The Effects of Mood on Retrieving Consumer Product Information

ABSTRACT - The effects of mood at retrieval on memory and judgment were investigated in an experiment in which college students were presented with reports about consumer products. It was found that mood was associated with better recall for mood-consistent than for mood-inconsistent information, but that this effect was not observed in recognition memory. Also, happy mood-induced subjects gave higher desirability ratings than did and subjects. These mood effects were not affected by whether retrieval took place immediately after presentation or two days later. Implications of mood effects for the consumer are discussed.

Citation:

Robert Lawson (1985) ,"The Effects of Mood on Retrieving Consumer Product Information", in NA - Advances in Consumer Research Volume 12, eds. Elizabeth C. Hirschman and Moris B. Holbrook, Provo, UT : Association for Consumer Research, Pages: 399-403.

Advances in Consumer Research Volume 12, 1985      Pages 399-403

THE EFFECTS OF MOOD ON RETRIEVING CONSUMER PRODUCT INFORMATION

Robert Lawson, Alfred University

[The author wishes to thank Robert Knowles and Christine Dekleva for the assistance provided in this study.]

ABSTRACT -

The effects of mood at retrieval on memory and judgment were investigated in an experiment in which college students were presented with reports about consumer products. It was found that mood was associated with better recall for mood-consistent than for mood-inconsistent information, but that this effect was not observed in recognition memory. Also, happy mood-induced subjects gave higher desirability ratings than did and subjects. These mood effects were not affected by whether retrieval took place immediately after presentation or two days later. Implications of mood effects for the consumer are discussed.

INTRODUCTION

The nature of the psychological interplay between emotion and cognition has been a popular topic in recent years. One way in which this issue is being explored is in studies which have looked at people's memory and judgment performance while they have been operating under the influence of a particular, experimentally-induced mood. From these studies, which are reviewed below, it is safe to conclude that mood can indeed affect cognitive performance. The importance of understanding the relationship between mood and cognition in consumer psychology is self-evident. Two general questions that are addressed in this paper are: (l) Does the nature of consumers' moods influence the type of information they remember about products? (2) How will their overall attitude toward the product be influenced by mood?

The Effects of Mood at the Time of Encoding

Studies in personality, cognitive and social psychology have analyzed the effects of mood according to whether the mood is manipulated at encoding or at retrieval. Experiments by Bower and his colleagues (Bower 1981; Bower, Gilligan and Monteiro 1981; Bower and Cohen 1982) have demonstrated the mood-congruency effect, in which the mood state at encoding has a differential effect on the type of material that is remembered best. For example, Bower, et. al. (1981) used post-hypnotic suggestion to induce either a happy or and mood in subjects who read a brief story about two characters, one of which was happy and the other sad. The next day, when they were in a neutral mood, the and subjects mostly recalled facts about the and character, and the happy subjects mostly recalled information about the happy character. Srull (1983a; Experiments 2 and 3) found a mood-congruency effect on judgment. Mood was induced by a procedure of having subjects recall the details of either a happy or and life event. They then read either a positive ad or a negative ad. It was found that desirability ratings were higher when subjects were put in a happy mood at encoding.

In addition to mood congruency effects, which focus on the compatibility between one's mood state and the nature of the to-be-learned material, there is some evidence which demonstrates a mood-valency effect. In particular, this effect occurs when learning and memory that is associated with a positive mood is better than that associated with a negative mood. For example, it is well-known that clinically depressed patients have poor memories, but this fact alone says little about the relationship between temporary mood state and memory. Leight and Ellis (1981) induced mood at encoding by having their subjects read a series of self-referent statements, and then gave them a learning task which required active organizational processing for maximum performance. They found that the depressed mood group performed more poorly during training than did either neutral or elated subjects, and that the encoding-depressed subjects did not benefit from being put in a neutral or elated moot at transfer two days later. Leight and Ellis discuss their findings in terms of the Lack of cognitive effort that is associated with depressed moods.

The Effects of Mood at the Time of Retrieval

A number of studies have explored the effects of induced mood at the time of retrieval. One notion is that of the "cognitive loop" hypothesis (Ibsen, Shalker, Clark and Karp 1978), in which a person's mood acts as a cue for retrieving information from memory that is consistent with the mood state. Isen, et. al. induced good moods by having subjects win a computer game, and, in another experiment, by giving subjects a free gift. They found that the positive mood-induced subjects were better able to recall positive than negative information and gave better evaluations of products they owned. Teasdale and Fogarty (1979) found that the reaction time to retrieve pleasant memories was longer for depressed subjects than for happy subjects, while there was no difference between the mood conditions in the speed of retrieval of unpleasant memories.

