UNIVERSITY OF 
 
 ILLINOIS LIBRARY 
 
 AT URBANA-CHAMPAIGN 
 
 BOOKSTACKS 
 
be charged a min «.. * eIow " Yo « may 
 
 JAN 42000 
 
 NOV 2 
 
 1999 
 
 L162 
 
Digitized by the Internet Archive 
 
 in 2011 with funding from 
 
 University of Illinois Urbana-Champaign 
 
 http://www.archive.org/details/mediatingprocess1532hast 
 

 
 BEBR 
 
 FACULTY WORKING 
 PAPER NO. 89-1532 
 
 Mediating Processes for 
 Source Credibility Effects 
 in Advertising: Review, 
 Implications, and Future 
 Research Directions 
 
 r HrT LIRRAl^ Of 
 
 FEB 2 3 1989 
 
 rtbil V Oh ILLIlxlUlS 
 
 Manoj Hastak 
 Jong-Won Park 
 
 College of Commerce and Business Administration 
 Bureau of Economic and Business Research 
 University of Illinois Urbana-Champaign 
 
BEBR 
 
 FACULTY WORKING PAPER NO. 89-1532 
 
 College of Commerce and Business Administration 
 
 University of Illinois at Urbana- Champaign 
 
 February 1989 
 
 Mediating Processes for Source Credibility Effects in Advertising: 
 Review, Implications, and Future Research Directions 
 
 Manoj Hastak, Assistant Professor 
 Department of Business Administration 
 
 Jong-Won Park 
 Graduate Student 
 
Abstract 
 
 Source credibility has been observed to produce favorable, neutral, and 
 sometimes even unfavorable effects on attitudes in persuasion contexts. These 
 diverse and conflicting findings can best be reconciled if it is first recognized 
 that effects due to variations in source credibility on attitude are likely to be 
 mediated by multiple distinct mechanisms or processes. In this paper, we discuss 
 several such mediating processes, and consider the conditions under which they 
 are likely to operate. We also discuss empirical evidence supporting the 
 existence of these processes, and explicate the implications of this research for 
 the use of credible sources in advertising. Finally, we suggest several issues for 
 future research. 
 
Mediating Processes for Source Credibility Effects in Advertising: 
 Review, Implications, and Future Research Directions 
 
 It has long been recognized that characteristics of the source of an 
 advertisement (either explicitly identified or implicitly understood) can influence 
 ad effects on the recipient. In particular, the use of a credible spokesperson in 
 advertising is commonplace, and clearly based on the assumption that source 
 credibility improves the persuasive impact of advertising messages. Given the 
 intuitive appeal and early empirical support for this assumption (Hovland and 
 Weiss 1951), it is not surprising that there has been minimal recent research 
 examining the effects of credibility on persuasion in advertising contexts (for 
 exceptions, see Frieden 1984; Friedman and Friedman 1979; Harmon and Coney 
 1982; Mizerski, Hunt, and Patti 1978). However, a growing body of evidence in 
 the psychology literature suggests that source credibility effects on persuasion 
 are far more complex than previously believed. Specifically, this literature 
 suggests that (a) source credibility may have favorable, neutral, and sometimes 
 even unfavorable effects on post exposure attitude towards the communication 
 topic, (b) these effects appear to be highly contingent on the specific levels of 
 other individual difference variables (e.g., involvement, prior knowledge) and 
 situational variables (e.g., modality, time compression, channel noise) that are 
 operating in the reception environment, and (c) a number of different 
 theoretical models and frameworks such as Kelman's functional approach to 
 social influence processes (Friedman and Friedman 1979; Mills and Harvey 1972), 
 the Yale reinforcement approach to persuasion (Maddux and Rogers 1980) , 
 cognitive response theory (Hass 1981; Sternthal, Dholakia, and Leavitt 1978), 
 attribution theory (Dholakia and Sternthal 1977), and the Elaboration Likelihood 
 Model (Petty and Cacioppo 1984, 1986a, 1986b) can account for these effects to 
 
varying degrees. Thus, advertisers are faced with an array of apparently 
 conflicting findings and competing explanations as they consider the merits of 
 using credible sources in their communications. 
 
 In this paper, we present a synthesis of recent research on source 
 credibility effects in persuasion, and consider the implications of this research 
 for advertising practitioners. Our focus here is primarily on attitudinal effects 
 (see Sternthal, Phillips, and Dholakia 1978 for a review of the literature on 
 behavioral effects due to source credibility) . Extant literature suggests that 
 source credibility can produce both direct and indirect effects on attitude 
 towards the advertised object. By direct effects, we mean that source credibility 
 influences final attitude without affecting processing of the ad message itself. 
 Indirect effects refer to the possibility that source credibility affects attitude 
 by modifying, changing, or otherwise altering message processing activity. In 
 the following sections, we discuss several direct and indirect mediating 
 mechanisms for credibility effects on attitude that have been proposed in the 
 literature. We also examine the conditions under which these mechanisms are 
 likely to operate from the perspective of two currently popular theoretical 
 frameworks proposed by Petty and Cacioppo (1986a) and Sternthal, Dholakia, 
 and Leavitt (1978). Finally, we discuss the implications of each of these 
 mediating mechanisms for source credibility effects in advertising, and isolate 
 several research issues that hold special promise for future research. 
 
 Direct Effects 
 
 Direct effects due to credibility on attitude in a persuasion context have 
 been most clearly explicated in research on central versus peripheral routes to 
 
persuasion by Petty, Cacioppo, and their colleagues (e.g., Petty, Cacioppo, and 
 Goldman 198.1). In this framework, recipients of a persuasive message can adopt 
 one of two distinct processing strategies as they examine the message. A 
 central route to persuasion is said to occur if the recipients carefully examine 
 and process those cues in the message that they believe are central to a 
 meaningful and logical evaluation of the communication object. In contrast, a 
 peripheral route to persuasion results when recipients evaluate the 
 communication object based on a rather cursory and superficial analysis of 
 readily available and salient cues in the communication, regardless of whether 
 or not these cues are meaningfully related to the communication object. 
 Specifically, attitudes may be formed or changed via peripheral processing either 
 because the object is associated with positive or negative cues, or because the 
 individual can make a quick evaluative inference about the object based on 
 simple cues in the persuasion context. 
 
