Effects of Source Likability on Attitude Change Through Message Repetition

Jean-Charles Chebat, University of Quebec at Montreal
Michel Laroche, Concordia University
Daisy Baddoura, University of Quebec at Montreal
Pierre Filiatrault, University of Quebec at Montreal
ABSTRACT - This paper reports on an empirical study of the effects of source likability, message repetition and the memorization of arguments on attitude change. The results indicate that likable sources enhance attitude change only through the mediation of message memorization and repetition. It is suggested that likable sources cannot compensate for inadequate argumentation.
[ to cite ]:
Jean-Charles Chebat, Michel Laroche, Daisy Baddoura, and Pierre Filiatrault (1992) ,"Effects of Source Likability on Attitude Change Through Message Repetition", in NA - Advances in Consumer Research Volume 19, eds. John F. Sherry, Jr. and Brian Sternthal, Provo, UT : Association for Consumer Research, Pages: 353-358.

Advances in Consumer Research Volume 19, 1992      Pages 353-358


Jean-Charles Chebat, University of Quebec at Montreal

Michel Laroche, Concordia University

Daisy Baddoura, University of Quebec at Montreal

Pierre Filiatrault, University of Quebec at Montreal

[The authors very gratefully acknowledge the financial support of the Social Sciences and Humanities Research Council of Canada.]


This paper reports on an empirical study of the effects of source likability, message repetition and the memorization of arguments on attitude change. The results indicate that likable sources enhance attitude change only through the mediation of message memorization and repetition. It is suggested that likable sources cannot compensate for inadequate argumentation.


Advertisers pay much attention to the choice of an adequate character who will introduce their product or service in the advertisement. One of the most desired characteristics of the personality is likability. Some studies have shown that likability has positive effect on attitudes toward the message, but these results are somewhat controversial; moreover, few studies have investigated the combined effect of attractiveness and repetition which is related to the following question: Do the positive effects of attractive sources wear out through repetition? If yes, advertising investments in personalities such as Madonna or Michael Jackson should be reconsidered.

This paper first reviews the main effects of likability on message acceptance and the effects of repetition on attitude change. Hypotheses will be presented on the basis of experimental results on the effects of repetition. Results from the empirical study are discussed and some managerial implications are proposed.

Effects of Likable Sources on Attitude Change

In the marketing literature, likability has often been confused with physical attractiveness. However, Kahle and Homer (1985) have suggested that there might not be any correlation between the two: for example, John McEnroe could be both physically attractive and unlikable due to his behavior.

A likable person may be defined as one who says pleasant things (Eagly and Chaiken, 1975). Based on this specific definition, Himmelfarb and Arazi (1974) found that, while an attractive communicator may be more persuasive than an unattractive one when subjects have no choice but to listen to the message, the unattractive communicator may be more effective when subjects have that choice. In a counter attitudinal behavior situation, Eagly and Chaiken (1975) found that, even if attractive communicators were more persuasive than unattractive ones when given undesirable positions, both were equally persuasive when given desirable positions.

Effects of Repetition on Attitude Change

Repeated exposure to a stimulus positively enhances its assessment (Zajonc, 1968). As shown by Moreland and Zajonc (1979), repeated exposure to a new stimulus brings about a loop reaction between two subjective elements: a feeling of familiarity and a predisposition toward the object of a message. Repetition of a given message makes the memorized elements more accessible to evaluation. Cacioppo and Petty (1979) have suggested that repeated exposure to a message triggers the cognitive elaboration of arguments: the number of counter-arguments declined then increased as the number of exposures increased whereas the number of support arguments followed the opposite pattern.

Krugman (1986) also showed that repetition can have negative effects. A message can be too familiar: when consumers are exposed to repeated messages that they have already memorized, even if they are positively assessed, these messages do not bring anything new and do not deserve their attention. This idea has been supported by a number of researchers. In particular, Berlyne's two-factor theory (1963) shows the existence of an inverted U-curve: increasing the number of exposures to a message from a low to a moderate level is expected to enhance its impact, whereas further exposures are expected to cause its effectiveness to decline. At lower levels of repetition, reduction of uncertainty brings about a rather positive assessment of the message; at higher levels, repetition is associated with boredom and leads to a negative assessment of the message. However there is also a substantial number of studies that show no relation between these two variables (e.g., Belch, 1982; Mitchell and Olson, 1981).