Other studies focusing on mood at retrieval have found effects on various judgment ratings. Isen and Shalker (1982) had subjects rate the pleasantness of slides of local scenes, and found that the better the mood the higher the rating. Bower and Cohen (1982) reported studies in which subjects in a good mood gave higher estimated probabilities of future positive events, and observed more positive behaviors of themselves and their interview partners than did subjects in a bad mood. Together, these studies support the idea that a person's mood can act as a retrieval cue which primes mood-consistent information in memory and influences the person to interpret events in a mood consistent manner .

State-Dependent Memory

If mood at retrieval supplies a cue for remembering information, and if it is assumed that mood is stored during initial learning, then it is reasonable to suppose that memory will be better for information when the moods at encoding and at retrieval match, rather than mismatch. This is the state-dependent memory effect, and there have been several studies demonstrating it. Bower, Monteiro and Gilligan (1978; Experiment 3) found it in a retroactive interference paradigm. Bartlett, Burleson and Santrock (1982) induced mood in kindergartners and third-graders with the result of finding the state-dependent effect. Leight and Ellis (1981) found the effect in a recognition memory test and Clark, Milberg and Ross (1983) observed it in a recall paradigm.

Conflicting Evidence and Boundary Conditions

Although considerable evidence has accumulated supporting the various mood effects described above, there have been some notable and interesting failures. In their study with children, Bartlett, et. al. (1982; Study 1) were not able to find the state-dependent effect when a relaxation technique was used prior to mood induction. Clark, et. al. (1983) found enhancement of state-dependency when subjects were highly aroused in addition to being in a particular mood, and suggested that lack of arousal may have been a reason why Bartlett, et. al. failed to find the effect. Bower, Monteiro and Gilligan (1978; Experiments 1 and 2) found no state-dependent effects in a recall paradigm with their hypnotised subjects. Eich (1980) and Bower and Cohen (1982) failed to find the effect in recognition memory studies. Taken as a whole the evidence suggests that the effectiveness of mood state as an aid to memory depends on the extent to which no other cues are available, as well as on the intensity and arousal-producing nature of the mood. All other factors being equal, mood at retrieval should be more effective in a difficult free recall situation than in a recognition memory task, because mood may be the only salient cue available in the former situation..

One previous study found results that were directly opposed to these mood retrieval effects. Srull (1983a), in his study on positive and negative information in ads, found that being in a happy mood at retrieval was associated with a tendency to have accessible in memory more negative information and a tendency to rate products less desirable, whereas being in a sat moot was associated with the opposite tendencies. These findings are difficult to reconcile with the rest of the literature. Srull attempts to explain his results by citing Watkins' (e.g., 1979) cue overload theory, which implies that a particular mood will cue a multitude of life events associated with that mood, and thereby overwhelm the target information, thus favoring the recall of information that is different from that implied by the mood.

Because the results of Srull's study are at odds with the literature concerning the effects of moot at retrieval, the present experiment was designed to further explore the issue of mood effects in the context of consumer information processing. The specific issues addressed were: (l) whether mood induced at retrieval facilitated or depressed memory for mood-congruent information, (2) whether increasing the retention interval resulted in enhancing mood effects, (3) whether recognition and recall tests gave similar patterns of results, and (4) what effects mood would have on product desirability ratings.

METHOD

Subjects and Design

Subjects were 44 college students recruited from the introductory psychology course. They participated in return for course credit. The subjects were randomly assigned, with the restriction that there be 11 subjects per group, to one of four conditions in a 2 x 2 factorial design. The two factors were Mood (Sat and Happy) and Retention Interval (Immediate and Delayed).

Materials

Subjects were presented with six brief objective reports, each of which described eight attributes of a product. The product brands chosen were names of actual products, but were judged to be ones with which the average college student would have little experience. The product categories were children's bookpacks, electric bug killers, electric citrus juicers, interior wall paint, gas barbecue grills, and bathroom scales. The product attributes were taken from Consumer Report articles and were integrated into the reports. Two of the reports presented two negative and six positive attributes (Favorable), two presented an equal number of negative and positive qualities (Balanced), and two presented six negative and two positive attributes (Unfavorable). Thus, across the six reports, 48 target pieces of information were presented, half of them negative and half of them positive.

There were three kinds of testing material: (1) a 20-point product desirability scale, (2) a recall test, in which the names of the six products were typed on paper as cues for written recall, and (3) a 48-item recognition memory test, which was structured so that half the items were "old" and the new items were negations of the other 24 pieces of information. There were an equal number of positive and negative attributes in both "old" and "new" categories. Subjects were required to rate their confidence in the correctness of their responses on a six-point scale.

