Skip to main content
NCBI home page
Wiley Open Access Collection logoLink to Wiley Open Access Collection
. 2025 May 7;17(3):e70030. doi: 10.1111/aphw.70030

The psychology of a thank‐you gift: Who gives it and why?

Shigehiro Oishi 1,✉, Minha Lee 2, Minkyung Koo 3, Youngjae Cha 1
PMCID: PMC12106001  PMID: 40342235

Abstract

Why do some people give a thank‐you gift, while others do not? What are the predictors of thank‐you gift‐giving? We conducted three studies to examine the relative predictive values of indebtedness, gratitude, fondness, and respect in gift‐giving. In Study 1 (N = 120), participants read two scenarios that depict two high school teachers and imagined how much gratitude, indebtedness, fondness, and respect they would feel toward each teacher and whether they would give a thank‐you gift upon graduation. A multiple regression analysis showed that indebtedness was the only predictor of a thank‐you gift. In Study 2 (N = 327, pre‐registered), we directly replicated the results of Study 1. In Study 3 (N = 561, pre‐registered), we asked participants to identify up to three high school teachers who wrote them recommendation letters and to state whether they gave them a thank‐you gift. Replicating Studies 1 and 2, a multilevel analysis showed that indebtedness uniquely predicted whom participants gave a thank‐you gift. A thank‐you gift seems to be best predicted by the emotion of indebtedness.

Keywords: a thank‐you gift, fondness, gratitude, indebtedness, respect

THE EMOTIONAL PREDICTORS OF A THANK‐YOU GIFT

Gift‐giving is an infrequent yet critical ingredient of most social relationships and is done on occasions ranging from birthdays, Mother's Day, and weddings, to graduations and anniversaries. Social scientists generally consider gift‐giving to be a universal behavior that helps build new relationships and cement existing ones (Dunn et al., 2008; Mauss, 1925/1990; Schwartz, 1967). The majority of gift‐giving research has centered on the characteristics of good vs. bad gifts (e.g., Belk, 1996; Galak et al., 2016), asymmetries between givers and recipients (e.g., Flynn & Adams, 2009; Kupor et al., 2017; Zhang & Epley, 2012), and relationship consequences (e.g., Aknin & Human, 2015; Dunn et al., 2008). Finally, many studies identified cultural norms and customs as predictors of who and when people exchange gifts (Joy, 2001; Mauss, 1925/1990). In contrast, it is unclear why some individuals give a gift, while others do not in less scripted situations than weddings and Mother's Day. More specifically, the literature on gift‐giving has not explored emotional predictors of gift‐giving in general and thank‐you gifts in particular. To advance our understanding of the psychology of gift‐giving, it is imperative to elucidate the emotional antecedents of gift‐giving and the role of gratitude and indebtedness, in particular.

The psychology of gift‐giving

Although emotional predictors of gift‐giving have not been explored, there is important work related to this issue. For instance, the economist Bradley Ruffle (1999) proposed a mathematical model of gift‐giving that included (hypothetical) emotional outcomes such as surprise, disappointment, embarrassment, and pride. In addition, using qualitative data, Wooten (2000) proposed a theoretical model regarding why gift‐givers often experience anxiety, while Givi and Galak (2017, 2019) empirically tested the roles of fear of getting the wrong gift and envy in choosing the right gift. Finally, Septianto et al. (2020) showed that powerful individuals are more likely to choose wine/chocolates as a gift when pride is emphasized in an advertisement, whereas powerless people tend to choose wine/chocolates as a gift when gratitude is emphasized. Unlike these articles, however, the present research explored the way in which emotions toward the potential recipient would predict whether an individual decided to give a thank‐you gift or not. This is an important context to investigate because the reciprocation of favors is the foundation of long‐term social relationships (Thibaut & Kelley, 1959).

Four emotional predictors of gift‐giving

According to previous research, the emotion that we would most expect to be associated with a thank‐you gift is gratitude, since a thank‐you gift, by definition, reflects the expression of thanks (Emmons, 2007). The most relevant study (Algoe et al., 2008) investigated the role of gratitude in the context of college sorority Big Sister weeks (when Big Sisters give gifts to Little Sisters). Little Sisters felt more gratitude when they thought that Big Sisters' gifts were thoughtful. Furthermore, Little Sisters' gratitude predicted their feelings of integration into the sorority and relationship quality with their Big Sisters one month later. This study suggests that gratitude should predict whether people give a thank‐you gift or not.

There is, however, another emotion closely associated with a thank‐you gift: indebtedness. Like gratitude, indebtedness is an interpersonal emotion that is experienced when something special is done for someone (Greenberg & Westcott, 1983; Watkins et al., 2006). A norm of reciprocity (Goulder, 1960) states explicitly that “people should help those who have helped them” (p. 170). Given that a favor evokes both gratitude and indebtedness in most cases (Greenberg & Westcott, 1983), the norm of reciprocity theory predicts that the feeling of indebtedness (the feeling that a person owes something to someone) should encourage the recipient of a benevolent act to repay their debt via a thank‐you gift (see Goranson & Berkowitz, 1966; Regan, 1971 for returning a favor).

Peng et al. (2018) conducted three studies to explicitly test distinct psychological mechanisms underlying indebtedness versus gratitude, using scenarios. They found that indebtedness leads more to a subjective sense of obligation to reciprocate the help one has received (but does not lead to proximity seeking, or seeking to be closer to the benefactor), whereas gratitude leads to proximity seeking (but does not lead to an obligation to reciprocate). What is not yet clear is whether gratitude or indebtedness would be more likely to give rise to giving a thank‐you gift, considering that a gift‐giving is an overt action that signals an interest in relationship formation and maintenance (Caplow, 1984; Schwartz, 1967). The sense of indebtedness may be a particularly strong predictor of gift‐giving behavior because the feeling of being indebted generally suggests wanting to return a favor. Obviously, gift‐giving is a concrete way of repaying someone.

Beyond the distinction between gratitude and indebtedness, previous research has explored the distinction between gratitude and other positive emotions such as amusement, happiness, joy, and contentment (e.g., Bartlett & DeSteno, 2006; McCullough et al., 2004; Watkins et al., 2003; Watkins et al., 2018). Yet, given that gratitude is a positive interpersonal emotion, it is also important to distinguish it from other positive interpersonal emotions. Fondness and respect are other positive interpersonal emotions (Fehr, 1988). Similar to gratitude, fondness and respect are also characterized by positive valence, other‐focused attention, and intrinsic motivation to engage in prosocial behaviors (Frei & Shaver, 2002). Because gratitude is an emotional reaction to another's action (Ortony et al., 1988), the grateful person is likely to feel fondness towards the other person as well. Fondness (liking) leads to interpersonal attraction (Aronson & Worchel, 1966) and is one of the most general and fundamental emotions in forming interpersonal relationships (Collins & Miller, 1994). Interpersonal attraction, furthermore, is related to more spontaneous gift‐giving in a romantic relationship (Belk & Coon, 1993). Thus, it is important to distinguish gratitude from the general positive interpersonal emotion of fondness.

