The Impact of Emotional Valence and Intensity on Ad Evaluation and Memory

Karen Russo France, West Virginia University
Reshma H. Shah, University of Pittsburgh
C. Whan Park, University of Pittsburgh
ABSTRACT - This study examines the impact of specific program-induced emotions at varying levels of intensity on the memory for and evaluation of embedded advertisements. Consistent with previous research, we found that a positive emotion-inducing program facilitates ad evaluation while a negative emotion-inducing program debilitates ad evaluation. Furthermore, we found that highly intense positive emotions elicited by a television program do not further facilitate ad evaluation when compared to programs eliciting lower intensity positive emotions. Highly intense negative emotions, however, further debilitate ad evaluation when compared to programs that elicit negative emotions at lower levels of intensity. Ad memory was not significantly impacted by the television program's emotional valence or intensity. We offer several processing explanations for these findings.
[ to cite ]:
Karen Russo France, Reshma H. Shah, and C. Whan Park (1994) ,"The Impact of Emotional Valence and Intensity on Ad Evaluation and Memory", in NA - Advances in Consumer Research Volume 21, eds. Chris T. Allen and Deborah Roedder John, Provo, UT : Association for Consumer Research, Pages: 583-588.

Advances in Consumer Research Volume 21, 1994      Pages 583-588

THE IMPACT OF EMOTIONAL VALENCE AND INTENSITY ON AD EVALUATION AND MEMORY

Karen Russo France, West Virginia University

Reshma H. Shah, University of Pittsburgh

C. Whan Park, University of Pittsburgh

ABSTRACT -

This study examines the impact of specific program-induced emotions at varying levels of intensity on the memory for and evaluation of embedded advertisements. Consistent with previous research, we found that a positive emotion-inducing program facilitates ad evaluation while a negative emotion-inducing program debilitates ad evaluation. Furthermore, we found that highly intense positive emotions elicited by a television program do not further facilitate ad evaluation when compared to programs eliciting lower intensity positive emotions. Highly intense negative emotions, however, further debilitate ad evaluation when compared to programs that elicit negative emotions at lower levels of intensity. Ad memory was not significantly impacted by the television program's emotional valence or intensity. We offer several processing explanations for these findings.

Given a television program's ability to induce valenced affective states (Averill 1969, Axelrod 1963, Gardner and Wilhelm 1987, Goldberg and Gorn 1987, Gouaux 1971, Gouaux and Summers 1973, Mattes and Cantor 1982, Pavelchak, Antil, and Munch 1988), a relevant area of research might consider the impact of these varying emotions on viewers' reactions to commercials embedded in a program. Although previous research has considered the dichotomous effect of positive and negative program-induced moods (c.f. Goldberg and Gorn 1987), it has neglected the array of differentially valenced emotions. In general, previous research has supported the hypothesis that positive program-elicited moods enhance commercial effectiveness whereas negative ones debilitate effectiveness. These findings have been applied to ad placement by managers who want viewers to be in a generally positive state when their advertisement is aired (Schultz 1979, Schwerin 1958).

The present research argues that examining the impact of positive and negative feeling states may be appropriate when studying the impact of moods which are milder in intensity than emotions (Cohen and Areni 1991, Clark and Isen 1992, Gardner 1985). However, the simple distinction between positive/negative moods may not be adequate to explain the role of different types of program-induced emotions on commercial effectiveness. Emotions are more stimulus specific than moods, comprise more of a cognitive interface than moods, and tend to interrupt goal-directed activities (Cohen and Areni 1991, Clark 1982). Moreover, unlike moods, emotions may vary greatly in their intensity level (Mehrabian and Russell 1973). Thus, emotions may impact commercial effectiveness differently than do moods. Specifically, differences in commercial effectiveness may be more pronounced at lower levels of emotional intensity than at higher levels. At lower levels of emotional intensity, the impact of valence may be more pronounced; whereas higher levels of emotional intensity may tend to disrupt or interfere with ongoing attention and processing of embedded ads, thereby diminishing the impact of emotional valence.