Two opposite effects are at play here: on the one hand, a decreased interest for the message when repeated beyond a given threshold and on the other, a positive relationship between the memorized content and the emotion generated by a likable source. How do these two factors interact? It is suggested that a likable source may soften the negative effects of repetition: when the repeated message is associated with a likable source, the rate of advertising wearout, i.e., the decreasing interest, is lower. Consequently, when the message is associated with a likable source, its repetition may have significantly more effect on attitude change than in the case of a neutral source. Memorization of the arguments in the message can be attributed to the fact that the message has triggered the receiver's attention. What remains to be answered are the following two questions:

(1) Does memorization enhance the production of counterarguments and create a negative attitude when the message is frequently repeated?

(2) Does memorization lead to a positive attitude when the message is associated with a likable source that creates a positive emotional bond with the audience?


a) Main Effects

H1 A likable source generates more attitude change than a neutral source.

H2 Repetition generates attitude change in a curvilinear way (inverted U shaped curve).

H3 Attitude change is higher for respondents who have memorized at least one argument than for those who have not.

b) Two-way Interaction Effects

H4 The effects of repetition on attitude change are stronger in the case of respondents exposed to a likable source than in the case of respondents exposed to a neutral source.

H5 The effects of memorized arguments on attitude change are significantly stronger when associated with a likable source than when associated with a neutral source.

H6 The effects of repeated arguments on attitude change are significantly stronger when memorized by respondents than when not memorized.

No three-way effect hypothesis is proposed due to the fact that the literature is silent on this relationship.


The hypotheses were tested on the basis of a factorial design experiment. Each dimension of the factorial design was conceived as properly representing a single concept. The essential purpose of the design was to understand the interactive effects of affective variables. The factorial design coupled with the analysis of variance were chosen to combine best the experimental technology and the data analysis.

Dimensions of the Factorial Design

For the purpose of this research, the factorial design was constructed around two major dimensions:

- The level of message repetition

- The level of likability (or attractiveness) of the source

- The complete design used six experimental conditions:

- Three levels of repetition: 1, 3 and 5

- Two levels of likability: high and neutral

Each of these dimensions is further explained in the following sections.


The rationale for varying the level of repetition from 1 to 3 to 5 for the same advertising message was to design an experiment which would as much as possible simulate real-life situations. In a normal television advertising sequence, a given message never appears more that five times; thus, 1, 3 and 5 seemed the appropriate frequencies to use, 1 being the minimum, 5 the maximum, and 3 the intermediate frequency. These are also the levels of frequency used in previous studies on repetition; for instance, Belch (1982) tested the effects of 1-3-5 repetitions on message acceptance. The commonly accepted rationale for three frequencies is that most studies on the effects of repetition show a curvilinear relationship, generally known as "wear-out effects" (Chestnut 1980).


This construct was operationalized in terms of a dichotomy. The same professional actor was instructed to play two different roles, i.e., a likable individual and a neutral one. The likable source used a soft and pleasant voice and showed a great deal of consideration for the audience, while the neutral source spoke in a natural fashion, with a rhythm in his speech intended to convey neither pleasantness nor aggressiveness. This is in accordance with the definition given by Eagly and Chaiken (1975). The actor was directed by a professional film director.

Experimental Stimuli

All messages were created specifically for the purpose of the research and used a team of professionals (e.g., an actor, a director, cameramen, editors) to produce very realistic television commercials, all being 30 seconds long.

Verbal Messages

The scenario for all the commercials was a financial executive sitting at his desk in a nicely decorated office. He introduced himself as an executive in a given financial institution. Then he gave a message on credit cards which contained five central arguments and he closed with an appeal to apply for the credit card offered by his financial institution.


These commercials were inserted into a real television program dealing with economics and finance. This program lasted 30 minutes and it contained one 150-second commercial pod. Subjects in a given experimental condition were exposed to only one type of commercial (one of the two versions). The program contained one, three or five repetitions of the same commercial inserted into the existing commercial pod at the same position. For all the experimental conditions, the rest of the program was identical, including the other (non-experimental) commercials.


Each of the six experimental conditions provided an average of 150 respondents (25 for each level of repetition) for a total of 918 subjects. All were college graduates, 48 percent were male, between the ages of 21 and 42. All were university students attending the same mandatory management course at the same level in their curriculum.


The subjects were administered a questionnaire whose purpose was to measure attitude change toward credit cards as a result of the experimental stimuli. Before viewing the videotape, subjects were asked a series of questions to measure a number of variables including their attitude toward credit cards (the `before' measure). The same questions were also asked after viewing the videotape but they were organized in a different order which was randomly determined. After being exposed to the videotape, the subjects were requested to write down all the elements of the message they could recall spontaneously. These elements could be formal (such as the speaker's tie or physical appearance), or substantive, i.e., related to the arguments in the message. In this study, only the number of arguments recalled was used: 19% of the subjects could not recall any argument, and 81% could recall at least one of the five arguments contained in the message.