 Petty, Cacioppo, and Goldman (1981; Petty and Cacioppo 1986a) have 
 proposed that motivation and ability to process a communication are the key 
 determinants of whether central or peripheral processing will occur in a 
 particular situation. Motivation to process depends on such antecedent variables 
 as involvement with the communication issue, need for cognition, and 
 forewarning about the persuasive intent of a communication, while ability may 
 be determined by both intrinsic factors such as prior knowledge and expertise, 
 and situational factors such as level of distraction and noise in the message 
 reception environment (see Petty and Cacioppo 1986b; Batra and Ray 1986; 
 Andrews 1988 for a detailed discussion on the antecedents of motivation and 
 ability to process persuasive communication) . High motivation and ability to 
 process (e.g, high involvement with and knowledge about the communication 
 
issue) leads to central processing, whereas peripheral processing results if either 
 the motivation or the ability to process the communication is low. 
 
 The "two routes to persuasion" framework suggests that source 
 characteristics such as credibility can have direct effects on attitude under both 
 central and peripheral processing. To illustrate, consider an advertisement for 
 brand X cereal which shows a physician eating the cereal, and also makes the 
 claims that (a) brand X is high in fiber, and (b) brand X stays crunchy in milk. 
 Under central processing, information about the source as well as other stated 
 claims in the ad are likely to be considered relevant to judging the true merits 
 of brand X. Thus, source credibility and believability of the stated claims should 
 independently contribute to the overall evaluation of the brand. In this 
 instance, the source is acting as a persuasive argument (i.e., a central cue), and 
 is processed in a manner akin to other arguments or claims in the message. 
 Under peripheral processing, recipients are not expected to diligently process 
 the stated claims in the message. However, the picture of a credible source 
 consuming the product could be used to rapidly generate a favorable evaluation 
 of brand X. In this case, the credible source acts as a peripheral cue that 
 triggers an overlearned heuristic (i.e., expert sources should be trusted). 
 There is now considerable empirical evidence to support the claim for 
 direct credibility effects on attitude -- especially under peripheral processing. 
 For instance, Petty and Cacioppo (1986a) have shown that when subjects are 
 uninvolved with an advocacy message (and hence unmotivated to process the 
 message), attitude is strongly affected by a peripheral cue such as source 
 credibility, but is unaffected by the quality of arguments in the message. 
 Similarly, Kiesler and Mathog (1968) obtained strong effects due to source 
 credibility on attitude only when distraction was high -- i.e., when subjects 
 
were in a peripheral processing mode. In contrast, evidence supporting 
 credibility effects under central processing is more limited (see Dean, Austin, 
 and Watts 1971) suggesting, perhaps, that these effects are weaker, and harder 
 to isolate. A likely reason for this is that persuasive arguments in the message 
 are perceived to be more relevant to judging the true merits of an advocacy, 
 and hence overshadow the effects of other variables such as source credibility 
 when involvement is high. 
 
 The preceding analysis generates relatively straightforward implications for 
 the use of credible sources in advertising. Direct effects of source credibility 
 are expected to follow the intuitively appealing experto crede phenomenon -- 
 credible sources should consistently produce more favorable attitudinal effects 
 than sources lacking in credibility. The timing of source introduction should not 
 influence the strength of these effects since the source does not exert its 
 influence on attitude by first affecting message processing. Consequently, the 
 advertiser has considerable latitude in deciding whether a source should be 
 introduced early or late in a commercial message. 
 
 Note that although source credibility is expected to produce similar 
 (positive) direct effects under both central and peripheral processing, there are 
 differences which have significant consequences for advertising. Attitudinal 
 effects induced through central processing are based on a detailed assessment of 
 message content, and are thus likely to be more enduring, and less susceptible 
 to counterattack than changes induced via the peripheral route (see Petty and 
 Cacioppo 1986a for details) . Consequently, credibility effects under peripheral 
 processing would need to be frequently augmented (perhaps via repetition, and 
 through point of purchase reminders) unless only short term attitudinal and 
 behavioral impact is desired. In contrast, relatively few exposures should be 
 
sufficient to maintain effects through central processing, although the 
 magnitude of these effects is likely to be more modest. 
 
 The variables that moderate direct effects due to source credibility on 
 attitude will also markedly differ depending on which of the two routes to 
 persuasion is being followed. Under central processing, the effects of a cue on 
 attitude are contingent on the persuasive quality of that cue, i.e., the extent to 
 which that cue is considered relevant to logically assessing the communication 
 object. Thus, the source cue should compete with other "central" cues in the ad 
 reception environment (e.g., persuasive quality of message arguments) for impact 
 on attitude. Indeed, the available evidence suggests that very strong or very 
 weak messages may dilute, or even completely eliminate credibility effects when 
 involvement is high (e.g., Petty, Cacioppo, and Goldman 1981). Thus, it may be 
 fruitful for advertisers to employ credible sources only when other central cues 
 such as message quality are either non-existent or at moderate levels. Under 
 peripheral processing, the relative impact of peripheral cues in the ad (such as 
 credibility, source attractiveness, spontaneous emotional responses, ad 
 attractiveness, etc.) should depend on the relative salience and vividness of 
 these cues -- i.e., the ease with which these cues can be attended to and 
 processed. This suggests, for instance, that if the spokesperson in an 
 advertisement is both attractive and credible, we would expect attitudinal 
 effects due to attractiveness to be stronger because of the vividness of this 
 cue. Evidence supporting this claim was obtained in a study on endorser effects 
 by Pallak, Murroni, and Koch (1984). Support for effects due to vividness of a 
 source cue under peripheral processing is also provided by Chaiken and Eagly 
 (1983) who obtained stronger source effects for audio- visual messages (vivid 
 source cue) than for print messages (less vivid cue) . 
 