The current study examines the impact of specific program-induced emotions at varying levels of intensity on the processing and evaluation of embedded advertisements. In pursuing these research issues, this study replicates and extends the work of Goldberg and Gorn (1987). We use Goldberg and Gorn's (1987) study as a benchmark for the following reasons: (1) it was the first study available that offered a specific mood-based explanation (cue-accessibility) for the impact of program-induced emotions on ad effectiveness, (2) it did not examine the impact of emotional intensity which may prove to have a differential impact on processing (Mehrabian and Russell 1973), and (3) while Goldberg and Gorn rely on the cue accessibility explanation, their cognitive response data reveal that other mechanisms may also operate. We test some of these other explanations. The present study is, thus, designed to shed light on some of the unresolved issues raised by Goldberg and Gorn, and to further extend findings of previous program-context effects research.

THEORY AND HYPOTHESES

While much of previous program-context effects research has examined the impact of positive/negative program-induced moods on commercial effectiveness, it may be more appropriate to examine the impact of program-induced emotions. Although researchers have found differences to exist between moods and emotions, these two terms have been used interchangeably in affect research, with little consideration given towards their unique differences. There has been an implicit assumption in previous research that the effects of emotions may be understood to operate similarly to moods on ad effectiveness. This assumption is evident in the Goldberg and Gorn study which utilized a mood-based explanation-the cue accessibility hypothesis-to predict the impact of positive and negative emotions (i.e. happy and sad) on commercial effectiveness.

According to the cue accessibility hypothesis, programs that elicit positive/negative affective states prime similarly valenced material stored in memory. In other words, an affective state may influence one's evaluations of stimuli because material stored in memory that is congruent with that feeling state will be more accessible, and consequently more likely to come to mind than it would at another time (Clark 1982, Clark and Isen 1982, Isen 1984, Isen 1975, Isen, Shalker, Clark, and Karp 1978). Previous research has found cue accessibility to operate in the case of positive emotions, but has neglected to examine whether it operates in the case of negative emotions (Clark 1982, Clark and Isen 1982, Isen 1984, Isen 1975, Isen, Shalker, Clark, and Karp 1978).

In the context of television commercials, consumers are likely to have an array of stored memories regarding brands and products. Therefore, affectively positive responses elicited by a television program would increase the accessibility of positive material stored in memory regarding the advertised brand, product, or product category. Goldberg and Gorn (1987) find support for the cue accessibility hypotheses in that the happy television program tended to lead to greater perceived commercial effectiveness, more affectively positive cognitive responses, and somewhat better recall than the sad television program. The cue accessibility hypothesis also predicts better recall in the case of positive programs since positive affective states lead to broader, more integrated cognitive categories (Isen et al. 1978, Isen 1984). These larger, well-integrated categories enhance information processing and memory through a spreading activation process. Positive affective states also tend to increase the extent of information processing (Hoffman 1986, Mamo 1990), thereby potentially increasing recall (Isen, Daubman, and Gorgoglione 1984).

Negative emotions may operate differently than positive emotions. Since negative emotions tend to induce more thought (Clark 1982, Goldberg and Gorn 1987), subjects may be distracted from attending to and processing embedded advertisements. When negative emotions are elicited by a television program, particularly from a negative program that is issue-oriented, it is likely that a viewer will be distracted from processing ad information. For example, in the Goldberg and Gorn (1987) study, 54 percent of those viewing the negative program reported still thinking about the program as they viewed the advertisement compared to only 21 percent who viewed the positive program. Thus, in the case of negative emotions, we would expect ad evaluation and memory to decrease not because of cue accessibility, but rather due to distraction/interference.

In light of the above discussion, we hypothesize that:

H1: A positive emotion-inducing program facilitates ad memory and evaluation while a negative emotion-inducing program debilitates ad memory and evaluation.

Clark (1982) proposes a link between arousal (emotional intensity) and affect valence. Specifically, experiencing intense emotions in the present will help to prime material previously stored in memory linked with the level of intensity. To the extent that a program induces strong emotional intensity, it facilitates retrieval of similarly valenced thoughts related to the brand, product, or product category. The greater the emotional intensity a person experiences with a specific emotion, the more intense the priming of similarly valenced material from memory, and consequently, the stronger the influence on brand evaluations. The cue accessibility hypothesis predicts that emotions, as a stronger form of moods, may better facilitate retrieval of similarly valence memories than do moods. While this may be a reasonable assumption for lower intensity emotions, the same may not hold for higher intensity emotions.