The five following statements related to the attitudes toward credit cards were provided to the S's before and after they were exposed to the commercials:

- In general I like credit cards;

- I prefer to use credit cards instead of cash;

- To use credit cards is pleasant;

- I like the idea of owning a credit card;

- I prefer using a credit card to writing a check.

Respondents were asked to indicate their degree of agreement with each statement on a 6-point Likert scale. The Cronbach's alpha for these five items is 0.83, which is very good.

The average score of these five items was computed as a `before' and `after' attitude measure, and the difference between these two attitude scores was computed for the attitude change measure.


Manipulation Checks

First, a manipulation check was performed on the experimental conditions.

Source Likability

The level of likability of the source was manipulated so that respondents were exposed to only one of two levels of source likability: high and neutral (which serves as a control condition). Respondents were asked to indicate to which degree they agreed with the following statement: "The character (named) in the commercial could become a friend of mine." On a 6-point scale the neutral likability group scored 2.82 and the high likability scored 4.52. This difference is highly significant (t = 10.2; p = 0.00).

Tests of Hypotheses

An analysis of variance (ANOVA) was performed with attitude change toward credit cards as the dependent variable and with repetition, likability and memorization as the three independent variables.


Table 1 shows the results of the analysis of variance (ANOVA), and the results are summarized next.

Main Effects

H1 is rejected: the effects of likability are not significant (F = 1.03; p = 0.31).

H2 is rejected: the effects of repetition are not significant (F = 1.65; p = 0.19).

H3 is supported: the respondents having memorized at least one argument significantly changed their attitude more toward credit cards than those not having memorized any argument (F = 4.95; p = 0.03).

Two-Way Interaction Effects

H4 is supported: the individuals exposed to a likable source more than once showed more positive attitude change than those exposed to a neutral source (F = 3.76; p = 0.02). Moreover, as shown in Figure 1, for those exposed to the ad three times, the likable source induced significantly more attitude change toward the financial service than the neutral source (F = 5.59; p = 0.02). For both those exposed to the ad five times (F = 0.71; p = 0.40) and those exposed only once (F = 1.31; p = 0.25), no significant difference is shown between the likable and the neutral sources. The relationship between repetition and attitude change is quadratic in the case of the likable source (F = 4.48; p = 0.03 for the weighted quadratic term), and linear for the neutral source (F = 5.68; p = 0.02 for the weighted linear term).

H5 is supported. While memorization has no significant effect on respondents exposed to the neutral source, it has a significant effect on those exposed to the likable source (F = 3.65; p = 0.06). More specifically, as shown in Figure 2, in the case of consumers having memorized no argument the effects of likability are remarkably insignificant (F = 0.00; p = 0.99): the average scores for the two groups are identical (X = 3.45). However, for those who have memorized at least one argument, likability has a significant effect on attitude change (F = 3.45; p = 0.07).

H6 is rejected (F = 0.27; p = 0.76).








Contrary to previous findings (e.g., Himmelfarb and Arazi, 1974), likable sources do not score higher on attitude change than neutral sources; however, our results are in line with what Eagly and Chaiken (1975) found, i.e., if the topic is not controversial, both sources are equally persuasive.

Repetition has no direct significant effect, contrary to previous findings.

Memorization of at least one of the message arguments significantly enhances attitude change. Memorization enhances attitude change in a way which is congruent with the arguments.

Repetition and likability strongly interact: the likable source brings about more attitude change than the neutral source when the message is repeated three times. At both one and five exposures, likability does not bring about significant attitude change. What is most remarkable is that the Berlyne's curvilinear effects (inverted U shape) are found only for the likable source and not for the neutral source. In the latter case, the relation is linear: likability counterbalances the negative effects of repetition. In other words, the tedium effects generated by repetition show up earlier when the source is neutral, as predicted. The results can be interpreted as follows: from 1 to 3 exposures, repetition increases the feeling of familiarity with the message and reinforces the link between the positive affect toward the source and acceptance of the arguments (Moreland and Zajonc, 1979); from 3 to 5 exposures, familiarity turns into boredom: all the formal cues (such as the speaker's physical appearance) serve to confirm that this is the same ad once again. If, as suggested by Hirschman and Holbrook (1982), a commercial is a "psychological reward," its utility decreases after a certain threshold of repetition. Thus, excessive repetition leads to a reduction in the positive effects of likability but not in the memorization of the arguments, which explains why the 3-way interaction is not significant (F = 1.09; p = 0.33).

