The Persuasive Effect of SourceCredibility: Tests ofCognitive Response

BRIAN STERNTHALRUBY DHOLAKIACLARK LEAVITT*

Two experiments are reported identifying the circumstances In whichhigh credibility either facilitates, inhibits, or has no effect on the communica-tor's persuasiveness in relation to a less credible source. These data pro-vide support for the cognitive response view of information processingand suggest the importance of message recipient's initial opinion as adeterminant of persuasion.

In experimental investigations of the persuasive ef-fect of source credibility, it has been frequently

demonstrated that highly trustworthy and expertspokespeople induce a greater positive attitude towardthe position they advocate than do communicatorswith less credibility (cf. Stemthal, Phillips, andDholakia in press). This finding can be explained interms of cognitive response (cf. Greenwald 1968,1970; Petty, Ostrom, and Brock 1978). According tothis formulation, a message recipient's initial opinionis an important determinant of influence. In responseto a persuasive appeal, individuals rehearse their issue-relevant thoughts, as well as those presented to them.Message rejection occurs when people opposed to thecommunicator's advocacy review counterarguments toassertions made to the message. If a highly crediblesource inhibits counterarguing. whereas a less crediblesource does not, cognitive response predicts the su-perior persuasive power of a highly credible com-municator. Consistent with this interpretation. Cook(1969) reported less counterargumentation in responseto a competent source than to an incompetent source.

Despite the substantial number of studies indicatingthat a highly credible source is more persuasive than alow credibility source, this finding is less than univocal.

• Brian Stemthal is Assistant Professor, Depanment of Marketing,Nonhwestem University, Evanston, IL 60201. Ruby Dholakiais Assistant Professor of Marketing, Indian Institute of Manage-ment, Calcutta, India. Clark Leavitt is Professor. Department ofMarketing, Ohio State University, Columbus, OH 43210. The au-thors wish to thank Bobby Calder, Anthony Greenwald, AliceTybout. and Peter Wright for their comments on drafts of this paper.

McGinnies (1973) reported that a highly credible sourcewas more persuasive than a less credible communica-tor when a message recipient's initial opinion towardthe advocacy was very negative, but no credibilityefTect emerged when they held a less negative initialopinion. Dholakia and Sternthal (1977) found nosystematic credibility effect, even though the highlycredible source was perceived to be more expert andtrustworthy than the low credibility communicator.Dean, Austin, and Watts (1971) reported that theirhighly credible source induced greater persuasion whenmessage recipients were adamantly opposed to thecommunicator's position, but the less credible sourcewas more influential when the issue was one towardwhich individuals were likely to have had a favorablepredisposition (e.g., counseling against an annualx-ray). Similarly, Bochner and Insko (1966) observedthat a highly credible source was more persuasivethan a moderately credible source when the advocacywas highly discrepant, but that the moderately crediblesource induced somewhat greater persuasion when theposition advocated was relatively close to their initialopinion. Finally, Bock and Saine (1975) found that alow credibility source was more persuasive than ahighly credible source when research participantsfavored the advocacy.

Although the majority of these investigations werenot conceived as tests of cognitive response (an ex-ception is Dholakia and Sternthal 1977), their find-ings can be interpreted in terms of this formulation.Specifically, the finding that a highly credible sourceexhibits greater persuasive power than one of lowercredibility is consistent with the cognitive response

JOURNAL OF CONSUMER RESEARCH • Vol. 4 • March 1978

EFFECT OF SOURCE CREDIBILITY 253

prediction If it is assumed that message recipients hada negative predisposition toward the advocacy. Inthis situation, a credible source is more likely to inhibitcounterargumentation than a source whose credibilityis moderate or low. In turn, the reduction of counter-argumentation stimulates persuasion.

The finding that a source who lacks credibility ismore persuasive than a highly credible communicatoris explained by cognitive response if it is assumed thatindividuals have a positive predisposition toward themessage issue. In cognitive response terms, thosefavoring the advocacy will feel a greater need toinsure that a position with which they agree is beingadequately represented when the source's credibilityis low than when the source's credibility is high.Therefore, it is expected that these individuals willgenerate more support arguments and will be morepersuaded by a source lacking in credibility. The find-ing that credibility has no systematic persuasiveeffect is consistent with cognitive response if it isassumed that individuals had a neutral initial opiniontoward the issue. In this case, they are likely togenerate a similar number of support arguments andcounterarguments in response to the appeal.

In sum. by making reasonable assumptions about amessage recipient's initial opinions, cognitive responsecan order the persuasive main effect of credibilityreported in previous investigations. A more compellingtest of the cognitive response explanation of the credi-bility-persuasion relationship requires:

1. the systematic manipulation of source credibility.