A great deal of research in psychology and consumer behavior has found a disruptive influence on information processing resulting from higher levels of emotional intensity (Eysenck 1976, 1982, Hasher and Zacks 1979, Park and Young 1986, Pavelchak, Antil and Munch 1988, Sanbonmatsu and Kardes 1988). Similarly, intense emotions resulting from program viewing may decrease one's motivation to attend to and process ad information (Bryant and Comisky 1978, Kennedy 1971, Soldow and Principe 1981, Television Audience Assessment 1984, Thorson and Oberman 1985, Thorson and Reeves 1986). Viewers' ongoing program-related thoughts may distract them from paying attention to the information presented in the advertisement.

As noted earlier, negative emotions tend to involve more issue-oriented thought. As such, higher intensity negative emotions should lead to increased interference/distraction. Positive emotions, on the other hand, tend to involve less program-oriented thought and, thus, should not disrupt processing of embedded commercials to the same extent as negative emotions. However, when the emotions felt during a television program are intensely positive, their impact on ad effectiveness is not clear. The cue accessibility hypothesis predicts that ad evaluation should be more favorable. According to this hypothesis, we may expect that strong positive emotions elicited by the program may, in fact, facilitate retrieval of similarly positive memories regarding the advertised brand, product, or product category. This should increase ad evaluation. However, other research on emotional intensity has demonstrated that intense emotions tend to distract viewers, thereby decreasing processing of subsequent stimuli. This should, in turn, decrease ad evaluation. Expecting the possible presence of both explanations, we propose that stronger positive emotions do not facilitate ad effectiveness to a greater degree than lower intensity positive emotions. From this discussion, it is hypothesized that:

H2: More intense negative emotions elicited by a television program decrease ad memory and evaluation; whereas more intense positive emotions elicited by a television program do not facilitate ad memory and evaluation when compared to lower intensity positive emotions.

METHODOLOGY

Design and Subjects:

To test the impact of program valence and intensity on advertisement effectiveness, a 2 X 2 full-factorial design was used in which program emotional valence (positive or negative) and felt emotional intensity (low or high) were manipulated between subjects. A total of 89 Masters of Business Administration students at a large, eastern university were recruited as subjects. Subjects represented a diverse group in terms of age, sex, and occupation.

Program and Advertisement Selection:

The two programs that were used to operationalize emotional valence and intensity were selected through expert judgement and several rounds of pretesting (see discussion below). The movies chosen for this study were Dirty Dancing (positive valence) and Midnight Express (negative valence). The stimuli for the current study were similar to those used in the Goldberg and Gorn (1987) study in that the negative program tended to induce more program-related thoughts in the elicitation of emotion while the positive program did not. It was necessary to use stimuli similar to that used by Goldberg and Gorn to test the competing hypotheses proposed.

A series of informal pretests, with approximately five subjects per cell, were conducted prior to the main pretest. During these informal pretests, subjects viewed the program segments and then completed a series of questions regarding the valence and intensity of emotions elicited by the program. Also, to ensure that the sequence of rearranged segments did not undermine understanding of the movie's plot, these subjects were questioned about the flow of the programs. After each informal pretest, the movies were re-edited in order to ensure that the programs varied along the appropriate dimensions.

The final version of the program stimuli was edited to approximately eleven minutes in duration and was pretested to ensure continuity as well as the intended manipulations. We manipulated emotional intensity at high and moderate levels of intensity since a low level of emotional intensity would decrease subject's involvement with the stimuli. Therefore, in this paper, we refer to low and high levels of intensity in a relative sense. To induce high levels of positive emotions, scenes from Dirty Dancing that included upbeat music and/or those in which the characters were portrayed as experiencing positive emotions (e.g., happiness, warmth, excitement) were used. Low intensity positive emotions were elicited with scenes that were void of music and in which the characters were experiencing less intense levels of positive emotions. Inducing high levels of negative emotions was accomplished with scenes from Midnight Express which were very graphic (e.g., a fight between two prison inmates) and/or portrayed the characters as experiencing intensely negative emotions (e.g., despair, loneliness, anger). Low intensity negative emotions were elicited with scenes in which there was less violence and in which the characters experienced less intense levels of negative emotions.