Finally, a person experiences gratitude when another does something special and praiseworthy (Ortony et al., 1988). Thus, gratitude is likely to entail not only fondness but also appreciation of another person's praiseworthy action (Ortony et al., 1988). Such appreciation could be elevated to admiration and respect (Berger, 1975) when another person's action comes with some cost, as “respecting x (in the evaluative sense) is like esteeming x” (Hudson, 1980, p. 72). Respect is especially likely to be felt in the context of a hierarchical relationship, such as a teacher‐student and a parent–child relationship (Mullin, 2016), when a person in power does something special for them. A thank‐you gift, then, could be given out of respect that the giver feels toward the recipient. In sum, gratitude is closely intertwined with other interpersonal positive emotions such as fondness and respect.

Previous research has rarely examined thank‐you gifts in hierarchical relationships. This provides a rare opportunity to test the relative importance of four interrelated emotions (gratitude, indebtedness, fondness, and respect) that represent different theories of prosocial behaviors. Gratitude theories emphasize that gratitude alone should give rise to prosocial behaviors (Algoe, 2012; Tsang, 2006). The norm of reciprocity theories (Goulder, 1960) predict that a sense of indebtedness plays a central role in prosocial behaviors. The interpersonal attraction theories (Collins & Miller, 1994) assumes that liking will lead to prosocial behaviors. Finally, respect theories (Berger, 1975; Mullin, 2016) suggest that respect plays a key role in prosocial behaviors. Previous research tested only the relative importance of two theories (e.g., indebtedness vs. liking in Regan, 1971). The current research tested all four theories simultaneously in the context of a thank‐you gift, providing critical information regarding why some people give a thank‐you gift, but others do not. This investigation also helps provide a more nuanced understanding of the role of gratitude versus indebtedness and more general interpersonal emotions such as fondness and respect in the maintenance of social relationships.

The present research

We conducted three studies to provide a critical test of the conceptual distinctions between gratitude and indebtedness on one hand, and the distinctions between gratitude and fondness/respect on the other hand, in the context of thank‐you gifts. In Studies 1 and 2 (the latter pre‐registered), we manipulated the costly favor (letter writing) and tested which of the four emotions of gratitude, fondness, respect, and indebtedness, would predict the decision to give a thank‐you gift.

Study 3 (pre‐registered) was an ecological extension of Studies 1 and 2. Specifically, we prompted participants to list up to three college recommendation letter writers and asked the degree of gratitude, fondness, respect, and indebtedness they felt toward them at that time, as well as whether they wrote a thank‐you card and/or gave a thank‐you gift to each of them. In addition, we asked the degree to which they were still in touch with each of the letter writers. We expected that indebtedness would predict whether participants gave a thank‐you gift to a letter writer. In terms of relationship continuity, we predicted that gratitude, fondness, and respect would play a larger role than indebtedness. In sum, our studies will clarify whether gratitude plays a unique role in thank‐you gift‐giving, beyond closely related interpersonal emotions: indebtedness, fondness, and respect.

Previous cross‐cultural research found that, when helped by someone, East Asians tend to feel indebtedness more intensely than Americans (Hitokoto et al., 2008). Gratitude also evokes indebtedness to a greater degree among East Asians than among Americans (Oishi et al., 2019). These findings suggest that Asian students might feel more indebted toward a letter writer than other students, and therefore more likely to give a thank‐you gift to a letter writer than others. We tested these possibilities in our studies.

Across three studies, we report how we determined the sample size, exclusions (if any), all manipulations, and all measures. All studies were approved by the host institution's Institutional Review Board (Protocol Number: 3961). For all studies, the materials, data, codes, and outputs are openly available on the Open Science Framework repository at the following link: https://osf.io/rfbwg/?view_only=655483c6c09546a4b2de55c4b74fb82d.

STUDY 1: A THANK‐YOU GIFT TO A LETTER WRITER OR AN INSPIRING TEACHER

Previous research has shown that Americans feel grateful in a variety of circumstances (Emmons, 2007), but that they do not necessarily feel indebted in response to small favors (Hitokoto, 2016; Shen et al., 2011). In order to evoke a sense of indebtedness, it is critical for Americans to recall a situation in which someone engaged in a favor of magnitude or value to them. One easily identifiable situation of this sort is receiving a letter of recommendation. Virtually all American college students have had the experience of asking someone to write a recommendation letter for their college applications.

With this situation in mind, in Study 1, we created two gratitude‐evoking scenarios with varying degrees of indebtedness and interpersonal fondness/respect. Specifically, we described two teachers. One teacher wrote a recommendation letter (a costly favor) but did not inspire students as much as other teachers (low fondness, low respect). The other teacher did not write a recommendation letter (no costly favor) but inspired students (high fondness, high respect). We expected that either writing a recommendation letter or inspirational teaching would evoke gratitude because both teachers were doing something special. Thus, these two scenarios were intended to simulate two situations: (a) gratitude + indebtedness (letter writer) and (b) gratitude + fondness/respect (inspiring teacher). By manipulating these factors, we attempted to test the role of indebtedness in the decision to give a thank‐you gift, above and beyond related emotions.

METHOD

Participants

For this first study, we anticipated small to medium effect sizes (d = .3 t), typical of personality and social psychology (Gignac & Szodorai, 2016; Richard et al., 2003). An a priori power analysis using G*Power (Faul et al., 2007) showed that, in a within‐subject design, the appropriate sample size to yield a statistical power of .95 for the effect size of d = .35 is 109. To surpass this level of statistical power, we collected data from 120 participants and did not exclude any participants. They were all undergraduate students (75 males, 45 females) at a large Midwestern public university in the U.S. They participated in the study in exchange for partial course credit. Out of the 120 students, 70 self‐identified as European American, 34 self‐identified as Asian, 8 self‐identified as Hispanic, 2 self‐identified as African American, and 6 self‐identified as being of some other ethnic origin. Sensitivity power analyses using G*Power revealed that the final sample size had 80% power to detect an effect size of d = .26 to test a paired t‐test or one‐sample t‐test (α = .05, two‐tailed). We conducted all data analysis only after data from all of the samples had been collected. All the measures included in this study are reported below.

Materials and procedure

Participants first read the two scenarios (below), then answered a series of questions regarding the two teachers.

Teacher A

Think of a high school teacher whom you really liked and admired. You gained a lot of knowledge, and he/she changed the way you understand yourself and the world. You took only one class from this teacher, so you did not ask this teacher to write a recommendation letter. With the recommendation letters from others, you were able to get into most of the schools you applied to.

Teacher B

Think of a high school teacher who really enjoyed you as a student. Although you did not admire him/her as much as some other teachers, you took multiple courses with this teacher in four years. This teacher wrote a letter of recommendation for you. In the end, you were able to get into most of the schools you applied to.

After reading both scenarios, participants were asked to answer the following questions: first, how likely is it that you would give a thank‐you gift to Teacher A after graduation (1 = not at all likely to 7 = extremely likely)? How likely is it that you would give a thank‐you gift to Teacher B after graduation (1 = not at all likely to 7 = extremely likely)? On graduation day, if you were to give only one thank‐you gift, which teacher would you give the gift to, Teacher A or Teacher B? To what extent would you experience the following feelings toward Teacher A? Fondness, gratitude, respect, indebtedness (1 = not at all to 7 = extremely). To what extent would you experience the following feelings toward Teacher B? Fondness, gratitude, respect, indebtedness (1 = not at all to 7 = extremely).