2. a knowledge of the individual s initial opiniontoward the communication issue, and

3. the measurement of thoughts as well as attitudesin response to an appeal.

The purpose of the present investigation was to pro-vide such a test. To this end, two experiments wereconducted. In Experiment I, the persuasive impact of ahigh and a moderate credibility source were examinedin a context where research participants were likelyto have a favorable initial opinion toward the message.In Experiment II, this test of cognitive response wasextended by investigating the source credibility effectfor subjects who had a negative initial opinion. Thepersuasive effect of credibility for all subjects, regard-less of their initial opinion, was also examined inExperiment II.

This analysis was motivated by the desire to deter-mine whether Dholakia and Sternthal's (1977) failureto observe a credibility main effect was attributableto the aggregation of responses from individualsvarying in initial opinion within a credibility treat-ment. In both experiments, measures of attitudes andbehavior as well as measures of support argumentsand counterarguments were administered.

EXPERIMENT I: METHOD

OverviewExperiment I employed a 2 x 2 factorial design.

Subjects who were likely to have a favorable initialopinion toward the issue were presented an appealthat was attributed to either a high or moderatecredibility source. For half the subjects in each treat-ment the communicator was identified prior to the mes-sage, while for the remainder his identification was de-ferred until the end of the communication. Afterreading the appeal, its persuasive impact was meas-ured on five indices: attitudes, intentions, supportarguments, counterarguments, and behavioral re-sponse.

On the basis of cognitive response, it is hypothesizedthat the interaction between source credibility andtiming of the source's identification will be signifi-cant. Specifically, it is expected that a moderatelycredible source will induce greater persuasion than ahighly credible source when communicator identifica-tion precedes the message. As noted earlier, in thissituation the moderately credible source is expectedto stimulate greater support argumentation and, there-fore, greater acceptance than the highly credible source.In contrast, when source identification follows themessage, credibility cues are made available too lateto affect the message recipient's thought generationprocess. As a result, it is predicted that the messagewill serve as the only influence cue, and sourcecredibility will have no systematic persuasive effect.

The timing variable was chosen because it provides away of determining the reliability of the credibilityeffects obtained in this experiment. Previous investiga-tions have demonstrated that there is no systematiceffect of credibility when source identification followsthe message (cf. Ward and McGinnies 1974). If thisfinding is replicated, then confirmation of the cogni-tive response prediction that a source who lackscredibility will induce greater persuasion than a highlycredible communicator when source identificationprecedes the message is unlikely to be due to someunknown factor peculiar to the present study.

IssueGiven the objective of the present research, the com-

munication issue had to be one toward which messagerecipients evinced a positive initial disposition. At thesame time, it was important that subject's predisposi-tion was not so favorable as to preclude attitudechange because of a ceiling effect. In a pretest, itwas found that passage of the Consumer ProtectionAgency Bill (CPAB) fulfilled these criteria. Forty sub-jects, drawn from the same pool as those who partici-pated in the main study, exhibited a favorable predis-position toward the bill on two 7-point Likert-typescales. Specifically, subjects indicated agreement with

254 THE JOURNAL OF CONSUMER RESEARCH

"freedom of private business results in the exploita-tion of consumers," (AT = 4.73. 5.D. = 1.39), and dis-agreement with "consumer interests are most ef-fectively protected by business," {X = 1.40, S.D.= .65), two of the items employed in the main studyto identify the attitudinal effects of the independentvariables. For both these initial opinion items, meanattitudes were significantly more positive than the mid-point of the scale."

SubjectsThe 56 participants in this study were recruited from

two sections of an undergraduate management course.During one of the regular class hours, the studentswere asked by their class instructor (who was not theexperimenter) to participate in a study of federallegislation. Participation was voluntary. The sponsorsof the study were described as an external groupwho had sought the assistance of the course instructorto administer the study. All the students agreed toparticipate.

The use of a convenience sample comprised of stu-dents is appropriate, given that the principal aim of thepresent research is to detect the relation between vari-ables of theoretical significance. In such theory centeredresearch, there are three requirements for external va-lidity (Kruglanski 1975). First, the theoretical variablesmust be captured in the experimental setting (constructvalidity). Second, the operationalizations of the inde-pendent variables must be causally related to theoperationalizations of the dependent variables (internalvalidity). Third, the causal relationships observedshould not occur by chance (statistical conclusionvalidity). In contrast, those aspects of the situationthat are theoretically irrelevant, in this instance samplerepresentativeness, need not be considered. In fact,given that the present study entails theory-orientedresearch where individual differences are not oftheoretical interest, between-subject variance is treatedas random error that is appropriately kept to a minimumby selecting homogeneous samples (Kruglanski 1975).Therefore, the selection of a student sample for thepresent study in no way undermines the generation ofinferences regarding the persuasive impact of sourcecredibility or the adequacy of cognitive response inordering the observed effects.