In addition to selecting various scenes from each movie, instructions to the subjects were used to induce different levels of felt emotional intensity. Subjects in the high emotional intensity condition were told that it was important for them to try to experience the emotions of the main character(s) as it was relevant to the completion of a questionnaire following the viewing. Subjects in the low emotional intensity condition were directed only to try to gain an understanding of the movie's contents. Finally, so as not to confound program involvement with emotional intensity, subjects in both the low and high emotional intensity groups were asked to pay close attention to the movie segments and were told they would be questioned on what they had viewed.

Advertisements were selected on the basis of two criteria: (1) they had to be unfamiliar to the subjects, and (2) they had to be generally positive in affect. As indicated by previous research, it was important to use commercials with which the subjects were unfamiliar to ensure no effects of ad memory on advertisement evaluation. Also, since advertisements are, in most cases, positive in nature (Batra and Ray 1986), it was necessary to use ads that met this criteria so as to not cause subjects to focus an unusual amount of attention on the advertisements. A non-local commercial for Meritor Bank was selected as the target commercial. A second commercial, for Gorton's Fish , was used as a filler.

Program and Advertisement Pretest:

To assess the extent to which the program segments varied along the critical dimensions, 68 subjects similar in profile to those used in the main study participated in a pretest. The pretest confirmed that the programs manipulated both emotional valence and level of intensity in the intended direction. Emotional responses to the program were measured via a sixteen-item, nine-point, semantic differential scale (1="felt not at all," 9="felt very strongly") that was modified from Edell and Burke (1987). Positive emotion items included: happy, hopeful, amused, cheerful, warmhearted, enthusiastic, lively, and energetic. Negative emotion items consisted of: angry, sad, apprehensive, anxious, distressed, depressed, disgusted, and offended. Composite scores were created by combining positively (negatively) valenced emotion dimensions. Cronbach's alpha reliabilities equaled 0.96 and 0.93 for the positive and negative composite scores, respectively. Results strongly suggest that subjects felt positive emotions when viewing the positive programs (Xlow-pos=5.28, Xhigh-pos=6.80, Xlow-neg=2.12, Xhigh-neg=2.06; F=128.53, p=0.00) and felt negative emotions when viewing the negative programs (Xlow-pos=2.20, Xhigh-pos=2.78, Xlow-neg=4.45, Xhigh-neg=6.27; F=56.34, p=0.00). A larger mean on the positive (negative) composite score indicates that the respondents felt positive (negative) emotions more intensely when viewing the program segment.

A composite measure, "NETMOVIE", was calculated by subtracting the mean score of the negative emotional scales from the mean score of the positive emotional scales. As such, NETMOVIE may be used to assess both emotional valence and emotional intensity. That is, the larger the NETMOVIE score, the more intensely felt were the positive emotions; whereas the smaller the NETMOVIE score, the more intensely felt were the negative emotions. Means on NETMOVIE for Midnight Express were -2.28 and -4.18 for low and high emotional intensity, respectively; while, means for Dirty Dancing were 3.41 and 4.79 for low and high emotional intensity, respectively. In the case of Dirty Dancing we find an increase in NETMOVIE from low to high levels of intensity, while a decrease is found for Midnight Express. A main effect of emotional valence on NETMOVIE (F=172.85, p=0.00) and a valence x intensity interaction (F=8.86, p=0.00) were obtained. Since NETMOVIE was calculated by subtracting the mean of negative emotion scores from the mean of the positive emotion scores, the significant interaction effect indicates that both emotional intensity and valence were successfully manipulated.

Subjects were asked to assess the degree to which they found the program to induce emotions while also provoking a great deal of program-related thought by indicating their agreement on three, nine-point scales ranging from "Disagree" (1) to "Agree" (9). The scales assessed the degree to which subjects felt the programs (1) compelled one to think about the issues of right or wrong, (2) made one think about the issues raised in the segment, (3) could be characterized as being heavy and thought-provoking. On all of these scales, subjects indicated that they found the versions of Dirty Dancing (positive program) to be less thought provoking than Midnight Express (negative program) (See Table 1).

