central and peripheral routes to advertising effectiveness

Central and Peripheral Routesto Advertising Effectiveness:The Moderating Role of Involvement

RICHARD E. PETTYJOHN T. CACIOPPODAVID SCHUMANN*

Undergraduates expressed their attitudes atx)ul a product atter being exposed toa magazine ad under conditions of either high or low product involvement. Thead contained either strong or weak arguments for the product and featured eitherprominent sports celebrities or average citizens as endorsers. The manipulationof argument quality had a greater impact on attitudes under high than low involve-ment, but the manipulation of product endorser had a greater impact under lowthan high involvement. These results are consistent with the view that there aretwo relatively distinct routes to persuasion.

O ver the past three decades, a large number of studieshave examined how consumers' evaluations of issues,

candidates, and products are affected by media advertise-ments. Research on the methods by which consumers" al-titudes are formed and changed has accelerated at a pacesuch that Kassarjian and Kassarjian were led lo the conclu-sion that "attitudes clearly have become the central focusof consumer behavior research" (1979, p. 3). Not only arethere a large number of empirical studies on consumer at-titude formation and change, but there are also a large num-ber of different theories of persuasion vying for the attentionof the discipline (see Engel and Blackwell 1982; Kassarjian1982).

In our recent reviews of the many approaches to attitudechange employed in social and consumer psychology, wehave suggested that—even though the different theories ofpersuasion possess different terminologies, postulates, un-derlying motives, and panicular "effects" that they spe-cialize in explaining—these theories emphasize one of twodistinct routes to attitude change (Petty and Cacioppo 1981.1983). One, called the central route, views attitude changeas resulting from a person's diligent consideration of in-formation that s/he feels is central to the true merits of a

'Richard E. Petty is Associate Professor of Psychology ai the Universityof Missouri, Columbia, MO 65211. John T. Cacioppo is Associate Pro-fessor of Psychology at the Universiiy of Iowa, Iowa City. lA 52242.David Schumann is a graduate student in psychology at ihe University ofMissoun, The authors are grateful lo Rob Greene. Nancy Stabler, jmBayer, Karen King. Brian Kinkade, Edith Meredith. Tim Nash, and ToddNixon for their considerable help in conducting the experiment reponedhere, and to the University of Missouri Research Council for grant support

135

particular attitudinal position. The theoretical approachesfollowing this route emphasize factors such as (1) the cog-nitive justification of attitude discrepant behavior (Cum-mings and Venkatesan 1976; Festinger 1957); (2) the com-prehension, learning, and retention of issue- or product-relevant information {Bettman 1979; Hovland, Janis, andKelly 1953; McGuire 1976); (3) the nature of a person'sidiosyncratic cognitive responses to external communica-tions (Cacioppo and Petty 1980a; Greenwaid 1968; Petty.Ostrom. and Brock 1981; Wright 1980); and (4) the mannerin which a person combines and integrates issue- or prod-uct-relevant beliefs into an overall evaluative reaction{Ajzen and Fishbein 1980; Lutz and Bettman 1977; Trout-man and Shanteau 1976). Attitude changes induced via thecentral route are postulated to be relatively enduring andpredictive of behavior {Cialdini. Petty, and Cacioppo 1981;Petty and Cacioppo 1980).

A second group of theoretical approaches to persuasionemphasizes a more peripheral route to attitude change.Attitude changes that occur via the peripheral route do notoccur because an individual has personally considered thepros and cons of the issue, but because the attitude issueor object is associated with positive or negative cues—orbecause the person makes a simple inference about themerits of the advocated position based on various simplecues in the persuasion context. For example, rather thandiligently considering the issue-relevant arguments, a per-son may accept an advocacy simply because it was pre-sented during a pleasant lunch or because the source is anexpert. Similarly, a person may reject an advocacy simplybecause the position presented appears to be too extreme.These cues (e.g.. good food, expert sources, extreme po-

C JOURNAL OF CONSUMER RESEARCH • Vol 10 • Scpiember 1983

136THE JOURNAL OF CONSUMER RESEARCH

sitions) and inferences (e.g., "If an expert says it, it mustbe true") may shape attitudes or allow a person to decidewhat attitudinal position to adopt without the need for en-gaging in any extensive thought about issue- or product-relevant arguments. The theoretical approaches followingthe peripheral route emphasize factors such as (1) whethera simple attitudinal inference can be made based on ob-sen'ing one's own behavior (Bem 1972; Scott 1978); (2)whether the advocacy falls within one's latitude of accep-tance or rejection (Newman and Dolich 1979; Sherif,Sherif, and Nebergall 1965); (3) whether some transientsituational utility is associated with adopting a particularanitude (Schlenker 1978, 1980); and (4) whether an ad-vocated position or product is classically conditioned tobasic but issue-irrelevant cues, such as food and pain (Janis.Kaye, and Kirschner 1965; Stemthal and Craig 1974). oris associated with secondary cues, such as pleasant picturesand attractive endorsers (Kelman 1961; Mitchell and Olson1981; Mowen 1980). Attitude changes induced under theperipheral route are postulated to be relatively temporaryand unpredictive of behavior.'

Unfortunately, none of the unique theories of persuasionhas yet provided a comprehensive view of attitude change.For example, cognitive response theory—an approach thatfalls under the central route—assumes thai people are usu-ally interested in thinking about and elaborating incomingInformation, or in self-generating issue- or product-relevantthoughts (Brock and Shavitt 1983). Yet. as Miller and hiscolleagues have noted. "It may be irrational to scrutinizethe plethora of counterattitudinal messages received daily.To the extent that one possesses only a limited amount ofinformation processing time and capacity, such scrutinywould disengage the thought process from the exigenciesof daily life." (Miller. Maruyama. Beaber. and Valone1976, p. 623). Haines (1974), in fact, has proposed a prin-ciple of information-processing parsimony according towhich consumers seek to process as little data as necessar>'in order to make decisions.

The accumulated research on persuasion clearly indicatesthat neither the central nor the peripheral approach alonecan account for the diversity of attitude-change results ob-served. Thus, a general framework for understanding atti-

'Our caicgorizalion of the iradUional iheoretical approaches under oneor the other roule to persuasion is mean! to be suggestive rather thanabsolute. For example, the iheorelical process of self-perccplion (Bem1972) mighi generally lead to altitude change because of a simple inference(peripheral route), bui might also be capable of initiating extended issue-relevant thinking in other circumstances (e.g.. when personal relevanceis high; see Liebhan 1979). Additionally, we note that the view that ihereare different "kinds" of persuasion can be traced back lo Aristotle's Rhet-oric, and thai tbe distinction we have made between the central and pe-ripheral routes to attitude change has much in common with Kelman's(1961) earlier view of "intemalization" vs. "identification" and with thereceni psychological distinctions between "deep" vs. "shallow" pro-cessing (Craik and Lockhan 1972). "controlled" vs. "automatic" pro-cessing (Schneider and Shiffrin 1977). "systematic" vs. "•heuristic" pro-cessing (Chaiken 1980). "thoughtful" vs. "scripted" or "mindless"processing (Abelson 1976; Langer et al. 1978). and others. For moredetails on similarities and differences among the approaches, see Pcuyand Cacioppo (forthcoming a).

tude change must consider thai in some situations peopleare avid seekers and manipulators of information, and inothers they are best described as "cognitive misers" whoeschew any difficult intellectual activity (Bumkrant 1976;McGuire 1969). An important question for consumer re-searchers then is: when will consumers actively seek andprocess product-relevant information, and when will theybe more cursory in their analysis of ads? Recent researchin consumer behavior and social psychology has focusedon the concept of "involvement" as an important moder-ator of the amount and type of information processing elic-ited by a persuasive communication (see Burnkrant andSawyer 1983; Petty and Cacioppo 1981, 1983). One majorgoal of the experiment reported in this paper was to test thehypothesis that under "high involvement." attitudes in re-sponse to an advertisement would be affected via the centralroute, but that un(3er "low involvement," attitudes wouldbe affected via the peripheral route.