We counterbalanced the order in which the teachers appeared. That is, 60 participants read about and answered all the questions about Teacher A first, whereas the other 60 read about and answered all the questions about Teacher B first. The order did not have any effect on degree of fondness, gratitude, respect, or indebtedness toward Teacher A or B, |t|s < 1.54, ps > .11, |d|s < .28. The order did not affect the likelihood of giving a thank‐you gift to Teacher A, t(118) = .37, p = .711, d = .07. However, the likelihood of giving a thank‐you gift to Teacher B was higher when they were rated first than second, t(118) = 2.02, p = .046, d = −.37. This means that we found one significant difference out of the 11 significance tests conducted (or 9.09%), which is not far from the typical Type I error rate of 5%. However, to be sure, we included the order in the hypothesis testing described below. The inclusion or exclusion of the order effect did not affect the key findings.

RESULTS AND DISCUSSION

First, we checked whether our manipulation worked. Our manipulation check showed that our manipulation worked as intended. Namely, a series of paired t‐tests showed that participants said that they would feel more fondness (M = 5.78, SD = 1.34 vs. M = 4.79, SD = 1.24), and respect (M = 6.17, SD = 1.04 vs. M = 5.47, SD = 1.17) toward inspiring Teacher A than toward letter writing Teacher B, t(119) = 6.27, p < .001, d = .57 1 ; t(119) = 6.25, p < .001, d = .57, respectively. Also as expected, participants reported that they would feel more indebted to Teacher B than to Teacher A (M = 4.35, SD = 1.55 vs. M = 5.23, SD = 1.45), t(119) = −5.52, p < .001, d = −.50. They felt the same amount of gratitude toward Teachers A and B (M = 5.58, SD = 1.33 vs. M = 5.64, SD = 1.26), t(119) = −.46, p = .65, d = −.04.

Hypothesis testing

Next, we tested whether participants said that they would be more likely to give a thank‐you gift to Teacher B than to give one to Teacher A. We tested our hypotheses in two ways: first by comparing the likelihood estimates for Teachers A and B, and second by examining the forced‐choice item between Teachers A and B. A paired t‐test showed that, as predicted, participants said that they would be more likely to give a thank‐you gift to letter‐writing Teacher B than to inspiring Teacher A (M = 5.22, SD = 1.47 vs. M = 4.79, SD = 1.71), t(119) = −2.64, p = .009, d = −.25. The forced‐choice item was tested against an even split using a one‐sample t‐test. It showed that participants were marginally more likely to give a thank‐you gift to Teacher B than to Teacher A (70 chose Teacher B, while 50 chose Teacher A), t(119) = 1.84, p = .068, d = .17. Thus, despite the fact that participants felt far more fondness and respect toward Teacher A than toward Teacher B, they were more likely to give a thank‐you gift to Teacher B than to Teacher A. Thus, we found that receiving a costly favor contributed more toward giving a thank‐you gift than did interpersonal fondness.

Table 1 shows simple correlations among the key variables. All four emotions were significantly associated with the likelihood of giving a thank‐you gift. Namely, the more fondness, respect, gratitude, and indebtedness participants imagined they would feel toward the teacher, the more likely they said they would be to give a thank‐you gift to the teacher.

TABLE 1.

Correlations among emotions and intention to give a thank‐you gift in study 1.

FondA RespA GratA IndeA FondB RespB GratB IndeB GiftA GiftB Gift AvsB
Fondness A ― .74 ** .69** .31** .10 .35** .37** .30** .24** .32** .04
Respect A ― .73** .32** .26** .39** .45** .27** .26** .30** −.01
Gratitude A ― .49 ** .11 .33** .24** .20* .30** .36** −.08
Indebtedness A ― .06 .04 −.03 .32** .32** .17+ −.18*
Fondness B ― .63 ** .59** .32** .14 .37** .28**
Respect B ― .61** .45** .20* .45 ** .24**
Gratitude B ― .49** .11 .47** .41**
Indebtedness B ― .06 .39** .34**
Open in a new tab

Note: GiftA indicates the likelihood of giving a thank‐you gift to Teacher A. GiftB indicates the likelihood of giving a thank‐you gift to Teacher B. Gift AvsB indicates the dichotomous choice of giving a thank‐you gift to Teacher A (1) vs. Teacher B (2).

*

p < .05.

**

p < .01.

As can be seen in Table 1, however, all four emotions were positively correlated with one another. Thus, we next tested the relative predictive value of the four emotions in predicting the likelihood of giving a thank‐you gift to letter‐writing Teacher B, relative to inspiring Teacher A, using a simultaneous regression analysis. Specifically, we regressed the likelihood estimate of giving a gift to Teacher B relative to Teacher A (i.e., the likelihood estimate of giving a gift to Teacher B – the likelihood estimate of giving a gift to Teacher A) on the fondness felt toward Teacher B relative to Teacher A (fondness toward Teacher B ‐ fondness toward Teacher A), gratitude felt toward Teacher B relative to Teacher A (gratitude toward Teacher B – gratitude toward Teacher A), respect felt toward Teacher B relative to Teacher A (respect toward Teacher B – respect toward Teacher A), and indebtedness felt toward Teacher B relative to Teacher A (indebtedness toward Teacher B – indebtedness toward Teacher A). We also included participants' gender (male = 1; female = 2), race (two race dummy codes: race dummy 1: European American = 1, the rest = 0; race dummy 2: Asian = 1, the rest = 0), and the scenario order (1 = A‐B order; 2 = B‐A order) to control for demographic variables and the order effect. 2 It should be noted that although the four emotions were significantly correlated with one another, multicollinearity was not likely to be an issue in our regression analyses below, as the variance inflation factor (VIF) ranged from 1.48 to 2.11 (well within the common recommendation for a VIF < 10, O'Brien, 2007) and tolerance ranged from .475 to .675 (well above the common recommendation for tolerance > .20, Menard, 1995).

As predicted, a greater degree of indebtedness felt toward Teacher B relative to Teacher A predicted a greater likelihood of giving a thank‐you gift to Teacher B relative to Teacher A, f 2  = .078 (see Table 2 for the full results). In contrast, fondness, gratitude, and respect were not associated with the greater likelihood of giving a thank‐you gift to Teacher B relative to Teacher A. There were no gender, racial, or order differences.

TABLE 2.

Two multiple regression analyses predicting likelihood of giving a thank‐you gift to letter writing teacher B relative to inspiring teacher A and a dichotomous choice (giving a thank‐you gift to teacher A vs. B) in study 1.

DV = likelihood DV = choice
B (SE) β t p B (SE) Wald p
Constant −.720 (.802) −.900 .371 −.024 (1.143) .000 .983
Order .479 (.305) .136 1.570 .119 .716 (.442) 2.620 .106
Gender −.167 (.321) −.046 −.518 .605 −.093 (.474) .039 .844
Race1 .667 (.481) .187 1.388 .168 −.644 (.765) .707 .400
Race2 .229 (.522) .059 .438 .662 −1.434 (.819) 3.065 .080
Fondness .039 (.114) .038 .344 .732 .076 (.174) .191 .662
Respect .076 (.150) .053 .506 .614 .062 (.211) .086 .769
Gratitude .005 (.138) .005 .038 .970 .342 (.227) 2.270 .132
Indebtedness .332 (.113) .331 2.953 .004 .496 (.177) 7.879 .005
R 2 (f 2 ) .178 (.146) .271 (.198)
Open in a new tab

Note: Likelihood = Likelihood of giving a gift to Teacher B – likelihood of giving a gift to Teacher A. In the logistic regression, the DV was the dichotomous choice (0 = Teacher A; 1 = Teacher B), thus higher numbers indicate a greater probability of choosing Teacher B. Race 1: European Americans = 1; other = 0. Race 2: Asians = 1, other = 0. Fondness, Respect, Gratitude, and Indebtedness all indicate respective ratings: Teacher B – Teacher A.