ProcedureSubjects were given booklets containing the inde-

pendent and dependent variables by their instructor.The experimental task involved reading a one-pagemessage dealing with the CPAB (S.707) that was pend-ing before the United States Senate. The message pre-

U = 2.80, df = 38. p < .01 for statement one and t = 21.67.df = 38, p < .001 for statement two. All t tests are one-tailedunless otherwise noted.

sented arguments in favor of the bill and stated thefunctions of the proposed agency. Depending upon thetreatment to which subjects were assigned, the messagewas attributed to either a high or moderate credibilitysource, who was identified either before or after themessage. These experimental conditions were ad-ministered in both classes to avoid confounding theeffects of the independent variables. Once they hadread the communication, subjects were asked to com-plete a questionnaire that included the dependentmeasures.

Independent VariablesTwo independent variables were manipulated in the

study: source credibility and timing of source identifica-tion. The source was described as a supporter of thebill and the reasons for his advocacy were given.

In the high credibility condition, the source was de-scribed as a Harvard-trained lawyer with extensiveexperience in the area of consumer issues and arecognized expert whose advice was widely sought.His support of this particular bill was made distinctiveby describing him as a person who generally did notfavor government controls, but who supported theCPAB. The less credible source was portrayed as anindividual with no special expertise, but one whowas interested in consumer protection because of ajob opportunity as consumer lobbyist. He was de-scribed as holding socialistic views with strong opin-ions in favor of all government controls.

The second independent variable involved manipu-lating the timing of the source's identification. Forhalf the subjects, the source description appearedbefore the written communication, while for the re-maining subjects the source description was deferreduntil the end of the communication.

Measuring InstrumentsFive measures of persuasion, as well as a credibility

manipulation check, served as the dependent variables.Message recipient's attitude toward the bill was deter-mined first by administering the two attitudinal itemsused in the pretest, plus two items that focused moreexplicitly on the position advocated in the appeal:"Passage of the Consumer Protection Agency Billwould make me feel good" and "It is my belief thatconsumers do not need protection." Responses to thesequestions were measured on a seven-point scale rang-ing from strongly agree to strongly disagree. The fourattitude questions were summed lo form an overallscore because they were highly interrelated (Cronbaeha = .64). These attitude measures were followed by abehavioral intention item where subjects indicatedtheir likelihood of supporting the bill on a lOO-pointscale.

The next measure tapped message recipients' sup-port arguments and counterarguments. Argumentation

EFFECT OF SOURCE CREDIBILITY 255

was determined by having subjects list all the thoughtsthat came to mind in response to the appeal. Threeminutes were given to complete this task. Subjectsthen categorized their thoughts as either favorable orunfavorable to passage of the bill. For each subject,the thoughts listed that were consistent with the posi-tion taken in the message were summed to yield asupport argument score, while the thoughts listed thatopposed the message position were summed to yielda counterargument score.

The argumentation measurements were followed bytwo source evaluation measures. One measured sub-jects' perception of the trustworthiness and expertiseof the source. Six items (three related to expertiseand three related to trust), each rated on a seven-point semantic differential scale, were used for this pur-pose. Items included: expert — not expert; experienced—not experienced; trained—untrained; trustworthy—not trustworthy; moral — immoral; good — bad.

A second set of dependent variables which was notmanipulated in the present study was related to attrac-tiveness of the source. These nonequivalent sourcecredibility items included: attractive — unattractive;dynamic — not dynamic: aggressive — not aggressive.These items are nonequivalent in the sense that un-like responses to the expertise and trust scales, re-sponses to the attractiveness items should not be sensi-tive to the dependent variable. They were administeredto determine whether demand characteristics ac-counted for the subject's perception of the source'scredibility (Orne 1969). If the source credibility in-duction was effective and not attributable to demandcharacteristics, then subjects would perceive thehighly credible source to be more trustworthy andexpert than the moderately credible source, but theyshould not find one source more attractive than theother.

Behavioral compliance was determined by whethersubjects signed a petition that was lo be sent to theirsenators in Washington urging a vote for the bill.

