An Investigation of the Effects of Repetition on Cognitive and Affective Reactions T0 Humorous and Serious Television Commercials

George E. Belch, San Diego State University
Michael A. Belch, San Diego State University
ABSTRACT - A laboratory experiment is used to examine the effects of three levels of exposure on cognitive and affective reactions to humorous and serious television commercials. Attention is also given to the use of message variation at high exposure levels as a way of reducing advertising wearout. The results indicate differences in the pattern of effects due to repetition for the two types of messages. Also, the findings indicate that wearout due to high levels of exposure can be reduced by the use of varied message execution. Particularly for a humorous commercial.
[ to cite ]:
George E. Belch and Michael A. Belch (1984) ,"An Investigation of the Effects of Repetition on Cognitive and Affective Reactions T0 Humorous and Serious Television Commercials", in NA - Advances in Consumer Research Volume 11, eds. Thomas C. Kinnear, Provo, UT : Association for Consumer Research, Pages: 4-10.

Advances in Consumer Research Volume 11, 1984      Pages 4-10

AN INVESTIGATION OF THE EFFECTS OF REPETITION ON COGNITIVE AND AFFECTIVE REACTIONS T0 HUMOROUS AND SERIOUS TELEVISION COMMERCIALS

George E. Belch, San Diego State University

Michael A. Belch, San Diego State University

ABSTRACT -

A laboratory experiment is used to examine the effects of three levels of exposure on cognitive and affective reactions to humorous and serious television commercials. Attention is also given to the use of message variation at high exposure levels as a way of reducing advertising wearout. The results indicate differences in the pattern of effects due to repetition for the two types of messages. Also, the findings indicate that wearout due to high levels of exposure can be reduced by the use of varied message execution. Particularly for a humorous commercial.

INTRODUCTION

The use of humor as a basis for advertising messages has become a popular form of marketing communication over the years. The increasing use of humor in advertising implies a belief that this ad form enhances the effectiveness of a persuasive message. While advertisers continue to use humor as a basis for persuasive appeals, empirical evidence supporting the faith in this form of advertising is scant. Research examining the impact of humor on the effectiveness of advertising messages per se has been limited, while studies examining the effects on persuasion have produced mixed results (cf. Sternthal and Craig 1973; Markiewicz 1974).

While empirical evidence supporting the effectiveness of humor on persuasion is in itself limited and equivocal, there is virtually no research on the issue of how reactions to humorous messages change with repeated exposures. Research on advertising wearout has shown that even when commercials are initially effective, repeated exposure may cause effectiveness to level off and eventually even decline (Grass and Wallace 1969; Appeal 1971; Craig, Sternthal and Leavitt 1976; Craig and Sternthal 1980). It may be that humor loses its impact after a few exposures and wears out quickly, perhaps even more so than serious messages. On the other hand, humor may not be as irritating as serious messages and may stand up well over repeated exposures.

The purpose of this study is to examine the impact of repetition on the communication effectiveness of humorous and serious commercials. More specifically, this study will attempt to determine if the repetition response pattern differs for humorous and serious commercials. Also, this investigation examines the use of variation of humorous and serious messages as a possible way of retarding wearout at high exposure levels.

RELEVANT LITERATURE

Effects of Humor

As noted above, research on the effectiveness of humor has produced mixed results. The major conclusion that comes from the humor literature is that humor does not influence persuasion differently from serious communications. In a review of humor studies, Markiewicz (1974) found that out of thirteen studies, ten failed to find evidence for any difference in persuasion due to message humor, two found tentative support for greater effectiveness of humor, and one found a serious message to be more persuasive.

Results for retention and comprehension also failed to show any support for the use of humor. Overall, empirical research in this area has failed to support the notion that humor enhances communication effectiveness (cf. Sternthal and Craig 1973 Markiewicz 1914: Gruner 1976)

There is reason to believe that humor should enhance the effectiveness of a persuasive message. Several theoretical explanations have been offered as to why the use of humorous messages should be an effective form of communication. A very practical explanation that advertisers would appear to agree on is that humor enhances attention to the message, at least on the first few exposures. Obviously, attention is prerequisite for further message processing and subsequent communication effectiveness. However, attention does not guarantee message comprehension and reception and, in fact, it has been suggested that humor may interfere with message comprehension and thus impair reception and learning (Sternthal and Craig 1973).