Procedure

All subjects went through the following sequence of events. First, they were instructed to process the reports for the purpose of comprehending them. Then they listened to a tape recording of the reports while they read along, after which they reread them silently for an additional three minutes. Following that, they evaluated their mood during presentation by filling out a mood adjective checklist, and they completed the desirability ratings. The mood adjective checklist, which consisted of 21 negative moot adjectives and 11 positive mood adjectives, was a modified version of Lubin's (1965) List A. At this point, subjects in the Delay groups were dismissed and returned approximately 48 hours later, while the other subjects continued in the same session.

Mood was then induced by having subjects think of the details of either a happy or and life event for a duration of six minutes. Subjects were reminded to continue remembering details two or three times during that interval. This mood induction procedure was intended to duplicate that of Srull (1983a). Following mood induction, the subjects were immediately read some instructions, and then given the desirability rating task (for the second time) and the recall task. This was followed by more instructions, then the recognition memory task, and finally the mood adjective checklist (for the second time) with the instructions that it should apply to the time of the memory tests. Before they were dismissed, the subjects were debriefed and the experimenter assured himself that the "sad" subjects were not inordinatelY depressed.

RESULTS

Mood Manipulation Checks

The adjective checklists were scored by giving one point for every negative mood adjective checked and for every positive mood adjective not checked. Thus, the higher the score, the worse the mood. The maximum score was 32, and a score of 11 indicated a neutral mood. the checklist revealed that there were no systematic differences among the four groups before mood induction; they all seemed to be in a moderately good mood with group mean scores ranging from 8.18 to 9.27. Following mood induction, the Immediate-Happy (IH) and Delayed-Happy (DH) groups scored 6.73 and 5.27, respectively, while the Immediate-Sad (IS) and Delayed-Sad (DS) groups had scores of 15.64 and 15.55, respectively. According to this measure, the procedure was successful in bringing about the intended error.

Memory Data

The recall task was scored on a generous, meaning-based criterion according to the number of target attributes recalled. The mean number recalled (out of 24) of the positive and negative attributes for each of the four groups is given in Table 1. Overall, immediate recall is better than delayed recall (F(1,40) = 19.86, p < .01), and subjects recall more negative than positive information (F(1,40) = 15.56, p < .01). Most importantly, a mood retrieval effect was revealed in the mood x information type interaction (F(1,40) = 14.57, p < .01). Here the and subjects recalled more negative than positive information, while the happy subjects recalled about the same levels of negative and positive items. Retention interval, however, did not significantly interact with the type of information presented (F(1,40) = 3.64). This reflected a tendency for the mood-congruency effect to occur at both delay intervals.

TABLE 1

MEAN NUMBER OF TARGET ITEMS RECALLED

When the recall data were organized according to report type, the mean number of items recalled (out of 16) was 8.9 for the Unfavorable reports, 7.4 for the Balanced reports and 7.2 for the Favorable reports. This pattern was not affected by moot state. In other words, there was no tendency for happy subjects to recall more information from Favorable reports and for sat subjects to recall more information from Unfavorable reports.

Compared to the recall data the recognition memory data are less clear-cut. Table 2 shows the mean number of sentences correctly classified as True or False for each of the four groups. The only significant effect is that of Retention Interval, with performance being better for the Immediate groups than for the Delayed groups (F(1,40) - 4.15, p < .05). The tendency for the Happy groups to recognize more positive than negative information relative to the and groups was not reliable (F(1,40) - 2.58). Thus, an informal comparison of subjects' performance on the recall and recognition tasks suggests that the recall task is more sensitive to showing the mood retrieval effect than is the recognition test.

TABLE 2

MEAN NUMBER OF SENTENCES CORRECTLY RECOGNIZED

Following the data analysis reported by Srull (1983a), in zn effort to obtain an indication of memory accessibility, retrieval ratios were calculated. Essentially, this measure shows the relative frequency with which an attribute is recalled, given that it is recognized with high confidence. Thus, the higher the retrieval ratio, the greater the accessibility of the information. Retrieval ratios were calculated for positive and negative attributes for each of the four groups, and are presented in Table 3. Only true recognition items which were correctly indicated with confidence ratings of 4, 5 or 6 were included. The result of interest is the interaction between information type and mood. Higher retrieval ratios are associated with congruent as opposed to incongruent mood states. For example, when averaged over retrieval interval, a and mood state is more likely to lead to recall of negative than positive information, given that that information was already available in memory. This result parallels the finding based on the number correctly recalled data.