Needed Research. Many of the practical implications for source credibility 
 
 effects that were drawn in the previous section are untested, and deserve 
 future research attention. For instance, several studies designed to investigate 
 the "sleeper effect" phenomenon have shown that credibility-induced attitudinal 
 effects decay rapidly over time (Hovland and Weiss 1951; Kelman and Hovland 
 1953; Shulman and Worrall 1970; Watts and McGuire 1964; see Gillig and 
 Greenwald 1974 for a review, Maddux and Rogers 1980 for an exception) . 
 However, these data cannot be interpreted as support for the transient nature 
 of credibility effects under peripheral processing since there is no evidence to 
 indicate that the subjects in these studies engaged in peripheral processing 
 during message exposure. Only one study (Chaiken 1980) has demonstrated that 
 source effects under peripheral processing decay relatively rapidly over time, 
 but this study manipulated source attractiveness, not credibility. Furthermore, 
 there is no evidence to support the predictions regarding differential effects 
 due to ad repetition or point of purchase information on credibility -induced 
 attitudes under central versus peripheral processing. 
 
 Also note that virtually all of the evidence on direct effects due to 
 credibility under peripheral processing that we have cited was obtained in lab 
 studies where source credibility was the only peripheral cue available to 
 subjects (see Maddux and Rogers 1980; Pallak, Murroni, and Koch 1984 for 
 exceptions) . In contrast, endorsers used in real advertisements frequently 
 provide multiple peripheral cues to audiences. Peripheral cues are also available 
 from other ad features such as background music or scenery. For instance, a 
 print ad might use a credible celebrity in conjunction with attractive 
 background scenery and a lengthy message. Here, source credibility, source 
 celebrity, source attractiveness, ad attractiveness, and message length could all 
 
potentially influence attitude through peripheral processing. Little is currently 
 understood about the relative strength of credibility effects in such realistic 
 multiple cue environments. 
 
 Second, future research needs to focus more carefully on the proposed 
 mediators of direct effects due to source credibility on attitude. For instance, 
 since uninvolved audiences are expected to rely on the credibility cue to bypass 
 message processing, one would expect such audiences to (a) generate fewer 
 message-related cognitive responses, (b) generate more source -related cognitive 
 responses, and (c) exhibit lower levels of message recall compared to involved 
 audiences. Unfortunately, cognitive response and recall data are rarely obtained 
 in source credibility research. Furthermore, those studies that have measured 
 these variables have produced ambiguous results. For instance, Pallak, Murroni, 
 and Koch (1984) obtained evidence supporting the predictions for message- 
 related, but not source-related cognitive responses, while Moore, Hausknecht, 
 and Thamodaran (1906) obtained the reverse pattern of results. Also, Petty, 
 Cacioppo and Goldman (1981) and Johnson and Scileppi (1969) found no 
 differences in message recall between involved and uninvolved subjects. In sum, 
 extant research provides strong support for direct effects due to source 
 credibility, but not for the mechanisms hypothesized to mediate these effects. 
 
 Third, virtually all of the evidence for source credibility effects has 
 concerned attitudes; effects on individual beliefs about brand attributes have 
 rarely been examined (for an exception, see Mizerski, Hunt, and Patti 1978). 
 However, belief measures can provide useful and relatively sensitive tests for 
 source credibility effects, and particularly so under conditions of central 
 processing when these effects are expected to be weak. More importantly, belief 
 measures may provide indirect tests of the mechanisms mediating these effects. 
 
 8 
 
Since subjects in a peripheral processing mode are not expected to process 
 attribute specific information in the ad, it is conceivable that their brand 
 attitude could produce strong halo effects on subsequent belief measures. If so, 
 source credibility should produce parallel effects on virtually all belief measures 
 as well on attitude. By contrast, subjects in a central processing mode may only 
 form those belief that are directly based on, or can be logically inferred from 
 the brand information provided in the ad. Thus, credibility should only influence 
 a select subset of (apriori identifiable) beliefs, and these effects may not 
 necessarily parallel the effect on attitude. 
 
 Fourth, virtually all of the source credibility research has had a persuasion 
 or yielding focus; effects on earlier information processing stages such as 
 attention have been virtually ignored. However, a credible source may play a 
 crucial role in the ad effects sequence by attracting attention to the ad 
 message. These attentional effects may be particularly crucial for uninvolved 
 audiences who may otherwise choose not to view and process the ad altogether. 
 Research that examines the effects of source credibility in natural viewing 
 environments (rather than in lab settings where attention is forced) and that 
 examines effects for ads with and without a credible source (rather than ads 
 endorsed by credible versus noncredible sources) should prove useful in isolating 
 credibility effects on attention, and also in examining the extent to which 
 attention levels moderate the credibility-attitude relationship. 
 
 Also note that an implicit assumption in virtually all of the research on 
 credibility effects is that audiences are persuaded (i.e., form attitudes towards 
 the advocacy object) during exposure to a communication, and differ only in the 
 persuasion route (central versus peripheral) they follow for judging attitude. 
 This may not be a valid assumption for uninvolved audiences. For instance, 
 
Mitchell (1981, 1906) has argued that uninvolved audiences may completely 
 bypass brand evaluation by engaging in a nonbrand processing strategy, i.e., by 
 focussing attention on brand unrelated cues in the advertisement (also see 
 Gardner, Mitchell, and Russo 1985; Gardner 1985; Krugman 1965). Lichtenstein 
 and Srull (1985, 1987) have shown that subjects who do not form a brand 
 evaluation during ad exposure subsequently rely on recalled information from 
 the ad to construct a brand attitude when required to do so (also see Hastie 
 and Park 1986) . This suggests that the strength of credibility effects for 
 uninvolved audiences will depend on the accessibility of source versus message 
 cues at the time an attitudinal judgment is made (e.g., during purchase). The 
 strength of these effects may also depend on the level of involvement at the 
 time such a judgment is made. For instance, some audiences may be uninvolved 
 with an advertised brand during ad exposure, but they may become much more 
 involved with the brand as they approach purchase. If so, then one would 
 expect them to retrieve brand-related information from memory, and engage in 
 central processing during the purchase episode. In contrast, audiences who are 
 uninvolved with the brand during the ad viewing as well as the purchase 
 episode may be much more likely to minimize cognitive effort by retrieving and 
 relying on a peripheral cue such as source credibility to judge attitude prior to 
 purchase. Future research needs to take a dynamic view of involvement, and 
 consider the effects of change in involvement between the ad viewing and 
 purchase episodes to better understand the mechanisms mediating source 
 credibility effects on attitude that occur after ad exposure and processing. 
 