Perhaps the most popular theoretical explanation offered for the effects of humor is the distraction hypothesis. This explanation suggests that a humorous message distracts the audience, thus inhibiting the production of counterarguments, particularly for those opposed to the arguments advanced in the message. Reduced counterargumentation should result in increased message acceptance and persuasion (Osterhouse and Brock 1970; Baron, Baron and Miller 1973). Although it offers an appealing explanation for the communication effects of humorous messages, the distraction process may not adequately explain humor's effect on persuasion. First of all, past research has shown that humorous messages are no more effective than serious messages in inducing persuasion. This failure to find increased persuasion due to humor suggests that the occurrence of a distraction-counterargumentation disruption process is questionable. Also, the counterargument-reduction explanation of distraction effects is predicated on two conditions: an initial motivation to counterargue against the message, and a capability of counterarguing by the message recipient. It is questionable whether these conditions are met by most low involvement advertising situations. Since past research has studied the effects of humor in terms of overall outcome measures rather than examining the process underlying its effect, the distraction explanation remains rather tenuous.

Another theoretical explanation that is relevant to the effects of humor on persuasion is based on learning theory. This explanation suggests that humor functions as a positive reinforcer and messages presented in this context may be more effective in conditioning attitudes. Markiewicz (1974) also suggests that humor may operate as an unconditioned stimulus in a classical conditioning sense and a message paired with humor might elicit a positive response. She suggests that this associative process might be well-suited for simple messages in which one concept or argument is presented several times in connection with humor.

Humor may also enhance message effectiveness by increasing source credibility and ultimately the persuasiveness of the source's appeal. Markiewicz (1974) has noted that the effect of humor on source perceptions tends to be inconsistent: however, humor generally appears to enhance rather than lower evaluations of the source. Sternthal and Craig (1973) suggest that the addition of humor to a message may enhance the audience's perception of the message source, particularly when the message is a dull communication. They note that to the extent that most commercials are perceived to be dull, the addition of humor may augment source credibility.

Repetition Effects

As noted earlier, the occurrence of wearout due to repeated exposures to a commercial has been demonstrated in a number of studies. The wearout phenomenon whereby attention, awareness, and recall initially increases, levels off, and then ultimately declines has been shown in both laboratory and field experiments. Calder and Sternthal (1980) have suggested two possible causes of wearout. One is inattention whereby, with increased exposure, recipients no longer attend to a message and the message loses its effectiveness as forgetting sets in. This explanation was supported in a study by Craig, Sternthal and Leavitt (1976). They found a significant decline in brand name recall when exposure levels exceeded the number needed to learn the brand names. However, experimental inducement of attention resulted in an elimination of the previous wearout.

The second explanation for wearout offered by Calder and Sternthal deals with active information processing. This explanation focuses on the cognitive processing which occurs with increased exposure to a persuasive communication. Several theoretical explanations for repetition effects have been offered which focus on the underlying processing. Cacioppo and Petty (1979) offer an explanation for repetition effects based on a two-stage attitude modification model. According to this motel, repeated exposure to a message provides recipients with more opportunity for cognitive elaboration upon message content and to realize the favorable implications and cogency of the arguments. At high exposure levels, however, tedium or reactance is likely to develop leading to an attack against the message and a decline in affect.

A similar explanation for repetition effects is offered by Berlyne's (1970) two-factor theory. According to this approach, two separate and opposing psychological processes (positive habituation and tedium) operate simultaneously. Positive habituation is similar to a reduction in response competition; exposure results in a reduction in arousal due to conflict and uncertainty and thus increases affect. Tedium also increases with exposure and results in negative affect. Berlyne suggests that the relative strength of each factor varies as a function of exposure to the stimulus with the habituation process dominating initially, while tedium and disliking takeover at higher exposure levels. Berlyne suggests that stimulus complexity and sequence heterogeneity slow the positive habituation process; thus tedium occurs at higher exposure levels for complex, varied stimuli and at relatively low exposure level for simple, nonvaried stimuli.