Likability and memorization significantly interact. Memorization enhances attitude change in the case of consumers exposed to a likable source but not in the case of those exposed to a neutral source. One interpretation is that the likable source acts as a node, in the sense used by Bower (1981): pleasant feelings toward the source are associated with the arguments contained in the message. The association, when it exists (i.e., the arguments are memorized and are provided by a likable source), makes the arguments more persuasive since their effect is positive. The relationship gives the arguments the same emotional tone as the source itself. In the case of the attractive source, it could be assumed that the receivers' moods were enhanced by their positive identification with the source which then could have focused their attention on the positive elements of the message. Thus, as suggested by Forgas and Bower (1988), better integration and recall of the arguments would result; the positive emotion may bias the recalled arguments to produce a positive attitude. Memorization alone is not a panacea: it enhances positive attitude only when the message arguments are associated with a likable source.

Memorization and repetition do not interact significantly because repetition enhances memorization (c2 = 17.82; p = 0.00). Their effects are additive. However, for the groups exposed five times to the same message, the non-memorization group scored significantly higher than the memorization group (F = 3.95; p = 0.05). It can be surmised that, as suggested by Cacioppo and Petty (1979), repeated arguments which are also memorized are mentally reviewed and criticized: counter-arguments are generated which reduce the level of attitude change. Repetition makes the counter-arguments more available which were left latent at lower levels of repetition.


A likable source enhances attitude change only through the mediation of memorized arguments or repetition of the arguments. It has no direct effects. Likable sources are no substitute for adequate argumentation: advertisers cannot simply invest in likable characters and presume their persuasive strategy to be a probable success. Likable sources can counterbalance the negative effects of repetition but only to a certain extent: beyond three repetitions, the effects of likability appear to be cancelled.

Memorization of the arguments in the message is not a panacea either: repetition may be more likely to generate counterarguments, if some arguments are memorized and more easily available for a critical assessment. When associated with a likable source, memorization appears to lead to a more positive attitude toward the message.


Belch, G.E. (1982), "The Effects of Television Commercial Repetition on Cognitive Response and Message Acceptance," Journal of Consumer Research, 9 (June), 56-65.

Berlyne, D.E. (1963), "Motivational Problems Raised by Exploratory and Epistemic Behavior," in Psychology: A Study of a Science, Vol. 5, ed. S. Koch, New York: McGraw-Hill, 284-364.

Bower, G.H. (1981), "Mood and Memory," American Psychologist, 36 (February), 129-148.

Cacioppo, J.T. and R.E. Petty (1979), "Effects of Message: Repetition and Position on Cognitive Response, Recall, and Persuasion," Journal of Personality and Social Psychology, 37 (1), 97-109.

Eagly, A.H. and S. Chaiken (1975), "An Attribution Analysis of the Effect of Communicator Characteristics on Opinion Change: the Case of Communicator Attractiveness," Journal of Personality and Social Psychology, 32, 136-144.

Forgas, J.P. and G.H. Bower (1988), "Affect in Social Judgments," Australian Journal of Psychology, 40, 125-145.

Himmelfarb, S. and D. Arazi (1974), "Choice and Source Attractiveness in Exposure to Discrepant Messages," Journal of Experimental Social Psychology, 10, 515-527.

Hirschman, E.C. and M.B. Holbrook (1982), "Hedonic Consumption: Emerging Concepts, Methods and Propositions," Journal of Marketing, 46 (Summer), 92-101.

Kahle, L.R. and P.M. Homer (1985), "Physical Attractiveness of the Celebrity Endorser: A Social Adaptation Perspective," Journal of Consumer Research, 11, 554-561.

Krugman, H.E. (1986), "Low Recall and High Recognition of Advertising," Journal of Advertising Research (February/March), 79-86.

Mitchell, A.A. and J.C. Olson (1981), "Are Product Attribute Beliefs the Only Mediator of Advertising Effects on Brand Attitude?" Journal of Marketing Research, XVIII (August), 318-332).

Moreland, R.L. and R.B. Zajonc (1979), "Exposure Effects May Not Depend on Stimulus Recognition," Journal of Personality and Social Psychology, 37 (6), 1085-1089.

Zajonc, R.B. (1968), "Attitudinal Effects of Mere Exposure," Journal of Personality and Social Psychology, Monograph Supplement, 9, 1-87.