TABLE 3

RETRIEVAL RATIOS FOR INFORMATION CONFIDENTLY RECOGNIZED

Desirability Ratings

Subjects gave product desirability ratings twice during the experiment, once before they were mood-induced and once after. Overall, they rated the products in a manner consistent with the numbers of positive and negative attributes described. The mean ratings were 14.89 for the Favorable reports, 10.43 for the Balanced reports, and 4.55 for the Unfavorable reports. When summed across all six product reports the mean desirability ratings before mood induction ranged from 10.20 to 10.76 for the four groups, indicating that they did not give systematically different ratings at that time. However, after mood induction, the mean ratings were IH = 10.74, DH - 10.45, IS = 8.58, and DS = 8.14 (F(1,40) - 8.50, p < .01). This indicates that mood at retrieval did color subjects' overall evaluations of the products in the direction that is consistent with their mood. This is especially true with the saddened subjects, and this effect did not depend on the retention interval.

DISCUSSION

The results of this experiment are generally consistent with the well-established effects of mood state on judgement and memory, and represent a demonstration that mood can be studied in a consumer behavior situation. The major finding that mood at retrieval is associated with congruent effects of recall and evaluation is consistent with Isen's et. al. (1978) cognitive loop hypothesis, which contends that a person's mood at retrieval can prime mood-congruent information in memory, thus leading to higher levels of recall and accessibility of such information, and to producing mood-congruent evaluations. People in and moods at the time of retrieval tend to recall relatively more negative information and produce relatively lower desirability ratings than do people in happy moods.

It is interesting to note that the failure to find moot retrieval effects with recognition memory is consistent with Bower's (1981) associative network theory. He states that such mood effects should occur most clearly in situations in which there are no other available cues to use in searching for the target information. Therefore one should not expect mood effects to surface in a recognition task because the test item itself functions as a powerful cue. However, according to the same reasoning, we should have observed an intensification of the mood effect with the Delay groups because memory fades over time. Instead we found no significant effect of delay interval on any dependent measure. Thus it does not appear that mood functions as a more salient retrieval cue as the retention interval increases;

The results of the present experiment appear to be in conflict with Srull's (1983a, Experiments 3 and 4) findings of retrieval mood-incongruent effects for evaluative judgments and for accessibility as measured by retrieval ratios. The reason for the discrepancies in the results between these two studies is not obvious but probably lies somewhere in the differences in the procedural details. For example, Srull induced mood both at encoding and at retrieval, did not employ a mood verification procedure, and presented different product information. Because of these and other differences, the two studies are not directly comparable. Nonetheless the conflicting findings indicate that extreme caution is called for in arriving at general conclusions concerning the effects of mood at retrieval.

Another source of caution about interpreting the results of this study is shared by all studies involving experimentally induced mood states. It lies in the double danger of, first of all, failing to create the target mood in the subject, and, secondly, of inadvertently creating a demand artifact. The present procedure, that of having subjects recall the details of an emotionally charged event from their past, has apparently been successful in eliciting the emotional state (Srull, 1983a, 1983b, 1984). This effectiveness was verified in the present study by use of the mood adjective checklist. However, regardless of the moods actually created, the inherent demand characteristic possibility of the situation remains. Subjects realize that they are expected to achieve a certain mood state. This observation applies not only to the present procedure, but to other methods of reading self-referent statements and of hypnotism that have been used as well. (See Gardner and Vandersteel, 1984, for a methodological discussion.) Whether or not this possibility of a demand artifact actually occurs is another question whose answer depends on the complexities of the situation.

In the present study, it is doubtful that demand characteristics influenced the memory results for the following reasons: (1) There was no mood effect on recognition, whereas there was one on recall. (2) The mood-congruency effect on the recall data was observed on two fairly independent measures; number of attributes recalled and retrieval ratios. (3) The retrieval ratio itself is a very obscure measure from the subject's viewpoint. (4) Debriefing sessions with the subjects revealed that they were unaware of the purpose of the study. However, the desirability ratings, which were transparent and did not call for maximum performance, may indeed have been artifactual. Neither the procedure used here nor any of the procedures reported in the literature are guaranteed to be immune from the demand artifact. This methodological problem will continue to be a major consideration in any study which manipulates mood.

While the literature has focused on the effects of mood on judgment and memory, there is little information which bears on the practical importance of this variable as it relates to actual consumer behavior. For example, should marketers be concerned with consumers' moods at encoding and at retrieval, and if so, how can they influence mood? There is little to indicate, in this study or in others, that mood is a crucial variable. In general, the effects, when they occur, are modest. If a consumer is in a happy mood at encoding, that is no guarantee that he or she will eventually recall positive as opposed to negative information about the product, although the literature suggests a trend in that direction. As for the consume - ' s mood at the time of retrieval, the effects are even more subtle, insofar as the literature presents conflicting results.

