Gift-Wrapping Effects on Product Attitudes: A Mood-Biasing Explanation
JOURNAL OF CONSUMER PSYCHOLOGY, 1(3), 197-223 Copyright (g) 1992, Lawrence Erlbaum Associates, Inc. Gift-Wrapping Effects on Product Attitudes: A Mood-Biasing Explanation Daniel J. Howard Department of Marketing Edwin L. Cox School of Business Southern Methodist University In four experiments, I examined the effects of gift wrapping on product attitudes. Two questions were addressed. First, does gift wrapping an item have a favorable influence on attitudes toward owning what is received? Results of all four experi- ments consistently support an affirmative answer to that question. Second, what explains the attitudinal results? I argued that gift wrapping, through repeated pairing with joyous events in people's lives, has utility in cuing a happy mood which, in turn, positively biases attitudes. Results of the last three experiments support this mood biasing position by demonstrating that a happy mood consist- ently mediates gift-wrapping effects on attitudes. The results are consistent with an encoding specificity view of mood retrieval and a mood maintenance explana- tion of attitude formation. The encoding specificity view was supported by find- ing stronger effects of gift wrapping on mood retrieval in conditions arguably present when the relation between gift wrapping and happy mood was estab- lished in the lives of subjects, such as the receipt of a personal gift (Experiment 2), the receipt of a gift wrapped in traditional gift-wrapping paper (Experiment 3), and the receipt of a gift-wrapped present on subjects' birthdays (Experiment 4). The mood maintenance process was supported by finding parallel effects of gift wrapping on mood and attitude and by finding the mediational effects of happy mood on attitude strengthened as subjects felt happier. These results are consistent with the premise that the happier the mood, the more subjects sought to maintain that state through the development of favorable attitudes toward owning the gift they received. Salesperson: This is very nice. I 'm sure she'll be happy with it. Take it to the back counter and we'll wrap it up for you. Requests for reprints should br sent to Daniel J. Howard, Department of Marketing, Ed- win L. Cox School of Business, Southern Methodist University, Dallas, TX 75275. "] 98 HOWARD Customer: Salesperson: Customer: Salesperson: Oh, there's no need for that. Sure there is. You should have it wrapped. Why? Because she'll like it more. This conversation is a paraphrased version of a dialogue I happened to over- hear a couple of years ago in a gift shop. It made an impression because of its suggestiveness about the nature of human reactions to a common cultural phenomenon: gift wrapping. Since then, I have conducted a series of informal discussions with people on this issue. Most people concede that when they receive a gift, with the exception of things like flowers and some food items, they prefer to have it wrapped. Equally clear, however, is the opinion that receiving a gift that is wrapped has no influence on how people feel about owning it. When asked why they prefer to have gifts wrapped, many people simply reply, "gifts are supposed to be wrapped." This suggests that perhaps people may simply like the idea of a present that is wrapped, independent of their feelings for what is inside, because gift wrapping is consistent with cultural norms and prior expectations. On the other hand, consistent with the gist of the argument offered by the salesperson, is it possible that gift wrapping may indeed have a favorable influence on evaluations of the object itself?. Further, if such an effect can be demonstrated, what theoretical position best explains the outcome? This investigation reports the results of four experi- ments designed to address these questions. The gift-giving process has been examined across a variety of disciplines, encompassing sociological (Gouldner, 1960; Neisser, 1973), anthropological (Belshaw, 1965; Mauss, 1954), economic (Camerer, 1988; Kerton, 1971), psy- chological (Jones, 1964; Schwartz, 1967), and consumer behavior perspectives (Belk, 1976, 1979). However, consideration of gift wrapping as a variable in the gift exchange process has been ignored to date. This is unfortunate given the extensive nature of the practice, especially in Western cultures. Last year Americans spent an estimated $1.3 billion on gift-wrapping paper, bows, and ribbons (Manufacturing USA, 1989). Cialdini (1980) suggested that much can be learned by observing the practices of those who make a living engineering and implementing influence techniques in the marketplace. Through such observations, means of testing those techniques in a controlled manner can be developed, which will allow us a better understanding of the practices. This series of experiments represents an extension of that view: I tested a commonly accepted cultural practice to see whether that practice is persuasive and what accounts for its effectiveness. These experiments should help us understand the dynamics underlying an aspect of our common cultural heritage as gift givers, gift receivers, and consumers. The first experiment was designed to determine whether wrapping a gift has an influence on attitude toward owning the item received. GIFT-WRAPPING EFFECTS 199 EXPERIMENT 1 Method Procedure and materials. Subjects were informed that a national, spe- cialty merchandising firm was interested in obtaining evaluations of a variety of new and existing products for use in developing market strategies for the coming year. Subjects were told that the sponsoring firm will award them with a gift upon completion of their tasks to thank them for their time. Each subject was individually tested. Subjects were taken to a room contain- ing four different products (a lensatic compass, a giant snow/ice remover, chimes, and a magnifying glass), each in an unsealed box and accompanied by a brief product description. On top of each box was a manila envelope contain- ing a questionnaire. Subjects were told there would be fifth product for them to evaluate, but the sponsor was late in delivering it. However, it was expected that the fifth product would arrive by the time they were through with the first four. First, subjects were told to examine each product closely and then answer each respective questionnaire. They were allowed to reexamine the products if they wished when answering the questions and were told their responses will remain anonymous. The first question on each questionnaire asked subjects their attitude toward owning each product. (This question was asked to be consistent with the target object, as discussed later.) Each questionnaire con- tinued with a series of open-ended questions, particular to each product, designed to support the cover story. Subjects were told to contact the experi- menter, who was in an office at the end of the hall, when they finished. This part of the experiment took approximately 60 rain to complete. As subjects arrived at the experimenter's office, they were informed that the fifth product still had not arrived. Then subjects were seated as the experi- menter explained that he had to call a contact person in the sponsoring firm to determine the problem. Then the experimenter made a call to an empty office with subjects far enough away to prevent them from detecting that the experimenter was actually holding a conversation with himself. Throughout the conversation, the experimenter was busily writing down information. This action, along with the gist of the experimenter's statements and replies on the phone, was designed to be consistent with the explanation subsequently pro- vided to subjects. After the call, the experimenter explained that the contact person, who was a researcher for the firm, was still waiting for one of the product managers to bring the fifth product and questionnaire to him and so he was unable to deliver it. However, to avoid losing valuable consumer input, he asked that subjects evaluate the product being used as a gift, because this product was also manu- factured by the firm. Then the experimenter asked subjects if they would mind 200 HOWARD spending the added time, and no one objected. Next, he took a questionnaire for one of the previous products and modified the questions (by hand) to correspond to the gift being awarded. For the critical attitude measure, this modification involved changing only the product name in the question. Throughout this process, the experimenter explained to subjects what he was doing while he was doing it. The modified questionnaire was then placed in a manila envelope. On the basis of pretesting, this experimental procedure was found to be necessary to alleviate subject suspicions about receiving a gift and then being asked to evaluate it. The procedure, therefore, was developed to make it appear as if the evaluation of the target product was unplanned. The independent variable manipulation was then introduced. In the wrapped condition, the experimenter went to a drawer and retrieved a box in blue and white gift-wrapping paper with a matching bow and ribbon.