RESULTS

Manipulation CheckA manipulation check determined whether the

source credibility induction was effective. Subjects whoreceived the message from the highly credible sourceperceived the communicator to be significantly moretrustworthy and expert than did subjects who receivedthe message from the moderately credible person.^There was no difference between the high and moderatecredibility sources on the nonequivalent attractivenessmeasures.^ Since the credibility manipulation involved

' High credibility: X = 28.86. S.D. = 8.64, n = 28. Low credi-bility: A* = 20.57. S.D. = 8.01, ;i = 28, f = i.n.df = 54. p < .001,

> High credibility: X = 13.07. S.D. = 4.52. Low credibility: X= 13.68, 5.Z). = 3,90.

varying trustworthiness and expertise but not attrac-tiveness, these data indicate that the credibility induc-tion was effective and unlikely to be attributable todemand characteristics.

Although the means for the highly and moderatelycredible sources differ significantly on the trustworthi-ness and expertise dimensions, the moderately crediblesource was not perceived in a totally negative light.In fact, there was no difference between the moderatelycredible source and the scale midpoint (/ < 1). Giventhat the experiment entailed having a source ask forcompliance with a request to sign a petition, it wasnecessary that the less credible source have somefavorable attributes. Indeed, in pretesting the credi-bility induction it was found that subjects perceivedthe situation to be contrived when the moderatelycredible source's biography was less favorable than theone used in this experiment.

Attitudinal ResponseSubjects' attitudina! responses categorized by inde-

pendent variables are reported in Table 1. To deter-mine the effects ofthe experimental variables on sub-jects' attitudes toward the CPAB, an analysis ofvariance was performed employing the sum ofthe atti-tudinal responses as the dependent measure. It wasfound that both source credibility (F < 1) and timingof the source's identification (F = 1.34, df- 1,52,p > .20) did not have a significant effect on attitudes.As predicted, however, a significant disordinal interac-tion between source credibility and the time of thesource identification was found (f" = 15.97,1^/= 1,52,p < .001). The moderately credible source was morepersuasive than the high credibility communicatorwhen the credibility cue was presented before the mes-sage, whereas the highly credible source induced amore positive attitude when the credibility cue followedthe communication as shown in the figure.

Although a significant source credibility x timinginteraction is necessary to confirm cognitive responsepredictions, it is not sufficient. It must also be demon-strated that:

1. the moderately credible source is significantlymore persuasive than the highly credible sourcewhen the source credibility cues precede themessage, and

2. the high and moderate credibility sources do notdiffer significantly in their persuasive powerwhen the credibility cues appear after the mes-sage.

Using the Newman-Keuls procedure to make thesecontrasts, the above predictions were confirmed. Whensource identification preceded the message, the moder-ately credible source induced a more positive attitudethan the highly credible source (q = 4.96, df= 52,p < .05), whereas there was no systematic credibility

256 THE JOURNAL OF CONSUMER RESEARCH

TABLE 1

MEAN (X) AND STANDARD DEVIATIONS (S.D.) FOR ATTITUDES, SUPPORT ARGUMENTS COUNTERARGUMENTS ANDINTENTIONS CATEGORIZED BY INDEPENDENT VARIABLES, SOURCE CREDIBlLlPi', AND

TIMING OF THE CREDIBILITY INDUCTION

Timing of thecredibility induction

Source before message (SM)

Source after message (MS)

Dependentmeasure

AttitudeSupport argumentsCounterargumentsIntention

AttitudeSupport argumentsCounterargumentsIntention

X

19,640.711.43

62.14

22.711.290.79

78.29

Source

High

S.D.

5.171.201.34

28.06

2.891.821.05

15.23

credibility

Moderate

X

25.001.860.93

89.00

19.430.931.57

59.64

S.D.

3,041.921,00

11.90

4.601.211.87

27.16

effect when the source identification was deferred untilafter the message (<? = 2.97, df -= 52, p > .05). Fur-thermore, the high credibility source induced a sig-nificantly more positive attitude toward the bill whenthe source was identified after—rather than beforethe message {q = 2.78, df = 52, p = .05). However,the moderately credible source was significantly moreinfluential when introduced before the message (q= 5.\5, df= 52, F < .01).

Argumentation-CounterargumentationA more compelling test of cognitive response re-

quires operationalizing the variables presumed to medi-

FIGURE

SOURCE CREDIBILITY x TIMING OF CREDIBILITY INDUCTION

Attitude

25242322212019

Source-Message Message SourceTiming of the SourceCredibility Induction

• - High Credibility Source

A-Moderate Credibility Source

ate persuasion — support argumentation and counter-argumentation. To address this issue, the effects onargumentation of source credibility and timing of thesource identification were examined (See Table 1).The source credibility x timing of source identifica-tion was marginally significant for support argumenta-tion {F = 3.19, df = 1,52, p = .08) and counterargu-mentation (F = 3.13, df= 1,52, p = .08). As pre-dicted, when source identification preceded themessage, subjects generated more support argumentsin response to the moderately credible source than thehighly credible one (/ = 2.74, df = 54, p < .01). Alsocongenial to the cognitive response formulation was thefinding of no differences in support argumentationattributable to source credibility when source identifica-tion occurred after the message (/ < 1) and the ob-servation of no credibility effects on counterargu-mentation (identification before message, / = 1.58,df = 54, /? > .10; identification after message, t= ].95, df= 54, p > .05).-