INVOLVEMENT ANDATTITUDE CHANGE

Methods of Stuiiying Involvement

Although there are many specific definitions of involve-ment within both social and consumer psychology, there isconsiderable agreement that high involvement messageshave greater personal relevance and consequences or elicitmore personal connections than low involvement messages(Engel and Blackwell 1982; Knigman 1965; Petty and Caci-oppo 1979; Shetif and Hovland 1961).- Various strategieshave been employed in studying involvement. For example,hoth social (Hovland et al. 1957) and consumer (Newmanand Dolich 1979) researchers have investigated existinggroups that differed in the extent to which an issue or prod-uct was personally important, or have employed designsallowing subjects to assign themselves to high and low in-volvement groups. These correlational methods may behigh in external validity, but they confound involvementwith all other existing differences between the high and lowinvolvement groups (attitude extremity, amount of priorinformation, and so on), and thus compromise intemal va-lidity (Kiesler. Collins, and Miller 1969). Other social(Rhine and Severance 1970) and consumer (Lastovicka andGardner 1979) researchers have defined involvement interms of the specific issue or product under consideration.This procedure, of course, confounds involvement with as-pects of the issue or product that may be irrelevant to itspersonal importance. Finally, some researchers have stud-ied involvement by varying the medium of message pre-

In the present paper, we use the term involvement to refer to "issue"or "product" involvement rather than "response" involvement. In theformer, the attitude issue or the product itself has some direct personalrelevance or consequence, and people art concerned with forming a rea-soned opinion (Petty and Cacioppo 1979), In the latter, the attitude re-sponse is imponam. and people arc more concerned with expressing analtitude that will produce immediate situational rewards (such as gainingfavor witb others) than with forming a veridical opinion (Zimbardo 1960t.

INVOLVEMENT AND PERSUASION 137

sentation. Interestingly, some investigators have argued thattelevision is a more involving medium than prim (Worchel.Andreoli, and Eason 1975), whereas others have arguedjust the opposite (Krugman 1967).

A preferred procedure for studying involvement wouldbe to hold recipient, message, and medium characteristicsconstant and randomly assign participants to high and lowinvolvement groups. Apsler and Sears (1968) employed aningenious method to manipulate involvement: some partic-ipants were led to believe that a persuasive proposal hadpersonal implications for them (an advocated change inuniversity regulations would take effect while the studentparticipants were still in school), while others were led tobelieve that it did not (i.e., the change would not take effectuntil after the students had graduated). A variation of thisprocedure was developed by Wright (1973. 1974) to ma-nipulate involvement in an advertising study. Participantsin the high involvement group were told that they wouldsubsequently be asked to evaluate the product in an adver-tisement they were about to see. and were given some ad-ditional background information. Participants in the lowinvolvement group did not expect to evaluate the productand were given no background information. The back-ground information provided to the high involvement sub-jects explained the relevance of their product decisions to"their families, their own time and effort, and their per-sonal finances'* (Wright 1973, p. 56). However, it is some-what unclear to what extent this background informationmade certain product-relevant arguments salient or sug-gested appropriate dimensions of product evaluation forhigh but not low involvement subjects.

In the present experiment, participants in both the highand low involvement groups were told that they would beevaluating advertisements for products, but subjects in thehigh involvement group were led to believe that the exper-imental advertised product would soon be available in theirlocal area, and that after viewing a variety of advertisementsthey would be allowed to choose one brand from the ex-perimental product category to take home as a gift. Lowinvolvement participants were led to believe that the ex-perimental advertised product would not be available intheir local area in the near future, and that after viewing theads they would be allowed to take home one brand froma category of products other than the experimental category.

Theories of Involvement

In addition to the methodological differences that haveplagued the involvement concept, another area of disagree-ment concems the effects on persuasion that involvementis expected to have. Perhaps the dominant notion in socialpsychology stems from the Sherifs' social judgment theory(Sherif et al. 1965). Their notion is that on any given issue,highly involved persons exhibit more negative evaluationsof a communication because high involvement is associatedwith an extended "latitude of rejection." Thus, incomingmessages on involving topics are thought to have an en-hanced probability of being rejected because they are more

likely to fall within the unacceptable range of a person'simplicit attitude continuum. Kmgman (1965) has proposedan alternative view that has achieved considerable recog-nition among consumer researchers. According to thisview, increasing involvement does not increase resistanceto persuasion, but instead shifts the sequence of commu-nication impact. Krugman argues that under high involve-ment, a communication is likely to affect cognitions, thenattitudes, and then behaviors, whereas under low involve-ment, a communication is more likely to affect cognitions,then behaviors, then attitudes (see also Ray et al. 1973).

As noted earlier, a focal goal of this study is to assessthe viability of a third view of the effects of involvementon consumer response to advertisements. This view stemsfrom our Elaboration Likelihood Model (ELM) of attitudechange (Petty and Cacioppo 1981). The basic tenet of theELM is that different methods of inducing persuasion maywork best depending on whether the elaboration likelihoodof the communication situation (i.e.. the probability ofmessage- or issue-relevant thought occuning) is high orlow. When the elaboration likelihood is high, the centralroute to persuasion should be particulariy effective, butwhen the elaboration likelihood is low. the peripheral routeshould be better. The ELM contends that as an issue orproduct increases in personal relevance or consequences,it becomes more important and adaptive to forming a rea-soned and veridical opinion. Thus, people are more moti-vated to devote the cognitive effort required to evaluate thetrue merits of an issue or product when involvement is highrather than low. If increased involvement increases one'spropensity to think about the true merits of an issue orproduct, then manipulations that require extensive issue- orproduct-relevant thought in order to be effective shouldhave a greater impact under high rather than low involve-ment conditions. On the other hand, manipulations that al-low a person to evaluate an issue or product without en-gaging in extensive issue- or product-relevant thinkingshould have a greater impact under low rather than highinvolvement.

Research in social psychology has supported the viewthat different variables affect persuasion under high and lowinvolvement conditions. For example, the quality of thearguments contained in a message has had a greater impacton persuasion under conditions of high rather than low in-volvement (Petty and Cacioppo 1979; Petty. Cacioppo. andHeesacker 1981). On the other hand, peripheral cues suchas the expertise or attractiveness of a message source (Chai-ken 1980; Petty. Cacioppo. and Goldman 1981; Rhine andSeverance 1970) have had a greater impact on persuasionunder conditions of low rather than high involvement. Insum, under high involvement conditions people appear toexert the cognitive effort required to evaluate the issue-relevant arguments presented, and their attitudes are a func-tion of this in format ion-process ing activity (central route).Under low involvement conditions, attitudes appear to beaffected by simple acceptance and rejection cues in thepersuasion context and are less affected by argument quality(peripheral route). Although the accumulated research in

138 THE JOURNAL OF CONSUMER RESEARCH

social psychology is quite consistent with the ELM, it isnot yet clear whether or not the ELM predictions wouldhold when involvement concerns a product (such as tooth-paste) rather than an issue (such as capital punishment),and when the persuasive message is an advertisement ratherthan a speech or editorial.

Central and Peripheral Routes toAdvertising Effectiveness

One important implication of the ELM for advertisingmessages is that different kinds of appeals may be mosteffective for different audiences. For example, a personwho is about to purchase a new refrigerator (high involve-ment) may scrutinize the product-relevant information pre-sented in an advertisement. If this information is perceivedto be cogent and persuasive, favorable attitudes will result,but if this information is weak and specious, unfavorableattitudes will result (central route). On the other hand, aperson who is not considering purchasing a new refrigeratorat the moment (low involvement) will not expend the effortrequired to think about the product-relevant arguments inthe ad. but may instead focus on the attractiveness, credi-bility, or prestige of the product's endorser (peripheralroute). Some evidence in consumer psychology is consis-tent with this reasoning. For example, Wright (1973. 1974)exposed people to an advertisement for a soybean productunder high and low involvement conditions (see earlier de-scription) and measured the number of source comments(derogations) and message comments (counterarguments)generated after exposure. Although Wright (1974) predictedthat involvement would increase both kinds of comments,he found that more message comments were made underhigh rather than low involvement, but that more sourcecomments were made under low involvement conditions.This finding, of course, is consistent with the ELM.