Support for the two-stage attitude modification model was found in studies by Cacioppo and Petty (1979, 1980). However, a study by Belch (1982) did not fully support these theoretical accounts of repetition affects. Belch did not find a significant effect due to repetition on either attitude or purchase intention although he did find an increase in negative cognitive processing between moderate and high exposure levels, thus suggesting the occurrence of tedium or reactance at high exposure levels.

Purpose and Overview

The effects of humor in advertising remains basically unanswered from an empirical perspective. Most of the extant studies have not directly examined the use of humor in an advertising context but rather have studied the use of other forms of humor in communication (mostly satire). While research is needed to better understand the effectiveness of humorous advertising versus serious appeals, advertisers must also be concerned about the wearout rate for humorous messages. Advertising campaigns generally call for constant repetition of a message and what is funny and entertaining initially may lose its impact after a few exposures and thus wearout rather quickly. After a few exposures to a humorous message. the receiver may feel there is little to be gained from further viewing (due to reduced complexity) and may tune out the message or even be irritated by it. A serious message, on the other hand, might continue to hold the viewer's attention and interest over multiple exposures (complexity remains high) although it also would eventually begin to show signs of wearout.

An internally valid test of the effects of humorous versus serious messages and the wearout of each requires that equivalent versions of the advertisement be used. Indeed this has been a stumbling block to research on humor in advertising as it is difficult to manipulate the humor level in a message without affecting some other aspect of the message. Costs and creative limitations also impede researchers from producing similar humorous and serious versions of television commercials. This study was also constrained by these limitations. However, it was possible to find two commercials (humorous and serious) for a service that though different in creative execution, were similar in length, number of arguments, and complexity of arguments. These commercials, though different in execution style, were felt to be similar enough to be used in testing the effects of repetition on humorous and serious messages. This study is not designed to be a test of humor versus serious advertising messages per se. but rather to provide insight into the effects of repetition on both of these ad forms. The use of existing commercials also allows for a variation in the message pool at high exposure levels, thus making it possible to determine if wearout can be reduced by using different executions of the same basic theme.

METHOD

A laboratory study was designed for the purposes of testing the research hypotheses concerning the effects of repeated exposures to humorous and serious commercials. between subjects design was used with type of message (humor or serious) and exposure level (one, three, or five) as the factors. Also, two additional five-exposure conditions (one humorous and one serious) were used whereby the advertisements varied. In these cells the stimulus commercial was seen three times and another commercial (humorous or serious), for the service, was shown twice.

The advertisements used in the study were 30-second "Federal Express" commercials promoting the speed and reliability of the firm's overnight delivery service. [One potential problem of using existing commercials is the lack of control of the respondents' prior exposure level to the Federal Express ads. Subjects may have entered the experimental situation with varying levels of prior exposure to the stimulus commercials. In order to control for this factor, commercials were chosen that had not been aired for at least six months prior to the experiment. Debriefing of the subjects revealed that there was very little, if any, prior awareness or familiarity with the stimulus commercials.] The commercials used in the study were chosen after pretesting a number of Federal Express messages with respect to perceived humor. In the pretest, 105 subjects were asked to rate a number of Federal Express commercials on a 10-point scale ranging from "not at all humorous" to "very humorous." The mean humor ratings for the serious commercial was 2.62, while the humorous commercial scored a 7.33. The mean scores for the serious and humorous commercials used as fillers in the five-exposure varied execution conditions were 1.52 and 8.34. respectively.

The stimulus commercials were embedded in a half hour television program about the recruitment of employees for a new Playboy club in the local area. The program was chosen since it was very much of interest to the student group used in the study and was thus effective in maintaining attention and interest. The Federal Express commercial(s) was the only advertisement in the program. In the single exposure conditions the stimulus commercial was shown near the end of the program, while in the multiple exposure conditions the ads were interspersed throughout the program in the same positions.