To evaluate the mediating role of support argu-ments and counterarguments further, the effect ofthetreatments on attitude was determined when the num-ber of each of these thought types was treated as acovariate. When the number of support argumentsserved as a covariate, the independent variables stillhad a significant effect on attitudes (F = 4.62, df= 3,51, p < .01), Nevertheless, support arguments

accounted for a substantial and statistically significantportion (38 percent) ofthe explained attitude variance(F = 3.63, df= 3,51, p < .05). Similarly, when thenumber of counterarguments was treated as a co-variate, the treatment-attitude relationship was main-tained (F = 5.01, df = 3,51, p < .005), though coun-terarguments accounted for a significant portion oftheexplained variance (F = 3.87, df= 3,51,/? < .05).

* Two-tailed tests v^ere used in making these contrasts, becauseno source credibility effects were predicted from cognitive re-sponse.

EFFECT OF SOURCE CREDIBILITY 257

Intention and BehaviorAnalysis of effects on behavioral intention yielded

essentially the same results as just reported for alti-tudes. An analysis of variance indicated that only thesource credibility x timing interaction was significant(F = \5.H,df= \,52,p < .001). Systematic compari-son of treatments using the Newman-Keuls procedurerevealed that the significant interaction was attributableto the moderately credible sources inducing a signifi-cantly greater intention to support the bill than thehighly credible source when the credibility cuespreceded the message {q - 4.62, df= 52, p < .01).When the source's identification followed the message,there was no systematic effect of source credibility onsubjects' behavioral intentions iq = 3.20, df - 52, p> .05).

Finally, neither source credibility nor the timing ofthe source credibility induction had a significanteffect on subjects' behavioral responses (x̂ = 3.64,df = 3,p > .30). However, the behavioral data exhibita pattern similar to that obtained for attitudinal andintention responses. Subjects complied more with themoderately credible source when the source induc-tion preceded the message, whereas the highly crediblesource induced greater compliance when the sourceinduction was deferred until after the message.

EXPERIMENT II: METHODThe results of Experiment I provide support for the

cognitive response formulation where message recip-ients were likely to have been favorably predisposedto the communication issue. To provide a strongertest of cognitive response predictions, the effect ofsource credibility upon favorably and negatively pre-disposed individuals was investigated in Experiment II.Subjects were drawn from the pool used in Experi-ment I. The question battery included five Likert-typescales pertaining to participants' attitudes toward theissue.^ Subjects' responses on these five scales weresummed to yield an initial opinion score. Subjectswhose score was at or above the median were classi-fied as having a "positive initial opinion" toward theissue; those with scores below the median were classi-fied as having a "negative initial opinion."®

Approximately five weeks later, a second experi-menter requested that subjects participate in a studyto determine individuals' feelings about the CPAB.

* The scales included: "It is more important to enforce theconsumer protection legislation that is already on the books than lopass more laws to protect consumers." "It is my belief thatconsumers do not need protection." "Consumers" interests are mosteffeclively protected by business." "Freedom of private businessfrom government regulation results in exploitation of consumers.""The only way to ensure that consumers are noi exploited is topass laws protecting consumers." Subjects responded to eachstatement on a five-point scale.

* A "neutral" initial opinion group was not established be-cause it would have made cell sizes very small.

The 37 participants followed essentially the procedureused In Experiment I, although the timing of the source'sidentification was not manipulated in Experiment II.Subjects read a communication attributed to either ahighly or moderately credible source. The credibilityinduction appeared at the outset of the message. Then,the dependent measures were administered. First, thesubjects completed the thought sampling question bylisting and rating their support arguments and counter-arguments relevant to passage of the CPAB. Threeminutes were given to complete this task. Next, sub-jects rated their attitudes toward the bill on ten 7-pointLikert-type scales. Five of these scales were the sameas those administered in the pretest, while the remain-ing five were specific to the bill.' This was followed bythe nine 7-point semantic differential scales pertainingto the source's attributes of character used in Experi-ment I. Finally, subjects completed the request tosign the petition by signing it or indicating their re-fusal to support the bill.