In an initial attempt to provide a specific test of the utilityof the ELM for understanding the effectiveness of adver-tising messages (Petty and Cacioppo 1980). we conducteda study in which three variables were manipulated: (1) thepersonal relevance of a shampoo ad (high involvement sub-jects were led to believe that the product would be availablein their l(x:al area, whereas low involvement subjects werenot); (2) the quality of the arguments contained In the ad;and (3) the physical attractiveness of the endorsers of theshampoo. Consistent with the ELM predictions, the qualityof the arguments contained in the advertisement had agreater impact on attitudes when the product was of highrather than low relevance. Contrary to expectations, how-ever, the attractiveness of the endorsers was equally im-portant under both the high and low involvement condi-tions. In retrospect, in addition to serving as a peripheralcue under low involvement, the physical appearance of theproduct endorsers (especially their hair) may have ser\edas persuasive visual testimony for the product's effective-ness. Thus, under high involvement conditions, the phys-ical attractiveness of the endorsers may have ser\'ed as acogent product-relevant argument.

The present study was a conceptual replication of pre-vious work (Petty and Cacioppo 1980), except that we em-ployed a peripheral cue that could not be construed as aproduct-re levant argument. In the current study, partici-pants were randomly assigned to high and low involvementconditions and viewed one of four different ads for a fic-titious new product. "Edge disposable razors." The ad waspresented in magazine format and was embedded in an ad-vertising booklet along with 11 other ads. Two features ofthe Edge ad were manipulated: the quality ofthe argumentsin support of Edge (strong or weak), and the celebrity statusof the featured endorsers of Edge (celebrity or average cit-izen). It is important to note that preliminary testing re-vealed that for most people, the celebrity status of the en-dorsers was irrelevant to an evaluation of the true merits ofa disposable razor, but that because the celebrity endorserswere liked more than the average citizens, they could stillserve as a positive peripheral cue.

We had two major hypotheses. First, we expected thequality of the arguments presented in the ad to have agreater impact on product attitudes under high rather thanlow involvement conditions. Second, we expected the ce-lebrity status of the product endorsers to have a greaterimpact on product attitudes under low rather than high in-volvement conditions. If these hypotheses were supported,it would provide the first evidence that the ElaborationLikelihood Model can contribute to understanding the ef-fects of involvement on attitudinal responses to advertise-ments.

METHOD

Subjects and DesignA total of 160 male and female undergraduates at the

University of Missouri-Columbia participated in the ex-periment to earn credit in an introductory psychologycourse; 20 subjects were randomly assigned to each of thecells in a 2 (involvement: high or low) x 2 (argumentquality: strong or weak) x 2 (cue: celebrity or noncelebritystatus) factorial design. Subjects participated in groups ofthree to 15 in a very large classroom. The subjects wereisolated from each other so that they could complete theexperiment independently, and subjects in a single sessionparticipated in different experimental conditions. In fact,if enough subjects were present it was possible to conductall eight experimental conditions simultaneously. This pro-cedure avoided confounding session with experimental con-dition.

ProcedureTwo booklets were prepared for the study. The first con-

tained the advertising stimuli and the second contained thedependent measures. The first page ofthe advertising book-let explained that the study concerned the evaluation ofmagazine and newspaper ads and that the psychology de-partment was cooperating with the journalism school in this


endeavor. The first page also contained pan of the involve-ment manipulation (see below). It was explained that eachad in the booklet was preceded by an introductor>' statementthat lold a little about the advertisement that followed (e.g.,"The company of Paris, France has just opened anAmerican office in New York City. This ^Hte men's cloth-ing company originally sold clothing only in Europe, butis now in the process of attempting to enter the Americanmarket. The ad on the next page is one that they will betesting soon in Tampa, Florida before running the ads inother major cities that will eventually carry their prod-ucts"). The instructions told subjects to continue throughthe booklet at their own pace and to raise their hands whenfinished. The ad booklet contained 10 real magazine adsfor both relatively familiar (e.g.. Aquafresh toothpaste) andunfamiliar (e.g.. Riopan antacid) products, and two bogusads. The sixth ad in each booklet was the crucial fictitiousad for Edge razors (the nature of the other bogus ad wasvaried but is irrelevant to the present study). When subjectshad completed perusing their ad booklets, they were givena questionnaire booklet to complete. Upon completion ofthe questionnaire, the subjects were thoroughly debriefed,thanked for their participation, and dismissed.

Independent Variables

Imolvement. Involvement was embedded in two placesin the ad booklet. First, the cover page offered subjects afree gift for participation in the experiment. Subjects wereeither informed that they would be allowed to choose aparticular brand of disposable razor (high involvement withthe fictitious Edge ad) or thai they would be allowed tochoose a brand of toothpaste (low involvement with Edge).A toothpaste ad did appear in the ad booklet, but it was thesame ad for all subjects. To bolster the involvement ma-nipulation, the page that introduced the Edge ad also dif-fered in the high and low involvement conditions. Highinvolvement subjects were told that the advenisement andproduct would soon be test-marketed in medium-sized citiesthroughout the Midwest, including their own city (Colum-bia, Missouri); low involvement subjects were told that theadvertisement and product were being test-marketed onlyon the East Coast. Thus high Involvement subjects werenot only led lo believe that they would soon have to makea decision about the product class, they were also led tobelieve that the product would be available in their area inthe near future. Low involvement subjects, on the otherhand, did not expect to make a decision about razors (butdid expect to make one about toothpaste), and were led tobelieve that Edge razors would not be available for purchasein their area in the forseeable future.

Argument quality. A variety of arguments for dispos-able razors were pretested for potency on a sample of un-dergraduates. In the strong argumenls ad, the razor wascharacterized as •'scientifically designed." and the follow-ing five statements were made about the product:

• New advanced honing method creates unsurpassed sharp-ness

• Special chemically formulated coating eliminates nicksand cuts and prevents rusting

• Handle is tapered and ribbed to prevent slipping• In direct comparison tests, the Edge blade gave twice as

many close shaves as its nearest competitor

• Unique angle placemeni of the blade provides the smooth-est shave possible

In the weak arguments version of the ad, the razor wascharacterized as "designed for beauty." and the followingfive statements were made about the product:

• Floats in water with a minimum of rusi• Comes in various sizes, shapes, and colors• Designed with the bathroom in mind• In direct comparison tests, the Edge blade gave no more

nicks or cuts than its competition• Can only be used once but will be memorable

Peripheral cue. In the "famous endorser" conditions,the headline accompanying the advertisement read "Profes-sional Athletes Agree: Until you tr\' new Edge disposablerazors you'll never know what a really close shave is." lnaddition, the ad featured the pictures of two well known,well-iiked golf (male) and tennis (female) celebrities, ln the"nonfamous endorser" conditions, the headline read"Bakersfield, Califomia Agrees: ." and the ad fea-tured pictures of average looking people who were unfa-miliar to the subjects. The average citizens in the ad weremiddle-aged and characterized as coming from Californiato minimize perceptions of similarity lo the subjects (Mis-souri college students). Figure A depicts two of the fourEdge ads used in the present study.

Dependent Measures

On the first page of the dependent variable booklet, sub-jects were asked to try to list all of the product categoriesfor which they saw advertisements, and to tr>' to recall thebrand name of the product in thai categor>'. On the nextpage, subjects were given descriptions of the 12 productcategories and were asked to select the correct brand namefrom among seven choices provided. .Although we had nospecific hypotheses about brand recall and recognition,these measures were included because of their practical im-portance and for purposes of comparison with the attitudedata.