 Finally, note that much of the evidence for direct effects due to source 
 credibility is based on studies in which involvement with the advocacy issue was 
 manipulated to induce central versus peripheral processing. Unfortunately, 
 
 10 
 
involvement is an individual difference variable that cannot be manipulated by 
 an advertiser to suit his communication objectives. However, recent research 
 suggests that other antecedents of motivation to process information such as 
 message tone (Pallak, Murroni, and Koch 1984), and mood induced by an 
 advertisement (Worth and Mackie 1987) , and ability to process information such 
 as media type (Chaiken and Eagly 1983), and time compression (Moore et al. 
 1986) could also moderate source credibility effects by inducing central versus 
 peripheral processing on the part of respondents. Research which examines the 
 simultaneous effects of source credibility and these types of variables should 
 prove useful because it would give advertisers insights about how they could 
 alter recipients' message processing strategy so as to maximize attitudinal 
 effects due to credibility. 
 
 Indirect Effects 
 
 Source credibility can also affect attitude indirectly by first influencing 
 the way in which people process and evaluate claims made in the persuasive 
 message. We consider two possibilities here, namely that credibility could either 
 influence the magnitude of message processing, or influence the (evaluative) 
 direction of processing. Since these two mediating mechanisms are predicted by 
 different theoretical perspectives, we consider each one separately. 
 Effects on Amount of Processing. The "central versus peripheral processing" 
 
 framework nicely accounts for source credibility effects under extreme levels of 
 motivation and ability to process message arguments. However, most persuasion 
 contexts are probably not characterized by extreme motivation and ability 
 levels. In recent years, Cacioppo and Petty (1984; Petty and Cacioppo 1986a, 
 
 11 
 
1986b) have proposed an extension of the "central/peripheral routes" framework 
 termed the elaboration likelihood model to account for persuasion processes 
 under moderate levels of involvement and ability. The ELM suggests that when 
 motivation to process a communication is at moderate levels, cues such as 
 credibility of the source will act neither as a message argument nor as a 
 peripheral cue. Rather, credibility will influence the amount of message 
 processing that audiences engage in (see Petty and Cacioppo 1986a for details). 
 For example, consider a situation in which a communication is somewhat 
 counterattitudinal, but the recipient is only moderately involved with the 
 advocacy issue and hence unsure about the extent to which (s)he should process 
 the communication. A counterattitudinal message clearly represents a threat to 
 the recipient's current beliefs and attitudes. A highly credible source intensifies 
 this threat, and should thus induce detailed evaluation of the message 
 arguments. In contrast, the threat perceived from a counterattitudinal message 
 should be lowered if the message is attributed to a source of low credibility. 
 This should allow recipients to assess the implications of the communication 
 without a detailed examination of its contents. Note that the effects of high 
 versus low credibility sources on the intensity of message processing will be 
 reversed if the communication is proattitudinal. Recipients will perceive a 
 greater threat if they receive a message they agree with from a source that 
 they do not trust. Consequently, a low credibility source should lead to greater 
 message processing for proattitudinal messages. 
 
 Empirical support for credibility effects on amount of message processing 
 comes from a study by Heesacker, Petty, and Cacioppo (1983) which examined 
 the effects of message quality (strong versus weak) and source credibility (high 
 versus low) on attitude towards a moderately involving and counterattitudinal 
 
 12 
 
topic (i.e., an issue whose consequences for the subjects were uncertain). As 
 expected, the credible source induced subjects to more deeply process message 
 claims, and thus intensified the effects of message quality on attitude. In 
 contrast, message quality had no effect on attitude when the message was 
 attributed to a source of low credibility. Stated differently, the low credibility 
 source diluted the effects of message quality by reducing recipients' motivation 
 to carefully scrutinize the message. 
 
 In sum, the ELM framework suggests that effects due to source credibility 
 under moderate involvement levels are contingent on the quality of the message 
 arguments as well as initial opinions of the audience members. If an 
 advertisement is targeted primarily at an unfavorable audience, then a highly 
 credible source should only be used if the claims made in the advertisement can 
 stand up to close scrutiny. If these claims are vacuous, then a credible source 
 would actually be dysfunctional since it would intensify message processing and 
 hence amplify the negative effects due to uncompelling arguments in the 
 message. It would also not be advisable to use a low credibility source, since 
 that would allow recipients to reject the message without engaging in message 
 processing. Instead, advertisers would do well to rely on other positive 
 peripheral cues (such as an attractive source or pleasant music) to create direct 
 attitudinal effects. 
 
 These recommendations are reversed for audiences who are initially 
 favorable. Specifically, favorable audiences will be more persuaded by a 
 compelling message if it is coupled with a source of questionable credibility, 
 since such a source would lead to more careful message scrutiny. A highly 
 credible source would only be advisable if it is desirable that the audience not 
 engage in detailed message processing. Such would clearly be the case when the 
 
 13 
 
advertised brand has no distinctive advantage over its competitors, and is hence 
 being supported by relatively vacuous claims. 
 
 Effects on Direction of Processing. Sternthal, Dholakia, and Leavitt (1981) 
 
 (also see Sternthal, Phillips, and Dholakia 1978) suggest an alternative way in 
 which source credibility may affect attitude in a situation of moderate 
 involvement. These authors examine credibility effects within a cognitive 
 response model of persuasion. The cognitive response model (Greenwald 1968; 
 Wright 1973, 1980) asserts that the effects of a persuasive message on attitude 
 are mediated by the spontaneous thoughts or cognitive responses generated by 
 recipients during message exposure. If these responses are primarily negative 
 (i.e., counterarguments) then negative attitudes result. On the other hand, 
 predominantly positive responses (i.e., support arguments) lead to a more 
 favorable attitude towards the advocacy object. 
 