Finally, we conducted a logistic regression, regressing the dichotomous choice between Teachers A and B on the same set of predictor variables. Again, indebtedness felt toward Teacher B relative to Teacher A predicted the gift‐giving choice to Teacher B, whereas fondness, respect, and gratitude did not predict (see Table 2 for the full results). There were no gender, race, or order effects.

In sum, these analyses showed that indebtedness was uniquely associated with the likelihood of giving a thank‐you gift. Although fondness, respect, gratitude, and indebtedness were all associated with a greater likelihood of giving a thank‐you gift when examined individually, it was only indebtedness that predicted the willingness to give a thank‐you gift when all four factors were examined together. Combined with the fact that participants felt a great deal of gratitude toward both teachers, our findings suggest that it is indebtedness combined with gratitude that is crucial for prompting thank‐you gift‐giving.

STUDY 2: A PRE‐REGISTERED REPLICATION OF STUDY 1

There were two goals for Study 2. First, we tested to see if the findings from Study 1 would be replicated in a pre‐registered study. Second, because Study 1 did not have a sufficient number of Asian students to test potential cultural differences between Asians and European Americans, the secondary goal of Study 2 was to see if Asian students were more likely to give a gift to the teacher who did them a costly favor, over the inspiring teacher who did not write them a letter. We were not able to conduct a parallel cross‐cultural study of letter writers in East Asia, in part because in East Asian countries (e.g., Japan, Korea, China, Taiwan), college admission is based mostly on entrance exam scores and does not require recommendation letters. Thus, we collected data in the U.S.

METHOD

Participants

In Study 1, the average correlation between indebtedness and giving a thank‐you gift was .25. Furthermore, Study 1's effect size for Teacher B vs. A (paired t‐test) was d = .25, while f 2  = .078 for indebtedness in the multiple regression analysis. A priori power analyses using G*Power showed that N = 197 for the correlational analysis, N = 210 for the paired t‐test, and N = 169 for the multiple regression would provide 95% power. In addition, we wanted to test a potential cultural difference between Asians and European Americans. Assuming a modest difference (d = .40), 328 participants would provide 95% power. Thus, we aimed to obtain 330 participants for the pre‐registered replication study (https://osf.io/zq58w?view_only=72c8ac58b8db4ee29510f878172d2662w. We successfully obtained data from 327 students (157 men, 170 women) at a large public university in the U.S. They received partial course credit for participating in this study and we did not exclude any participants. Out of 327 participants, 128 self‐identified as White, 9 self‐identified as Black, 33 self‐identified as Hispanic, 146 self‐identified as Asian, 6 self‐identified as multiracial, and 5 chose “other.” Sensitivity power analyses using G*Power revealed that the final sample size had 80% power to detect an effect of d = .16 in a paired t‐test or one‐sample t‐test (α = .05, two‐tailed). All data analysis was run after the data collection was completed.

Procedures and materials

The procedures and materials were identical to those used in Study 1.

RESULTS AND DISCUSSION

First, we checked whether our manipulation worked. Like Study 1, our manipulation checks showed that our manipulation worked as intended. Namely, a series of paired t‐tests showed that participants reported feeling more fondness (M = 6.15, SD = 1.10 vs. M = 5.19, SD = 1.33, t[326] = 11.57, p < .001, d = .64) and respect (M = 6.31, SD = 1.00 vs. M = 5.87, SD = 1.17, t[326] = 6.85, p < .001, d = .38) toward inspiring Teacher A than toward letter writing Teacher B. Also as expected, participants reported that they would feel more indebted to Teacher B than to Teacher A (M = 5.07, SD = 1.70 vs. M = 5.57, SD = 1.36), t(326) = −4.86, p < .001, d = .27. Like in Study 1, they felt the same amount of gratitude toward Teachers A and B (M = 6.03, SD = 1.14 vs. M = 6.00, SD = 1.16), t(326) = .37, p = .71, d = .02.

Next, we tested whether participants said that they would be more likely to give a thank‐you gift to Teacher B than to Teacher A. Like in Study 1, participants said that they would be more likely to give a thank‐you gift to letter writing Teacher B than to inspiring Teacher A (M = 5.22, SD = 1.62 vs. M = 5.62, SD = 1.42), t(326) = −4.09, p < .001, d = −.23. The forced‐choice item also indicated that participants were more likely to give a thank‐you gift to Teacher B than to Teacher A (205 or 62.7% chose Teacher B, while 122 or 37.3% chose Teacher A); one sample t‐test against 50% (326) = 4.74, p < .001, d = .26.

In addition, we tested whether Asians were more likely to give a thank‐you gift to letter‐writing Teacher B over inspiring Teacher A. Indeed, Asians were significantly more likely to give a thank‐you gift to Teacher B (71.2%) than to Teacher A (28.8%). This was to a greater extent than European Americans, of whom 58.6% said they would give the gift to Teacher B and 41.4% said they would give it to Teacher A, χ 2 (1, 274) = 4.81, p = .028, Cohen's w = .13.

Table 3 shows simple correlations among the key variables. Replicating Study 1, all four emotions were significantly associated with the likelihood of giving a thank‐you gift. We next tested the relative predictive value of the four emotions in predicting the likelihood of giving a thank‐you gift to letter‐writing Teacher B relative to inspiring Teacher A, using a simultaneous regression analysis. Like in Study 1, multicollinearity was not likely to be an issue in our regression analyses below, as the variance inflation factor (VIF) ranged from 1.03 to 1.74 (well within the common recommendation for a VIF < 10, O'Brien, 2007) and tolerance ranged from .58 to .97 (well above the common recommendation for tolerance > .20, Menard, 1995).

TABLE 3.

Emotions and intention to give a thank‐you gift to teachers A and B in study 2.

FondA GratA RespA IndeA FondB GradB RespB IndeB GiftA GiftB Gift AvsB
Fondness A ― .64** .66** .28** .25** .46** .46** .40** .29** .26** −.07
Gratitude A ― .75** .45** .27** .32** .40** .30** .33** .20** −.17**
Respect A ― .35** .32** .42** .42** .28* .26** .27** −.06
Indebtedness A ― .23** .05 .15** .28** .34** .06 −.25**
Fondness B ― .58** .55** .43** .21** .35** .25**
Gratitude B ― .73** .48** .16* .43** .29**
Respect B ― .43** .18** .36** .19**
Indebtedness B ― .13* .42** .26**
Open in a new tab

Note: Gift A indicates the likelihood of giving a thank‐you gift to Teacher A. Gift B indicates the likelihood of giving a thank‐you gift to Teacher B. Gift A vs B indicates the dichotomous choice of giving a thank‐you gift to Teacher A (1) vs. Teacher B (2).

*

p < .05.

**

p < .01.