Finally, pretest results revealed that Midnight Express and Dirty Dancing differed in terms of both familiarity (F=9.59, p=0.00) and involvement (F=5.11, p=0.03). Familiarity was measured on a nine-point scale ranging from 1 (Very Unfamiliar) to 9 (Very Familiar). Similarly, involvement was assessed on a nine-point scale with end anchors ranging from "Uninvolved" (1) to "Involved" (9). Subjects tended to be more familiar with Dirty Dancing; whereas, they found Midnight Express to be more involving. While the intention was to control involvement through the experimental instructions, it appears that the Midnight Express stimuli created higher levels of felt involvement. Therefore, familiarity and involvement are used as covariates in analyzing the data for the main experiment.

A second pretest of twenty-five students was conducted to assess the target commercial used in our study. Pretest results revealed that subjects found the Meritor Bank commercial to be positive in nature (composite mean of eight, nine-point positive emotion scales X=4.9). Subjects also felt very few negative emotions while watching this commercial (composite mean of eight, nine-point negative emotion scales=2.5). Moreover, subjects evaluated the Meritor Bank commercial positively based on the composite mean for two, nine-point scales (1="Dislike" to 9="Like", and 1="Negative" to 9="Positive") (X=6.24). Additionally, subjects were unfamiliar with the Meritor Bank commercial (X=2.16 on a nine-point scale ranging from (1) Unfamiliar to (9) Familiar).

Experimental Procedure

The experiment was conducted during class time. The experimenter introduced him/herself and stated in the cover story that this was a research experiment being conducted by faculty members who were interested in assessing how people process television programs. Subjects were then told that they would be viewing a videotape of Dirty Dancing [Midnight Express]. Following this brief introduction, the experimenter distributed instruction sheets and read them aloud while the subjects were asked to read along. To substantiate the cover story and to better approximate a realistic viewing environment, subjects were instructed to "pay as much attention" as they could to the movie. However, it was thought that subjects might become suspicious about the true intent of the study for the following two reasons: (1) the appearance of commercials in the program, and (2) for anyone familiar with the movies, the rearranged sequence of the movie. To reduce demand effects, subjects were told that the movie they would be watching was edited in order to be aired on television in another part of the country. Thus, subjects were led to expect some changes in movie sequence and the appearance of commercials.

After the instructions were read, subjects watched a videotape of one of the two versions of either Midnight Express or Dirty Dancing. The videotapes were prepared with the commercials embedded at the end of the program segment. Immediately following exposure to the programs, subjects completed the questionnaire booklet.

TABLE 1

Experimental Measures

Respondents' evaluations of advertisements were inferred from measures of attitude toward the ad (AAD) and attitude toward the brand (ABR). Advertisement processing was assessed through measures of ad recall, attention to the ad, and distraction from it. Ad attitudes were measured on a two-item, nine-point semantic differential scale with anchors ranging from "Dislike" to "Like" and "Negative" to "Positive". Brand attitudes were measured using a two-item, nine-point semantic differential scale that asked subjects to indicate whether they found the advertised product or service to be "Bad" or "Good" and "Unappealing" or "Appealing". The AAD and ABR scales described below, were computed by the average of the summated items. Attention to the ad was assessed by asking subjects to indicate their level of attention to the ad on a scale ranging from 1 (Very Little) to 9 (Very Much). The degree to which subjects found the programs to distract them from paying attention to the ad was assessed on a nine-point scale ranging from 1 (Movie Did Not Distract Me) to 9 (Movie Distracted Me). Finally, direct measures of recall of the brand name and advertised product were taken.

RESULTS

Manipulation Checks

The manipulation checks of the selected programs yielded results that were highly consistent with those found in the pretest. The composite measure of program valence and intensity, NETMOVIE, revealed a main effect of emotional valence (F=253.14, p=0.00) and a significant emotion by intensity interaction (F=13.54, p=0.00). Means on NETMOVIE were as follows: Dirty Dancing Xlow=2.30, Xhigh=3.80; Midnight Express Xlow=-2.84, Xhigh=-4.52.