Next, subjects responded to some questions about one ofthe legitimate ads in the booklet; this was followed by thecrucial questions about Edge razors. The questions aboutEdge were placed relatively early in the booklet to avoidsubject fatigue and boredom and lo maximize the effec-tiveness of the manipulations. Subjects were first asked torate, on a four-point scale, how likely it would be thai theywould purchase Edge disposable razors "the next time youneeded a product of this nature." The descriptions for each


FIGURE A

EXAMPLE MOCK ADS

PROFESSIONAL ATHLETES AGREE BAKERSFIELD, CAUFORNIA AGREESUntil TWi trr n*wEDGE dJMpomablurmMon youTl a€v»rImowwbMta"naltf doM Mbmn

lm.

rmamw youtl oavmrkotpwrnrhmtm

• New •dvinced honing meihod cteitei uniuiptHrt «h«rpn»ti

• Sp»ci»l chemiciUv (ormul»[«l coaling climiiMi** mchi Md nJruciins.

• Handl« IS lapcietl and ribbed lo prevent ilipping

• In duBct compatiion i*»n the EDGE blade gave twice a* many do«e ihavMal lU nearest competitor

• Unique mngle placemeni ol the bl.de provide* the .moothe.i »ha«B pOMible

• noat) in water vnih a minimum o( ruit

• Comei tn vanoui ( U M . ihape*. and colon

• Deitgned with the balhioom in mind

• In du*ct companion i*(i* the EDGE blade gav* no moie niclii or CUKthan IU competition

• Can only be uied one* bui wiU be memorable

GET THE EDGE DIFFERENCE! ( a r THE E D ^ DIFFERENCE!

NOTE- Lett panel shows celebrity efKJorwr ad lor Eflge razofs employing Ihe Strong arguments R.gtit panel shows flverage Mi ien endo-sor ad lo. Edge razors employing the weak

arguments. PiOures ol celebfities arxJ crtaens have been blacked out lo preserve propriety areJ arwnymity.

scale value were: I = "1 definitely would not buy it." 2= "I might or might not huy it," 3 = "I would probablybuy it," and 4 = "1 would definitely buy it." Followingthis measure of purchase intentions, subjects were asked torate their overall impression of the producl on three nine-point semantic differential scales anchored at - 4 and +4(bad-good, unsatisfactory-satisfactory, and unfavor-able-favorable). Since the intercorrelations among thesemeasures were very high (average r = 0.86). responseswere averaged to assess a general positive or negative at-titude toward the product.

Following some additional questions that were consistentwith the cover story, subjects were instructed to list thethoughts that crossed their minds as they examined the adfor Edge disposable razors. These thoughts were subse-quently scored on several dimensions by trained judges.Since subjects listed very few thoughts about the product(M = 1.18) and since the manipulations failed to affectthis measure, it will not be discussed further. This "cog-nitive response" measure would probably have been moresensitive if it had been administered immediately after ex-posure to the Edge ad rather than after exposure to all !2ads, but in the present study this would have compromisedthe cover story (for an extended discussion ofthe reliability,validity, and sensitivity of the thought-listing measure inpersuasion research, see Cacioppo and Petty 1981).

After listing their thoughts, several questions were asked

to check on the experimental manipulations, and subjectswere asked to try to list as many of the attributes mentionedin the ad about Edge razors as they could recall. Followingthe questions about Edge were several questions about someof the other products and ads in the booklet. As a check onthe involvement manipulation, the ver>' last question in Ihebooklet asked subjects to recall the free gift they had beentold to expect.

RESULTS

Manipulation Checks

In response to the last question in the dependent variablebooklet asking subjects what gift they had been told toexpect, 92.5 percent of the subjects in the high involvementconditions conectly recalled that they were to select a brandof disposable razor. In the low involvement conditions,none of the subjects indicated a razor and 78 percent cor-rectly recalled that they were to select a brand of toothpaste.Thus, subjects presumably realized what product they weresoon to make a decision about as they examined the adbooklet.

To assess the effectiveness ofthe endorser manipulation,two questions were asked. First, subjects were asked if theyrecognized the people in the ad for the disposable razor.When the famous athletes were employed. 94 percent in-


dicated "yes," whereas when the average citizens wereemployed, 96 percent indicated "no ." In addition, subjectswere asked to rate the extent to which they liked the peopledepicled in the ad on an 11-point scale, where 1 indicated"liked very little" and 11 indicated "liked very much."An analysis of this measure revealed that the famous en-dorsers were liked more {M = 6.06) than the average cit-izens (M = 3.64; F (I , 143) = 40.81. p < 0.0001); onaverage, women reported liking the endorsers more (M =5.32) than did men (M = 4.44; F (I, 143) = 5.25, p <0.03).

As a check on the argument-persuasiveness manipula-tion, two questions were asked. The first required respon-dents to "rate the reasons as described in the advertisementfor using EDGE" on an 11-point scale anchored by "un-persuasive" and "persuasive"; the second question askedthem to rate the reasons on an 11-point scale anchored by"weak reasons" and "strong reasons." On the first mea-sure, subjects exposed to the strong arguments rated themas significantly more persuasive (Af = 5.46) than did sub-jects exposed to the weak arguments (M = 4.03; F (I, 139)= 12.97, p < 0.0004), Additionally, a main effect forgender was found such that women rated the arguments asmore persuasive (M = 5.26) than did men (M = 4.28; F(1, 139) = 5.25. p < 0.02). Finally, an Arguments xGenderinteraction emerged (F(I . 139) = 5.43.p < 0.02).indicating that the tendency for females to fmd the argu-ments more persuasive than males was greater for the strongthan for the weak arguments. On the second manipulationcheck measure, subjects rated the strong arguments as"stronger" (M = 5.58) than the weak ones (M = 4.13:F (1, 138) = 14.31. p < 0.002). Again, an ArgumentsX Gender interaction occurred, indicating that females es-pecially tended to rate the strong arguments more highlythan did males. In short, all of the variables were manip-ulated successfully. The tendency for females to be morepositive in their ratings of both endorsers and the argumentsin the ads is generally consistent with previous psycholog-ical research portraying women as more concerned withsocial harmony than men (Eagly 1978). Importantly, thesesex differences did not lead to any significant gender effectson the crucial measures of attitude and purchase intention.

Attitudes and Purchase ltitentionsThe Table presents the means and standard deviations for

each cell on the attitude index. A number of interestingmain effects emerged. First, involved subjects were some-what more skeptical of the product (M = 0.31) than wereless involved subjects (M = 0.99; F (I. 148) = 6.64. p< 0.01). Second, subjects iiked the product significantlymore when the ad contained cogent arguments (M = 1.65)than when the arguments were specious (M = - 0.35; F(1, 148) = 57.81. p < 0.0001). Third, subjects tended tolike the product more when it was endorsed by the famousathletes (M = 0.86) than hy the average citizens of Bak-ersfield, California (M = 0.41; F (I, 148) = 2.91, p <0.09).

TABLE

MEANS AND STANDARD DEVIATIONS FOR EACHEXPERIMENTAL CELL ON THE ATTITUDE INDEX

Citizenendorser

Celebrityendorser

Low involvement

Weakarguments

- .12(1.81)

1.21(2.28)

Strongarguments

.96(1.52)

1.85(1.59)

High involvemeni

Weakarguments

-1.10(1.66)

-1.36(1.65)

Strongarguments

1.98(1.25)

1.80(1-07)

NOTE: ARiiude scores reprBsent Ihe averags rating ol Ihe pfOdud on three nine-poiotsemantic drflerentiai scales anOwM al - 4 and • 4 (bad-good, unsatslaOorv-satisfacKxv.srvj unlavorsbte-favorable). Standard deviations are <n pareniheses.