~ The box contained a sheepskin seat cover for a bike, which was in a plastic bag provided by the manufacturer.' A brief description (professionally typeset and com- prised of actual promotional copy used by the firm) accompanied the gift and was placed in the plastic bag: Sheepskin bike seat cover is comfortable and cool. On a long bike ride, every additional mile makes that seat feel more and more like a rock! I f you had one of these soft, shearing sheepskin covers on that seat, you'd be off and away. . , in total comfort. The cover does not retain either heat or cold and does not feel damp from perspiration. A drawstring holds the cover in place. May be dry-cleaned or washed by hand in warm, sudsy water. In an envelope, a small white card containing the words thank you was also placed in the plastic bag. In the unwrapped gift condition, the seat cover was placed in the manufacturer's plastic bag along with the product description and the thank you card. The box, wrapping paper, bow, and ribbon were not used. The experimenter led subjects back to the room where the other product evaluations were completed. Subjects were reminded their responses would remain anonymous. Upon completion of the seat cover evaluations, subjects were debriefed and asked not to discuss the study with anyone. Measured variables. Attitude toward owning the product served as the critical dependent variable. Subjects were asked to complete the statement ~The wrapping was selected on the basis of pretesting, which indicated that it was perceived as gender neutral in appearance. 2"I'he seat cover was selected on the basis of pretesting. The goal was to select a product that was favorably perceived as a gift but did not run the danger of ceiling effects on the attitudinal measures attenuating the results. GIFT-WRAPPING EFFECTS 201 "owning a sheepskin bike seat cover is" by using three 9-point scales, ranging from undesirable (1) to desirable (9), bad (1) to good (9), and foolish (1) to wise (9). These items measured a single construct reliably (a = 0.87) and were summed and averaged to yield an attitude index. Subjects and design. Subjects were 45 university students who par- ticipated in exchange for course credit. They were randomly assigned to the two experimental conditions (gift wrapping: wrapped vs. unwrapped). Thus, a one-factor, between-subjects design was implemented. Results and Discussion One-way analysis of variance (ANOVA) results for gift wrapping were signifi- cant. Subjects receiving a gift-wrapped item had a more favorable attitude toward owning it (M = 7.14) than those who received the unwrapped item (M = 6.06), F(I, 43) = 7.00, p < .01. These findings clearly support the conclusion that wrapping a gift has a significant influence on how people feel about owning it. The next question to be addressed is why? The theoretical position examined is that gift wrapping can serve as a cue for the retrieval of positive affect, specifically a happy mood, associated with the receipt of gifts. Once elicited, a happy mood results in a more favorable attitude toward owning the product received because of a desire for mood maintenance or protection (e.g., Isen & Levin, 1972; Isen, Means, Patrick, & Nowicki, 1982; Isen & Simmonds, 1978). The possibility of gift wrapping serving as a retrieval cue for positive affect is consistent with the view proposed by Bower (1981). Bower suggested that mood can be represented as a node, or unit, within an associative network model of memory. Within such a network, a particular mood may become linked or associated with events that occur during one's life when the mood was experienced. Given these relations, when someone is in a happy mood they may be more likely to retrieve and become consciously aware of events that occurred at a previous time when the happy mood was experienced. Bower (1981) referred to this as mood-state dependent memory. However, a reversal of this effect should also be plausible: taking a particular event associated with a mood--such as the receipt of a wrapped giftmand using it as a cue to retrieve the mood per se. Stayman and Batra (1991) recently provided support for this process of cued mood retrieval. To explore the plausibility of this hypothesis in a gift receipt context, a survey of 123 respondents (from the same population as the subjects used in this series of studies) was conducted. The survey revealed a variety of joyous events on which they typically receive gifts. The vast majority of respondents stated that most or all of their gifts are usually wrapped on those gift-giving occasions. For example, 96% of respondents stated that they usually receive 202 HOWARD most or all of their gifts wrapped on birthdays and Christmas, and the same percentage reported that this was true throughout their childhood years. These findings, although not surprising, are congruent with the suggestion that the gift wrapping of presents is associated with occasions on which happy moods were likely experienced. Given that gift wrapping has consistently been paired with joyous occasions, gift wrapping should have the potential to be effective as a retrieval cue for happy moods. Clearly, the receipt of a product as a gift, in itself, should stimulate a happy mood. However, the mood should be hap- pier when the product is gift wrapped because of its consistent association with joyous occasions in the life of the recipient. These ideas were examined in Experiment 2. EXPERIMENT 2 I did not expect that gift wrapping is a strong enough cue to retrieve mood under any and all circumstances. Gift wrapping may be too weak to cue activation of a mood because it is likely to be associated with many other things as well, resulting in cue overload (Watkins, 1979). Tulving (1983; also see Tulving & Thompson, 1973) discussed the factors that operate to determine the effectiveness of a retrieval cue. The theory of encoding specificity suggests that the better the match, that is, the greater the similarity, between the original conditions of encoding and the factors present at the time of attempted retrieval, the higher the likelihood that retrieval will be successful. This similarity can involve properties of the stimulus (e.g., Fisher & Craik, 1977; Geiselman & Glenny, 1977), recipient states (e.g., Bower, 1981; Eich et al., 1975), or the context in which the learning occurred (e.g., Godden & Baddeley, 1975). Given these arguments, the effectiveness of gift wrapping in cuing a happy mood should be greater under conditions that more closely match the factors operating when an individual previously experienced the happy mood. One obvious candidate is receiving a gift intended for oneself (labeled personal gift in Experiment 2), as opposed to a gift intended for someone else (labeled other's gift in Experiment 2). I argue that, for college-age students at least, most joyous experiences associated with wrapped girls have been receiving personal gifts. On the basis of this reasoning, it is expected that gift wrapping (wrapped vs. unwrapped) should interact with ownership (personal gift vs. other's gift) in its effects on mood. A happier mood should be found in the gift wrapped than in the unwrapped condition, but only when a personal gift is received. However, the elicitation of a happy mood is not a sufficient condition for attitudinal biasing to occur. Researchers have found that, under certain condi- tions, a happy mood sometimes results in a less favorable attitude (Worth & Mackie, 1987), and may not result in a more favorable attitude (e.g., Batra & GIFT-WRAPPING EFFECTS 203 Stayman, 1990; Mackie & Worth, 1989). Thus, a theoretical position is needed to explain why the elicitation of a happy mood should be expected to result in favorable attitudinal biasing in a gift receipt context. The rationale is the expectation that subjects will desire to maintain their happy moods once they are retrieved. There is ample evidence to suggest that when people experience positive affect, they engage in activities or behaviors designed to maintain those feelings (e.g., Isen & Levin, 1972; Isen et al., 1982; Isen & Simmonds, 1978). In this context, a mood protection rationale for attitudinal biasing is as follows. Developing unfavorable attitudes about owning something that sub- jects believe is theirs to keep would be likely to disrupt or destroy the happy mood being experienced, making such attitudes less likely. On the other hand, developing favorable attitudes about owning something that subjects believe is theirs to keep should help maintain their happy mood, making such attitudes more likely. Such a mood protection strategy is quite consistent with the view that one function of attitudes is to maximize rewards and minimize punish- ments (Katz, 1960). I further argue that the happier the mood, the stronger this mood protection tendency should be. The assumption here is that a happy mood is an intrinsically gratifying state, and strategies designed to maintain such a state should vary with the level of gratification being experienced (Berlyne & Madsen, 1973). Based on these arguments, three outcomes are predicted. First, attitude scores should run parallel to mood scores. The happier subjects feel, the greater the drive to maintain that mood, and thus they should be more likely to develop favorable product attitudes, given the compatibility between feeling happy and evaluating the ownership of one's gift as good, desirable, and wise (i.e., the attitude scale items). Thus, gift wrapping should interact with item ownership in its effects on attitude. A more favorable attitude toward owning the object should be found in the gift-wrapped condition than in the un- wrapped condition but only when a personal gift is received. Second, happy mood should statistically mediate the effects of gift wrapping on attitude scores. This outcome follows from the argument that gift-wrapping effects on product attitudes stem from the influence of happy moods on those attitudes. Mediational tests were conducted using Baron and Kenny's (1986) method (see Batra & Stayman, 1990, for a review of the use of this method in an experimental mood analysis). Baron and Kenny stated that three regression equations must be estimated to establish a