 A variable such as source credibility can influence attitudes by first 
 affecting the mix of counter/support arguments generated during the message 
 viewing episode. If the message is counterattitudinal, then recipients are 
 primarily predisposed to counterargue with the message regardless of the 
 credibility of the source. However, it is more difficult to counterargue with 
 statements made by a credible or expert source. Thus, a credible source should 
 inhibit counterargumentation and hence lead to a more favorable attitude. In 
 contrast, a proattitudinal message will primarily predispose recipients to support 
 argue. If the message is attributed to a source lacking in credibility, recipients 
 will believe that the source is not qualified to adequately represent the issue 
 that they support. Consequently, a source of moderate or low credibility will 
 actually bolster the recipient's natural tendency to engage in support 
 argumentation, and hence further polarize the already favorable attitude. In 
 
 14 
 
sum, source credibility is expected to influence attitude by first affecting the 
 direction rather than the intensity of thinking. 
 
 It should be noted that the framework discussed above was originally 
 presented as a general representation of source credibility effects regardless of 
 level of involvement. Subsequently, Sternthal, Phillips, and Dholakia (1978) 
 argued that credibility would likely affect message processing only when 
 involvement was not at extremely high levels. Since the cognitive response 
 model has generally been recognized as not adequately dealing with low 
 involvement message processing situations, it appears that the Sternthal et al. 
 framework is primarily suitable in situations of moderate issue involvement. 
 
 A key prediction of the Sternthal et al. framework is that credible sources 
 will be persuasive if recipients have an initial negative opinion towards the 
 advocacy issue, but will actually operate as a persuasive liability for initially 
 positive recipients . This predicted interaction between source credibility and 
 initial opinion has been supported in a number of studies (Bock and Saine 1975; 
 Harmon and Coney 1982; Sternthal, Dholakia, and Leavitt 1978). Note, however, 
 that the ELM model makes an identical prediction if the quality of persuasive 
 messages is assumed to be high. Since all three of the studies listed above 
 employed reasonably compelling arguments in their experimental communications, 
 these studies do not differentiate the ELM model from the framework proposed 
 by Sternthal, Dholakia, and Leavitt (1978). 
 
 Although the ELM model and the Sternthal et al. framework generate 
 identical predictions when message quality is strong, it is worth emphasizing 
 that the mediating mechanisms for credibility effects postulated by the two 
 approaches are conceptually distinct. The ELM suggests that source credibility 
 influences the extent to which a message is processed. Consequently, credibility 
 
 15 
 
simply serves to amplify or weaken the effects of other variables (such as 
 message quality) on attitude. In contrast the Sternthal et al. framework 
 proposes that credibility modifies the way in which message arguments are 
 interpreted independent of the quality of these arguments. Consequently, a 
 compelling test of the two frameworks requires an examination of credibility 
 effects for strong as well as weak messages. Sternthal, Dholakia, and Leavitt 
 (1978) predict no difference in the source credibility by initial opinion 
 interaction as a function of message quality, while the ELM framework predicts 
 a three-way interaction, i.e., a source of high (low) credibility is expected to 
 polarize effects due to variations in message quality when initial opinion is 
 negative (positive). Evidence from Heesacker, Petty, and Cacioppo (1983) thus 
 appears to support the ELM framework. However, more research is clearly 
 needed before the precise mechanism for source effects under moderate 
 involvement levels is clearly understood. 
 
 Finally, both frameworks make similar predictions regarding the durability 
 of credibility effects, and the appropriate timing for source introduction in the 
 message. Source credibility effects on attitude under moderate involvement are 
 expected to be based on the amount and direction of message processing. 
 Consequently, these effects should be durable and resistant to counterattack, 
 and should be observed only when the identity of the source is revealed early 
 in the communication. 
 
 Needed Research. Since the frameworks proposed by Petty and Cacioppo 
 (1986a) and Sternthal, Dholakia, and Leavitt (1978) propose different mediating 
 mechanisms for source credibility effects, an important priority for future 
 research should be to develop critical tests which pit these frameworks against 
 each other. As noted earlier, an experiment which manipulates initial opinion, ad 
 
 16 
 
message quality, and source credibility can provide such a test. Additional tests 
 can be generated by examining cognitive response data. Since ELM predicts that 
 source credibility effects are mediated by the amount of message processing 
 that audiences engage in, credibility should influence the total number of 
 message-related cognitive responses, but should not influence the mix of these 
 CRs (i.e., counter versus support arguments). By contrast, credibility effects on 
 the direction of processing (as predicted by Sternthal et al) should be reflected 
 in the mix of message-related CRs. 
 
 Even stronger tests for the two competing frameworks can be generated 
 by examining the moderating influence of variables which are known to 
 influence either the amount of ad message processing, or the direction of ad 
 message processing (but not both). One such variable may be ability to process. 
 Since ability to process (as determined by intrinsic factors such as prior 
 knowledge, or extrinsic factors such as distraction) is likely to influence the 
 magnitude, but not the direction of message processing, ability should interact 
 with source credibility under ELM, but not under the Sternthal et al framework. 
 For instance, Figure 1 presents the results we might expect under the ELM 
 framework (panels a, b, and c) and under the Sternthal, Dholakia, and Leavitt 
 (1978) framework (panels d, e, and f) from a hypothetical experiment in which 
 moderately involved subjects are exposed to a counterattitudinal message in a 2 
 (high versus low source credibility) by 2 (strong versus weak message quality) 
 by 3 (high, moderate, and low ability to process) factorial design. Under ELM, 
 source credibility should strongly polarize the effects of message quality on 
 attitude when ability is high. As ability drops to moderate levels, so should the 
 strength of polarizing effects due to source credibility. Finally, at low levels of 
 ability, subjects should switch into a peripheral processing mode, and thus 
 
 17 
 
manifest a simple credibility main effect. In contrast, the Sternthal et al. 
 framework predicts identical credibility effects regardless of level of the ability 
 factor. 
 