Each of these main effects must be qualified and inter-preted in light of two important two-way interactions. First,an Involvement x Endorser interaction (F( l . 148) = 5.94,p < 0.02) revealed that the nature of the product endorserhad a significant impact on product attitudes only under lowinvolvement (F (I, 148) = 5.96, p < 0.02). but not underhigh involvement (f < I; see top panel of Figure B). Onthe other hand, an Involvement x Arguments interaction( F ( l . 148) = 18.47,/J < 0.0001) revealed that althoughargument quality had an impact on product attitudes underboth low involvement (F (1. 148) = 5.40, p < 0.02) andhigh involvement (F {\, 148) = 71.36, p < 0.000!). theimpact of argument quality on attitudes was significantlygreater under high rather than low involvement (see bottompanel of Figure B). Neither the Endorser x Arguments northe three-way interaction approached significance (F =0.14 and 0.54, respectively).

Two significant effects emerged from the question askingsubjects to rate their likelihood of purchasing Edge dispos-able razors the next time they needed a product of thisnature. Subjects said that they would be more likely to buythe product when the arguments presented were strong {M= 2.23) rather than weak (M = l .68 ;F( l , 152) = 25.37.

p < 0.0001). Additionally, an Involvement x Argumentsinteraction emerged (F (I. 152) = 4.25, p < O.CM). Thisinteraction paralleled that obtained on the altitude measureand indicated that argument quality was a more importantdeterminant of purchase intentions under high rather thanlow involvement.

The correlation between attitudes and purchase intentionsfor low involvement subjects was 0.36; and for high in-volvement subjects it was 0.59. Although both correlationsare significantly different from zero (ps < 0.001). it isinteresting to note that the low involvement correlation isconsiderably smaller than the high involvement correlation(p < 0.07). The fact that the argument quality manipulationaffected behavioral intentions while the endorser manipu-lation did not (although it did affect attitudes)—and the factthat attitudes were better predictors of behavioral intentionsunder high rather than low involvement—provide somesupport for the ELM view that attitudes formed via the


FIGURE B

PRODUCT ATTITUDES

0)-a

1.2 •

.6 •

0 •

- - 6 •

S. -1-2

0)

to 1.8 -

1.2 -

0 -

-.6 -

1

• — • —

•

" --.. FarTKJus endorsers

Non-famous endorsers ^ ^

- — ^Strong arguments

^•\^Weak arguments

^ \

Lowinvolvement

HighInvolvement

NOTE: Top panel shows interactive effect of inwolvemenl and enao'ser siaius on atliiudestoward Edge Tabors. Boltom panel shows inietaciive eHect ol invotvemerit and arguniierAquslily on aTtrtuOes loward Edge razors.

central route will be more predictive of behavior than at-titudes formed via the peripheral route.

Recall and Recognition Measures

Subjects were asked to list al! of the products for whichthey saw ads and all of the brand names they encountered.Following this, all subjects were told thai they had seen anadvertisement for a disposable razor and were asked to se-lect the correct brand name from a list of seven (Gillette,Wilkinson, Schick, Edge. Bic, Schaffer, and Remington).The proportion of subjects showing correct recall or rec-ognition was calculated for each cell. These proportionswere then subjected to an arcsin transformation (Winer1971) and analyzed by the procedure recommended by Lan-ger and Abelson (1972).

The involvement manipulation had a significant Impacton free recall of the product categor>'. with more high in-volvement subjects (81 percent) recalling the product cat-egorj' than low involvement subjects (64 percent; Z = 2.4,p < 0.02). Additionally, exposure to the famous endorserincreased recall of the product category under low involve-ment conditions (from 52 percent to 75 percent; Z = 2.14,

p < 0.03), but had no effect on product category recallunder high involvement (80 versus 82 percent).

Involvement affected free recall of the brand name of theproduct, increasing it from 42 percent in the low involve-ment conditions to 60 percent in the high involvement con-ditions (Z = 2.28, p < 0.01). There was also an effect forgender on this measure, with males showing greater brandname recall (61 percent) than females (39 percent; Z =2.78, p < 0.007). The endorser manipulation had a mar-ginally significant effect on brand name recall, with thefamous endorsers tending to enhance recall over averagecitizens from 43 to 58 percent (Z = 1.89, p < 0.06).

On the measure of brand name recognition, an interactionpattern emerged. Under low involvement, the use of famousendorsers reduced brand name recognition from 85 to 70percent, but under high involvement, the use of famousendorsers improved brand name recognition from 77 to 87percent (Z = 1.96, p < 0.05).'

To summarize the recall and recognition data thus far,it appears that increasing involvement with the product en-hanced recall not only of the product category, but also ofthe brand name of the specific product advertised. The ef-fects of the endorser manipulation were more complex anddepended on the level of involvement. In general, underlow involvement a positive endorser led to increased recallof the product category but reduced brand name recogni-tion. Thus, people may be more likely to notice the productsin low involvement ads when they feature prominent per-sonalities, but because of the enhanced attention accordedthe people in the ads and the general lack of interest inassessing the merits of the product (due to low involve-ment), reductions in brand recognition may occur. Thisfinding is similar to the results of studies on the use ofsexually oriented material in ads for low involvement prod-ucts—the sexual material enhances recognition of the ad.but not the brand name of the product (e.g.. Chestnut.LaChance, and Lubitz 1977; Steadman J969). Under highinvolvement, however, the use of prominent personalitiesenhanced brand name recognition.When people are moreinterested in the product category, they may be more mo-tivated to assess what brand the liked personalities are en-dorsing. The manipulation of argument quality had no ef-fect on recall of the product category, brand name recall,or brand name recognition.

A final recall measure assessed how many of the specific

authors have bugpcsicd thai it may be appropnatc tu analyzedichotomous data usinf analy.sis of variance without biasing (he resuhsgreatly (eg . Winer 1971). We subjected our recall and recognition data(scored 0 or I) to ANOVA, and Ihc following si{:nificant effects wereobtained. On the measure of recall of the product catcgor), a main effectof involvement ( f ( 1 . 152) = 6,42. p < 0.01) and an involvement xendorser interaction (F ( I , 152) = 3.28. p < 0.07) were obtained. Onthe measure of brand name recall, main effects for involvement (F ( I .145) = 6.34. p < 0.01). gender ( F ( l . 145) = 7.20, p < 0 008). andendorser [F (\. 145) = 3.49, p < 0.06) were obtained. On the measureof brand name recognition, an involvement x endorser interaction wasobtained ( f ( 1 , 152l = 4,04./^ < 0.05). This pattern of sijinilicant effectsis identical to the significant pattern of effects rcponed in the text.


arguments for Edge razors the subjects could spontaneouslyrecall after they had examined the entire ad booklet. Over-all, subjects were able to correctly reproduce only 1.75 ofthe five arguments presented. This was not affected by anyof the experimental manipulations.

Clearly, the manipulations produced a very different pat-tern of effects on the recall and recognition measures thanon the attitude and purchase intention measures. In addi-tion, the recall and recognition measures were uncorrelatedwith attitudes or intentions toward Edge razors. This findingis consistent with a growing body of research indicatingthat simple recall or recognition of information presentedabout an attitude object is not predictive of attitude for-mation and change (e.g., Cacioppo and Petty 1979; Green-wald 1968; Insko, Lind. and LaTour 1976).

The present data also argue against using measures ofbrand name recall or recognition as the sole indicants ofadvertising effectiveness. For example, in the presentstudy, enhancing involvement led to a significant improve-ment in brand name recall, but increasing involvement ledto a decrement in attitude toward the brand when the ar-guments presented were weak.