mediational model and the following effects must hold: (a) a significant effect of regressing the mediator on the independent variable; (b) a significant effect of regressing the dependent varia- ble on the independent variable; and (c) when regressing the dependent varia- ble on both the mediator and the independent variable, the effect of the independent variable must be weaker than in Equation 2. If all three of these conditions hold, the mediational model is supported. Baron and Kenny noted 204 HOWARD that "the strongest demonstration of mediation" (p. 1176) occurs when the independent variable "drops out" (is reduced to nonsignificance) in Equa- tion 3. Third, mediational effects should be strongest when the mood protection tendency is operating to the fullest extent. This outcome is anticipated because if mood state is argued to mediate attitude scores because of mood protection tendencies, those mediational effects should be most apparent in a condition in which the mood protection tendency is operating relative to another condi- tion in which this tendency is not operating or operating to a lesser degree. Thus, once general mediational support is provided (i.e., a main effect of mediation), a test of differential mediational strength (i.e., a mediational inter- action effect) is evidenced by regressing the dependent variable (attitude) on the mediator (happy mood) separately within levels of the independent varia- ble (gift wrapped vs. unwrapped)? Hence, assuming evidence for the media- tional hypothesis is provided, happy mood should be most predictive of attitude in the wrapped personal gift condition. Finally, although the hypothesizing focused on the effects of gift wrapping on happy mood, it could be argued that the opposite outcomes should be found for sad mood. Thus, a wrapped personal gift should not only result in more happy feelings but fewer sad feelings. This may be expected if happy and sad moods reflected opposite poles on a single mood continuum. However, evi- dence suggests that such a possibility may not be an accurate reflection of reality (see Isen, 1984, for a review). This issue is discussed later in more detail. Sad mood was examined in the second experiment for reasons of completeness. Method Procedure and materials. The opinion booklets for the four products that the subjects evaluated prior to the target gift were modified in this experi- ment: Prior to the three attitude items, six additional scaled items were pre- sented. These items changed from product to product. For one of the products, six mood items were used: elated, calm, enthusiastic, drowsy, peppy, and nervous. This was the opinion booklet that the experimenter then modified (by hand, as described in Experiment 1) for the gift item. These six mood items were changed by the experimenter to those described in the Measured Varia- bles section. qn Experiments 2, 3, and 4, the first two equations in Baron and Kenney's (1986) tests of mediation are supported and the effect of the independent variable is reduced to nonsignificance (or marginal significance) in Equation 3. Given these results, the argument that regressing the dependent variable on the mediator within each level of the independent variable provides a test of differential mediational strength is true by definition because the main mediational model will have already been supported across those levels. GIFT-WRAPPING EFFECTS 205 The procedure was the same as in Experiment 1 except for the introduction of a second independent variable: Whether subjects thought the gift item they were evaluating was a gift for them (personal gift) or for someone else (other's gift). This involved changes in the experimenter's explanation to the subjects after his phone call to the company contact person. In the personal gift condition, the procedure was the same as in Experiment 1. In the other's gift condition, subjects were told that to avoid losing valuable consumer input, the sponsoring firm requested that participants evaluate one of the products being used as a gift, because each of those products were also marketed by the firm. The experimenter explained that this would not create a problem because there were some extra gifts left. Subjects were also told that the gift they would receive was different from the one they were being asked to evaluate. The experimenter then asked if they would mind spending the extra time, and no one objected. The rest of the procedure was basically as described in Experiment 1. 4 Measured variables. Attitude toward owning the product (a = 0.84) was again examined as the critical dependent measure. Mood was assessed with a six-item list, modified from Wood, Saltzberg, and Goldsant's (1990) study. Subjects were asked to "place an 'x' in the space that corresponds most closely to how you feel now" and presented with 9-point scales ranging from not at all (1) to extremely (9) for the following descriptions: depressed, happy, sad, hopeful, sorry, and cheerful. A happy mood scale was developed by summing and averaging the happy, hopeful, and cheerful items (a = 0.81); a sad mood scale was developed using the depressed, sad, and sorry items (a = 0.78). The correlation between the happy and sad mood scales was - .48; R 2 = 0.23. The correlation of attitude with happy mood was .46 and with sad mood was --.31. Subjects and design. Eighty-two university students were randomly as- signed to the four experimental conditions represented by the gift-wrapping and ownership variables. Thus, a 2 X 2 (Wrapped and Unwrapped )< Personal Gift and Other's Gift) between-subjects factorial design was implemented. 'There was one variation from Experiment 1. In the wrapped conditions, when the experi- menter led subjects back to the examination room and placed the product on the table he stated, " I 'm going to let you open this up by yourself since this is a scientific study and you opened boxes for all the other products. I 'm supposed to tell you that for reasons of control." These comments were made because preliminary testing revealed that some subjects in the wrapped other's gift condition wondered why the experimenter did not unwrap the product before giving it to them to examine. These comments were found to alleviate that problem without directly mentioning the gift-wrapping stimulus and without causing confusion in the wrapped personal gift condition. Subjects in the unwrapped conditions were told, " I 'm going to leave you alone now. Remember, this is a scientific study. I 'm supposed to tell you that for reasons of control." 206 HOWARD Results Moan differences. ANOVA results revealed a main effect of ownership for the attitude index, F(1, 78) = 19.85, p < .001. Those who evaluated the seat cover when they thought it was theirs to keep (the personal gift conditions) had a more favorable attitude toward owning it (M = 6.31) than those who thought the gift was for someone else (the other's gift conditions; M = 4.82). A main effect of gift wrapping also emerged, F(1, 78) = 4.95,p < .03. Subjects had a more favorable attitude toward owning the gift-wrapped seat cover (M = 5.90) than the unwrapped seat cover (M = 5.18). However, the interaction effect was not significant, F(1, 78) = 1.59, p < .21. Given the a priori expectations, the hypothesis was tested using two orthogonal F tests compar- ing the wrapped with the unwrapped treatments within each of the ownership conditions. The means are plotted in Figure 1. For the personal gift condition, attitude toward owning the gift-wrapped seat cover (M = 6.90) was signifi- cantly more favorable, F(1, 78) = 6.03, p < .03, than attitude toward owning the unwrapped item (M = 5.72). However, no significant difference was seen (F < 1) between the wrapped. (M = 4.98) and unwrapped (M = 4.65) treatments for the other's gift condition. For happy mood, a main effect of gift ownership was seen, F(I, 78) = 22.68, p < .001, with subjects feeling happier in the personal gift (M = 6.64) than in the other's gift condition (M = 5.21). The main effect of gift wrapping was marginally significant, F(1, 78) = 3.76, p < .06, with subjects reporting happier moods when they received a wrapped (M = 6.17) versus an un- wrapped (M = 5.63) gift. The interaction effect was not significant, F(1, 78) At tkude 7.o &.9 6.a 6.7 6.6 6.5 6.4 6.3 6.2 6.1 6-o $.9 5.1 5.7 $.6 s.s $.4 5.3 5.2 5.1 5.0 4.9 4.8 4.7 4.6 4.5 OTHER'S GIFT F IGURE 1 Effects of gift wrap t I ping and gift ownership on product GIlT WR~I ; D UNWRAPPED attitudes. GIFT-WRAPPING EFFECTS 207 = 2.46, p < . 12. Again, however, given the a priori expectations, the hypothe- sis was examined using two orthogonal contrasts. The means are plotted in Figure 2. For the personal gift condition, subjects were significantly happier, F(1, 78) = 6.19, p < .03, when they received a wrapped (M = 7.17) versus an unwrapped (M = 6.10) gift. However, no significant difference was seen (F < 1) between the wrapped (M = 5.27) and unwrapped (M = 5.15) treatments for the other's gift condition. For sad mood, a marginally significant main effect of gift ownership was seen, F(1, 78) = 3.34, p < .07, with subjects feeling less sad in the personal gift (M = 3.23) than in the other's gift condition (M = 3.84). The main effect of gift wrapping, F( l , 78) = 1.18, p < .28, and the interaction effect (F < 1) were both nonsignificant. However, cell comparisons revealed that subjects were marginally less sad, F(1, 78) = 2.10, p < .10, one-tailed, when they received a gift-wrapped item (M = 2.88) as opposed to an unwrapped (M = 3.58) item in the personal gift condition. No significant difference (F < 1) was found between the wrapped (M = 3.82) and unwrapped (/14 = 3.87) treat- ments for the other's gift condition. Tests of mediation. The results up to this point reveal that gift wrapping has strong and parallel effects on both attitude and mood, particularly happy mood, when subjects evaluate an item they believe is theirs to keep. These findings are consistent with what one would expect if gift wrapping a personal present influences mood which, in turn, positively biases object attitudes. Tests of mediation were conducted next to provide additional evidence that mood is a factor influencing the attitudinal effect observed. FIGURE 2 Effects of gift wrap- ping and gift ownership on happy mood. Happy Mood 7.