Hypotheses Testing

Hypothesis 1 predicts that a program that elicits positive emotions facilitates ad memory and evaluation; whereas a negative program decreases ad memory and evaluation. In terms of AAD, we found a significant main effect of valence (p=0.00); whereas, emotional intensity and the interaction effect were not significant (p=0.11 and p=0.80, respectively). In other words, AAD was significantly decreased when the ad was embedded in the negative (X=4.21) as compared to the positive program (X=6.17; F=10.78, p=0.00); thus supporting Hypothesis 1. A similar decrease resulted on ABR from the positive to the negative program (Xpos=6.32, Xneg=4.79 respectively; F=12.13, p=0.00). The negative program distracted subjects from paying attention to the ad (X=6.26) more so than the positive program (X=4.11; F=16.26, p=0.00). Likewise, subjects paid more attention to the commercial when it was embedded in the positive program (X=6.87) than when it was embedded in the negative program (X=5.58; F=6.54, p=0.01). However, as was found by Goldberg and Gorn (1987), recall of the advertised brand names and product was not significantly impacted by valence of the television program. These results are consistent with the findings of previous research regarding the facilitating effects of positive programs on ad effectiveness.

According to Hypothesis 2, more intense negative programs should decrease ad evaluation and memory more so than less intense negative emotions; whereas more intense positive emotions should not facilitate ad evaluation and memory. In the case of negative emotions, we found directional support for this hypothesis in that AAD was somewhat lower for the high intensity condition as compared to the low intensity condition (Xlow=4.70, Xhigh=3.69; t=1.52, p=0.07, one-tailed). The impact on ABR did not, however, significantly differ between the two programs (Xlow=4.98, Xhigh=4.60; t=0.71, p=0.2, one-tailed). Additionally, ad processing was reduced as indicated by a decrease in attention paid to the ad when it was embedded in the high intensity negative program (X=5.05) when compared to the low intensity negative program (X=6.09; t=1.75; p=0.04, one-tailed). Respondents also reported that the highly intense negative programs distracted them from thinking about the commercial (Xlow=5.36, Xhigh=7.19, t=2.47, p=0.00, one-tailed). However, brand and product recall were not impacted by the level of emotional intensity. Therefore, these results did not support Hypothesis 2 in the case of negative emotions.

In the case of positive emotions, we found no significant differences on AAD (t=1.27, p=0.21) between the low intensity (X=6.63) and high intensity (X=5.85) conditions. Similar results were found for ABR (Xlow=6.29, Xhigh=6.33; t=0.09, p=0.93). Subjects indicated that they paid equal attention to the advertisement in both the low (X=7.32) and high intensity (X=6.56) conditions (t=1.54, p=0.13). Moreover, respondents were not differentially distracted from thinking about the ad in the low (X=3.95) versus high intensity (X=4.22) conditions (t=0.39, p=0.7). Recall of the brand and product also were unaffected by the intensity of the positive emotions. These results support our contention that high intensity positive emotions do not facilitate ad effectiveness when compared to low intensity positive emotions. Overall, Hypothesis 2 is, thus, partially supported.

DISCUSSION

This study examined the impact of program induced emotions and their intensity on advertising effectiveness. Consistent with previous research, the results show that positive programs facilitate ad effectiveness whereas, negative programs are found to be debilitating. Moreover, the results of the current study suggest that when emotional intensity (low and high) is introduced, highly intense positive programs tend not to further facilitate ad effectiveness in comparison with low intensity positive programs. In the case of negative programs, high levels of emotional intensity did not significantly debilitate ad effectiveness when compared to lower levels. While these results constitute important findings in the area of advertising research, they also offer several important implications for future research.

First, the results of the present study did not support Hypothesis 2. Specifically, in the case of negative programs, although we found a significant increase in distraction from and decrease in attention to the ad, from low to high levels, standard levels of significance on AAD, ABR and recall were not obtained. AAD and ABR may not have differed between the high and low negative intensity conditions since 1) the experimental commercial (see the liking rating in the methods section) was highly well liked and 2) the advertised product (free checking) was likely to be appealing to respondents, thereby further augmenting respondent's liking of the commercial. Moreover, to the extent that a commercial captures a viewer's attention, the impact of program emotional intensity may be lessened. The commercial, with its humor, was able to draw audience attention despite interference from the program, thereby mitigating the expected distraction effects. Additionally, since this study was conducted in an experimental setting, recall may not have been affected by emotional intensity as respondents might have been sensitive to questions asking them to recall specific details of the experimental stimuli.