DISCUSSIONAs we noted earlier, previous research on attitude for-

mation and change has tended to characterize the persuasionprocess as resulting either from a thoughtful (though notnecessarily rational) consideration of issue-relevant argu-ments and product-relevant attributes (central route), orfrom associating the attitude object with various positiveand negative cues and operating with simple decision rules(peripheral route). Over the past decade, investigators inboth social psychology and consumer behavior ha\e tendedto emphasize the former process over the latter. Considerthe recent comments of Fishbein and Ajzen (1981. p. 359):

The general neglect ofthe informalion contained in a message. . . is probably the mosi serious problem in communicationand persuasion research. We are convinced that the persua-siveness of a communication can be increased much moreeasily and dramatically by paying careful attention to its con-tent . . . than by manipulation of credibility, attractiveness. . . or any of the other myriad factors that have caught thefancy of investigators in the area of communication and per-suasion.

The present study suggests that, although the informa-tional content of an advertisement may be the most impor-tant determinant of product attitudes under some circum-stances, in other circumstances such noncontent manipulationsas the celebrity status (likeability) or credibility of the prod-uct endorsers may be even more important. Specifically,we have shown that when an advertisement concerned aproduct of low involvement, the celebrity status of the prod-uct endorsers was a very potent determinant of attitudesabout the product. When the advertisement concerned aproduct of high involvement, however, the celebrity statusof the product endorsers had no effect on attitudes, but thecogency of the information about the product contained in

the ad was a powerful determinant of product evaluations."These data clearly suggest that it would be inappropriatefor social and consumer researchers to overemphasize theinfiuence of issue-relevant arguments or product-relevantattributes and ignore the role of peripheral cues. Each typeof attitudinal influence occurs in some instances, and thelevel of personal involvement with an issue or product ap-pears to be one determinant of which type of persuasionoccurs.

According to the Elaboration Likelihood Model, personalrelevance is thought to be only one determinant of the routeto persuasion. Personal relevance is thought to increase aperson's motivation for engaging in a diligent considerationof the issue- or product-re levant infonnation presented inorder to form a veridical opinion. Just as different situationsmay induce different motivations to think, different peoplemay typically employ different styles of information pro-cessing, and some people will enjoy thinking more thanothers (Cacioppo and Petty 1982, forthcoming). However,a diligent consideration of issue- or product-re levant infor-mation requires not only the motivation to think, but alsothe ability to process the information. Thus situational vari-ables (e.g., distraction; Petty. Wells, and Brock 1976) andindividual difference variables (e.g.. prior knowledge; Ca-cioppo and Petty 1980b) may also be important moderatorsof the route to persuasion. In the present study, subjects"ability to think about the product was held at a high levelacross experimental conditions—that is. the messages wereeasy to understand, the presentation was self-paced, and soon. Thus the primary determinant ofthe route to persuasionwas motivational in nature.^

*A!ihough not tested in the present study, the ELM predicts thai undermoderate involvemeni conditions, source infonnation serves neither as asimple cue (as under low involvement) nor is it ignored (as under highinvolvement). Instead, source informalion helps a person determine howmuch thinking to do about the message (Petty and Cacioppo 1981. forth-coming a).

*An anonymous reviewer of this artiete took issue with our motivationalinterpretation of the effects of involvement and suggested thai perhaps oureffects resulted because our experimental task overtaxed our subjects'cognitive abilities. This suggestion assumes that subjects lacked the abilityto evaluate both the source and the message, and iherefore had to chtxtseone over the other. We find this explanation implausible for several rea-sons Firsl. since the subjects paced themselves through the ad txHiklei.they could spend as much time as they wished evaluaiing each ad the'•overtaxed" explanation may thus be more plausible for research in whichthe message is extemally paced. Second, our experinieni included severalchecks on whether or not subjects attended to ihe source and messageinformation. For example, all subjects were asked if they recognized andliked the endorsers appearing in the ad. If the reviewer's suggestion iscorrect, we would expect subjects in the high involvement group (whodiligently processed the message conient) to be less likely to report rec-ognizing the endorsers in the ad. and hence to show less liking for the adendorsers. However, the involvement manipulation failed to affecl eiiherthe recognition or the liking measure. In fact, subjects in ihe high in\olvc-mem group reponcd slightly ([hough nol sipnilicantlyi greater recognitionand liking of the famous endorsers than did the low invoKemeni subjects.Thus high invotvemeni subjects WCR- not overtaxed They recognized andliked the famous endorsers to the same exieni as did low invohcmentsubjects, [t is Just that the product attiludes of ihe hiph involvement sub-jects were noi affected by ihis liking, while the produci attiiudcs of thelow involvement subjects were.


It is important to note that although our "peripheral"manipulation was a source variable presented visually andour "central" manipulation was a message variable pre-sented verbally, neither the source/message nor the visual/verbal dichotomy is isomorphic with the central/peripheralone. Thus a source variable may induce persuasion via thecentral route, and a message variable may serve as a pe-ripheral cue. For example, in one study described previ-ously (Petty and Cacioppo 1980), we obsened that a phys-ically attractive message endorser might serve as a cogentproduct-re levant argument for a beauty product. In anotherstudy (Petty and Cacioppo, forthcoming b), we found thatthe mere number of message arguments presented may ac-tivate a simple decision rule (the more the better) under lowinvolvement, but not under high involvement, where ar-gument quality is more important than number. Similarly,a "central" manipulation may be presented visually—e,g.,depicting a kitten in an advertisement for facial tissue toconvey the product-relevant attribute "softness" (Mitchelland Olson 1981)—and a "peripheral" manipulation maybe presented verbally—e.g., providing a verbal descriptionof a message source as an expert or as likeable (Chaiken1980; Petty, Cacioppo, and Goldman 1981). The criticalfeature of the central route to persuasion is thai an attitudechange is based on a diligent consideration of informationthat a person feels is central to the true merits of an issueor product. This information may be conveyed visually,verbally, or in source or message characteristics. In theperipheral route, attitudes change because of the presenceof simple positive or negative cues, or because of the in-vocation of simple decision rules which obviate the needfor thinking about issue-relevant arguments. Stimuli thatserve as peripheral cues or that invoke simple decision rulesmay be presented visually or verbally, or may be part ofsource or message characteristics.

In the present study, the overall pattern of results on theattitude and purchase intention measures is more consistentwith the Elaboration Likelihood Model formulation thanwith the Sherif. Sherif, and Nebergall (1965) social judg-ment model or with Krugman's (1965, 1967) sequencemodel of involvement. Although increasing involvementdid produce a main effect on the attitude measure (moreresistance to the product under high rather than low in-volvement), as anticipated by social judgment theory, themore complicated interactions of endorser and argumentquality with involvement cannot be accounted for by thetheory. Thus the social judgment theory view that it is moredifficult to change attitudes under high involvement is, atbest, only partially correct, and is unable to account for thecomplete pattern of attitude data. The attitude and behav-ioral data are generally inconsistent with Krugman's se-quence formulation, Kxugman suggested that under highinvolvement, attitude change preceded behavior change,but that under low involvement, behavior change precededattitude change. This reasoning would suggest that on im-mediate measures, attitudinal effects should be easier lodetect than behavioral effects under high involvement,while behavioral effects should be easier to detect than at-

titudinal effects under low involvement. In the presentstudy, both attitudinal and behavioral (intention) effectswere observed under high involvement, which is consistentwith both models. Under low involvement, however, ef-fects were obtained on the measure of attitude but not onthe measure of behavioral intentions. This finding is incon-sistent with Krugman's formulation, which expects strongerbehavioral effects under low involvement than under highinvolvement, but it is consistent with the ELM, which pos-tulates a greater correspondence between attitudes and be-haviors under high involvement (central route) than underlow involvement (peripheral route).