$ 7.4 7.2 7.2 7.1 7.0 lt.g I;.8 6.7 6.6 S.$ 6.4 S.2 6.Z l i . I 6.0 5.9 S.8 5.7 $.6 5.5 $.4 $.3 S,Z $,I S,0 OTHER'S GIFT t I GIFT WRAPgEO UNWRAPPED 208 HOWARD Mediational tests were conducted separately for the personal gift and other's gift conditions because the mediational hypothesis was specific to the former conditions. For the personal gift conditions, gift wrapping was found to have a significant effect on happy mood (p < .01) but not on sad mood (p > . 10). Gift wrapping was also found to have a significant effect on attitudes (p < .01). Finally, regressing attitude on both happy mood and gift wrapping revealed a significant effect only for happy mood (p < .0003); the gift-wrap- ping effect became nonsignificant (p > . 10). For the other's gift conditions, regressing attitude on gift wrapping revealed a nonsignificant effect, and thus there was nothing to mediate. These results support the position that happy mood mediates the effects of gift wrapping on product attitudes for the receipt of a personal gift. Given this support for the mediational hypothesis, next I examined whether the media- tional effects were equally strong within the two personal gift conditions (gift wrapped vs. unwrapped). Results revealed happy mood to be a significant predictor of attitude, F(1, 18) = 6.34, p < .02 (R' = 0.26), in the wrapped personal gift condition but a nonsignificant predictor in the unwrapped per- sonal gift condition (p > . 10, two-tailed). The attitude-happy mood correla- tion was .36 (p < . 10, one-tailed). Discussion The results are again consistent with the suggestion that wrapping a gift has a favorable influence on evaluations of what is inside. Similar to the results of the first study, the second experiment found that when subjects received a personal gift that was wrapped, they had a more positive attitude toward owning it than subjects who received the same item that was unwrapped. In addition, when subjects received a personal gift that was wrapped, it elicited a happier mood state, in comparison to the unwrapped condition. However, gift wrapping had no clear effect on the attitude and mood measures when subjects thought the seat cover was not theirs to keep. These results are congruent with the view that gift wrapping serves as an effective cue for the retrieval of a happy mood only when the retrieval situation closely matches the conditions of original encoding, consistent with encoding specificity princi- ples. Specifically, receiving a gift-wrapped present that is theirs to keep more closely replicates the typical condition in which college students have experi- enced happiness with previously received gifts, compared to the relatively unfamiliar position of receiving a wrapped present not intended for them. Thus, gift wrapping was effective at eliciting a happy mood in the former, but not the latter condition. The suggestion that the retrieval of a happy mood accounts, in turn, for the attitudinal effects observed was supported by finding mood to mediate gift wrapping effects statistically on attitude in the personal gift condition. Further, the parallel pattern of mood and attitude scores and finding mood GIFT-WRAPPING EFFECTS 209 to mediate attitude more strongly in the wrapped than in the unwrapped personal gift condition are findings congruent with what one would expect if subjects engaged most strongly in a mood protection strategy in the former condition. These findings are consistent with the premise that the happier subjects felt, the more they sought to maintain their happy mood through the development of favorable attitudes toward owning the gift they received. The second experiment helped to eliminate two alternative explanations for the results. First, it could be argued that the more favorable attitudes in the gift-wrapped condition in Experiment 1 may reflect equity-based evaluations (e.g., Adams, 1965) if subjects perceived greater care and effort, and therefore personal investment, on the part of the gift giver by wrapping the product. However, this explanation does not appear viable given the results of the second experiment. Equal investment in the gift-wrapping effort should have been perceived in both the personal and other's gift conditions, but the more favorable attitudes were found only in the personal gift treatment. Further, the equity-based evaluation explanation does not account for the pattern of mood effects observed. Second, it might be argued that the results are a function of demand artifacts. However, if that were true, gift wrapping should again have served to elicit a happy mood and bias attitudes even in the other's gift condition, which was not found. Further, it is difficult to see how demand artifacts could account for the pattern of mediational results observed. Two points need to be discussed concerning the strength of the findings in the second experiment. First, although the pattern of mean differences was as expected (i.e., a significant difference between the wrapped and unwrapped conditions for the personal gift but a lack of a difference for the other's gift conditions), the omnibus F-statistic for the interaction effect was not signifi- cant. The literature presents differences of opinion as to the wisdom of inter- preting cell comparisons in the absence of a significant interaction effect. Some experimentalists warn against such interpretations because the effects may not be reliable (e.g., Kennedy, 1978). Others argue that nonsignificant interactions often mask reliable treatment differences (e.g., Rosenthal & Rosnow, 1985). In this context, reliability assessment via the test-retest method supports this latter view. Specifically, the critical wrapped versus unwrapped cell contrasts for a personal gift were highly significant in both the first and second experi- ments. Additional testing revealed the pattern of cell means for the other's gift condition to also replicate? Thus, despite the lack of a significant omnibus 'Using the same materials and procedures as in Experiment 2, 38 subjects were randomly assigned to wrapped and unwrapped other's gift treatments. No significant differences were seen for the attitude (a = 0.80, F < 1, Ms = 4,77 and 4.60), happy mood (a = 0.78, F < 1, Ms = 5.04 and 4.82), or sad mood (a = 0.85, F < 1, Ms = 3.68 and 3.79) measures between the wrapped and unwrapped conditions, respectively. Attitude correlated 0.18 and 0.15 and --0.21 and --0.20 with happy and sad mood for the wrapped and unwrapped treatments, respectively. None of the correlations were significant. 210 HOWARD F-statistic for the interaction effect, the treatment differences appear reliable. Second, clearly the results for happy mood were stronger than for sad mood, in terms of both mean differences and relationship with attitude. Isen (1984) discussed at length how findings for positive affect are often not sym- metrical with those for negative affect. In the second experiment, this lack of symmetry was revealed by an attenuated impact of the manipulations on sad mood relative to happy mood. Isen suggested that one reason for that lack of symmetry is the possibility that happiness and sadness are not opposite poles on a single continuum (i.e., they represent distinct feeling states). This possibil- ity is congruent with the observed correlation between the happy and sad mood scales, which revealed only 23% common variation. The possibility that happi- ness and sadness reflect distinct feeling states also fits well with the mood retrieval and mood maintenance explanations observed in the second experi- ment. If happiness and sadness are distinct feeling states, then the retrieval of more of a happy mood would not necessarily mean that less of a sad mood was also retrieved. Further, if happiness and sadness are distinct feeling states, then a positive correlation of happy mood with attitude would not necessarily imply that sad mood should have an equivalent but opposite correlation with atti- tude. These arguments are consistent with the results observed in the second experiment. EXPERIMENT 3 A third experiment was conducted to provide additional evidence on the mood retrieval and mood maintenance explanations for the effects of gift wrapping on object attitudes. The encoding specificity principle holds that the likelihood of retrieval of some learned event varies as a function of the similarity between encoded and cue information. The results of Experiment 2 were consistent with this explanation and suggests that the receipt of a personal gift is a dominant cue in the retrieval environment because the effects of gift wrapping on mood occurred only when subjects believed the gift was their own. Yet, the results of Experiment 2 did not answer whether it is the act of unwrapping a personal gift or the appearance of the wrapping per se that accounts for the effects observed. Because the receipt of a gift-wrapped present involves stimu- lation by the appearance of the package and then by the act of unwrapping, both factors may serve as potential retrieval cues for associated affect. A third experiment was