Consumer Gift-Giving: Opening the Black Box

ABSTRACT - The three studies in this session on consumer gift-giving are characterized by a "black box" orientation and fail to reveal underlying determinants of gift-giving behavior. Future research needs to more systematically address the issue of motivations for gift-giving behavior.


Richard J. Lutz (1979) ,"Consumer Gift-Giving: Opening the Black Box", in NA - Advances in Consumer Research Volume 06, eds. William L. Wilkie, Ann Abor, MI : Association for Consumer Research, Pages: 329-331.

Advances in Consumer Research Volume 6, 1979      Pages 329-331


Richard J. Lutz, University of California, Los Angeles


The three studies in this session on consumer gift-giving are characterized by a "black box" orientation and fail to reveal underlying determinants of gift-giving behavior. Future research needs to more systematically address the issue of motivations for gift-giving behavior.


Today we have heard the results of three studies pertaining to gift-giving behavior. Two of these studies were empirical in nature, while the third was essentially a literature review leading to a conceptual model of the gift-giving phenomenon. The question that must be addressed after hearing these three papers is the following: What do we know now about gift-giving that we didn't know prior to hearing these papers? The remainder of this discussion will attempt to answer that question.


As noted by Banks in her paper, the search process for gifts has not been the subject of much past research. Thus, it is laudable that both the empirical papers presented today dealt with consumer search effort in the purchase of a gift versus search effort expended when purchasing for oneself. While the two studies are not directly comparable in all the issues they addressed, both studies did examine the amount of consumer effort expended in purchasing a "personal gift for a close friend" (Heeler, et al.); the gift situation reported on by Clarke and Belk was "a gift for a good friend." In each case, the gift purchase situation was contrasted with the purchase of the same item for "personal use."

Heeler, et al. concluded that there was no difference in search effort between the two situations, while Clarke and Belk observed a significant main effect, such that "effort expenditures for gifts [were] reported as higher than expenditures for personal product use." How can these apparently conflicting findings be resolved? In order to answer that question, we must examine in detail the studies which led to those conclusions.

Recap of the Heeler, Francis, Okechuku and Reid Study

The study reported by Heeler, et al, really consists of two separate two-group experiments. In the first experiment, which will not be discussed at length here, students were asked to select information about a blender from an information display board (IDB), under one of two conditions--for their own use, or for a wedding gift. The rationale was that the wedding gift occasion constituted a "distant" gift-giving situation, although this assumption was never verified. In the second experiment, which is of more direct relevance to the present discussion, students again used an(IDB)to examine information regarding watches either as a gift for a close friend or for their personal use. This "close friend" manipulation was intended to represent a "close" gift-giving situation which again was an untested assumption, In the absence of manipulation checks, we have no evidence to support the contention that the two experiments in truth examined differences in two gift-giving situations differing in "proximity."

Subjects in Heeler, et al's. "close friend" vs. "personal use" experiment were male and female students at a "Polytechnic Institute." They were shown an IDB displaying 16 makes of watches with 10 attributes per watch. Eight of the watches were men's and eight were women's models, so that each subject presumably selected a friend from whichever sex he/she preferred in the "close friend" condition. We are given no information as to what proportion selected men's vs. women's watches. In any case, the effective size of the IDB was 10 attributes by eight brands for any given subject, regardless of sex or the sex of the imaginary recipient of the gift. Subjects selected information cell by cell until they reached a choice of watch.

Four dependent measures relevant to consumer search effort were recorded by an observer: (1) number of cells accessed, (2) search time, (3) number of brands consulted, and (4) number of attributes consulted. The former two measures showed no significant differences between the two cells. The latter two measures moved in opposite directions, but no tests of significance were reported for either measure. Based on these results, Heeler, et al. concluded that there were no differences in search effort between the "personal use" and "close friend" situations.

Analysis of the Heeler, et al. Study

Internal validity. As noted above, there was no evidence presented with respect to the "proximity" manipulation. Similarly, there was no evidence to suggest that subjects in the two conditions of the watch experiment clearly perceived themselves to be in truly different situations. While the instructions have face validity (i.e., personal use vs. gift for close friend) one wonders just how much "mundane realism" the experimental conditions contained. Subjects were presented with a highly artificial task with no real expectations that either they or their "close friends" would actually receive the item in question. Thus, the two conditions may have, in fact, simply been viewed as a "game" in which the subject was trying to be a good player. In the absence of any realistic expectations regarding the consequences of their choices, would one really expect the subjects in the two conditions to behave differently?