In sum, the present study has provided support for theview that different features of an advertisement may bemore or less effective, depending upon a person's involve-ment with it. Under conditions of low involvement, pe-ripheral cues are more important than issue-relevant argu-mentation, but under high involvement, the opposite is true.The realization that independent variables may have differ-ent effects, depending on the level of personal relevance ofa message, may provide some insight into the conflictingpattern of results that is said to characterize much attituderesearch. It may well be that attitude effects can be arrangedon a continuum, depending on the elaboration likelihoodof the particular persuasion situation. This continuumwould be anchored at one end by the peripheral route andat the other end by the central route to persuasion. Fur-thermore, these two routes may be characterized by quitedifferent antecedents and consequents. If so, future workcould be aimed at uncovering the various moderators of theroute to persuasion and at tracking the various consequentsof the two different routes.

[Received March 1982. Revised April 1983.]

REFERENCESAbclson. Robert P. (1976). "Script Proccssing in Aiiitude For-

mation and Decision Making," in Cognition and Social Be-havior, eds. John S. Carroll and John W. Payne. Hillsdale.NJ: Lawrence Eribaum.

Ajzen. Icek and Martin Fishbein (1980), Understanding Aniiudesand Predicting Social Behavior. Englewood Cliffs. NJ: Pren-tlee-Hall.

Apsler. Robert and David O. Scars (1968). •"Waming. PersonalInvolvement, and Attitude Change." Journal of Personal ityand Social Psychology. 9 (June), 162-166.

Bem. Daryl J. (1972). "Self-Percept ion Theor>'." in Advances inExperimental Social Psychology. Vol. 6, ed. Leonard Ber-kowitz. New York: Academic Press, 2-57.

Bettman. James R. (1979), "Memory Factors in ConsumerChoice: A Review," Journal of Marketing. 43 (Spring).37-53.

Brock. Timothy C. and Sharon Shavitt (1983), "Cognitive Re-sponse Analysis in Advertising," In Advertising and Con^sumer Psychology, eds. LaiT>' Percy and Arch Woodsidc,Lexington. MA: Lexington Books. 91-116.

Bumkrant. Robert E. (1976), "A Motivational Model of Infor-mation Processing Intensity." Journal of Consumer Re-seareh. 3 (June). 21-30.


and Alan C. Sawyer (1983). "Effects of Involvemeni andMessage Content on Informalion Processing Intensity." inInformation Processing Research in Advertising, ed. RichardHarris. Hillsdale. NJ: Lawrence Eribaum.

Cacicppo. John T. and Richard E. Petty (1979). "Effects of Mes-sage Repetition and Position on Cognitive Responses. Recall,and Persuasion." Journal of Personality and Social Psy-chology, 37 (January), 97-109.

and Richard E. Petty (1980a). "Persuasiveness of Com-munications is Affected by Exposure Frequency and MessageQuality: A Theoretical and Empirical Analysis of PersistingAttitude Change," in Current Issues and Research in Ad-vertising, eds, James H. Leigh and Claude R. Martin. AnnArbor. MI : University of Michigan, 97-122.

and Richard E, Petty (1980b). "Sex Differences in Influ-enceability: Toward Specifying the Underlying Processes."Personality and Social Psychology Bulletin. 6 (December).651-656."

and Richard E, Petty (1981). "Social Psychological Pro-cedures for Cognitive Response Assessment: The ThoughtListing Technique." in Cognitive Assessment, eds, ThomasV, Merluzzi. Carol R. Glass, and Myles Genest. New York:Guilford Press. 309-342.

and Richard E, Petty (1982). "The Need for Cognition."Journal of Personality and Social Psychology." 42 (Janu-ary), 116-131.

and Richard E, Petty (forthcoming). "The Need for Cog-nition: Relationships to Social Influence and Self Influence."in Social Perception in Clinical and Counseling Psychology.eds. Richard P. McGIynn. James E, Maddux. Cai D. Stol-tenberg, and John H. Harvey, Lubbock. TX: Texas TechPress.

Chaiken. Shelly (1980). "Heuristic Versus Systematic Informa-tion Processing and the Use of Source Versus Message Cuesin Persuasion," Journal of Personality and Social Psychol-ogy. 39 (November). 752-766.

Chestnut. Robert W.. Charles C. LaChance. and Amy Lubitz(1977). "The Decorative Female Model: Sexual Stimuli andIhe Recognition of Advertisements." Journal of AdvertisingResearch. 6. 11-14.

Cialdini. Robert B.. Richard E. Petty, and John T. Cacioppo(I98I) . "Attitude and Attitude Change." Annual Review ofPsychology. 32, 357-404.

Craik. Fergus M. and Robert S. Lockhart (1972). "Levels ofProcessing: A Framework for Memory Research." Journalof Verbal Learning and Verbal Behavior. 11 (October).671-684.

Cummings. William H, and M. Venkatesan (1976), "CognitiveDissonance and Consumer Behavior: A Review of the Evi-dence." Journal of Marketing Research. 13 (August).303-308.

Eagly, Alice H, (1978). "Sex Differences in lnfluenceability,"Psychological Bulletin. 85 (January). 86-116.

Engel. James F. and Roger D. Blackwell (1982). Consumer Be-havior. Hinsdale. IL: Drj'den Press,

Festinger. Leon (1957). A Theory of Cognitive Dissonance. Stan-ford. CA: Stanford University Press,

Fishbein. Martin and Icek Ajzen (1981). "Acceptance. Yielding,and Impact: Cognitive Processes in Persuasion." in CognitiveResponses in Persuasion, eds, Richard E, Petty. ThomasOstrom. and Timothy C. Brock, Hillsdale. NJ: LawrenceEribaum. 339-359.

Greenwald. Anthony G. (1968). "Cognitive Leaming. CognitiveResponse to Persuasion, and Attitude Change." in Psycho-

logical Foundations of Attitudes, eds. Anthony G. Green-wald. Timothy C. Brock, and Thomas Ostrom. New York:Academic Press. 147-170.

Haines. George H. (1974). "Process Models of Consumer De-cision Making." in BuyerlConsumer Information Process-ing, eds. G. David Hughes and Michael L. Ray. Chapel Hi l l .NC: University of North Carolina Press. 89-107.

Hovland. Carl I . . O. J. Harvey, and Muzifer Sherif (1957)."Assimilation and Contrast Effects in Reactions to Com-munication and Attitude Change." Journal of Abnormal andSocial Psychology. 55 (September). 244-252.

, Irving Janis. and Harold Kelley (1953). Communicationand Persuasion. New Haven, CT: Yale University Press.

Insko, Chester A. , E. Allen Lind. and Stephen LaTour (1976),"Persuasion, Recall, and Thoughts," Representative Re-search in Social Psychology. 7, 66-78,

Janis, Irving L., D. Kaye, and P. Kirschner (1965). "FacilitatingEffects of 'Eating while Reading' on Responsiveness to Per-suasive Communication." Journal of Personality and SocialPsychology. I (Febniary), 181-186.

Kassarjian, Harold H. (1982). "Consumer Psychology," AnnualReview of Psychology. 33. 619-649.

and Waltraud M. Kassarjian (1979), "Attitudes underLow Commitment Conditions." in Attitude Research Playsfor High Stakes, eds. John C. Maloney and Bernard Silver-man. Chicago: American Marketing Association. 3-15.

Kelman. Herbert C. (1961).."Processes of Opinion Change."Public Opinion Quarterly. 25 (Spring). 57-78.

Kiesler, Charles A. . Barry E. Collins, and Norman Miller (1969).Altitude Change: A Critical Analysis of Theoretical Ap-proaches. New York: John Wiley.

Krugman. Herbert E. (1965). "The Impact of Television Adver-tising: Learning without Involvement." Public OpinionQuarterly. 29 (Fall). 349-356.

(1967). "The Measurement of Advertising Involvement."Public Opinion Quarterly. 30 (Winter), 583-596,

Langer. Ellen J. and Robert P. Abelson (1972), "The Semanticsof Asking a Favor: How to Succeed in Getting Help withoutReally Dying. "Journa! of Personality and Social Psychol-ogy, 24 (October). 26-32.