conducted to attempt to tease apart the influence of these two factors. If the similarity principle underlying the theory of encoding specificity holds, then receipt of a personal gift in traditional gift wrapping (labeled the traditionally wrapped condition in Experiment 3) should have a stronger effect than receipt of a personal gift that is wrapped but not with gift-wrapping paper (labeled the nontraditionally wrapped condition in Experiment 3). This should GIFT-WRAPPING EFFECTS 211 be true because the former condition involves cuing of associated affect via appearance identification (i.e., the gift "looks like" previously wrapped gifts subjects have received), as well as by the act of unwrapping (i.e., all previously wrapped gifts subjects received required unwrapping to observe the contents). However, the latter condition should involve only cuing of associated affect via the act of unwrapping if the item is not wrapped in a manner that identifies it as gift wrapping. Thus, this reasoning hypothesizes the presence of a signifi- cant linear trend across three conditions. Subjects receiving the unwrapped, nontraditionally wrapped, and traditionally wrapped gifts should be the least happy, more happy, and most happy groups, respectively. Specifically, the difference in happy moods between the traditionally wrapped and unwrapped conditions should be greater than the difference between the nontraditionally wrapped and unwrapped conditions. The same reasoning was examined for sad mood as well but with opposite and attenuated effects anticipated based on the results of Experiment 2. If the mood maintenance reasoning again holds, the attitude scores should parallel the happy mood scores for the same reasons discussed in Experiment 2. Specifically, developing a favorable attitude about owning an object that subjects believe is theirs to keep should help maintain their happy mood. The happier subjects feel, the greater the drive to maintain that mood and hence the more likely they should be to develop favorable attitudes, given the consist- ency between feeling happy and positively evaluating the ownership of their gift. Thus, a significant linear trend in mean attitude scores across groups should be found. Subjects receiving the unwrapped, nontraditionally wrapped, and traditionally wrapped gifts should have the least favorable, more favor- able, and most favorable attitudes toward owning the product, respectively. Specifically, the difference in attitudes between the traditionally wrapped and unwrapped conditions should be greater than the difference between the non- traditionally wrapped and unwrapped conditions. Finally, happy mood is again expected to mediate the effects of gift wrap- ping on attitude scores, and those mediational effects should be stronger in conditions with greater mood protection bias. The stronger the biasing effects of a mood on attitude, the more predictive of attitude the mood should be. Thus, assuming the mediational hypothesis is supported, the regression of attitude on mood should be weakest, stronger, and strongest in the unwrapped, nontraditionally wrapped, and traditionally wrapped conditions, respectively. Method Pretest. Thirty-six university students were recruited for a gift evaluation study they were told was being sponsored by a local firm and were randomly assigned to two different conditions. In both conditions, subjects were told they would be asked to evaluate the appearance of wrapping used for a gift. In one 212 HOWARD condition, subjects were presented with the gift-wrapped packaged used in the first two experiments. In the other condition, the box was wrapped in brown packaging paper (the type used to wrap boxes for mailing) and no bow or ribbon was used. Subjects were asked to examine the packages and respond to three 5-point Likert-type items on a scale ranging from strongly agree (1) to strongly disagree (5): "the package is neatly wrapped," "the wrapping makes the package look like a gift," and "the package is professionally wrapped." One-way ANOVA results revealed no significant differences on the first and third items, indicating that subjects perceived no differences in appearance in terms of how well wrapped the packages were. However, a strong difference was observed for the second item, F(1, 35) = 1028.50; p < .00001 (Ms = 1.0 vs. 4.67). All 18 subjects in the gift-wrapped condition strongly agreed that the wrapping made the package look like a gift. All 18 subjects either disagreed or strongly disagreed that the brown packaging paper made that package look like a gift. These findings support the contention that the appearance of the gift in traditional gift wrapping should serve as an effective retrieval cue for memories associated with prior gift receipt occasions, but the appearance of the gift in brown packaging paper should not. Procedure andmateria/s. The same procedure was used as in Experiment 1. The gift-wrapped and unwrapped treatments were again used in Experiment 3 but were called the traditionally wrapped and unwrapped conditions. A third treatment group received a nontraditionally wrapped gift. That condition was identical to the traditionally wrapped treatment except that the box was wrapped in plain brown packaging paper and no bow or ribbon was used. Measured variables. Attitude toward owning the product (a = 0.79) was again examined as the critical dependent measure. Mood was assessed using the same items as in Experiment 2 (for the happy scale, ct = 0.77; for the sad scale, a = 0.76). The correlation between the happy and sad mood scales was --0.58 (R 2 = 0.33). The correlation between attitude and happy mood was 0.50 and between attitude and sad mood was --0.31. Subjects and design. Sixty university students were randomly assigned to the three experimental conditions (traditionally wrapped, nontraditionally wrapped, and unwrapped). Thus, a one-factor between-subjects design was implemented. Results Mean differences and linear trends. One-way ANOVA results were significant for the attitude index, F(2, 57) = 5.14, p < .01, and the means were in the expected direction. Cell comparisons were performed using t tests with GIFT-WRAPPING EFFECTS 213 critical value adjustments based on the Bonferroni procedure (Rosenthal & Rosnow, 1985). Attitude toward owning the seat cover when nontraditionaUy wrapped (M = 6.32) was more favorable, t(57) = 2.18, p < .05, one-tailed, than when unwrapped (M = 5.45). However, attitude toward owning the seat cover when traditionally wrapped (M = 6.70) was even more strongly favor- able than when unwrapped, t(57) = 3.13, p < .01. A test of the linear trend was significant, F(1, 57) = 9.78, p < .01. ANOVA results for the happy mood were significant, F(2, 57) = 5.44, p < .01, and paralleled the attitudinal results. Those receiving the nontradition- ally wrapped gift were happier (M = 6.18) than those receiving the unwrapped gift (M = 5.48), t(57) = 2.16,p < .05, one-tailed. However, those receiving the traditionally wrapped gift expressed an even happier mood (M = 6.53) than those in the unwrapped condition, t(57) = 3.24,p < .01. The linear trend was again significant, F(I, 57) = 10.50, p < .01. ANOVA results for the sad scale were marginally significant, F(2, 57) = 2.38, p < . 10. However, the linear trend for the three means was significant, F(1, 57) = 4.73, p < .05. Those receiving the nontraditionally wrapped gift reported being less sad (M = 3.35) than those receiving the unwrapped gift (M = 3.73) but not significantly so, t(57) = 1.06, p > .10. Those receiving the traditionally wrapped gift (M = 2.95) were significantly less sad, t(57) = 2.17, p < .05, one-tailed, than those receiving the unwrapped gift. Tosts of mediation. Mediational tests were run separately for the two effects of interest: the effect of nontraditional gift wrapping (assessed across the nontraditionally wrapped and unwrapped conditions) and the effect of tradi- tional gift wrapping (assessed across the traditionally wrapped and unwrapped conditions). For the nontraditional gift-wrapping effect, gift wrapping was found to have a significant effect on happy mood (p < .03) but not on sad mood (p > .10). Gift wrapping was further seen to have a significant effect on product attitudes (t9 < .05). Finally, regressing attitude on both gift wrapping and happy mood revealed only a significant effect for happy mood (p < .007). The gift-wrapping effect was reduced to nonsignificance (p > .10). For the traditional gift-wrapping effect, gift wrapping had a significant effect on happy mood (p < .005) but not on sad mood (p > . 10) and a significant effect on attitude (p < .004). However, when regressing attitude on both happy mood and gift wrapping, only happy mood was a statistically significant predictor (p < .01). The gift-wrapping effect was reduced to marginal signifi- cance (p < .08). These results support the position that happy mood mediates both the traditional and nontraditional gift-wrapping effect on product attitudes ob- served in Experiment 3. Given this support for the mediational hypothesis, I examined whether the mediational effects were equally strong within each of the conditions examined in the third experiment. Results revealed happy mood 214 HOWARD to be a nonsignificant predictor of attitude in the unwrapped condition (again p > .10, two-tailed), but the attitude-happy mood correlation was 0.30 (p < . 