Replicating Study 1, greater indebtedness felt toward Teacher B relative to Teacher A predicted a greater likelihood of giving a thank‐you gift to Teacher B relative to Teacher A, f 2  = .125, whereas fondness, gratitude, and respect were not. There were no main effects of gender, race, or order (Table 4).

TABLE 4.

Two multiple regression analyses predicting likelihood of giving a thank‐you gift to letter writing teacher B relative to inspiring teacher A and a dichotomous choice (giving a thank‐you gift to teacher A vs. B) in study 2.

DV = likelihood DV = choice
B (SE) β t p B (SE) Wald p
Constant −.061 (.307) −.199 .842 .305 (.463) .435 .510
Order .198 (.183) .056 1.081 .280 .356 (.280) 1.617 .203
Female .018 (.181) .005 .098 .922 −.198 (.277) .510 .475
Race1 −.040 (.098) −.021 −.404 .686 −.255 (.152) 2.827 .093
Race2 .079 (.083) .050 .952 .342 .044 (.123) .131 .718
Fondness .088 (.065) .075 1.346 .179 .349 (.109) 10.330 .001
Respect −.072 (.092) −.048 −.782 .435 −.086 (.156) .305 .581
Gratitude .149 (.087) .113 1.704 .089 .519 (.162) 10.223 .001
Indebtedness .337 (.054) .355 6.275 < .001 .517 (.099) 27.231 < .001
R 2 (f 2 ) .190 (.154) .381 (.236)
Open in a new tab

Note: Likelihood = Likelihood of giving a gift to Teacher B – likelihood of giving a gift to Teacher A. In the logistic regression, the DV was the dichotomous choice (0 = Teacher A; 1 = Teacher B), thus the higher numbers indicate a greater probability of choosing Teacher B. Race 1: European Americans = 1; Asians = −1; others = 0. Race 2: European Americans = 1; Asians = 1, other = −2. Fondness, Respect, Gratitude, and Indebtedness all indicate respective ratings: Teacher B – Teacher A.

Finally, we conducted a logistic regression, regressing the dichotomous choice between Teachers A and B on the same set of predictor variables. Again, replicating Study 1, indebtedness felt toward Teacher B relative to Teacher A predicted the choice of Teacher B. Unlike Study 1, fondness and gratitude felt toward Teacher B relative to Teacher A also predicted giving the gift to Teacher B over Teacher A (see Table 4 for the full results). There were no gender differences. Asian participants were marginally more likely to give the gift to Teacher B over Teacher A above and beyond gender, order, and the four emotions.

In sum, Study 2 replicated key findings from Study 1. Although fondness, respect, gratitude, and indebtedness were all associated with a greater likelihood of giving a thank‐you gift when examined individually, indebtedness consistently predicted the willingness to give a thank‐you gift when all four factors were examined together, in both likelihood ratings and the dichotomous choice. We found some evidence that Asian participants were more likely than their European American counterparts to give a thank‐you gift to a person who did them a costly favor than to a person they admire.

STUDY 3: A THANK‐YOU GIFT TO LETTER WRITERS?

There were two main limitations of Studies 1 and 2. First, they utilized a scenario method. Although the scenario method allowed us to control for various factors and isolate the effect of indebtedness on gift‐giving, it was hypothetical. It was thus unclear whether students actually do give thank‐you gifts to people they feel a greater degree of indebtedness toward. Second, Studies 1 and 2 explored only gift‐giving. Many people thank others by writing a thank‐you card. Unlike a thank‐you gift, sending a thank‐you card (without a gift) is a pure expression of thanks without any explicit monetary repayment.

To address these limitations, in Study 3, we asked participants to identify up to three high school teachers who wrote them recommendation letters and to state whether they had given them thank‐you gifts and/or cards and whether students stayed in touch with letter writers after graduating from high school. We expected that fondness, respect, and gratitude would predict whether students maintained relationships with letter writers, because fondness, respect, and gratitude are positive interpersonal emotions that signal interest in the continuation of relationships, whereas indebtedness might predict repayment in the form of a gift.

In sum, we conducted Study 3 to extend Studies 1 and 2 to real‐life gift‐giving, while assessing other modes of thanking (i.e., card writing) as well as another important relationship outcome (the continuation of the relationship).

METHOD

Participants

Studies 1 and 2's main effect size for indebtedness predicting a thank‐you gift was r = .37, whereas the effect size for the relative likelihood of giving a thank‐you gift to the letter‐writing teacher vs. the inspiring teacher was d = .25. G*Power analysis showed that 84 participants would provide 95% power to detect the effect size r = .37, and that information about whether cards or gifts had been given to 892 letter writers would be required to detect the effect size of d = .25. Assuming that participants would list 2.5 letter writers on average, we would need at least 357 participants to gather information about 892 letter writers in total. We first collected data from 180 students at a large Southern public university and then recruited 425 more students at a large Midwestern public university to boost the final sample size. All the students participated in this study in exchange for course credit. Out of the 425 students who took the survey at the Midwestern university, 44 were excluded because they did not mention any letter writers (the majority of them were transfer students who were not required to turn in recommendation letters at this university). In sum, our final sample from the two universities was 561, with information about 1,471 letter writers (i.e., 2.62 letter writers per respondent). The final sample consisted of 369 females, 191 males, and 1 not specified; 301 self‐identified as White, 19 as Black, 31 as Hispanic, 182 as Asian, and 26 as “other” (2 did not specify). The mean age was 20.32 (SD = 1.55).

Procedures and materials

The materials for Study 3 were pre‐registered (for the masked material files for this submission, see Materials folder at https://osf.io/rfbwg/?view_only=655483c6c09546a4b2de55c4b74fb82d; the original pre‐registration was done in 2019). Participants were first asked to list up to three people who wrote letters of recommendation when they applied for colleges. For each person, we then asked them to indicate how much they felt each of the following emotions toward the letter writer on a 9‐point scale (1 = not at all; 9 = extremely): fondness, gratitude, respect, and indebtedness. Then, we asked them to indicate (1) whether they subsequently wrote a thank‐you card to any of them (a Yes/No response for each letter writer), and, if yes, roughly how much time they spent writing the card in minutes; (2) whether they sent a thank‐you gift (a Yes/No response for each letter writer), and, if yes, roughly how much they spent on the gift in dollars (for each letter writer). Then, we asked whether they were still in touch with each of the letter writers on a 5‐point scale (1 = completely lost touch, 2 = rarely contact, 3 = contact sometimes, 4 = contact regularly, 5 = contact frequently). The participants then completed demographic information. All the measures included in this study are reported below.

Minutes spent on writing a thank‐you card (M = 3.16, SD = 8.73, range = 0 to 120) and dollars spent on a thank‐you gift (M = 3.35, SD = 11.30, range = 0 to 200) were severely positively skewed: skewness = 7.09, kurtosis = 74.67 for the thank‐you card; skewness = 9.95, kurtosis = 144.03 for the thank‐you gift. According to West et al. (1995), skewness greater than 2 and kurtosis greater than 7 are considered severe deviations from normality. Thus, we log‐transformed these two variables, which removed the severe violations of the distributional assumption (skewness = 1.33; kurtosis = .72 for the thank‐you card; skewness = 1.57, kurtosis = 1.42 for the thank‐you card). The analyses regarding the amount of time spent on the thank‐you card and the amount of money spent on the thank‐you gift were conducted using the transformed variables.

