause of its largely cognitive base thatattempts to explain the persuasive powerof a message in terms of thoughtelaboration. Thus, the ELM seems to beemphasising cognition over emotion inthe process of persuasion. This study hasattempted to show that the basic tenetthat the ELM proposes may be incorrectand that affect needs to be given a muchmore significant role and not be kept onthe periphery of the ELM.

Research in neuro scientific sphereshas re-affirmed the importance ofemotion, and the results of this studyseem to support this research. It is hopedthat this study will add further insight tothe role that affect plays in a persuasiveprocess.

References1 Petty, R. E. and Cacioppo, J. T. (1986a)

‘Communication and persuasion: Central andperipheral routes to attitude change’,Springer-Verlag, New York, NY.

2 Lippmann, W. (1922) ‘Public opinion’, Macmillan,New York, NY.

3 Lasswell, H. (1927) ‘Propaganda techniques in theWorld War’, Peter Smith, New York, NY.

4 Petty, R. E. and Priester, J. R. (1994) ‘Mass mediaattitude change: Implications of the elaborationlikelihood model of persuasion’ in Bryant, J. andZillmann, D. (eds) ‘Media effects: Advances intheory and research’, Erlbaum, Hillsdale, NJ, pp.91–122.

5 Hyman, H. H. and Sheatsley, P. B. (1947) ‘Somereasons why information campaigns fail’, PublicOpinion Quarterly, Vol. 11, pp. 412–423.

6 Katz, D. and Lazarsfeld, P. R. (1955) ‘Personalinfluence’, Free Press, New York, NY.

7 Hovland, C.I., Lumsdaine, A. and Sheffield, F.(1949) ‘Experiments on mass communication’,Princeton University Press, Princeton, NJ.

8 Allport, G. (1935) ‘Attitudes’, in Murchison, C.(ed.) ‘Handbook of social psychology’, ClarkUniversity Press, Worcester, MA, p. 2.

9 Petty and Priester (1994) op. cit.10 Kiesler, C. A., Collins, B. E. and Miller, N. (1969)

‘Attitude change: A critical analysis of theoretical

for the study was restricted to a certaindemographic and product categories.This restricts generalisation of the resultsof this study beyond high-involvement,durable goods.

In addition, the data were recoded onthe basis of simple rules specially devisedfor the purpose of the study. These rulesmay need to be modified in order toconduct a more in-depth analysis of theparticipant responses. Furthermore, thenumber of coders involved limited thecoding procedure. The perception of thediverse responses may differ from coderto coder.

The purchase intent variable, too, canbe considered a limitation as there is noactual proof that a respondent from the‘emotional’ group actually did gothrough with the purchase.

Suggestions for future research

Future research may attempt to includedifferent demographic groups as well asdifferent media, such as print and radio,and also different types of products andcampaigns within this framework. Thereare a multitude of products and productclasses available that could be consideredfor research purposes. In addition, amore rigorous statistical analysis can beemployed to reveal even more significantresults between ‘cognitive’ and‘emotional’ groups.

It would be interesting to conduct astudy on other highly priced consumergoods that would require as considerablepre-purchase cognition on the part of theconsumer and identify the amount ofemotion involved in the purchase. Tomake this study even more intriguing, itcould be worthwhile to ask a consumerthe thoughts that went through theirmind immediately prior to and after thehigh-price purchase. These responsescould then be examined for theiremotional content and this could help in



American Psychology, Vol. 39, pp. 117–123.37 Ledoux, J. E. (1995) ‘Emotion: Clues from the

brain’, Annual Review of Psychology, Vol. 46, pp.209–235.

38 Damasio (1994) op. cit.39 Damasio, A. R. (1999) ‘The feeling of what

happens: Body and emotion in the making ofconsciousness’, Harcourt Brace, New York, NY, pp.41–61.

40 Davidson (2000) op. cit.41 Tooby, J. and Cosmides, L. (1990) ‘The past

explains the present: Emotional adaptations and thestructure of ancestral environments,’ Ethology andSociobiology, Vol. 11, pp. 375–424.

42 Ledoux (1995) op. cit.43 Davidson (2000) op. cit.44 Cohen, J. B. (1990) ‘Attitude, affect and consumer

behavior’ in Moore, B. S. and Isen, A. M. (eds)‘Affect and social behavior’, Cambridge UniversityPress, Cambridge, UK, pp. 152–206.

45 Petty and Cacioppo (1996) op. cit.46 Ibid.47 Ibid.48 Belch, G. E. and Belch, M. A. (1998) ‘Advertising

and promotion: An integrated marketingcommunications perspective’, 4th edn, McGraw-Hill,New York. NY.

49 Lavidge, R. C. and Steiner, G. (1961) ‘A model forpredictive measurements of advertising effectiveness’,Journal of Marketing, Vol. 25 (October), pp. 59–62.