Perhaps a more fruitful way to approach mood effects is to look at the specific situations in which mood might be an important variable. For example, Srull (1983b) found that mood at encoding had congruent effects on product judgments with subjects who were not familiar with the product, but no effect with high familiarity subjects. In another study, Srull (1984) found congruent mood-at-encoding effects on judgment only when subjects were instructed to form an impression of the product as the information was being presented, as opposed to being asked for that judgment afterwards, at the time of retrieval. In focusing on mood at encoding, Srull (1984) concludes that mood state will be incorporated into the judgment only when a person's opinion is not yet formed; such as when the consumer is confronted with information about a product of which he or she knows little, and has the specific goal in mind of forming an impression while the information is being presented. The mood-congruent retrieval effects found in the present study can be understood in this framework if it is assumed that the judgment is not made until retrieval, at which point the mood not only colors the type of attribute that will be recalled but also is incorporated into the product evaluation.

REFERENCES

Bartlett, James C., Georgia Burleson and John W. Santrock (1982), "Emotional Mood and Memory in Young Children," Journal of Experimental Child Psychology, 34, 59-76.

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

Bower, Gordon, H. and Paul R. Cohen (1982), "Emotional Influences in Memory and Thinking: Data and Theory," in Affect and Cognition, eds. Margaret S. Clark and Susan T. Fiske, Hillsdale, NJ: Laurence Erlbaum Associates.

Bower, Gordon, H., Stephen G. Gilligan and Kenneth P. Monteiro (1981), "Selectivity of Learning Caused by Affective States," Journal of Experimental Psychology: General, 110, 451-473.

Bower, Gordon, H., Kenneth P. Monteiro and Stephen G. Gilligan (1978), "Emotional Mood as a Contest for Learning and Recall," Journal of Verbal Learning and Verbal Behavior, 17, 573-585.

Clark, Margaret S., Sandra Milberg and John Ross (1983), "Arousal Cues Arousal-Related Material in Memory: Implications for Understanding Effects of Mood on Memory," Journal of Verbal Learning and Verbal Behavior, 22, 633-649.

Eich, J. Eric (1980), "The Cue-Dependent Nature of State-Dependent Retrieval, Memory and Cognition, 8, 157-173.

Gardner, Meryl P. and Marion Vandersteel (1984), "The Consumers Mood: An Important Situational Variable," in Advances in Consumer Research, Vol. 11, ed. Thomas C. Kinnear. Ann Arbor: Association for Consumer Research.

Isen, Alice M. and Thomas E. Shalker (1982), "The Effect of Feeling State on Evaluation of Positive, Neutral, and Negative Stimuli: When You Accentuate the Positive, 'Do You 'Eliminate the Negative'? Social Psychology Quarterly, 45, 58-63.

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, 1-12.

Leight, Kenneth A. and Henry C. Ellis (1981), "Emotional Mood States, Strategies and State-Dependency in Memory, Journal of Verbal Learning and Verbal Behavior, 20, 251-266.

Lubin, Bernard (1965), "Adjective Checklists for Measurement of Depression," Archives of General Psychiatry, 12, 57-62.

Srull, Thomas K. (1983a), "Affect and Memory: The Impact of Affective Reactions in Advertising on the Representation of Product Information in Memory," in Advances in Consumer Research, Vol. 10, eds. Richard P. Bagozzi and Alice M. Tybout, Ann Arbor: Association for Consumer Research

Srull, Thomas K. (1983b), "The Role of Prior Knowledge in the Acquisition, Retention, and Use of New Information," in Advances in Consumer Research, Vol. 10, eds. Richard P. Bagozzi and Alice M. Tybout, Ann Arbor: Association for Consumer Research.

Srull, Thomas K. (1984), "The Effects of Subjective Affective States on Memory and Judgment," in Advances in Consumer Research, Vol. 11, ed. Thomas C. Kinnear, Ann Arbor: Association for Consumer Research.

Teasdale, John D. and Sarah J. Fogarty (1979), "Differential Effects of Induced Mood on Retrieval of Pleasant and Unpleasant Events from Episodic Memory," Journal of Abnormal Psychology, 88, 248-257.

Watkins, Michael J. (1979), "Engrams as Cuegrams and Forgetting as Cue Overload: A Cueing Approach to the Structure of Memory," in Memory Organization and Structure, ed. Richard C. Puff, New York: Academic Press.

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