10, one-tailed). Happy mood was marginally predictive of attitude in the nontraditionally wrapped condition, F(1, 18) = 4.19, p < .06 (R 2 = 0.19). Finally, happy mood was significantly predictive of attitude in the traditionally wrapped condition, F(1, 18) = 7.49, p < .01 (R 2 = 0.29). Discussion The results of Experiment 3 replicate those of the first two experiments, indicating that gift wrapping has a favorable effect on product attitudes. They also further support the position that a mood retrieval and mood maintenance process explains the effects observed. According to encoding specificity principles, the more similar a cue is to factors operating at the time of encoding, the more effective the retrieval. Thus, one would expect the traditionally wrapped gift to be more effective than the nontraditionally wrapped gift in retrieving associated moods. This was sup- ported. Yet, the fact that the use of wrapping that did not look like gift wrapping still had an effect on moods suggests that the act of unwrapping is an important part of the retrieval cue mechanism. This is not surprising in that the act of unwrapping is obviously a part of the gift receipt ritual. However, the finding that traditional gift wrapping had a greater effect on moods indi- cates that the appearance of the gift wrapping played an added role in the retrieval process. Subjects receiving the traditionally wrapped gift were not only exposed to a stimulus that involved the act of unwrapping, but they were also exposed to a stimulus that looked like a gift. The incremental effect of the traditional over the nontraditional condition, therefore, can reasonably be attributed to the appearance of the gift wrapping in the former condition. The suggestion that retrieval of a happy mood accounts for the attitudinal effects observed was again supported by finding mood to statistically mediate gift- wrapping effects on those attitudes. It was again seen that the effects of sad mood were attenuated relative to happy mood. Although the two mood states were more highly correlated in Experiment 3 than in Experiment 2, the level of shared variance (R ' = 0.33) was again consistent with the possibility that the two may represent distinct feeling states. The attitudinal results of the third experiment paralleled the mood effects, particularly those for happy mood. This is again consistent with what one would expect under a mood maintenance-processing scenario. This possibility was further supported by the results of the mediational tests, which revealed happy mood has a stronger mediational effect on attitude when a mood mainte- nance process was more strongly anticipated. Thus, when subjects were the happiest, attitudes were also the most favorable, and mood mediated attitude GIFT-WRAPPING EFFECTS 215 most strongly. These findings are congruent with the position that the happier subjects felt, the more they sought to maintain that mood through the develop- ment of favorable attitudes toward owning the gift they received. The results of the third experiment also help to eliminate several alternative explanations for the results. First, it is difficult to see how demand artifacts can account for the linear trends in means observed, as well as the pattern of mediational effects observed. A second alternative explanation is concerned with a reason for the elicitation of a happy mood by gift wrapping. One may argue that gift wrapping made recipients feel special which made them happy. Perhaps subjects were moved by the apparent time, effort, and trouble the firm went through to gift wrap the item for them. This may have made those who received the gift-wrapped item feel special and unique (Snyder & Fromkin, 1980) which made them happier. A third explanation is that perhaps the gift wrapping of a personal present heightened the tendency of subjects to recipro- cate the favor (Gouldner, 1960) through the reporting of high evaluations of the object, along with the reporting of desirable mood states. In other words, perhaps subjects reasoned "because you went through the trouble of wrapping my gift so nicely, I'll evaluate it highly." A fourth explanation is perhaps subjects believed that more important or substantial gifts are typically wrapped, and this belief drove the observed effects independent of the theoreti- cal arguments presented here. However, although these explanations may account for the differences between the traditionally wrapped and unwrapped conditions, they cannot explain the differences between the nontraditionally wrapped and unwrapped conditions. The mood retrieval explanation, follow- ing encoding specificity principles, however, is able to account for the complete range of observed effects. EXPERIMENT 4 A fourth and final experiment was conducted to provide even more evidence on the mood retrieval explanation. Tulving (1983) stated that a stimulus serves as an effective retrieval cue to the extent to which it recreates an original encoding situation. Closely matched, more compatible cue and stored informa- tion increase the chances that the cue will actualize the memory. I have been arguing that gift wrapping is effective in eliciting a happy mood because of the consistent pairing over time of gift wrapping with joyous occasions in the fives of recipients--and thus with happy mood experiences. According to encoding specificity principles, presenting subjects with a gift-wrapped item on the day of one of those joyous occasions should be more effective in retrieving happy mood than presenting subjects with a gift-wrapped item on some other day. This should be true because the gift-wrapping stimulus would more closely replicate a previous event associated with the experience of a happy mood 216 HOWARD when gift wrapping is used on a joyous occasion than on some other occasion. For example, in our survey (see the Discussion section of Experiment 1), 96% of subjects stated that most or all gifts they received on their birthday were usually wrapped and that this was also true throughout their childhood years. The effectiveness of gift wrapping in retrieving mood when used on someone's birthday versus its effectiveness when used on some other day was examined in the fourth experiment. The expectations for attitude, the mediational rela- tion between mood and attitude, and the attenuated findings for sad mood all follow the same reasoning discussed in detail for Experiments 2 and 3. Thus, a happier mood, a more favorable attitude, and a stronger mediational relation between mood and attitude are expected for those receiving a gift-wrapped item on their birthday than those receiving a gift-wrapped item on some other day. However, three groups are examined in the fourth experiment: those who received a gift-wrapped item on their birthday (the gift-wrapped-birthday condition), those who received an unwrapped item on their birthday (the unwrapped-birthday condition), and those who received a gift-wrapped item not on their birthday (the gift-wrapped-other day condition). As just indi- cated, the critical comparison is between the gift-wrapped-birthday and gift- wrapped-other day conditions. The unwrapped-birthday condition was added to determine whether any differences observed in the critical compari- son are a function of both the gift-wrapping and birthday factors. I chose not to include an unwrapped-other day condition to eliminate unnecessary redundancy with prior experiments. The gift-wrapped versus un- wrapped effect on days other than birthdays was already clearly demonstrated in the first three experiments. 6 Thus, there is little need to demonstrate that effect again. Further, eliminating that condition does not result in confounded comparisons because the gift-wrapped-birthday and unwrapped-birthday conditions differ only on the gift-wrapping variable, whereas the gift-wrapped- birthday and gift-wrapped-other day conditions differ only on the occasion variable. Method Procedure and materials. The same procedure was used as in Experi- ment 1. Measured variables. Attitude toward owning the product (ct = 0.76) was the critical dependent variable. Mood was again assessed using the same items as in previous experiments (for the happy scale, ct = 0.75; for the sad ~A postexperimental review revealed that only two of the subjects in Experiments 2 and 3 participated on their birthday. This information could not be obtained for subjects who par- ticipated in the first experiment, but there is no reason to expect that factor to differ from the later studies. GIFT-WRAPPING EFFECTS 217 scale, ct = 0.81). The correlation between the happy and sad mood scales was --0.49 (R' = 0.24). The correlations of attitude with happy and sad moods were 0.58 and --0.42, respectively. Subjects and design. I did not use a completely random design due to the nature of the conditions. There were three comparison groups. Two of the groups were the gift-wrapped and unwrapped conditions on subjects' birth- days. Subjects who participated in the study on their birthday were randomly assigned to either the gift-wrapped or unwrapped conditions (the gift- wrapped-birthday and unwrapped-birthday conditions). None of subjects were aware that they were recruited because of their birthday. A third group was the gift-wrapped condition in which subjects participated on days other than their birthday (the gift-wrapped-other day condition). Approximately the same number of subjects (N = 15) were assigned to this condition as in the other two conditions (Ns = 15, 14). In total, 44 university students were assigned to the three conditions. Results Mean differences. One-way ANOVA results were significant for the attitude index, F(2, 41) = 4.92, p < .01. Cell comparisons were again per- formed using t tests with critical value adjustments based on the Bonferroni procedure (Rosenthal & Rosnow, 1985). Attitude toward owning the wrapped seat cover on subjects' birthdays (M = 7.87) was more favorable than owning an unwrapped seat cover on subjects' birthdays (M = 6.52), t(41) = 3.02, p < .005. A second comparison revealed that the receipt of the gift-wrapped item on subjects' birthdays resulted in more favorable attitudes than receipt of the gift-wrapped item on other days (M = 6.80), t(41) = 2.39, p < .05, one- tailed. The mood results paralleled the attitudinal findings. ANOVA results for happy mood were significant, F(2, 41) = 5.40, p < .01. Subjects reported being significantly happier, t(41) = 3.22, p < .005, in the gift-wrapped- birthday than the unwrapped-birthday conditions (Ms = 7.40 and 6.29, respectively). A second comparison further revealed that the gift-wrapped- birthday subjects were significantly happier, t(41) = 2.31, p < .05, one-tailed, than the gift-wrapped-other day subjects (M = 6.60). ANOVA results for sad mood were not significant, F(1, 41) = 1.96, p < .15. Tests of mediation. Tests of mediation were again conducted separately for the two effects of interest: the effect of gift wrapping (assessed across the gift-wrapped-birthday and unwrapped-birthday conditions) and the effect of occasion (assessed across the gift-wrapped-birthday and gift-wrapped-other day conditions). 