RESULTS AND DISCUSSION

Although we pre‐registered study materials, we did not pre‐registered the data analytic plan. We followed the same analytic procedures used in Studies 1 and 2 (correlation analyses, followed by multiple regression analyses), except that we employed multilevel modeling in Study 3 due to the nested nature of the data. To correct for the non‐independence of standard errors, we conducted multilevel modeling using HLM 6.04. Specifically, the giving of a thank‐you gift (dichotomous variable) was predicted from fondness, gratitude, respect, and indebtedness, each analyzed separately at Level 1 (within‐individual). Each emotion was group‐centered (centered around each participant's own mean). No Level 2 predictors were included. In other words, we ran four separate HLM analyses, each with a different emotion as a predictor variable. This is conceptually equivalent to running simple correlation analyses in Studies 1 and 2. Because the dependent variable was dichotomous, we used the Bernoulli model of HLM. Finally, HLM outputs included significance tests using both standard errors generated by the restricted maximum likelihood estimator, and robust standard errors generated by the Huber/White formula. The HLM results reported below are based on the robust standard errors, as they are less vulnerable to the violation of the homoscedasticity assumption (the assumption that errors are normally distributed), according to Raudenbush and Bryk (2002). However, the results using the standard errors were similar to those using the robust standard errors (for full results, see supplementary materials in Syntax and Outputs folder at https://osf.io/rfbwg/?view_only=655483c6c09546a4b2de55c4b74fb82d).

Overall, participants gave thank‐you gifts to 436 out of 1,471 letter writers (29.6%). Consistent with Studies 1 and 2, when examined separately, the intensity of fondness (b = .49, SE = .06, t(1449) = 7.98, p < .001, Odds Ratio (OR) = 1.63, 95% CI [1.44; 1.83]), respect (b = .58, SE = .08, t(1445) = 7.72, p < .001, OR = 1.79, 95% CI [1.54, 2.07]), gratitude (b = .54, SE = .08, t [1445] = 7.16, p < .001, OR = 1.72, 95% CI [1.48, 1.99]), and indebtedness all predicted whether participants gave a thank‐you gift (b = .51, SE = .07, t[1437] = 6.90, p < .001, OR = 1.66, 95% C. I [1.44, 1.92]).

Participants wrote thank‐you cards to 507 out of 1,460 letter writers (34.7%). We next tested whether fondness, respect, gratitude, and indebtedness were associated with sending a thank‐you card, respectively. Like the thank‐you gift, the intensity of fondness (b = .37, SE = .06, t(1443) = 6.19, p < .001, OR = 1.45, 95% CI [1.29, 1.63]), respect (b = .46, SE = .07, t(1439) = 6.61, p < .001, OR = 1.59, 95% CI [1.38; 1.82]), gratitude (b = .42, SE = .07, t[1439] = 6.16, p < .001, OR = 1.52, 95% CI [1.33; 1.74]), and indebtedness all predicted whether participants wrote a thank‐you card (b = .35, SE = .06, t[1431] = 5.66, p < .001, OR = 1.42, 95% CI [1.26; 1.60]).

Hypothesis testing

Thank‐you gift

Next, we tested which of the four emotions would predict giving a thank‐you gift by including all four emotions simultaneously. Although the four emotions were significantly correlated with one another, multicollinearity was not likely to be a major issue in our analyses below, as the variance inflation factor (VIF) ranged from 1.56 to 5.23 (well within the common recommendation for a VIF < 10, O'Brien, 2007) and tolerance ranged from .19 to .64 (mostly above the common recommendation for tolerance > .20, Menard, 1995).

As in Studies 1 and 2, we included participant gender and race as control variables. Specifically, the intercept and four slopes from Level 1 (within‐participant) were predicted by gender (female = 1; male = 0) and race dummy codes in which European Americans were the reference group.

Table 5 shows the results of this analysis. Replicating Study 2, Asian participants were more likely to give a thank‐you gift than European American participants. Also, replicating Studies 1 and 2, indebtedness was the only significant predictor of whether participants gave a thank‐you gift to the recommendation letter writer. Gratitude, fondness, and respect did not predict whether participants gave a thank‐you gift to the letter writer.

TABLE 5.

Results of HLM analysis: predicting a thank‐you gift from four emotions.

Fixed effect Coefficients (SE) t‐value p‐value Odds ratio (95% CI)
For intercept, P0
Intercept, B00 −1.41 (.22) −6.45 .00 .24 (.16; .38)
Female, B01 .19 (.23) .83 .41 1.21 (.77; 1.90)
Black, B02 −.14 (.65) −.22 .82 .87 (.24; 3.09)
Asian, B03 .50 (.24) 2.13 .03 1.66 (1.04; 2.63)
Hispanic, B04 .03 (.49) .06 .95 1.03 (.39; 2.71)
Other, B05 −.35 (.55) −.64 .52 .70 (.24; 2.06)
For fondness slope, P1
Intercept, B10 .19 (.13) 1.43 .15 1.21 (.93; 1.56)
Female, B11 .21 (.17) 1.26 .21 1.23 (.89; 1.71)
Black, B12 −.72 (.36) −2.00 .05 .49 (.24; .99)
Asian, B13 −.44 (.19) −2.33 .02 .64 (.44; .93)
Hispanic, B14 −.85 (.40) −2.13 .03 .43 (.20; .93)
Other, B15 .10 (.31) .33 .74 1.11 (.60; 2.03)
For gratitude slope, P2
Intercept, B20 −.09 (.23) −.40 .69 .91 (.58; 1.44)
Female, B21 .08 (.27) .28 .78 1.08 (.63; 1.84)
Black, B22 .51 (.36) 1.39 .16 1.66 (.81; 3.38)
Asian, B23 .42 (.29) 1.45 .15 1.52 (.86; 2.68)
Hispanic, B24 .45 (.48) .92 .36 1.56 (.60; 4.04)
Other, B25 −.08 (.60) −.13 .90 .92 (.28; 3.01)
For respect slope, P3
Intercept, B30 .14 (.17) .81 .42 1.15 (.82; 1.61)
Female, B31 .19 (.20) .92 .36 1.21 (.81; 1.80)
Black, B32 −.30 (.28) −1.09 .28 .74 (.43; 1.27)
Asian, B33 −.19 (.22) −.88 .38 .82 (.53; 1.27)
Hispanic, B44 .17 (.46) .38 .71 1.19 (.48; 2.91)
Other, B45 .66 (.84) .78 .43 1.94 (.37; 10.10)
For indebtedness slope, P4
Intercept, B40 .41 (.16) 2.58 .01 1.50 (1.10; 2.04)
Female, B41 −.48 (.20) −2.33 .02 .62 (.42; .93)
Black, B42 .38 (.50) .76 .45 1.47 (.55; 3.93)
Asian, B43 .29 (.21) 1.39 .16 1.34 (.89; 2.01)
Hispanic, B44 .33 (.41) .81 .42 1.39 (.63; 3.10)
Other, B45 .35 (.68) .51 .61 1.42 (.37; 5.41)
Open in a new tab

Note: Degrees of freedom for Intercept 1, P0 were 550, whereas degrees of freedom for the 4 slopes were 1,405.

There were some moderations by gender and race. First, indebtedness was more strongly associated with a thank‐you gift among men than women. Second, fondness was more strongly associated with a thank‐you gift among European Americans than African Americans, Asians, and Hispanics.