 Analogously, forewarning subjects about the persuasive intent of a 
 counterattitudinal message is likely to put them in an anticipatory 
 counterargumentative mode (Petty and Cacioppo 1977), i.e., influence the 
 direction of message processing rather than the magnitude of processing. 
 Consequently, the Sternthal et al framework predicts an interactive effect due 
 to forewarning and source credibility on attitude -- credibility effects should be 
 stronger for audiences who are forewarned. In contrast, The ELM predicts no 
 differences in credibility effects as a function of forewarning. In sum, future 
 experimental research of the type discussed here should provide the basis for 
 "strong inference" in the study of source credibility effects by supporting one 
 proposed explanation for these effects while simultaneously rejecting the other. 
 
 Conclusion 
 
 Our review of the literature suggests that source credibility can operate in 
 a persuasion environment in several distinct capacities -- as a persuasive 
 argument, as a peripheral cue, or as a variable that influences the intensity 
 and/or direction of active message processing. Figure 2 displays each of these 
 mediating mechanisms, while Table 1 summarizes the conditions under which 
 they are likely to operate, prior empirical support, and implications for the use 
 of credible sources in advertising. 
 
 The ELM framework, and the cognitive response framework proposed by 
 Sternthal, Dholakia, and Leavitt (1978) have emerged in only the last decade or 
 
 18 
 
so, and yet they provide a rich conceptual network of propositions and ideas 
 that can be used to develop extremely subtle and complex predictions 
 concerning credibility effects in persuasion contexts. Unfortunately, theoretical 
 development has far outpaced empirical research in this area --in many 
 instances, direct empirical support for well developed hypotheses is simply 
 nonexistent. 
 
 Perhaps the area in most pressing need of future research concerns 
 credibility effects under moderate involvement situations. It seems likely that 
 recipients of advertising messages are frequently uncertain about the 
 consequences of the advertised brand to their personal lives. Extant literature 
 suggests that credibility effects under such conditions may occasionally be 
 negative, and could be accounted for by more than one mediating mechanism. 
 Moreover, it is possible that the two proposed mechanisms may operate in 
 parallel, or one may dominate the other as a function of other (as yet 
 unspecified) variables in the ad reception environment. These important issues 
 have received virtually no attention in the literature. Indeed, we found very few 
 studies that were explicitly designed to examine the hows and whys of 
 credibility effects under moderate involvement. The study by Heesacker, Petty, 
 and Cacioppo (1983) is an exception in that it provides compelling support for 
 predictions derived from the ELM in a counterattitudinal situation. However, 
 predictions of this framework for credibility effects on an initially favorable 
 audience have never been tested. Furthermore, neither the ELM nor the 
 Sternthal et al. framework makes any predictions concerning credibility effects 
 for audiences that are neutral towards, or have no initial opinion about the 
 communication object. This is clearly a research area with tremendous 
 theoretical and practical consequences. 
 
 19 
 
Finally, it should be noted that much of our discussion in this paper is 
 based on persuasion studies reported in the psychology literature. There is 
 clearly no guarantee that effects and mediators uncovered in lab studies 
 involving simple, verbal messages will generalize to more complex ad reception 
 environments. There is a need for constructive replication designed to "fix" 
 these effects in environments relevant to advertising and marketing 
 practitioners. 
 
 20 
 
References 
 
 Andreoli, V. and S. Worchel (1978), "Effects of Media, Communicator, and 
 Message Position on Attitude Change," Public Opinion Quarterly , 42, 
 59-70. 
 Andrews, J. Craig (1988), "Motivation, Ability, and Opportunity to 
 
 Process Information: Conceptual and Experimental Manipulation Issues," 
 in Advances in Consumer Research , M. J. Houston, ed., 15, Ann Arbor, 
 MI: Associations For Consumer Research, 219-215. 
 Batra, R. and M. L. Ray (1986), "Situational Effects of Advertising 
 Repetition: The Moderating Influence of Motivation, Ability, and 
 Opportunity to Respond," Journal of Consumer Research , 12 (March), 
 432-445. 
 Bock, D. and T. Saine (1975), "The Impact of Source Credibility, Attitude 
 Valence, and Task Sensitization on Trait Errors in Speech Evaluations," 
 Speech Monograph , 37, 342-358. 
 Cacioppo, J. T. and R. E. Petty (1984), "The Elaboration Likelihood Model 
 of Persuasion," in Advances in Consumer Research , T. Kinnear, ed., 11, 
 Ann Arbor: Associations for Consumer Research, 673-675. 
 Chaiken, S. (1980), "Heuristic versus Systematic Information Processing 
 and the Use of Source versus Message Cues in Persuasion," Journal of 
 Personality and Social Psychology , 39, 752-766. 
 Dean, R. B., J. A. Austin, and W. A. Watts (1971), "Forewarning Effects 
 in Persuasion: Field and Classroom Experiments, " Journal of 
 Personality and Social Psychology , 18, 210-221. 
 Dholakia, R. and B. Sternthal (1977), "Highly Credible Sources: Persuasive 
 
 21 
 
Facilitators or Persuasive Liabilities, " Journal of Consumer Research , 
 
 3, 223-232. 
 Frieden, J. B. (1984), "Advertising Spokesperson Effects: An Evaluation of 
 
 Endorser Type and Gender on Two Audiences," Journal of Advertising 
 
 research , 24 (5), 33-41. 
 Friedman, H. H. and L. Friedman (1979), "Endorser Effectiveness by Product 
 
 Type," Journal of Advertising Research , 19 (5), 63-71. 
 Gardner, M. P. (1980), "Does Attitude Toward Ad Affect Brand Attitude under 
 
 a Brand Evaluation Set?" Journal of Marketing Research , 22, 192-198. 
 , A. A. Mitchell, and J. E. Russo (1985), "Low Involvement Strategies 
 
 for Processing Advertisements," Journal of Advertising , 44-56. 
 Gillig, Paulette M. and Anthony G. Greenwald (1974), "Is It Time to Lay the 
 
 Sleeper Effect to Rest?" Journal of Social Psychology , 29 (1), 132-139. 
 Greenwald, A. G. (1968), "Cognitive Learning, Cognitive Responses to 
 
 Attitude Change," in Psychological Foundations of Attitudes , A.G. 
 