50 Ibid.51 Krugman, H. E. (1965) ‘The impact of television

advertising: Learning without involvement’, PublicOpinion Quarterly, Vol. 29, pp. 349–356.

52 Houston, M. J. and Rothschild, M. L. (1978)‘Conceptual and methodological perspectives oninvolvement’ in Jain, S. C. (ed.) ‘Research frontiersin marketing: Dialogues and directions’, AmericanMarketing Association, Chicago, IL, pp. 184–187.

53 Vaughn, R. (1983) ‘How advertising works: Aplanning model’, Journal of Advertising Research, Vol.23, pp. 22–28.

54 Petty, R. E., Cacioppo, J. T. and Schumann, D. W.(1983) ‘Central and peripheral routes to advertisingeffectiveness: The moderating role of involvement’,Journal of Consumer Research, Vol. 10, pp. 134–148.

55 Preston, I. L. (1982) ‘The association model of theadvertising communication process’, Journal ofAdvertising, Vol. 11, No. 2, pp. 3–15.

56 McGuire, W. J. (1985) ‘Attitudes and attitudechange’ in Lindzey, G. and Aronson, E. (eds)‘Handbook of social psychology’, 3rd edn, RandomHouse, New York, NY, p. 2, pp. 233–346.

57 McGuire, W. J. (1989) ‘Theoretical foundations ofcampaigns’ in Rice, R. E. and Atkin, C. K. (eds)‘Public communication campaigns’, 2nd edn, Sage,Newbury Park, CA, pp. 43–65.

58 Petty, R. E., Baker, S. M. and Gleicher, F. (1991)‘Attitudes and drug abuse prevention: Implications ofthe Elaboration Likelihood Model of persuasion’ inDonohew, L., Sypher, H. E. and Bukoski, W. J.(eds) ‘Persuasive communication and drug abuse

analyses’, John Wiley and Sons, New York, NY.11 Petty, R. E. and Cacioppo, J. T. (1996) ‘Attitude

and persuasion: Classic and contemporaryapproaches’, Westview Press, Boulder, CO.

12 Insko, C. A. and Schopler, J. (1972) ‘Experimentalsocial psychology’, Academic Press, New York, NY.

13 Petty and Cacioppo (1996) op. cit.14 Morris, J. D., Woo, C. M., Geason, J. A. and Kim,

J. (2002) ‘The power of affect: Predicting intention’,Journal of Advertising Research, Vol. 42 (May/June),pp. 7–17.

15 Corey, S. M. (1937) ‘Professed attitudes and actualbehavior’, Journal of Educational Psychology, Vol. 28,pp. 271–280.

16 Lambert, J. (1980) ‘Social psychology’, Macmillan,New York, NY.

17 Fishbein, M. and Ajzen, I. (1974) ‘Attitudes towardobjects as predictors of single and multiplebehavioral criteria’, Psychological Review, Vol. 81, pp.59–74.

18 Fishbein, M. and Ajzen, I. (1975) ‘Belief, attitude,intention and behavior: An introduction to theoryand research’, Addison-Wesley, Reading, MA.

19 Ajzen, I. and Fishbein, M. (1977) ‘Attitude-behaviorrelations: A theoretical analysis and a review ofempirical research’, Psychological Bulletin, Vol. 84, pp.888–918.

20 Ajzen, I. and Fishbein, M. (1980) ‘Understandingattitudes and predicting social behavior’, PrenticeHall, Englewood Cliffs, NJ.

21 Fishbein and Ajzen (1975) op. cit.22 Petty and Cacioppo (1986a) op. cit.23 Ibid.24 Ibid.25 Petty, R. E., Cacioppo, J. T., Strathman, A. J. and

Priester, J. R. (1994) ‘To think or not to think:Exploring two routes to persuasion’ in Shavitt, S.and Brock, T. C. (eds) ‘Persuasion: Psychologicalinsights and perspectives’, Allyn and Bacon, Boston,MA, pp. 113–147.

26 Petty and Cacioppo (1986a) op. cit.27 Morris et al. (2002) op. cit.28 Izard, C. E. (1991) ‘The psychology of emotions’,

Plenum Press, New York, NY.29 Mowrer, O. H. (1960) ‘Learning theory and

behavior’, Wiley, New York, NY.30 Neisser, U. (1967) ‘Cognitive psychology’, Prentice

Hall, Englewood Cliffs, NJ.31 Damasio, A. R. (1994) ‘Descartes’ error: Emotion,

reason and the human brain’, Grosset/Putnam, NewYork, NY, pp. 18–19.

32 Davidson, R. J. (2000) ‘Cognitive neuroscienceneeds affective neuroscience (and vice versa)’, Brainand Cognition, Vol. 42, pp. 89–92.