The dependent variable strategy employed in Experi-ment II departs from that used in Experiment I inseveral ways. The number of attitude scales were in-creased in Experiment II to ensure the reliability ofthe attitude measure (a = .78). In addition, the orderin which the dependent variables were administeredwas altered. Rather than have subjects complete theattitude scales prior to administering the thoughtsmeasure, as was the case in Experiment I, subjects'support arguments and counterarguments were elicitedbefore their attitudes were measured in ExperimentII. This change eliminated the possibility that responseson the scaled attitude measures would affect thesupport arguments and counterarguments (Wright1977).

On the basis of cognitive response, two predictionswere made. It was hypothesized that subjects with apositive disposition toward the communication issuewould be more persuaded by the moderately crediblesource than by the high credibility source, replicatingthe result found in Experiment I. In contrast, for indi-viduals who were negatively predisposed to the issue,a highly credible source was expected to be morepersuasive.

RESULTS

Manipulation CheckTo check the source credibility induction, the rat-

ings given on the trust and expertise items by sub-jects assigned to the high credibility condition weresummed and compared with the summed ratings

' T h e attitude scales specific to the bill included: "Passage ofthe Consumer Protection Agency Bill (CPAB) would make me feelgood; ' "1 feel very negatively toward the CPAB." "i wouldpersonally support establishing the CPAB." '[t is my belief thaithe CPAB merits public support." "If asked. I would not bewilling to give my time lo ensure passage of the CPAB."

258 THE JOURNAL OF CONSUMER RESEARCH

TABLE 2

MEAN (X) AND STANDAflD DEVIATIONS {S.D.) FOR ATTITUDES, SUPPORT ARGUMENTS, AND COUNTERARGUMENTSCATEGORIZED BY SOURCE CREDIBILITY TREATMENT AND TYPE OF SAMPLE

Sample

All subjects

Positive initial opinion

Negative initial opinion

Measure

AttitudeSupport argumentsCounterarguments

AttitudeSupport argumentsCounterarguments

AttitudeSupport argumentsCounterarguments

High

X

46.171.111,89

41.50.75

2.00

49.901.401.80

Source

S.D.

9.711.051.29

7.38,63

1.50

9.741.111.08

credibility

X

41.161.632.16

47,782.112.00

35.201.202,30

Moderate

S.D.

11.161.811.76

6,991.641,49

11,131.461.85

given by subjects assigned to the moderate credibilityconditioti. It was found that the highly credible source(A-= 30.78,5.0. =5.2],n = 18) was perceived to bemore trustworthy and expert than the moderatelycredible communicator (A* = 20.47. S.D. = 4.29, n= 19; / = 6.57, df= 35, p < .01). No treatmentdifferences were obtained with respect to the non-equivalent attractiveness measures.** The sourcecredibility induction was also successful for the subsetof the message recipients who had a positive initialopinion toward the issue^ and who had a negativeinitial opinion."*

Effect of Source CredibilityThe data analysis involved the responses of the

entire sample, those with a positive initial opiniontoward the issue and those with a negative initialopinion. The means and standard deviations of thedependent variables for each of these groups, cate-gorized by the independent variable, are shown inTable 2. Analysis of the mean attitudinal responsesobtained for the entire sample indicated that the highlycredible source induced a more positive attitudetoward the CPAB than did the moderately crediblesource, though this difference did not reach statisticalsignificance (/ = 1,46, df = 35, p > .05). This findingreplicates the result obtained in an experiment using asimilar paradigm reported by Dholakia and Sternthal(1977). Furthermore, source credibility did not have asignificant effect upon support argument (/ = 1.08,df = 35. p > .20) or counterargument production

•High credibility, X = 12.79. S.D. = 2.94; low credibility X= n .50 .5 .D . = 1.95;/ = 1.59,^/= 35, p > .05,

* High credibility X = 29. S.D. = 4.61, n = 8; low credibilityX = 2], S.D. = 2.90./I = 9: f = 4,23, rf/= ]5,p < .Oi.

'". High credibility A* = 32.20, S.D. = 5.15, n = 10; low credi-bility A* = 20, S.D. = 5.20, n = 10; ; = 5.28, d / = 18 /i < 01

(r < 1). Also the communicator's credibility did notaffect compliance with the request to sign the petitionfavoring the CPAB; 28 percent of those receiving themessage from the highly credible source signed thepetition and 32 percent of those receiving the appealfrom the moderately credible source signed.