Subjects and Procedure

Subjects for this study were 184 undergraduate business students enrolled in introductory marketing courses. The subjects were randomly assigned to one of the eight experimental treatments resulting in 23 subjects per cell. Subjects were told that they were participating in a research project evaluating people's reactions to various aspects of television programming. The subjects were told that they would be asked questions about the program and the commercials in the show.

After completing some premeasures, including demographics and television viewing habits, the half hour program was shown. Immediately after the program ended the subjects were told that a sampling of their reactions to the advertising shown during the program was of interest. The subjects were requested to read the cognitive response instructions and were then given two minutes to list their thoughts. After completing the cognitive response task, the subjects were asked to complete a program evaluation form and a set of postmeasures concerning issues dealt with in the program. After completing these measures, subjects were asked to respond to a number of dependent measures pertaining to the commercial. These measures are described below.

Dependent Measures

In addition to the cognitive responses, a number of dependent measures were taken including message recall, advertiser credibility, attitude toward the commercial, attitude toward the service, and usage intention. Message recall was measured by having the subjects list as many claims as they could remember from the message. The recall score was then calculated by counting the number of correct claims listed by the subject. Advertiser credibility was assessed by having the subjects indicate their impression of the advertiser on seven-point scales measuring trustworthiness, credibility, and truthfulness of the advertiser.

Attitude toward the advertisement was measured by having subjects respond to five scales (favorable - unfavorable; informative - uninformative; enjoyable - unenjoyable; interesting - uninteresting; irritating - not irritating). Attitudes toward using Federal Express were measured on two semantic differential scales (favorable - unfavorable; good - bad), while behavioral intentions were measured by asking subjects what the probability is they might use Federal Express if they had the need for such a service. Their intentions were measured on two semantic differential scales (likely - unlikely; definitely would use definitely would not use). For the attitude and usage intention, the dependent measure was calculated by averaging across the two scales used to assess each variable.

The cognitive response classification scheme included three major categories of thoughts: product/message related evaluations, ad - execution related evaluations, and repetition related evaluations. The product message related evaluations included the categories of counterargument, support argument, source bolster, source derogation, simple affirmations, simple disaffirmations, and curiosity thoughts. Ad execution evaluations included thoughts directed at the execution of the commercial and included responses concerning the quality of the commercial, creative style used, voice tones, colors, visual effects, etc. These thoughts were coded as being either positive or negative in valence. It should be noted that advertising execution related thoughts are distinct from source derogations and source bolsters. The latter two types of thoughts reflect a "message discounting" process resulting from thinking about the message source or style of argument, while ad - execution type responses represent a reaction to the execution of the commercial rather than the message.

Repetition related responses included any thought that addressed the fact that the commercial was seen more than one time in the program. While these thoughts can be either positive or negative in valence, the latter is generally much more common.

The cognitive response protocols were coded by three judges who were given operational definitions of the response categories and were trained in the application of these definitions. Unanimous agreement among the three judges was achieved for 72% of the responses. Two of the three judges agreed on another 23% of the cognitive responses. Thus, 95% of the responses could be classified using a modal scoring convention. The remaining responses were classified after some discussion among the judges.

Manipulation Checks

Although the stimulus commercials used in this study were pretested for humor, it is necessary to determine whether the subjects in the experimental conditions differed in their perceptions of the degree of humor in the commercials. One of the scales used to measure the subjects' reactions to the commercial was a semantic differential anchored by serious/humorous. The mean scores on the humor scale for each treatment are shown in Table 1. Higher scores on this scale indicate a more humorous rating of the commercial.