Next, we tested whether the log‐transformed amount spent on gifts (in dollars) was predicted by fondness, respect, gratitude, and indebtedness and whether there were gender and cultural differences in the likelihood of giving a thank‐you gift. As the dependent variable was continuous, we used the default restricted maximum likelihood estimator. Asian participants spent more money on thank‐you gifts than European American participants (b 03  = .21, SE = .10, t[1358] = 2.16, p = .031, d = .12). Indebtedness was the only significant predictor of the amount spent on thank‐you gifts (b 40  = .12, SE = .05, t[1358] = 2.37, p = .018, d = .13). Fondness was marginally associated with the amount spent (b 10  = .06, SE = .04, t[1358] = 1.65, p = .099, d = .09), while respect (b 30  = .03, SE = .05, t[1358] = .61, p = .541, d = .03) and gratitude (b 20  = −.07, SE = .07, t[1358] = −.95, p = .345, d = −.05) were not. Indebtedness was also associated more with the amount spent on thank‐you gifts among men than women (b 41  = −.17, SE = .07, t[1358] = −2.44, p = .015, d = −.13).

Thank‐you card

We ran the same Bernoulli model for thank‐you cards instead of thank‐you gifts. Whereas Asian participants were more likely to give thank‐you gifts than European American participants, they were less likely to write thank‐you cards than European Americans (b 03  = −.54, SE = .26, t[1399] = −2.08, p = .038, OR = .58, 95% CI [.35; .97]). Like thank‐you gifts, indebtedness was the only predictor of writing thank‐you cards: the more indebtedness participants felt toward a letter writer, the more likely they were to write a thank‐you card (b 40  = .37, SE = .15, t[1399] = 2.51, p = .012, OR = 1.44, 95% CI [1.08, 1.92]). Fondness (t = 1.50, p = .133), gratitude (t = .21, p = .837), and respect (t = −1.21, p = .228) did not predict whether participants wrote a thank‐you card. In terms of moderators, fondness was associated more strongly with writing a thank‐you card among European Americans than Asians (b 13  = −.38, SE = .18, t[1399] = −2.13, p = .033, OR = .68, 95% CI [.48, .97]). In contrast, respect was more strongly associated with writing a thank‐you card among Asians than European Americans (b 33  = .54, SE = .23, t[1399] = 2.34, p = .019, OR = 1.72, 95% CI [1.09, 2.70]). Likewise, respect was more strongly associated with giving a thank‐you gift among African Americans than European Americans (b 32  = .91, SE = .29, t[1399] = 3.17, p = .002, OR = 2.49, 95% CI [1.42, 4.38]). In addition, indebtedness was more strongly associated with writing a thank‐you card among men than women (b 41  = −.55, SE = .17, t[1399] = −3.15, p = .002, OR = .58, 95% CI [.41; .81]), whereas respect was more strongly associated with it among women than men (b 31  = .47, SE = .22, t[1399] = 2.15, p = .031, OR = 1.60, 95% CI [1.04; 2.45]).

In terms of log‐transformed time spent on writing the thank‐you card, indebtedness was predictive (b 40  = .12, SE = .05, t[1380] = 2.40, p = .017, d = .13), whereas the other three emotions were not (ps > .100). Fondness was more strongly associated with time spent writing a thank‐you card among European Americans than Asians (b 13  = −.12, SE = .05, t[1380] = −2.24, p = .025, d = −.12). Finally, respect was more strongly associated with time spent writing a thank‐you card among Asians than European Americans (b 33  = .16, SE = .08, t[1380] = 2.10, p = .035, d = .11).

Relationship continuation

Finally, we explored which emotions might be associated with the continuation of the relationship with the letter‐writer. Women were marginally more likely to stay in touch with the letter writer than men (b 01  = .14, SE = .08, t[541] = 1.78, p = .076, d = .15). Asians were less likely to stay in touch with the letter writer than European Americans (b 03  = −.18, SE = .08, t[541] = −2.27, p = .023, d = −.20). Like the thank‐you gift and card, indebtedness was associated with relationship continuation (b 40  = .23, SE = .06, t[1377] = 3.99, p < .001, d = .22). In addition, the more respect someone felt toward a letter‐writer, the more likely they were to have stayed in touch with them (b 30  = .19, SE = .08, t[1377] = 2.38, p = .018, d = .13). Gratitude was more strongly associated with relationship continuation among European Americans than Hispanics (b 24  = −.77, SE = .36, t[1377] = −2.12, p = .034, d = −.11).

In sum, Study 3 largely replicated the key findings from Studies 1 and 2: indebtedness was the only predictor of giving a thank‐you gift, and Asian participants were more likely than others to give a thank‐you gift to letter writers. Furthermore, indebtedness predicted whether participants wrote a thank‐you card to a letter‐writer as well as whether they were still in touch with them.

GENERAL DISCUSSION

Existing theories on prosocial behaviors provide different predictions regarding who gives a thank‐you gift. Gratitude theories (e.g., Algoe, 2012; Tsang, 2006) predict that the more intensely a person feels gratitude toward a recommendation writer, the more likely that they give a thank‐you gift to the recommendation writer. The norm of reciprocity theory (Goulder, 1960) suggests that indebtedness should predict a thank‐you gift, whereas interpersonal attraction theories (Collins & Miller, 1994) suggest that fondness should predict it instead. Finally, respect theories (Berger, 1975; Mullin, 2016) suggest that respect might be a strong motivation for a thank‐you gift. We conducted three studies to explore the emotional predictors of a thank‐you gift, focusing in particular on gratitude, indebtedness, and two positive interpersonal emotions (i.e., fondness and respect).

In Studies 1 and 2, we manipulated the receipt of a costly favor (receiving a recommendation letter) and interpersonal liking using a scenario method. Teacher A was inspiring and admirable but the student did not ask for a recommendation letter from them, whereas Teacher B was not inspiring but the student asked them for a letter. As expected, participants felt more fondness and respect toward Teacher A than Teacher B, and more indebtedness toward Teacher B than Teacher A (with no difference in the degree of gratitude felt toward them). As predicted, participants were more likely to give a thank‐you gift to the letter‐writing Teacher B than the inspiring Teacher A. When all four emotions were simultaneously included in regression analyses, it was only indebtedness that was associated with the likelihood of giving a thank‐you gift.

In order to increase ecological validity, in Study 3, we asked participants to list the teachers who actually wrote them recommendation letters when they were applying for college and indicate whether they sent them a thank‐you card and/or gift, as well as how much they kept in touch with them. Examined together, indebtedness was again the only predictor of giving a thank‐you gift or card. Thus, using two divergent methods, we found that indebtedness is a critical emotional predictor of giving a thank‐you gift, at least in a teacher‐student relationship. Given that a thank‐you gift exemplifies the norm of reciprocity in action, our findings suggest that the norm of reciprocity (Goulder, 1960), more specifically, the recognition of a favor received and the need to repay it, plays a powerful role in a teacher‐student relationship. In many hierarchical relationships, ranging from parent–child to supervisor‐supervisee, indebtedness might be an understudied, but important emotion that helps predict engagement in reciprocal exchanges in their everyday lives.