Examination of the impact of source credibility forsubjects with a relatively positive initial opinion towardthe communication issue (Table 2) indicates that themoderately credible source induced a more positiveattitude toward the issue than the highly credible com-municator (/ = \.19,df= 15, p < .05). Furthermore,those who received a message attributed to the moder-ately credible source generated more support argu-ments than did subjects for whom it was attributed to ahighly credible source (/ = 2.19, df = 15, p < .025).When the number of support arguments were treatedas a covariate, the efTect of source credibility on atti-tudes was not significant {F = .84, df = 1, 14, p > .35),In contrast, source credibility had no systematiceffect on the generation of counterarguments [t < I),and treating the number of counterarguments as a co-variate did not have a substantial effect on the credi-bility-attitude relationship (F = 6.88, df= 1.14. p= .02). Finally, the source's credibility did not have asignificant influence on subject's response to the re-quest to sign the petition (x^ = 1.55, df = I, p> .20).although compliance was somewhat greater in re-sponse to the moderately credible source (56 percent)than to the highly credible communicator (25 percent).These findings essentially replicate the results reportedfor the treatment where source identification precededthe message in Expetiment I.

These analyses were repeated for message recipientswho had a relatively negative initial opinion towardthe communication issue. Consistent with the findingsreported in most previous credibility main effectstudies, the highly credible source induced a more

EFFECT OF SOURCE CREDIBILITY 259

positive attitude toward his advocacy than did the lesscredible communicator (/ = 3.14, df = 18, p < .01).However, source credibility did not have a systematiceffect on either counterargument (/ < 1) or supportargument (/ < 1) generation. And, treating the num-ber of thoughts as a covariate did not affect thecredibility-attitude relationship which was significantwhether support arguments ( f = 8.53, df = 1,17,p = .01) or counterarguments {F = 9.35. df= 1.17.p < .01) served as the covariate. Furthermore, credi-bility did not systematically affect compliance (x^= \.QS, df= ],p> .20), although those receiving thecommunication from the highly credible source showeda greater tendency to sign the petition than thosewho received it from ihe less credible source (30percent versus 10 percent).

DISCUSSION

Three major findings emerge from these experi-ments. First, when message recipients were favorablypredisposed to an issue and the communicator wasidentified prior to the message, the moderately crediblesource induced more agreement and support argu-mentation than did the highly credible communicator.This result was obtained in both Experiments I and II.Second, when the source was identified after the mes-sage, credibility had no systematic effect on attitudes,support argumentation, or counterargumentation(Experiment I). Third, when message recipients werenegatively predisposed to the communication issue,the highly credible source induced more agreementthan the less credible communicator (Experiment II).though source credibility did not affect thought genera-tion.

These results generally support the cognitive re-sponse predictions. In cognitive response terms, themoderate credibility source induced greater positiveattitude and support argumentation when identifiedat the outset of the communication, presumably be-cause message recipients felt a need to bolster supportfor a position they favored when the communicatorwas of questionable credibility. They felt less inclinedto engage in this cognitive work when a highlycredible source was presenting a favored position.In contrast, there was no source credibility effectwhen source identification followed the message be-cause the credibility cue was made available too lateto mediate the thought generation process.

The failure to observe a significant effect of sourcecredibility on counterargumentation in both of thetiming conditions examined in Experiment I is pre-dicted by the cognitive response position. Becausesubjects had a favorable disposition toward the issue,it was anticipated that the source credibility cuewould have its predominant effect on support argu-mentation rather than counterargumentation. Thesuperior persuasive power of the highly credible

source when message recipients were initially opposedto the advocacy is also predicted by cognitive re-sponse. In this situation, a highly credible sourceserves to inhibit counterargumentation, whereas aless credible source facilitates it.

Also of interest is the finding (Experiment II) thatsource credibility has a systematic effect on thoughtgeneration and on attitudes when subjects are cate-gorized by their initial opinion, but not when the re-sponses of these groups are aggregated. This resultmay explain the failure to observe a source credibilityeffect in a previous investigation conducted byDholakia and Sternthal (1977) using a paradigm similarto that in the present study. By aggregating the re-sponses from subjects differing in initial opinion,Dholakia and Stemthal may have washed out thecredibility effect.

Despite the general support for the cognitive re-sponse formation, several findings are at odds with thepredictions made from this theory. One pertains tothe effect of the independent variables on messagerecipient's compliance with the request to sign the peti-tion favoring the bill. Although the compliance datawere in the same direction as the other measures ofpersuasion in both experiments, source credibility didnot have a statistically significant effect on an individual'sbehavioral response. This result is not only inconsistentwith the cognitive response prediction, but it questionsthe importance of the source credibility variable whenconsumer's behavioral response is of focal concern.However, before it is concluded that source credibilitydoes not have a systematic effect on behavior, orthat cognitive response does not order this effect, thenegative finding must be demonstrated in situationswhere the communicator's attributes are highly salientto message recipients. The present investigation doesnot provide such a test. Salience of the source credi-bility may be achieved by presenting the communicatorin person. Indeed, when this approach was used,source credibility did have a significant effect on be-havior (Brock 1965; Woodside and Davenport 1976).