Second, since we used stimuli similar to that used by Goldberg and Gorn (1987), this study also has a confound between program valence and the amount of issue-oriented thinking. Midnight Express, the negative program, induced more issue-oriented thought than the positive program, Dirty Dancing. Additionally, these programs tapped into specific types of emotions (e.g., warmhearted and disgusting). Perhaps varying emotions impact ad processing and evaluation differentially. For example, the impact of excitement might be markedly different from the impact of warmheartedness. Similarly, a negative emotion such as sadness may impact ad effectiveness differently than anger. These limitations should be addressed in future research which replicates and extends the present study through the use of programs that induce different degrees of thought and tap into different positive and negative emotions.

Third, while the present study specified two possible mechanisms (cue accessibility and distraction) that explain the impact of emotional valence and intensity on ad effectiveness, other mechanisms may also have been operating. Specifically, as noted earlier the degree of issue-oriented thinking associated with a particular program was proposed to affect an audience's attention to the program. Issue-oriented thinking might also activate alternative processing mechanisms. This may be particularly true in the case of negative emotions, which as discussed earlier, tend to elicit issue-oriented thought more so than do positive programs.

One additional mechanism alluded to, but not tested in the present study is processing capacity resource constraints. This mechanism may decrease one's ability (versus their motivation) to process ad information when the majority of processing resources are devoted to the program. Eysenck (1976, 1982) and Hasher and Zacks (1979) explain that stimuli that is very affectively involving (i.e. higher in emotional intensity) interferes with the viewer's ability to encode ad information because the program may preempt some of the resources needed to process ad information. Schumann and Thorson (1990) discuss this phenomenon in terms of a "processing activity ratio." That is, if processing activity, or the amount of program-generated processing, favors the program, it is predicted that memory for the advertisement will be damaged. Future research may examine the impact of processing capacity and other explanations for the impact of program emotions and intensity on ad effectiveness. Finally, this study did not consider the dynamic relationship among mechanisms. For example, addressing the questions of when does one mechanism dominate the others, or when do mechanisms become complementary to one another, is important for advertisers in helping to alleviate the problems associated with negative emotion inducing programs.

REFERENCES

Averill, James R. (1969), "Autonomic Response Patterns During Sadness and Mirth," Psychophysiology, 5 (4), 399-414.

Axelrod, Joel N. (1963), "Induced Moods and Attitudes toward Products," Journal of Advertising Research, 3 (2), 19-24.

Batra, Rajeev and Michael L. Ray (1986), "Affective Responses Mediating Acceptance of Advertising," Journal of Consumer Research, 13 (September), 234-249.

Bryant, Jennings and Paul W. Comisky (1978), "The Effect of Positioning a Message Within Differentially Cognitively Involving Portions of a Television Segment on Recall of the Message," Human Communications Research, 5 (1), 63-75.

Clark, Margaret S. (1982), "A Role for Arousal in the Link between Feeling States, Judgments, and Behavior," in Affect and Cognition: The Seventeenth Annual Carnegie Symposium on Cognition, eds. M.S. Clark and S.T. Fiske, Hillsdale, NJ: Lawrence Erlbaum Associates, 263-290.

Clark, Margaret S. and Alice M. Isen (1982), "Toward Understanding the Relationship Between Feeling States and Social Behavior," in Cognitive Social Psychology, eds., A.H. Hastorf and A.M. Isen, New York, NY: Elsevier North Holland, Inc., 73-108.

Cohen, Joel B. and Charles S. Areni (1991), "Affect and Consumer Behavior," in Handbook of Consumer Behavior, eds. T.S. Robertson and H.H. Kassarjian, Englewood Cliffs, NH: Prentice Hall.

Edell, Julie A. and Marian Chapman Burke (1987), "The Power of Feelings in Understanding Advertising Effects," Journal of Consumer Research, 14 (December), 421-433.

Eysenck, Michael W. (1982), Attention and Arousal: Cognition and Performance, New York, NY: Springer-Verlag.

Gardner, Meryl Paula (1985) "Mood States and Consumer Behavior: A Critical Review," Journal of Consumer Research, 12 (December), 281-300.

Gardner, Meryl Paula and Frederick O. Wilhelm, Jr. (1987) "Consumer Responses to Ads with Positive vs. Negative Appeals: Some Mediating Effects of Context-Induced Mood and Congruency Between Context and Ad," in Current Issues and Research in Advertising, eds. J.H.Leigh and C.R. Martin, JR. Ann Arbor, MI: University of Michigan Press.