TABLE 1

PERCEPTIONS OF MESSAGE HUMOR FOR EXPERIMENTAL TREATMENTS

An analysis of variance was performed to determine if there were significant effects due to the humorous message. The results of the ANOVA showed a highly significant main effect for humor (F, 1,132 d,f. = 101.01, p < .001). However, perceptions of the amount of humor were not affected by repetition on the interaction of the message with humor. The ANOVA results show that the main effect of humor was significant indicating that overall the humorous message was perceived to be more jocular than the serious message. However, since the purpose of this study is to examine the effects of humorous and serious messages over repeated levels of exposure, it is important to determine whether the humorous message is perceived as being funnier than the serious message at each exposure level. To determine this, t-tests were conducted comparing the perceptions of humor for the humorous and serious messages at each exposure level. As can be seen by the t-test results in Table 1, the humorous message is indeed seen as being more humorous than the serious message at all three-exposure levels. The effectiveness of humorous message vs. the serious message also occurs in the five-exposure varied conditions. Thus, the humor vs. serious message manipulation was successful across the three-exposure levels.

RESULTS

To test the hypotheses concerning the effects of repetition on humorous and serious commercials, a two-way analysis of variance was performed for each of the dependent measures of interest with message style and repetition as the factors.

The purpose of the ANOVA was to determine if there were significant effects due to the type of message and to determine if the effects of repetition are different for humorous and serious messages. Separate comparisons were then made across the three-exposure levels for both the humorous and serious messages. The results for the various dependent measures are summarized below.

Advertiser Credibility

The results of the ANOVAs for the three measures of source credibility resulted in only one significant effect. There was a significant main effect of message type for the credibility measure (F = 4.75, 1/132, p < .03). This main effect was a function of the humorous message resulting in more favorable perceptions of advertiser credibility than the serious message. Perceptions of advertiser credibility did not change significantly across the three-exposure levels and there were no significant interactions between message type and exposure level.

Attitude Toward the Ad

The cell means for the five measures used to assess attitude toward the commercial are presented in Table 2. Several significant main effects were found for these measures, primarily for message type. Significant main effects were found for favorability (F = 7.46, 1/132 d.f., p < .001), enjoyableness, (F = 61.78, 11132 d.f., p < .001), interestingness (F = 22.94, 1/132, p < .001), and irritation (F = 5.95, 1/132, p < .01). For all four variables, the main effects were due to the humorous message being rated more positively than the serious message. The only measure which showed an effect for repetition was irritation where there was a marginally significant main effect [F (1,132) = 2.91, p < .06] resulting from increased perceptions of irritation across the three-exposure levels.

Significant interactions between message type and exposure were found for two of the measures: favorability and interestingness [F (2,132) = 4.81, p = < .01 and 3.56, D < .03. respectively]. For the favorability measure, the interaction resulted from the humorous message being rated most favorably in the three-exposure condition while the serious message was perceived the least favorably in the three-exposure condition. Regarding the interestingness measure, in the humorous condition evaluations increased between the one- and three-exposure level then declined in the five-exposure, while in the serious condition the ratings increased across the three-exposure levels.

TABLE 2

ATTITUDE TOWARD ADVERTISEMENTS FOR EXPERIMENTAL TREATMENTS

Separate one-way ANOVAs across the three-exposure levels for the two message groups revealed several differences. For the humorous message, favorableness increased significantly between the one- and three-exposure levels then declined between the three- and five-exposure levels [F (2,66) = 4.66, p < .01]. A similar pattern occurred for the interestingness measure (F = 5.93, p < .01). The enjoyableness measure also showed a significant repetition effect for the humorous message [F (2,66) = 3.75, p < .05]. The pattern of effects was similar to that found for the favorableness and interestingness measures; however, the increase in the enjoyableness rating between the one- and three-exposure conditions was not significant. The decline in the enjoyableness ratings between the three- and five-exposure conditions was significant. Separate analyses for the serious message conditions resulted in no significant differences due to repetition for any of the attitude toward the ad measures. However, this was due primarily to a decline in the ratings between the three- and five-exposure groups. Separate analyses for the serious message conditions resulted in no significant differences due to repetition for any of the attitude toward the ad measure.