The finding (Experiment II) that individuals ex-posed to the moderately credible source who werenegatively predisposed to the issue did not generatesignificantly more counterarguments than those re-ceiving the message from the highly credible sourceis also inconsistent with cognitive response. This re-sult may be attributable to the fact that subjects inthe moderate credibility-negative initial opinion condi-tion were unable to generate a substantial number ofcounterarguments because of their lack of familiaritywith the issue. Thus, while they exhibited a lessfavorable attitude toward the bill when source credi-bility was moderate, they were unable to verbalizemany thoughts related to that disposition.

Finally, although the credibility treatment has a sub-stantial effect on the thoughts generated by subjectsin Experiment I. contrary to the cognitive response

260 THE JOURNAL OF CONSUMER RESEARCH

prediction, treating thoughts as a eovariate still resultedin a significant credibility-attitude relationship. Thisfinding may be attributable to thought samples beingaffected by the attitude and intention dependent meas-ures which preceded them. Support for this contentionemerges from the analysis of responses from thosewho were favorably predisposed to the message issuein Experiment II. In this instance, thought sampleswere generated before other dependent measures.When this procedure was followed, treating supportarguments as a covariate yielded a nonsignificantsource credibility effect on attitude, as predicted bycognitive response. This finding suggests that thoughtsampling measures should be administered prior toother dependent measures in future tests of cognitiveresponse (cf. Wright 1977).

[Received February 1977. Revised September 1977.]

REFERENCESBochner, Stephen, and Insko, Chester (1966), "Communica-

tion Discrepancy, Source Credibility, and OpinionChange," Journal of Personality and Social Psychology,4, 614-21.

Bock, Donald, and Saine, Thomas (1975), "The Impact ofSource Credibility, Attitude Valence, and Task Sensi-tization on Trait Errors in Speech Evaluation," SpeechMonographs, 37, 342-58.

Brock, Timothy (1965), "Communicator-Recipient Simi-larity and I>ecision Cha^nge." Journal of Personality andSocial Psychology, 1,650-4.

Cook, Thomas (1%9), "Competence, Counterarguing, andAttitude Change," Journal of Personality. 37, 342-58.

Dean, Robert, Austin, John, and Watts, William (1971),"Forewarning Effects in Persuasion: Field and Class-

room Experiments," Journal of Personality and SocialPsychology. 18, 210-21.

Dholakia, Ruby and Stemthai, Brian (1977), "HighlyCredible Sources: Persuasive Facilitators or PersuasiveLiabilities?" The Journal of Consumer Research. 3,223-32.

Greenwald, Anthony (1968), "Cognitive Learning, Cogm-tive Response to Persuasion, and Attitude Change,"in Psychological Foundations of Attitudes, eds. An-thony Greenwald, Timothy Brock, and Thomas Ostrom,New York: Academic Press, 147-70.

Greenwald, Anthony (1970), "Initial Opinion and Responseto Persuasion," proposal to the National ScienceFoundation.

Kruglanski, Arie (1975), "The Two Meanings of ExternalValidity," Human Relations. 28, 653-59.

McGinnies, Elliot (1973), "Initial Attitude, Source Credi-bility, and Involvement as Factors in Persuasion,"Journal of Experimental Social Psychology. 9, 285-96.

Ome, Martin (1969), "Demand Characteristics and the Con-cept of Quasi-Controls," in Artifact in Behavioral Re-search, eds. R. Rosenthal and R. Rosnow, New York;Academic Press.

Petty, Roger, Ostrom, Thomas, and Brock, Timothy (1978),Cognitive Responses in Persuasive Communication: AText in Attitude Change, New York: McGraw Hill.

Stemthai, Brian, Phillips, Lynn, and Dholakia, Ruby (inpress), "The Persuasive Effect of Source Credibility: ASituational Analysis," Public Opinion Quarterly.

Ward, Charles, and McGinnies, Elliot (1974), "PersuasiveEffects of Early and Late Mention of Credible and Non-credible Sources," yoorna/o/PsycAo/ogy, 86, 17-23.

Woodside, Arch and Davenport, James (1976), "Effects ofPrice and Salesman Expertise on Customer PurchasingBehavior," The Journal of Business. 49, 51-9.

Wright, Peter (1977), "Research on Ad-Stimulated ThoughtProcesses; A Review," Faculty Working Paper, Gradu-ate School of Business. Stanford University.