 Greenwald, T.C. Brock, and T.M. Ostrom, eds., NY: Academic Press, 
 
 147-170. 
 Harmon, R. R. and K. A. Coney (1982), "The Persuasive Effects of Source 
 
 Credibility in Buy and Lease Situation," Journal of Marketing Research , 
 
 19 (May), 255-260. 
 Hastie, Reid and Bernadette Park (1986), "The Relationship Between Memory 
 
 and Judgment Depends On Whether the Judgment Task is Memory- Based or 
 
 On-Line," Psychological Review , 93 (3), 258-268. 
 Hass, R. G. (1981), "Effects of Source Characteristics on Cognitive 
 
 Responses and Persuasion," in Cognitive Responses in Persuasion , R. E. 
 
 Petty, T. M. Ostrom, and T. C. Brock Hillsdale, eds., NJ: Earlbaum, 
 
 22 
 
141-172. 
 Heesacker, M. , R. Petty, and J. Cacioppo (1983), "Field Dependence and 
 
 Attitude Change: Source Credibility Can Alter Persuasion by Affecting 
 
 Message-Relevant Thinking," Journal of Personality , 51, 653-666. 
 Hovland, C. and W. Weiss (1951), "The Influence of Source Credibility on 
 
 Communication Effectiveness," Public Opinions Quarterly , 15, 635-650. 
 Johnson, H. H., and J. A. Scileppi (1969), "Effects of Ego- Involvement 
 
 Conditions on Attitude Change to High and Low Credibility 
 
 Communicators," Journal of Personality and Social Psychology , 13, 
 
 31-36. 
 Kelman, H. C. and C. I. Hovland (1953), "'Reinstatement* of the 
 
 Communicator in Delayed Measurement of Opinion Change," Journal of 
 
 Abnormal and Social Psychology , 48, 327-335. 
 Kiesler, S. A. and R. Mathog (1968), "The Distraction Hypothesis and 
 
 Attitude Change: The Effects of Inference and Credibility," 
 
 Psychological Reports , 23, 1123-1133. 
 Krugman, H. (1965), "The Impact of Television Advertising: Learning 
 
 Without Involvement," Public Opinion Quarterly , 29, 349-356. 
 Lichtenstein, Meryl and Thomas K. Srull (1985), "Conceptual and 
 
 Methodological Issues in Examining the Relationship Between Consumer 
 
 Memory and Judgment," in Psychological Processes and Advertising 
 
 Effects: Theory, Research and Applications , L. Alwitt and A. A. 
 
 Mitchell, eds., Hillsdale, NJ: Earlbaum, 113-128. 
 and (1987), "Processing Objectives as a Determinant of the 
 
 Relationship Between Recall and Judgment," Journal of Experimental 
 
 Social Psychology , 23, 93-118. 
 
 ■ 
 
 23 
 
Maddux, J. E. and R. W. Rogers (1980), "Effects of Source Expertness, 
 Physical Attractiveness, and Supporting Arguments on Persuasion: A 
 Case of Brains over Beauty," Journal of Personality and Social 
 Psychology , 39 (2), 235-244. 
 
 McGinnies, E. (1973), "Initial Attitude, Source Credibility, and 
 
 Involvement as Factors in Persuasion," Journal of Experimental Social 
 Psychology . 9, 285-296. 
 
 Mills, J. and J. Harvey (1972), "Opinion Change as a Function of When 
 Information about the Communicator Is Received and Whether He Is 
 Attractive or Expert," Journal of Personality and Social Psycholog y, 21 
 (1), 52-55. 
 
 Mitchell, Andrew A. (1981), "The dimensions of Advertising Involvement," 
 
 in Advances in Consumer Research , K. B. Monroe, ed. , 8, Ann Arbor, MI: 
 Association for Consumer Research, 25-30. 
 
 (1986), "Theoretical and Methodological Issues in Developing an 
 
 Individual-Level Model of Advertising Effects," in Advertising and 
 Consumer Psychology , J.C. Olson and K. Sentis, eds., 3, 172-196. 
 
 Mizerski, R. W. , J. M. Hunt, and C. H. Patti (1978), "The Effects of 
 
 Advertising Credibility on Consumer Reactions to an Advertisement," in 
 Research Frontiers in Marketing: Dialogues and Directions , S. C. Jain, 
 ed., Series #43, Chicago, IL: American Marketing Association, 174-168. 
 
 Pallak, S. R., E. Murroni, and J. Koch (1984), "Communicator Attractiveness 
 and Expertise, Emotional versus Rational Appeals, and Persuasion: A 
 Heuristic versus Systematic Processing Interpretation," Social cognition , 
 2, 122-141. 
 
 Petty, Richard E. and John T. Cacioppo (1977), "Forewarning, Cognitive 
 
 24 
 
Responding, and Resistance to Persuasion," Journal of Personality and 
 
 Soical Psychology , 35 (9), 645-655. 
 and (1984), "Source Factors and the Elaboration Likelihood 
 
 Model of Persuasion," in Advances in Consumer Research , T. Kinnear, 
 
 ed., 11, Ann Arbor: Associations for Consumer Research, 668-672. 
 and (1986a), "The Elaboration likelihood Model of Persuasion," 
 
 in Advances in Experimental Social Psychology , L. Berkowitz, ed., 19, 
 
 Orlando, FL: Academic Press, 123-205. 
 and (1986b), Communication and Persuasion: Central and 
 
 Peripheral Routes to Attitude Change , New York: Springer- Verlag. 
 , , and R. Goldman (1981), "Personal Involvement as a Determinant 
 
 of Argument-Based Persuasion," Journal of Personality and Social 
 
 Psychology , 41, 847-855. 
 Rhine, R. J. and L. J. Severance (1970), "Ego-Involvement, Discrepancy, 
 
 Source Credibility, and Attitude Change," Journal of Personality and 
 
 Social Psychology , 16 (2), 175-190. 
 Shulman, G. I. and C. Worrall (1970), "Salience Patterns, Source Credibility, 
 
 and the Sleeper Effect," Public Opinion Qualterly . 34, 371-382. 
 Sternthal, B., R. Dholakia, and C. Leavitt (1978), " The Persuasive Effect 
 
 of Source Credibility: Tests of Cognitive Response," Journal of Consumer 
 
 Research , 4, 252-260. 
 , L. Phyllips, and R. Dholakia (1978), "The Persuasive Effect of Source 
 
 Credibility: A Situational Analysis," Public Opinions Quarterly , 42, 
 
 285-314. 
 Watts, W. A. and W. J. McGuire (1964), "Persistency of Induced Opinion 
 
 Change and Retention of the Inducing Message Contents," Journal of 
 
 25 
 
Abnormal Psychology , 68, 233-241. 
 Worth, L. and D. Mackie (1987), "Cognitive Mediation of Positive Affect in 
 
 Persuasion," Social Cognition , 5 (1), 76-94. 
 Wright, P.L. (1973), "The Cognitive Processes Mediating Acceptance of 
 
 Advertising," Journal of Marketing Research , 10, 53-62. 
 (1980), "Message-Evoked Thoughts: Persuasion Research Using Thought 
 
 Verbalizations," Journal of Consumer Research , 7 (2), 151-175. 
 