33 Nauta, W. H. (1971) ‘The problem of the frontallobe: A reinterpretation’, Journal of PsychiatricResearch, Vol. 8, pp. 167–187.

34 Damasio, A. R. (1989) ‘The brain binds entities andevents by multiregional activation from convergencezones’, Neural Computation, Vol. 1, pp. 123–132.

35 Davidson (2000) op. cit.36 Zajonc, R. B. (1984) ‘On the primacy of affect’,



behavior’, 3rd edn, Houghton Mifflin, Boston, MA,p. 57.

84 Cacioppo, J. T., Petty, R. E., Kao, C. F. andRodriguez, R. (1986) ‘Central and peripheral routesto persuasion: An individual difference perspective’,Journal of Personality and Social Psychology, Vol. 51,No. 5, pp. 1032–1043.

85 Ibid.86 Ibid.87 Lang, P. J. (1980) ‘Behavioral treatment and

bio-behavioral assessment: Computer applications’in Sidowski, J. B., Johnson, J. H. and Williams,T. A. (eds) ‘Technology in mental health caredelivery systems’, Ablex, Norwood, NJ, pp.119–137.

88 Morris, J. D., Bradley, M. M., Waine, C. A. andLang, J. B. (1992) ‘Assessing affective reactions toadvertisements with the self-assessment manikin(SAM)’. Paper presented at Southern MarketingAssociation Conference.

89 Mehrabian, A. and Russell, J. A. (1977) ‘Evidenceof a three-factor theory’, Journal of Research inPersonality, Vol. 11, pp. 273–294.

90 Havlena, W. J. and Holbrook, M. B. (1986) ‘Thevarieties of consumption experience: Comparingtwo typologies of emotion in consumer behavior’,Journal of Consumer Research, Vol. 13, pp. 394–404.

91 Morris, J. D. and Waine, C. A. (1993) ‘Managingthe creative effort: Pre-production andpost-production measures of emotional’ in Thorson,E. (ed.) ‘Proceedings of the 1993 Conference ofthe American Academy of Advertising’, Columbia,MO, pp. 158–176.

92 Morris, J. D., Woo, C. M. and Cho, C.-H. (2003)‘Internet measures of advertising effects: A globalissue’, Journal of Current Issues and Research inAdvertising, Vol. 25 (Spring), pp. 27–43.

93 Morris and Waine (1993) op. cit.94 Morris et al. (1992) op. cit.95 Morris et al. (2003) op. cit.96 Jreskog, K. G. and Srbom, D. (1997) ‘LISREL 8:

Structural equation modeling with the SIMPLIScommand language’, Scientific SoftwareInternational, Chicago, IL.

97 Embretson, S. E. and Reise, S. P. (2000) ‘Itemresponse theory for psychologists’, LawrenceErlbaum Associates, Mahwah, NJ.

98 Hair, J. F. Jr, Anderson, R. E., Tatham, R. L. andBlack, W. C. (1998) ‘Multivariate data analysis’,5th edn’, Prentice Hall, Upper Saddle River, NJ.

99 Johnson, R. A. and Wichern, D. W. (1992)‘Applied multivariate statistical analysis, 3rd edn’,Prentice Hall, Englewood Cliffs, NJ.

100 Rawlings, J. O., Pantula, S. G. and Dickey, D. A.(1998) ‘Applied regression analysis: A research tool’,2nd edn’, Springer-Verlag, New York, NY.

101 Batra and Staynman (1990) op. cit.102 Bohner and Apostolidou (1994) op. cit.103 Bohner et al. (1994) op. cit.104 Kuykendall and Keating (1990) op. cit.105 Morris et al. (2002) op. cit.106 Petty and Cacioppo (1986a) op. cit.

prevention’, Erlbaum, Hillsdale, NJ, pp. 71–90.59 Petty and Priester (1994) op. cit.60 Greenwald, A. G. (1968) ‘Cognitive learning,

cognitive response to persuasion, and attitudechange’ in Greenwald, A. G., Brock, T. andOstrom, T. (eds) ‘Psychological foundations ofattitudes’, Academic Press, New York, NY,pp. 147–170.

61 Petty, R. E., Ostrom, T. M. and Brock, T. C.(1981) ‘Cognitive responses in persuasion’, Erlbaum,Hillsdale, NJ.

62 Petty and Priester (1994) op. cit.63 Petty and Cacioppo (1996) op. cit.64 Taylor, S. E. (1981) ‘The interface of cognitive and

social psychology’ in Harvey, J. H. (ed.) ‘Cognition,social behavior and the environment’, Erlbaum,Hillsdale, NJ.