External validity. The population being studied was quite narrow, and one can question whether any of the subjects in the experiment had ever previously considered purchasing a watch for a friend. Also in question, however, is the generalizability of the findings to the immediate sample employed. Only one product out of a plethora of possible gift items was studied; would the results be the same for all other products? Similarly, there are many conceivable situations in which one might consider giving a gift to a close friend. Would the results of the present study remain consistent across a variety of gift-giving occasions? In order to consider these questions, one should sample systematically from the domain of gift items and the domain of gift-giving situations.

Measurement reliability and validity. Little is known about the reliability of IDB measures of search behavior. Would the same results be obtained if the exact same people were subjected to the experiment a second time? With respect to validity, it is unclear that an IDB can be used to address the question of search effort. By their very nature IDB's display information in such a way that subjects can obtain information with very little expenditure of effort. In contrast, actual shopping behavior is much more effortful in terms of time and inconvenience.

Thus, the true validity of IDB-derived measures is very much in question. The argument that only relative levels of effort were being examined rather than absolute levels of search pales in the face of possible test-treatment interaction effects. Consider a 2x2 experimental design in which the "treatment" factor is "personal use" versus "close friend," and the "test" factor is the shopping environment in which the treatment is administered, such that one level is an SDB and the other level is '!the real world." It seems quite likely that different patterns of search would occur between the two test environments.

In sum, the evidence pertaining to internal validity, external validity, and measurement reliability and validity is quite weak for the Heeler, et al. study. This calls into question their basic conclusion with respect to differential search effort. Let us now turn to the Clarke and Belk study.

Recap of the Clarke and Belk Study

Clarke and Belk ostensibly used a gift-giving versus personal use manipulation to study the effects of "task involvement" on consumer research effort. They also used four different product categories to vary "product involvement." Since neither manipulation was checked within the study, we can safely ignore the involvement issue and focus on gift-giving.

Subjects in the Clarke and Belk study were undergraduate female students presented with two experimental conditions, in random order. In each condition, they were asked to respond to a series of intentions measures with respect to the (1) amount of time, (2) number of stores, and (3) amount of money to be spent in making a purchase decision. Although the exact measures used were never described, two forms of each measure were used, a "real" measure and a "relative" measure. Thus, a total of six dependent measures were obtained from each subject in each of two experimental conditions. In one condition, the subjects were asked to respond as if they were buying for personal use; in the other condition they responded as though they were buying a "gift for a good friend." The four product categories, which were completely crossed with the "involvement" manipulation, were bubble bath, blankets, jeans and record albums. The former two were chosen to represent low product involvement, while the latter two presumably were higher in product involvement.

Based upon a series of six univariate 2x4 factorial ANOVA, [It appears that an incorrect statistical model was used in the analysis, since subjects were not nested within cells.] Clarke and Belk observed a general pattern of results in support of the hypothesis that high "task involvement" (i.e., gift) leads to more search effort than does low "task involvement" (i.e., personal use).

Analysis of the Clarke and Belk Study

Internal validity. Ignoring the issue of involvement, which was mentioned above, does the Clarke and Belk study effectively manipulate gift-giving versus personal use? While no direct manipulation checks were reported, the fact that rather large and systematic differences were observed in the dependent variables across the two conditions leads one to believe that something was manipulated. The instructions were not fully presented so there is no basis for determining their "strength;" however, they do have face validity. It should be noted that Heeler, et al. used quite similar instructions; the difference here is that the experimental task of filling out a questionnaire is not as exotic as that of playing with an IDB. Thus, a very similar manipulation may have worked better for Clarke and Belk because the task was not so powerful as to overwhelm it. However, the manipulation is still very sterile and lacking in mundane realism. In particular, one wonders how often products like bubble bath, blankets, and even jeans are given as gifts by college coeds. Only record albums would seem to be reasonably common as gift items out of the four product categories studied. To the extent that the products are not ones commonly thought of in terms of gift-giving, motivations other than that implied by the manipulation may be entering into subjects' responses. Additionally subjects may find it difficult to provide adequate estimates of such behaviors.