Goldberg, Marvin E. and Gerald J. Gorn (1987), "Happy and Sad TV Programs: How They Affect Reactions to Commercials," Journal of Consumer Research, 14 (December), 387-403.

Gouaux, Charles (1971), "Induced Affective States and Interpersonal Attraction," Journal of Personality and Social Psychology, 20 (1), 37-43.

Gouaux, Charles and Karen Summers (1973), "Interpersonal Attraction as a Function of Affective State and Affective Change," Journal of Research in Personality, 7, 254-260.

Hasher, Lynn and Rose T. Zacks (1979), "Automatic and Effortful Processes in Memory," Journal of Experimental Psychology: General, 108 (3), 356-388.

Hoffman, Martin L. (1986), "Affect, Cognition, and Motivation," in Handbook of Motivation and Cognition, eds. R.M. Sorrentino and E.T. Higgins, New York, NY: The Guilford Press, 244-280.

Isen, Alice M. (1984), "The Influence of Positive Affect on Decision Making and Cognitive Organization," in Advances in Consumer Research, Vol 11, ed., T. Kinnear, Provo, UT: Association for Consumer Research, 534-537.

Isen, Alice M. (1970), "Success, Failure, Attention, and Reactions to Others: The Warm Glow of Success," Journal of Personality and Social Psychology, 15, 294-301.

Isen Alice M., Kimberly A. Daubman, and Joyce M. Gorgoglione (1987) "The Influence of Positive Affect on Cognitive Organization," in Aptitude, Learning and Instruction: Conative and Affective Processes, eds., R.E. Snow and M.J. Farr, Hillsdale, NJ: Erlbaum.

Isen, Alice M., Thomas E. Shalker, Margaret Clark, and Lynn Karp (1978), "Affect, Accessibility of Material in Memory, and Behavior: A Cognitive Loop?" Journal of Personality and Social Psychology, 36 (1), 1-12.

Kennedy, John R. (1971), "How Program Environment Affects TV Commercials," Journal of Advertising, 11 (1), 33-38.

Mamo, Haim (1990), "Emotional States and Decision Making,' in Advances in Consumer Research, Vol 17, eds. M.E. Goldberg, G. Gorn, and R.W. Pollay, Provo, UT: Association for Consumer Research, 577-584.

Mattes, John and Joanne Cantor, (1982), "Enhancing Responses to Television Advertisements via the Transfer of Residual Arousal from Prior Programming," Journal of Broadcasting, 26 (2), 553-566.

Mehrabian, Albert and James A. Russell (1973), "A Measure of Arousal Seeking Tendency," Environment and Behavior, 5 (3), 315-333.

Park, C.W. and S. Mark Young (1986), "Consumer Response to Television Commercials: The Impact of Involvement and Background Music on Brand Attitude Formation," Journal of Marketing Research, 23 (February), 11-24.

Pavelchak, Mark A., John H. Antil, and James M. Munch (1988), " The Super Bowl: An Investigation into the Relationship Among Program Context, Emotional Experience, and Ad Recall," Journal of Consumer Research, 15 (December), 360-367.

Sanbonmatsu, David M. and Frank R. Kardes (1988), "The Effects of Physiological Arousal on Information Processing and Persuasion," Journal of Consumer Research, 15 (December), 379-385.

Schultz, D. E. (1979), "Media Research Users Want," Journal of Advertising Research, 19, 13-17.

Schwerin, Horace A. (1958), "Do Today's Programs Provide the Wrong Commercial Climate?" Television Magazine, 15 (8), 90-91.

Soldow, Gary F. and Victor Principe (1981), "Response to Commercials as a Function of Program Context," Journal of Advertising Research, 21 (2), 59-65.

Television Audience Assessment, Inc. (1984), "Commercial Effectiveness and Viewers' Involvement with Television Programs: A Literature Review," Cambridge, MA: Television Audience Assessment, Inc.

Thorson, Esther and Heiko Oberman (1985), "Program Involvement and the Processing of Television Commercials," Paper presented at the Association for Education in Journalism and Mass Communication Annual Conference, Memphis, August.

Thorson, Esther and Bryon Reeves (1986), "Effects Over-Time Measures of Viewer Liking and Activity During Programs and Commercials on Memory for Commercials," in Advances in Consumer Research, Vol 13, ed. R. Lutz, Provo, UT: Association for Consumer Research.

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