65 Petty and Cacioppo (1996) op. cit.66 Damasio (1999) op. cit.67 Ledoux (1995) op. cit.68 Davidson (2000) op. cit.69 Morris et al. (2002) op. cit.70 Petty et al. (1994) op. cit.71 Batra, R. and Staynman, D. M. (1990) ‘The role of

mood in advertising effectiveness’, Journal ofConsumer Research, Vol. 17, pp. 203–214.

72 Bohner, G. and Apostolidou, W. (1994) ‘Mood andpersuasion: Independent effects of affect before andafter message processing’, Journal of Social Psychology,Vol. 134, pp. 707–709.

73 Bohner, G., Chaiken, S. and Hyundadi, P. (1994)‘The role of mood and message ambiguity in theinterplay of systematic and heuristic processing’,European Journal of Social Psychology, Vol. 24, pp.207–221.

74 Kuykendall, D. and Keating, J. P. (1990) ‘Mood andpersuasion: Evidence for the differential influence ofpositive and negative states’, Psychology and Marketing,Vol. 7, pp. 1–9.

75 Chaiken, S. and Maheswaran, D. (1994) ‘Heuristicprocessing can bias systematic processing: Effects ofsource credibility, argument ambiguity, and taskimportance on attitude judgment’, Journal ofPersonality and Social Psychology, Vol. 66, pp. 460–473.

76 Edens, K. M. and McCormick, C. B. (2000) ‘Howdo adolescent process advertisements? The influenceof ad characteristics, processing objective and gender’,Contemporary Educational Psychology, Vol. 25,pp. 450–463.

77 Ibid.78 Petty and Cacioppo (1986a) op. cit.79 Ibid.80 Holsti, O. R. (1969) ‘Content analysis for the social

sciences and humanities’, Addison-Wesley Publishing,Reading, MA.

81 Vaughn (1983) op. cit.82 Petty, R. E. and Cacioppo, J. T. (1986b) ‘The

elaboration likelihood model of persuasion’ inBerkowitz, L. (ed) ‘Advances in experimental socialpsychology (Vol. 19)’, Academic Press, New York,NY, pp. 123–205.

83 Hoyer, W. D. and Maclnnis, D. J. (2004) ‘Consumer



• main ideas• spontaneous reactions (thoughts and

feelings)• elements found confusing• elements found hard to believe• AdSAM�• likes/dislikes• commercial profile — a battery of

attributes designed to get at thetonality of the commercials

• overall commercial rating• uniqueness of message• overall sedan rating• image scan — a battery of attributes

designed to determine whether explicitor implicit objectives are beingcommunicated

• increased interest rating• visitation intent• driver profile• slogan/tagline evaluation• character evaluation• music evaluation• overall profile battery• vehicle ownership

APPENDIX C: ADSAM�

The structural model of LISRELindicates that although motivated (intendsto purchase a new, not used, vehiclefrom a qualifying vehicle set within thenext two years) ability (age 35–54 yearswith $50,000+ household income)sample respondents see relevant (must beprincipal driver and primary/shareddecision maker of 1992 or newervehicle, bought new) car (highinvolvement/think product category inFoote Cone & Belding [FCB] quadrant)ad message, they strongly followed bigand wide peripheral route to purchaseintention. The peripheral affective routeis not always peripheral per se in theELM and it becomes the centralmessage.76

107 Lang (1980) op. cit.108 Morris et al. (1992) op. cit.109 Morris and Waine (1993) op. cit.110 Morris et al. (2002) op. cit.

APPENDIX A: ADVERTISEMENTSCommercial 1: ‘Listen’. This commercialcompares the sedan with its Europeancounterpart by letting the owner of theEuropean counterpart car test drive thesedan on a mountain road. The extensivebranding of the make and modelthroughout the execution emphasise thename of the sedan. The performance ofthe sedan thoroughly impresses theEuropean car owner and the commercialfocuses on the low price of the sedan bycommunicating it as a ‘whisper’ betweenfriends. The overall effect of thecommercial is to provide a perception ofaffordability without offering a stickerprice.

Commercial 2: ‘Side Road’. Thiscommercial communicates a morefocused message of performance andhandling of the sedan. By usingdescriptions of the sedan’s engine andcreating an exciting imagery ofperformance coupled with upbeat music,the commercial attempts to communicatethe superiority of the sedan as comparedwith its European counterparts. Thecommercial concentrates on the facets ofa comfortable, safe and secure ride andvarious useful features to make its casefor the sedan. Affordability is indicatedtoo, but not to the extent that it was inthe first commercial. The overall effectof the commercial is to generate overallpositive imagery for the sedan.

APPENDIX B: COPYTESTEDVARIABLE SETVariables tested in isolated re-exposure tosedan commercials:





Figure A1 AdSAM�

elaboration likelihood model: a missing intrinsic emotional … · 2017-08-23 · human history. it...

Documents