218 HOWARD For the first effect, gift wrapping was found to have a significant effect on happy mood (p < .001) and a marginal effect on sad mood (p < .07). Further, gift wrapping was seen to affect product attitudes significantly (p < .002). Finally, regressing attitude on both gift wrapping and happy mood revealed only a significant effect for happy mood (p < .0005); the gift-wrapping effect was reduced to nonsignificance (p > .10). However, regressing attitude on both gift wrapping and sad mood resulted in no clear reduction of the gift- wrapping effect (p < .01), and the effect of sad mood was only marginally significant (p < .08). Next, the effect of the occasion variable was assessed. Subjects' birthday was seen to have a significant effect on happy mood (p < .04) and a marginal effect on sad mood (p < .09). Further, the birthday factor had a significant effect on product attitudes (p < .02). Finally, the regression of attitudes on both the happy mood and occasion variables revealed only happy mood to be a signifi- cant predictor (p < .0009); the occasion variable was reduced to nonsignifi- cance (p > .10). However, the regression of attitudes on both the sad mood and occasion variables resulted in no clear reduction in the effect of subjects' birthday (p < .03), and the effect for sad mood was only marginally significant < .09). These results support the position that happy mood mediates the effects of the gift-wrapping and occasion variables on product attitudes. Given that support for the mediational model, attitude was regressed on happy mood within the three experimental conditions to test for the presence of differential mediational strength. Results revealed happy mood to be a nonsignificant predictor of attitude in the unwrapped-birthday condition (p > .10, two- tailed), but the attitude-happy mood correlation was 0.31 (p < .12, one- tailed). Happy mood was marginally predictive of attitude in the gift-wrapped-other day condition, F(1, 13) = 3.56, p < .08 (R 2 = 0.21). However, happy mood was significantly predictive of attitude in the gift- wrapped-birthday condition and had greater explanatory power, F(1, 13) = 7.38, p < .02 (R 2 = 0.36). Discussion There were two comparisons of interest in Experiment 4. First, I found that receiving a gift-wrapped item on subjects' birthdays elicited a happier mood and more favorable attitudes than an unwrapped item. This is not surprising given the results of Experiments 1, 2, and 3. It demonstrates that gift wrapping has an effect when used on people's birthday. Second, I found that a gift- wrapped item received on subjects' birthdays elicited a happier mood and more favorable attitudes than the same gift-wrapped item received on days other than their birthdays. This critical comparison indicates that the gift-wrapping and occasion variables work in conjunction with one another, consistent with GIFT-WRAPPING EFFECTS 219 encoding specificity principles. I argued that when people receive gifts they are typically wrapped and typically received on joyous occasions marked by happy moods. Thus, gift wrapping should have the ability to serve as a retrieval cue for those moods. However, if gift wrapping serves as a retrieval cue, it should be a more effective one when it is actually used on occasions like birthdays, because the retrieval conditions would more closely match the conditions of original encoding of the relation between gift wrapping and mood. The results were consistent with this expectation. The suggestion that retrieval of a happy mood accounts for the attitudinal results observed was supported by finding mood to mediate treatment effects statistically on those attitudes. What is impressive about the findings in Experiment 4 is that they were obtained without any reference being made to subjects' birthdays during the course of the experiment. Subjects had no knowledge that the experimenters were aware of their birthday; the issue of birthdays was never raised; and, in debriefing, no subject expressed concerns or suspicions regarding the date of their participation. That fact helps to again eliminate several alternative expla- nations for the results, such as demand artifacts, reciprocal evaluations, and feelings of being special. None of those explanations make sense given that subjects believed the provider of the gift was unaware of their birthday. Beliefs that more important or substantial wrapped gifts are typically received on one's birthday is also not viable for the same reason--subjects believed the provider of the gift was unaware of their birthday. Finally, the attitude and mood/attitude mediational data is again congruent with a mood maintenance scenario of attitude development: The happier mood resulted in more favorable attitudes and stronger mediational effects of mood on attitude. This suggests that the happier subjects felt, the more they sought to maintain their happy mood through the development of favorable attitudes toward owning the gift they received. The fact that Experiments 2, 3, and 4 reveal considerable stability in their pattern of expected findings suggests that the phenomenon examined here is a robust and reliable one. GENERAL DISCUSSION One area of fruitful study in social psychology has been the discovery and/or exploration of the persuasive power of different techniques of social influence (Cialdini, 1980). This series of studies has sought to assess whether gift wrap- ping qualifies as a tool of persuasion. Two questions were addressed. First, does the use of gift wrapping have an influence on attitude toward owning the product received? The answer to that question is yes, as long as ownership of the item is assumed by the person receiving the gift. Strong support for this finding was consistently observed across four separate experiments. Second, what explains the attitudinal results of the gift-wrapping manipula- 220 HOWARD tion? The last three experiments provided support for the contention that gift wrapping elicits a happy mood which, in turn, positively biases resulting attitudes. The mediational effects of sad mood were generally not supported. These studies were not designed to address the question of whether happy and sad moods represent opposite poles on a single mood continuum, although the data are consistent with Isen's (1984) suggestions that they may not be. Thus, it is possible that on joyous occasions increases in happy moods are more pronounced than decreases in sad moods. This may have been reflected in the strength of the association between gift wrapping and the two mood states, thus determining the differential effectiveness of gift wrapping in cuing those moods. One issue that deserves discussion is why gift wrapping is effective in eliciting a happy mood. I believe a general understanding of those results can be gained by considering the role and frequency of gift wrapping in the gift- giving process as a cultural phenomenon. Most people have had a lifetime of experience receiving things that are gift wrapped. My survey work revealed that when people receive gifts, the gifts are usually wrapped. Gift wrapping is part of our common cultural heritage and is closely tied to joyous occasions, including successes, celebrations, romance, gratitude, and other salient events laced with emotion throughout our childhood, adolescence, and adult years. Given this perspective, the results obtained here are easily understood because they are consistent with the expectation that the gift wrapping of presents should help cue the retrieval of happy moods experienced on such occasions. One limitation to the mood retrieval argument is that, due to the life-history explanation for the issue investigated, the experiments did not actively manip- ulate the acquisition of the gift wrapping-happy mood association. Thus, the position that the use of gift wrapping resulted in the retrieval of a happy mood rests first on the consistency of results with encoding specificity expectations and second on the experimental ability to eliminate other alternative explana- tions. Concerning the first issue, the consistency of the results with expecta- tions derived from existing principles of encoding specificity appears clear. Thus, in a theoretical sense, the more closely matched the retrieval situation was to conditions presumably present in the establishment of the relationship between gift wrapping and happy mood (e.g., the receipt of a personal gift, a gift in traditional gift wrapping, and a gift on one's birthday), the higher the likelihood of subjects