External validity. As in the Heeler, et al. study, a very select subject population was investigated. To what extent are Clarke and Belk's results generalizable to their own subjects? With respect to sampling the product domain, Clarke and Belk used four products rather than only one. This is an improvement in quantity, but not necessarily in quality if the products are not representative of gift items typically selected by the subjects. Similar to Heeler, et al., Clarke and Belk did not specify either the person or the gift occasion, leaving that free to vary across subjects. Thus, the generalizability of the observed pattern over gift situations or recipients has not been established.

Measurement reliability and validity. Clarke and Belk's measures were not described in enough detail to understand them fully. All six appear to be assessed via single-item scales, leaving serious questions regarding their reliability and validity. In particular, the two "money" scales are questionable even in terms of their face validity as measures of search effort. It should also be noted that the six measures were probably fairly highly correlated and could possibly have been used to construct a single, more reliable measure of search effort.

In summary, examination of the Clarke and Belk study leaves little encouragement that their conclusion is generalizable to other people, other products, or other gift occasions.

Summary of the Two Empirical Studies

From the above description and analysis, it is impossible to resolve the conflict between the Heeler, et al. and Clarke and Belk studies. The two experiments used virtually identical instructions with respect to "personal use" versus "gift for friend," but the task environments and dependent measures were quite different. Neither study demonstrated adequate internal or external validity, nor were the measures utilized satisfactorily documented. In short, the results of neither study should in any way color subsequent investigators' thinking about the substantive phenomena of gift-giving behavior.

Of course, it is unrealistic to expect any single study to adequately sample person, product, recipient, and occasion domains. But it is the responsibility of researchers in the gift-giving arena to recognize these domains and their potential impacts on their studies' findings. Only when these domains have been adequately sampled will we have good descriptive information upon which to build a basis for truly understanding gift-giving behavior.

A Paradigm for Gift-Giving Research

Banks, in her paper, begins to move away from purely descriptive studies of gift-giving behavior and into the domain of explanatory constructs. The notions of reciprocity, identity formation, and interaction all seem to hold promise for future research in this area. However, none of these explanations appears in the paradigm presented later in the paper. In fact, the paradigm as presented seems devoid of any motivational constructs which might explain why the giver is buying the gift.

It is not entirely clear that consumer gift-giving behavior must be studied from the perspective of a paradigm different from that applied to other forms of consumer behavior. For instance, a general expectancy-times-value model should capture significant motives influencing gift-giving. None of the studies reported here today address the underlying determinants of gift-giving. For marketers to engage in actions designed to modify consumer gift-giving behaviors, a thorough understanding of these behaviors is necessary. Such an understanding must incorporate influences due to the type of gift occasion, the nature of the recipient, and intra-individual variables characterizing the giver.

Both of the empirical studies reported here operated at the level of brand choice. It would appear that two decisions more fundamental to the gift-giving domain are (1) the budget decision (i.e., how much to spend), and (2) product class selection. In fact, the gift occasion may be an extremely complex one in the sense of delineating product/market boundaries. For example, if the giver's primary requirement is that the gift in some way be unique, then the uniqueness attribute may be housed in an incredible number of alternative product forms. This goes back to the necessity for understanding giver motivations, for only then can the range of possible choice alternatives be identified.


In conclusion, the three papers in this session have added somewhat to our knowledge of gift-giving in a descriptive sense. Further, they have pointed out some directions for future gift-giving research of a more explanatory nature. Hopefully future research will pay more attention to motivational constructs and help to reveal the contents of the black box of gift-giving behavior.



Richard J. Lutz, University of California, Los Angeles


NA - Advances in Consumer Research Volume 06 | 1979

Share Proceeding

Featured papers

See More


Unintended Customer Consequences of Corporate Lobbying

Gautham Vadakkepatt, George Mason University
Kelly Martin, Colorado State University
Neeru Paharia, Georgetown University, USA
Sandeep Arora, University of Manitoba, Canada

Read More


R1. How Consumers Deal With Brand Failure-An Individual Differences Approach

Melika Kordrostami, California State University-San Bernardino
Elika Kordrostami, Rowan University

Read More


How Framing Donor Match as Collaboration Impacts Donation: The Importance of In-Context Field Experiments In Fundraising

Indranil Goswami, SUNY Buffalo
Oleg Urminsky, University of Chicago, USA

Read More

Engage with Us

Becoming an Association for Consumer Research member is simple. Membership in ACR is relatively inexpensive, but brings significant benefits to its members.