Attitude and Cognitive Response

The cell means for the attitude toward using Federal Express, intention to use the service, and the cognitive response measures are presented in Table 3. Neither the attitudinal nor the intention measure showed any significant effects due to the experimental treatments. However, several of the cognitive response measures showed significant effects. For total thoughts, there was a significant main effect for repetition [F, (1,132) =-4.15, p < .023. As can be seen in Table 3, in both the humorous and serious conditions total thoughts increased between the one- and three-exposure levels and then declined between the three- and five-exposure levels. For both the total positive and total negative thoughts-measures, the main effect of message type was significant [F, (1,132) w 32.08, 5.44, respectively, p < .01]. These effects were a result of subjects viewing the humorous message generating more positive thoughts and less negative thoughts than those subjects viewing the serious message. For the total negative thoughts measure, there was also a significant main effect for repetition. As can be seen in Table 3, total negative thoughts increased between the one- and three-exposure levels for both types of messages. However, between the three- and five-exposure conditions total negative thoughts increased slightly for the humorous message but declined sharply for the serious message. Separate ANOVAs for the two message groups revealed only one significant effect due to repetition. The total negative thoughts measure was significant for the serious message [F (2,66) = 4.44, p < .02].

TABLE 3

ATTITUDE, INTENTION, COGNITIVE RESPONSE AND RECALL SCORES FOR EXPERIMENTAL TREATMENTS

Message Recall

Table 3 also shows the unaided message recall scores for the six treatments. These scores represent the mean number of correct claims recalled by the message recipients. The ANOVA results revealed no significant effects for the experimental variables. However, it is interesting to note that there are differences in the pattern of recall scores across the three exposure levels for the two types of messages. For the humorous message, the recall scores increase slightly between the one- and three-exposure conditions then show a slight decline between the three- and five-exposure conditions. For the serious message, the recall scores decline slightly between the one- and three-exposure conditions and then increase sharply between the three- and five-exposure conditions. A separate one-way analysis of variance for the two message groups revealed no significant differences across the three-exposure levels for the humorous message. However, for the serious message there was a marginally significant effect of repetition [F, (2,66) = 2.74, p < .07]

Effects of Varied Execution

The results presented thus far have shown the effects of exposure to humorous and serious messages across one-, three- and five-exposure levels. This section examines the effects of using varied executions in the five exposure condition. Of interest here are the ratings for the humorous and serious messages in the five-exposure varied and nonvaried conditions. Comparisons of the two five-exposure conditions were made using t-tests.

For the advertiser credibility measures, the t-tests revealed no significant differences in perceptions between the two five-exposure humorous conditions. However, significant differences in perceptions of advertiser credibility between the five-exposure serious nonvaried and varied conditions were found for trustworthiness (t -1.97, p < .05) and truthfulness (t = -2.33, p < .05) variables. For both variables, the use of varied execution in the five-exposure condition resulted in more favorable perceptions.

The results of the t-tests for the attitude toward the advertisement variables revealed several significant differences between the five-exposure humorous varied and nonvaried conditions including the informativeness (t = -2.00, p < .05), enjoyableness (t = -2.28, p < .05), and interestingness (t = -2.85, p < .01) measures. For all three measures the use of varied execution in the five-exposure humorous condition resulted in more favorable reactions toward the message than did five exposures to the same humorous message.

Comparisons of the attitude toward the commercial measures for the five-exposure serious conditions revealed only a marginally significant effect for the enjoyableness measure (t = -1.87, p < .07). This difference was a result of a more favorable rating in the five-exposure serious varied vs. nonvaried condition.

For the attitude toward using Federal Express measure, the difference between the five-exposure humorous varied and nonvaried conditions was significant (t = -2.40, p < .02) as the varied condition resulted in more favorable attitudes than the nonvaried condition. The differences in attitude were not significant for the five-exposure serious conditions nor were there any significant differences in the behavioral intention scores of the five-exposure condition groups.

The effects of varied execution in the five-exposure humorous and serious conditions on cognitive responses can be seen in Table 3. Examination of this table shows that varied execution in the five-exposure humorous condition resulted in more total thoughts, more positive thoughts, and fewer negative thoughts than in the nonvaried conditions. However, none of these differences were statistically significant. The only variable to approach significance was the total thoughts measure (t = -1.80, p < .08).