. Arthur Blank, and Benzion Chanowitz (1978). "TheMindlessness of Ostensibly Thoughtful Action: The Role of•Placebic' Information in Interpersonal Interaction." Journalof Personality and Social Psychology. 36 (June), 635-642.

Lastovicka, John L. and David M. Gardner (1979). "Componentsof Involvement." in Anitude Research Plays for High Stakes.eds, John C. Maloney and Bernard Silverman. Chicago:American Marketing Association. 53-73.

Liebhart, Emst H. (1979). "Information Search and Attribution:Cognitive Processes Mediating the Effect of False AutonomicFeedback." European Journal of Social Psychology. 9. (Jan-uary-March). 19-37,

Lutz. Richard J. and James R. Bettman (1977). "MultiattributeModels in Marketing: A Bicentennial Review." in Consumerand Industrial Buying Behavior, eds. Arch Woodside. Jag-dish Sheth. and Peter D. Bennett, New York: Elsevier North-Holland. 137-150.

McGuire. William J. (1969). "The Nature of Attitudes and At-titude Change." in The Handbook of Social Psychology. Vol.3, eds. Gardner Lindzey and Elliot Aronson. Reading, MA:Addison-Wesley. 136-314.

(1976). "Some Internal Psychological Factors InfluencingConsumer Choice. " Journal of Consumer Research. 2(March), 302-319.


Miller, Norman, Geoffrey Maruyama, Rex Julian Beaber, andKeith Valone (i976), "Speed of Speech and Persuasion."Journal of Personality and Social Psychology. 34 (October).615-624,

Mitchell, Andrew A, and Jeny C. Olson (1981), "Are ProductAttribute Beliefs the Only Mediator of Advertising Effectson Brand Attitude?" Journal of Marketing Research. 18(August). 318-332.

Mowen, John C. (1980), "On Product Endorser Effectiveness: ABalance Mode! Approach," in Current Issues and Researchin Advertising, eds. James H. Leigh and Claude R. Martin.Ann Arbor, MI: University of Michigan, 41-57.

Newman, Larry M. and Ira J, Dolich (1979), "An Examinationof Ego-Involvement as a Modifier of Attitude ChangesCaused from Product Testing," Advances in Consumer Re-search. Vol. 6, ed. William L, Wilkie, Ann Arbor, MI:Association for Consumer Research, 180-183.

Petty, Richard E. and John T. Cacioppo (1979). "Issue Involve-ment Can Increase or Decrease Persuasion by EnhancingMessage-Relevant Cognitive Responses." Journal of Per-sonality and Social Psychology, 37 (October). 1915-1926.

and John T. Cacioppo (1980), "Effects of Issue Involve-ment on Altitudes in an Advertising Context." in Proceed-ings of fhe Division 23 Program, eds. Gerald G. Gom andMarvin E. Goldberg. Montreal. Canada: American Psycho-logical Association. 75-79,

and John T, Cacioppo (1981). Altitudes and Persuasion:Classic and Contemporary Approaches. Dubuque. IA: Wil-liam C. Brown.

and John T. Cacioppo (1983). "Central and PeripheralRoutes to Persuasion: Application to Advertising," in Ad-vertising and Consumer Psychology, eds. Lany Percy andArch Woodside, Lexington, MA: Lexington Books, 3-23.

and John T. Cacioppo (forthcoming a). Attitude Change:Central and Peripheral Routes to Persuasion. New York:Springer/Verlag.

and John T. Cacioppo (fonhcoming b). "The Effects ofInvolvement on Responses to Argument Quantity and Qual-ity: Central and Peripheral Routes to Persuasion." Journalof Personality and Social P.sychotogy.

, John T. Cacioppo, and Rachel Goldman (1981). "Per-sonal Involvement as a Delerminant of Argument-Based Per-suasion," Journal of Personality and Social Psychology. 41(November), 847-855.

, John T. Cacioppo. and Manin Heesacker (1981). "TheUse of Rhetorical Questions in Persuasion: A Cognitive Re-sponse Analysis," Journal of Personality and Social P.ix-chology. 40 (March), 432-440.

. Thomas Ostrom, and Timothy C. Brock (1981). •His-torical Foundations of the Cognitive Response Approach toAttitudes and Persuasion," in Cognitive Responses in Per-suasion, eds. Richard E. Petty. Thomas Ostrom. and Tim-othy C, Brock, Hiilsdale, NJ: Lawrence Erlbaum. 5-29.

. Gary L. WelLs. and Timothy C. Brock (1976). "Dis-traction Can Enhance or Reduce Yielding to Propaganda:Thought Disruption Versus Effort Justification," Journal ofPersonalit\ and Social Ps\cholog\. 34 (November).874-884,

Ray, Michael L., Alan G, Sawyer. Michael L, Rothschild. RogerM. Heeler. Edward C. Strong and Jerome B. Reed (1973),"Marketing Communication and the Hierarchy of Effects,"in New Models for Mass Communication Research. Vol. 2,ed. Peter Clarke, Beverly Hills. CA: Sage Publications.147-176.

Rhine, Ramon, and Laurence J. Severance (1970). "Ego-Involve-ment. Discrepancy, Source Credibility, and AttitudeChange." Journal of Personality and Social Psychology. 16(October), 175-190.

Schlenker. Barry R. (1978), "Attitudes as Actions: Social IdentityTheory and Consumer Research." Advances in ConsumerResearch, Vol. 5, ed. H. Keith Hunt, Ann Arbor. MI: As-sociation for Consumer Research, 352-359.

(1980). Impression Management: The Self-Concept. So-cial Identity, and Interpersonal Relations. Monterey, CA:Brooks/Cole.

Schneider. Walter and Richard M. Shiffrin (1977). "Controlledand Automatic Human Information Processing: I. Detection.Search, and Attention," Psychological Review. 84 (January),1-66.

Scott, Carol A. (1978). "Self-Perception Processes in ConsumerBehavior: Interpreting One's Own Experiences," Advancesin Consumer Research. Vol, 5. ed. H. Keith Hunt. AnnArbor. MI: Association for Consumer Research, 714-720,

Sherif, Muzifer and Carl I. Hovland (1961). Social Judgment.New Haven: Yale University Press.

Sherif, Carolyn W.. Muzifer Sherif. and Roger E. Nebergall(1965). Attitude and Attitude Cbange. Philadelphia: Saun-ders.

Steadman, Major (1969), "How Sexy Illustrations Affect BrandR e c a l l , " Journal of Advertising Research. 9 ( 1 ) . 1 5 - 1 9 ,

Sternthal, Brian and C. Samuel Craig (1974). "Fear Appeals:Revisited and Revised." Journal of Consumer Research. 1(December), 22-34.

Troutman. C, Michael and James Shanteau (1976), "Do Con-sumers Evaluate Products by Adding or Averaging AttributeInformation." Journal of Consumer Re.\earch. 3 (E>ccem-ber). 101-106.

Winer. B. J, (1971), Statistical Principles in Experimental De-sign. New York: McGraw-Hill.

Worchel. Stephen. Virginia Andreoli. and Joe Eason (1975), "Isthe Medium the Message? A Study of the Effects of Media.Communicator, and Message Characteristics on AttitudeChange," Journal of Applied Social Ps\cholog\. 5(April-June), 157-172.

Wright. Peter L. (1973). "The Cognitive Processes MediatingAcceptance of Advertising." JoMrna/ of Marketing Research.10. (Fcbruar>'), 53-62,

(1974), "Analyzing Media Effects on Advcnising Re-sponses." Public Opinion Quarlerlv. 38 (Summer).192-205.

(1980), "Message-evoked Thoughts: Persuasion ResearchUsing Thought Verbalizations. Journal of Consumer Re-search. 7 (September). 151-175,

Zimbardo. Phillip (1960). "Involvement and Communication Dis-crepancy as Determinants of Opinion Conformity," Journalof Abnormal and Social Psychology, 60 (January). 86-94.

central and peripheral routes to advertising effectiveness

Documents