reexperiencing that affect. Concerning the second issue, the pattern of results supporting the mood retrieval explanation was found to be inconsistent with a variety of alternative explanations, including demand artifacts, equity based evaluations, reciprocal evaluations, feelings of being special or unique, and beliefs in the importance or substantial nature of gifts that are wrapped. One alternative explanation not excluded is that gift wrapping results only in the retrieval of cognitions which, in turn, stimulates a happy mood. The argument then becomes a "which came GIFT-WRAPPING EFFECTS 221 first, cognition or affect?" debate. The problem here is that it is unclear whether either position is unambiguously disconfirmable. Debates concerning the temporal primacy of cognition versus affect have been engaged for years without an end in sight (cf. Lazarus, 1984, with Zajonc, 1984), and some have even questioned the utility of expending resources attempting to resolve this issue (Peterson, Hoyer, & Wilson, 1986). Nevertheless, although not all alternative explanations for the observed results can be eliminated from consideration, many are. Further, the view that gift wrapping is associated with joyous events in the lives of recipients, and thus has utility in cuing a happy mood in theoretically predictable circum- stances, is a simple but intuitively appealing notion that was found to be consistent with the observed data across experiments. The suggestion that the retrieval of a happy mood accounts, in turn, for the attitudinal effects observed was consistently supported by finding mood to mediate treatment effects on attitudes in the last three experiments. In the last three studies, attitude was most favorable and mood, especially happy mood, was strongest in the critical gift-wrapped treatments. These findings are consistent with a mood maintenance explanation of attitudinal biasing. More favorable attitudes were developed in response to happier moods because the former supported the latter when the gift item was believed by subjects to be theirs to keep. Further, the happier the mood, the greater its mediational effects on attitude. This supports the contention that the happier subjects felt, the greater the reliance on attitudinal biasing to support that mood. Previous studies have provided evidence of mood maintenance princi- pally through the use of behavioral indicators (see Isen, 1984, and Manucia, Baumann, & Cialdini, 1984, for reviews). This study suggests that such a strategy can also be executed through attitudinal biasing. The possibility that people may manage moods through the process of attitude formation opens the door to future research examining just how common this strategy is. One limitation to the mood maintenance argument is that it involves making infer- ences about people's motivation on the basis of experiments that did not directly manipulate the need or desire to maintain the happy mood. One problem is that it is not clear how such a manipulation could be unambigu- ously employed. Past studies arguing both for and against motivational based mood explanations have relied on evidence concerning the consistency of data with theoretical expectations (see Isen, 1984, for a review; Mackie & Worth, 1989), similar to the arguments presented here, rather than active manipula- tions per se of the motivations underlying mood maintenance. Further, it is unclear that ability-based explanations provide a better fit to the data. The most common ability-based explanation for the facilitating effects of mood on persuasion is that positive mood reduces the ability to process a stimuli (e.g., Mackie & Worth, 1989), which does not appear to have strong explanatory potential for the results observed in this series of studies. On the other hand, 222 HOWARD the data are quite consistent with the motivational explanation offered. How- ever, I do not discount the possible interaetion of motivational and ability- related factors in accounting for the results. Overall, I believe the results presented here can help us understand a common facet of gift-giving (and gift-receiving) experiences. Given the perva- siveness of gift exchange in this country, I believe these findings also can help us better understand ourselves. REFERENCES Adams, J. S. (1965). Inequity in social exchange. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 2, pp. 1-40). New York: Academic. Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic and statistical considerations. Journal of Personal- ity and Social Psychology, 5L 1173-1182. Batra, R., & Stayman, D. M. (1990). The role of mood in advertising effectiveness. Journal of Consumer Research, 17, 203-214. Belk, R. A. (1976). It's the thought that counts: A signed digraph analysis of gift-giving. Journal of Consumer Research, 3, 155-162. Belk, R. A. (1979). Gift giving behavior. In .J. Sheth (Ed.), Research in marketing (Vol. 2, pp. 95-126). Greenwich, CT: JAI. Belshaw, C. S. (1965). Traditional exchange in modern markets, Englewood Cliffs, NJ: Prentice- Hall. Berlyne, D. E., & Madsen, K. B. (1973). Pleasure, reward, preference: Their nature, determinants and role in behavior. New York: Academic. Bower, G. H. (1981). Mood and memory. American Psychologist, 36, 129-148. Camerer, C. (1988). Gifts of economic signals and social symbols. American Journal of Sociology, 94 (Suppl.), 5180-5214. Cialdini, R. (1980). Full cycle social psychology. Applied Social Psychology Annual, 1, 21-45. Eich, J. E., Weingartner, H., Stillrnan, R. L., & Gillin, J. C. (1975). State dependent accessibility of retrieval cues in the retention of a categorized list. Journal of Verbal Learning and Verbal Behavior, 14, 408-417. Fisher, R. P., and Craik, F. I. M. (1977). Interaction between encoding and retrieval operations in cued recall. Journal of Experimental Psychology: Human Learning and Memory, 3, 701- 711. Geiselrnan, R. E., & Glenny, J. (1977). Effects of imaging speaker's voices on the retention of words presented visually. Memory & Cognition, 5, 499-504. Godden, D. R., & Baddeley, A. D. (1975). Context dependent memory in two natural environ- ments: On land and underwater. British Journal of Psychology, 66, 325-331. Gouldner, A. (1960). The norm of reciprocity: A preliminary statement. American Sociological Review, 25, 161-178. Isen, A. M. (1984). Toward understanding the role of affect in cognition. In R. Wyer & T. Srull (Eds.), Handbook of social cognition (pp. 179-236). Hillsdale, NJ: Lawrence Erlbaum Associ- ates, Inc. Isen, A. M. & Levin, P. F. (1972). The effect of fecling good on helping. Journal of Personality and Social Psychology, 21, 384-388. Isen, A. M., Means, B., Patrick, R., & Nowicki, G. (1982). Some factors influencing decision making strategy and risk taking. In M. S. Clark & S. T. Fiske (Eds.), Affect and cognition: The GIFT-WRAPPING EFFECTS 228 17th Annual Carnegie Symposium on Cognition` Hillsdale, NJ: Lawrence Erlbaum Associates, Inc. Isen, A. M. & Simmonds, S. F. (1978). The effect of feeling good on a helping task that is incompatible with good mood. Social Psychology, 41, 345-349. Jones, E. E. (1964). Ingratiation. New York: Appleton-Century-Crofts. Katz, D. (1960). The functional approach to the study of attitudes. Public Opinion Quarterly, 24, 163-204. Kennedy, J. J. (1978). An introduction to the design and analysis of experiments in education and psychology. Washington, D.C.: University Press. Kerton, R. K. (1971). An economic analysis of the extended family in the West Indies. Journal of Developmental Studies 9, 423-434. Lazarus, R. S. (1984). On the primacy of cognition. American Psychologist, 39, 124-129. Mackie, D. M., & Worth, L. T. (1989). Processing deficits and the mediation of positive affect in persuasion. Journal of Personality and Social Psychology, 57, 27-40. Manucia, G. K., Baumann, D. J., & Cialdini, R. B. (1984). Mood influences on helping: Direct effects or side effects? Journal of Personality and Social Psychology, 46, 357-364. Manufacturing USA. (1989). Industry analysis." Statistics and leading companiex Detroit: Gale Research. Mauss, M. (1954). The gift. London: Cohen and West. Ncisser, M. (1973). The sense of self expressed through giving and receiving. Social Casework, 54, 294-301. Petcrson, R. A., Hoyer, W. D., & Wilson, W. R. (1986). Reflections on the role of affect in consumer behavior. In R. A. Pctcrson, W. D. Hoycr, & W. R. Wilson (Eds.), The role of affect in consumer behavior (pp. 141-159). Lexington, MA: Heath. Roscnthal, R., & Rosnow, R. L. (1985). Contrast analysis." Focused comparisons in the analysis of variance. Cambridge, England: Press Syndicate. Schwartz, B. (1967). The social psychology of the gift. American Journal of Sociology, 73, 1 - 11. Snyder, C. R., & Fromkin, H. L. (1980). Uniqueness" The human spirit of difference. New York: Plenum. Stayman, D., & Batra, R. (1991). The encoding and retrieval of ad affect in memory. Journal of Marketing Research, 28, 232-239. Tulving, E. (1983). Elements of episodic memory. Oxford, NY: Clarendon Press. Tulving, E., & Thompson, D. M. (1973). Encoding specificity and retrieval processes in episodic memory. Psychological Review, 80, 352-373. Watkins, M. J. (1979). Engrams as cuegrams and forgetting as cue overload. In C. R. Puff (Ed.), Memory organization and structure (pp. 21-39). New York: Academic. Wood, J. V., Saltzberg, J. A., & Goldsant, L. A. (1990). Does affect induce self-focused attention? Journal of Personality and Social Psychology, 58, 899-908. Worth, L. T., & Mackie, D. M. (1987). Cognitive mediation of positive affect in persuasion. Social Cognition, 5, 76-94. Zajonc, R. B. (1984). On the primacy of affect. American Psychologist, 39, 117-123.