In the five-exposure serious conditions, Table 3 shows that varied execution resulted in slightly less total ideation and slightly more positive thoughts and negative thoughts than nonvaried execution. However, again none of these differences were statistically significant.

For message recall, varied execution results in a slight decline in the humorous condition although the difference was not significant. In the serious condition the recall scores were identical between the two groups.

DISCUSSION

The purpose of this study was to examine differences in the effects of repeated exposure to humorous and serious television commercials. Attention was also given to the effects of using variation in message execution at the highest exposure level as a possible way of retarding wearout. The results of this study showed several interesting findings.

As might be expected, the humorous message resulted in more favorable evaluations by the message recipients than the serious message. The humorous message resulted in more positive perceptions of advertiser credibility, more favorable attitudes toward the commercial, and more favorable cognitive responses than did the serious message. However, attitudes toward using the service and intention to use the service were not affected differently by the serious and humorous messages. This finding is consistent with past research which has failed to show a significant difference between humorous and serious messages with respect to persuasion. The absence of an effect for these two variables may be due to the fact that they are somewhat higher order response variables which would not be particularly relevant to the sample group used. It is unlikely that college students would have much of a need for an overnight delivery service, thus attitude toward using and intention to use would be rather abstract constructs which might not be differently affected by variations in message execution.

Regarding wearout of humorous vs. serious messages, there were several interesting findings. For the humorous message, evaluations of the commercial generally tended to become more favorable between the low and moderate exposure levels but then became more negative between the moderate and high exposure levels. This pattern of results occurred for the attitude toward the ad measures and for the attitudinal and intention measures although differences were not significant for the latter two measures. For the cognitive response measures the effects were slightly different as total positive thoughts remained constant between the low and moderate exposure levels then declined at the high exposure level. Total negative thoughts did not change significantly across the three-exposure levels.

The repetition function for the serious message differed from that of the humorous for many of the variables. Although the differences were not significant, most of the attitude toward the ad measures subjects' evaluations declined between the low and moderate exposure levels then remained constant or actually increased between the three- and five-exposure levels. The exceptions were the interestingness and irritation measures which showed an increase across the three-exposure levels. The pattern of a decline in evaluation between the low and moderate exposure levels followed by an increase between the moderate and high exposure levels was also found for the attitude and intention measure. The cognitive response measure showed a similar pattern. Positive thoughts remained constant between the one- and three-exposure levels then increased slightly between the three- and five-exposure levels. Total negative thoughts increased significantly between the low and moderate exposure levels but then declined between the moderate and high exposure levels. It is also interesting to note that message recall was relatively constant between the one- and three-exposure levels (actually declining slightly) but then increased between the three- and five-exposure levels.

The repetition results for the humorous and serious messages suggest different wearout functions for the two messages. The humorous message showed a repetition pattern similar to that predicted by Berlyne's two-factor theory. Reactions to the humorous message showed evidence of becoming more favorable between the low and moderate exposure levels then declining between moderate and high exposure levels. The serious message, on the other hand, was least effective at the moderate exposure level then tended to show an increase in effectiveness at the high exposure level.

It may be that once the humorous message had been seen few times, the recipient was satiated by the humor and thus becomes bored or irritated by subsequent exposures. Thus, tedium or reactance would result in more negative evaluations at the high exposure level. For the serious message, it appears that the tedium or reactance which might lead to negative evaluations did not occur at the high exposure level. Also, the increase in recall scores between the three- and five-exposure conditions suggests that the recipients did continue to attend to the serious message at the high exposure level. Of course it may be that there was something about the serious message itself (e.g., execution, delivery, etc.) that held the subjects attention and delayed wearout. Further research with another serious message would be needed in order to make any firm conclusions.