 26 
 
TABLE 1 
 Summary of Mechanisms Mediating Source Credibility Effects 
 
 Determining condition Empirical support 
 
 Implications 
 
 Direct Effects: Credibility as a persuasive argument 
 
 High motivation/ 
 High ability 
 
 Dean, Austin, and 
 Watts (1975) 
 
 1. 
 
 2. 
 
 3. 
 
 4. 
 
 Effects are always positive. 
 Effects are relatively 
 enduring. 
 Timing of source 
 identification will not 
 influence these effects. 
 Magnitude of effects 
 influenced by other central 
 cues (e.g., ad message 
 quality) . 
 
 Direct Effects: Credibility as a peripheral cue 
 
 Low motivation/ 
 Low ability 
 
 Johnson and Scileppi 
 (1969) 
 
 Rhine and Severance 
 (1970) 
 
 McGinnies 
 (1973) 
 
 Andreoli and Worchel 
 (1978) 
 
 Mizerski, Hunt, 
 and Patti (1978) 
 
 Petty, Cacioppo, 
 and Goldman (1981) 
 
 Pallak, Murroni, and 
 Koch (1984) 
 
 Worth and Mackie 
 (1987) 
 
 1. 
 
 2. 
 
 3. 
 
 4. 
 
 Effects are always positive 
 Effects are transient and 
 likely to decay rapidly. 
 Timing of source identific- 
 ation will not influence 
 these effects. 
 Magnitude of effects is 
 influenced by other 
 peripheral cues (e.g. 
 source attractiveness. 
 
TABLE 1 (continued) 
 
 Determining condition Empirical support Implications 
 
 Indirect Effects: Effects on the amount of processing 
 
 Moderate Heesacker, Petty, 1. Nature of effects (positive 
 
 motivation and Cacioppo or negative) is contingent 
 
 (1983) on (a) recipient's prior 
 
 opinion and (b) ad message 
 
 quality 
 
 2. Effects are relatively 
 enduring. 
 
 3. These effects will be 
 observed only when source 
 is identified before 
 message. 
 
 4. Magnitude of effects is 
 moderated by other 
 variables (e.g., ability 
 to process message) that 
 influence amount of ad 
 message processing. 
 
 Indirect Effects: Effects on the direction of processing 
 
 Moderate Bock and Saine 1. Nature of effects (positive 
 
 motivation (1975) or negative) is contingent 
 
 on recipient's prior 
 Sternthal, Dholakia, opinion 
 
 and Leavitt 2. Effects are relatively 
 
 (1978) enduring. 
 
 3. These effects will be 
 Harmon and Coney observed only when source 
 (1982) is identified before 
 
 message. 
 
 4. Magnitude of effects is 
 moderated by other 
 variables (e.g., 
 forewarning) that 
 influence direction of ad 
 message processing. 
 
FIGURE 1 
 
 Source Credibility Effects as a Function of Ability to Process 
 the Ad Message: Differential Predictions as Derived from the ELM 
 and Sternthal, Dholakia, and Leavitt (1978) Framework 1 
 
 ELM Predictions 
 
 Attitude 
 
 Attitude 
 
 SMQ 
 
 WMQ 
 I SC 
 
 Low High 
 (a) High ability 
 
 Attitude 
 
 SMQ 
 
 WMQ 
 
 SC 
 
 Low High 
 (b) Moderate ability 
 
 SMQ 
 * WMQ 
 
 SC 
 
 Low High 
 (c) Low ability 
 
 Sternthal et al. Predictions 
 
 Attitude 
 
 Attitude 
 
 Attitude 
 
 SMQ 
 
 WMQ 
 
 SC 
 
 Low High 
 (d) High ability 
 
 ■I — SC 
 
 Low High 
 (e) Moderate ability 
 
 ■I — SC 
 
 Low High 
 (f) Low ability 
 
 : Counterattitudinal message and moderate involvement situation 
 
 ote : 
 SC; Source Credibility 
 SMQ; Strong message quality 
 WMQ; Weak message quality. 
 
FIGURE 2 
 Alternative Mediation routes for Credibility Effects on Attitude 
 
 
 
 IV 
 
 
 Ad Message Processing 
 
 a c n 
 
 
 
 
 
 
 Amount 
 
 Direction 
 
 i\ , v_ , u 
 
 i 
 
 
 
 
 
 
 i 
 
 i i 
 
 : i 
 
 i 
 
 
 [_ 
 
 Ad message 
 and Source 
 
 
 ( 
 
 ) 
 
 Brand 
 Attitude 
 
 
 
 
 
 
 3 
 
 l 
 
 
 o 
 
 
 i 
 
 
 
 
 
 
 
 
 
 
 
 
 
 A 
 B 
 C 
 
 D 
 
 Direct effects (Credibility serves as a persuasive argument). 
 Direct effects (Credibility serves as a peripheral cue). 
 Indirect effects (Credibility influences the amount of ad 
 message processing) . 
 
 Indirect effects (Credibility biases the direction of ad 
 message processing) . 
 
HECKMAN Ul 
 BINDERY INC. |s| 
 
 JUN95 
 
 
 ft™J To.PI«»? N MANCHESTER. 
 Bom,I ° fk " INDIANA 46962 
 ^ *