The results of this study also suggest that wearout can be delayed somewhat by the use of varied messages. Particularly interesting is the fact that the use of varied execution was more effective for the humorous message than for the serious message. This suggests that while repeated exposure to the same humorous message may result in satiation and wearout, variation in the message execution can reduce the amount of fatigue and irritation. Thus, it may be important for advertisers who use a humorous campaign to maintain a pool of commercials which can be used rather than to use the same message over and over. This is not to suggest that variation in execution is not important for serious messages. As past research has shown, wearout can be reduced when different commercial executions for a brand are used (Grass and Wallace 1969). However, variation may be particularly important for humorous messages as there may be little to gain from the funny ad once the humor has been processed.

SUMMARY

This study suggests that there may be differences in the effects of repetition for humorous and serious messages. However, caution must be taken in generalizing these findings as there were differences in the execution styles of the two types of messages. It is worth noting that the results of this study indicate that even a very humorous commercial is proven to wearout after a high number of exposures. Thus, advertisers should not assume that a commercial can be used repeatedly to a target market because it is funny.

Of course more research is needed on the wearout pattern for humorous messages. Research along the lines of Calder and Sternthal (1980), which examines wearout on a more longitudinal basis, would be particularly valuable in examining wearout of humorous messages. Also, studies which use delayed measures of persuasion as well as immediate measures are needed . A study by Lammers, Leibowitz and Seymour (1983) of humorous and serious messages found no differences in the effectiveness of the two types of messages when measures were taken immediately after exposure. However, when cognitive response measures were delayed, the humorous appeal was superior to the serious message. Hopefully this study provides some interesting findings on the wearout of humor and will encourage others to give further research attention to the issue.

REFERENCES

Appel, Valentine (1971), "On Advertising Wearout," Journal of Advertising Research, 11 (February), 11-13.

Baron, Robert, Baron, P. and Miller, N. (1973), "The Relation Between Distraction and Persuasion," Psychological Bulletin, 80, 310-323.

Belch, George 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. (1970), "Novelty, Complexity and Hedonic Value, Perception and Psychophysics, 8, 279-286.

Cacioppo, James and Petty, Richard (1979), "Effects of Message Repetition and Position on Cognitive Response, Recall and Persuasion, Journal of Personality and Social Psychology, 37 (January), 97-109.

Cacioppo, James and Petty, Richard (1980), "Persuasiveness of Communications is Affected by Exposure Frequency and Message Quality: A Theoretical and Empirical Analysis of Persisting Attitude Change," in Current Issues and Research in Advertising, J.H. Leigh and C.R. Martin, Jr. (eds.), Ann Arbor: Division of Research, Graduate School of Business Administration, University of Michigan.

Calder, Bobby J. and Sternthal, Brian (1980), "Television Commercial Wearout: An Information Processing View, Journal of Marketing Research, 17 (May), 173-186.

Craig, C. Samuel, Sternthal, Brian and Leavitt, Clark (1976), "Advertising Wearout: An Experimental Analysis," Journal of Marketing Research, 13 (November), 365-372.

Grass, R.C. and Wallace, Wallace B. (1969), "Satiation Effects of T.V. Commercials," Journal of Advertising Research, 19, 47-57.

Gruner, C.R. (1976), "Wit and Humor in Mass Communication, in Humor and Laughter Theory, Research and Applications, A.J. Chapman and H.C. Foot (eds.), Wiley, New York.

Lammers, H. Bruce, Leibowitz, Laura and Seymour, George E. (1983), "Immediate and Delayed Effects of Humor on Cognitive Responses to Advertising Stimuli: A Trace Consolidation Approach," Journal of Business Research, in press.

Markiewicz, Dorothy (1974), "Effects of Rumor on Persuasion," Sociometry, 37, 407-422.

Osterhouse, Robert and Brock, Timothy (1970), "Distraction Increases Yielding to Propaganda by Inhibiting Counterarguing, Journal of Personality and Social Psychology, 15 (August), 344-358.

Sternthal, Brian and Craig, C. Samuel (1973), "Humor in Advertising," Journal of Marketing, 37 (October), 12-18.

----------------------------------------