Conceptual distinctions between gratitude and indebtedness

Our findings provide further support for the presence of an important distinction between gratitude and indebtedness in a new context: the giving of thank‐you gifts. As seen in Study 3, American college students are highly grateful to the people who write them college recommendation letters. However, only 30% of them expressed their gratitude in the form of a thank‐you gift, and 35% in the form of a thank‐you card. In this normatively gratefulness‐eliciting situation, it was indebtedness that successfully distinguished those who gave a thank‐you gift or card from those who did not.

In the past, prosocial aspects of gratitude have been emphasized (e.g., DeSteno et al., 2019; Dickens & DeSteno, 2016; Kates & DeSteno, 2021; Lee et al., 2019; Tsang, 2006). However, the previous studies that have linked gratitude to prosocial behaviors did not test whether indebtedness also plays a role. 3 Our findings raise a possibility that previous research on gratitude might have overstated its role in promoting prosocial behaviors. That is, a sense of indebtedness toward a target person might have also played a role in galvanizing various prosocial behaviors, above and beyond gratitude, and gratitude alone (without indebtedness) might not have produced such prosocial behaviors.

Another important finding was that gratitude did not predict the degree to which participants remained in contact with letter writers above and beyond the other emotions surveyed (fondness, respect, and indebtedness), whereas respect and indebtedness did. Consistent with previous theorizing (e.g., Algoe, 2012; Lambert et al., 2010), gratitude was indeed associated with the frequency with which participants contacted letter writers when analyzed separately. 4 However, once all four emotions were examined together, it was indebtedness and respect that uniquely predicted relationship continuation.

Our findings, then, suggest that it might not be gratitude per se, but indebtedness and respect combined with gratitude that are critical in motivating the choice to continue a relationship with a benefactor, at least among American college students. Gratitude is a response to specific favors that a target person has done, especially in the context of teacher‐student relationships. As the memory of a favor fades away with time (Converse & Fishbach, 2012), gratitude might not be as strong a predictor of relationship continuation as indebtedness and respect in the long run. In a rare case where a beneficiary feels gratitude but does not necessarily indebtedness or respect toward a benefactor, the current findings predict that the beneficiary might not seek relationship continuation, despite being grateful. However, this pattern might be specific to teacher‐student relationships. Because this is a critical aspect of theories of gratitude, it will be essential to distinguish fondness and respect from gratitude in the context of other close relationships in the future.

The cultural psychology of gratitude, indebtedness, and respect

Our findings also have important implications for the cultural psychology of emotion (Cohen, 2003; Mesquita & Leu, 2007; Tsai & Clobert, 2019). Previous research has found cultural differences in the predictors (Hitokoto et al., 2008) and outcomes (Shen et al., 2011) of indebtedness. Cultural psychologists have long theorized that East Asians are more prone than others to feeling indebtedness (Markus & Kitayama, 1991). To the extent that indebtedness is associated with giving thank‐you gifts, then, it is not surprising that our Asian participants were more likely than others to give a thank‐you gift to letter writers.

Recent research has identified various cultural differences in gratitude (e.g., Corona et al., 2020). For instance, gratitude is more strongly intertwined with indebtedness among Koreans than among Americans (Oishi et al., 2019). That is, when participants were asked to think about things they were grateful for, Koreans were more likely than Americans to feel indebtedness. Another recent study also found that Taiwanese express their gratitude via self‐improvement (e.g., by attempting to make themselves better people), whereas Americans express their gratitude behaviorally (e.g., by giving a hug, Chang & Algoe, 2020). In Study 3, we did not find any cultural differences in the degree to which gratitude is associated with giving a thank‐you gift or card, or with relationship continuation. Instead, we found that fondness was a uniquely important predictor of giving a thank‐you card among European Americans, more than among Asians, African Americans, or Hispanics. In the European American cultural context, writing a thank‐you card might be construed as a genuine expression of fondness. In contrast, respect was more strongly associated with giving a thank‐you card among Asians and African Americans than among European Americans. These associations appear to reflect the role of particular emotional and cultural scripts, norms, and ideals. They need to be explicated in a variety of settings in the future.

Despite the fact that Asians went the extra mile to give thank‐you gifts to their letter writers, they were less likely to stay in touch with them once they entered college. Again, this is consistent with the view of indebtedness as a sense of obligation to repay a debt. By giving a thank‐you gift, Asians might have felt that the debt had been repaid and thus that there was no need to stay in touch with their letter writers. In this sense, Shen et al.’s (2011) findings are instructive. They found that Asians refused to receive a small gift from a casual acquaintance because they did not want to feel indebted to them (i.e., feel obliged to repay it in the future). It is possible that Asians give thank‐you gifts to quickly dispense with the obligation to repay debts so that they can move on with their daily lives, unburdened. Likewise, it is also well‐known that Asians do not seek social support as much as European Americans do (Kim et al., 2008). This might also be motivated, in part, by Asians' efforts to avoid accumulating debts toward others. In other words, East Asians may feel burdened if they stay in relationships with one‐time benefactors, such as letter writers, because many of their built‐in relationships (e.g., with family) are obligation‐ridden and they constantly feel that they must repay others. These hypotheses need to be tested explicitly in the future.

Limitations and conclusion

It is important to recognize some limitations of our research. First, our participants were convenience samples of American college students. Future research should investigate whether our findings can be generalized to non‐college student populations. Second, our studies focused on teacher‐student relationships. It is important to examine emotional predictors of thank‐you gifts in other relationship contexts such as romantic relationships and friendships. Third, we used single‐item measures to reduce the burden on participants (as they were asked to rate their feelings toward each teacher). It is unlikely that differential reliability across the four emotions accounted for our results about the predictive value of indebtedness (one must assume that indebtedness was measured with far more reliability than gratitude, fondness, or respect). Future research should use multiple items and structural equation modeling to correct measurement errors.

Despite these limitations, the current research adds unique insight into the distinction between gratitude and indebtedness, in addition to the distinction between gratitude and other positive interpersonal emotions such as fondness and respect, in predicting thank‐you gifts and relationship continuation. Although indebtedness and respect have not received as much research attention as gratitude, they are important interpersonal emotions that have widespread implications for interpersonal behaviors such as gift‐giving. It is crucial to examine indebtedness and respect along with gratitude in the future.

AUTHOR CONTRIBUTIONS

S.O conceived study ideas, designed the study, analyzed the data, and wrote a first draft of the paper. M.K. pre‐registered Studies 2 and 3. M. L and M. K collected data, read and revised the paper. Y.C. checked data analyses, read and revised the paper.

CONFLICT OF INTEREST STATEMENT

The authors declare that they have no conflict of interests.

ACKNOWLEDGMENTS

We would like to thank Liz Gilbert, Jane Derk, Erin Westgate, and Kate Kaufman for their invaluable comments on earlier versions of this paper.

Oishi, S. , Lee, M. , Koo, M. , & Cha, Y. (2025). The psychology of a thank‐you gift: Who gives it and why? Applied Psychology: Health and Well‐Being, 17(3), e70030. 10.1111/aphw.70030

ENDNOTES

1

The effect size ds for the paired t‐tests were calculated using the pooled standard deviation and the correction for the correlation between two measures, using www.psychometrica.de/effect_size.html.

2

We added gender as previous research (e.g., Dunn et al., 2008) found gender differences in the context of gift giving. We also added race as previous research showed that there are racial differences in propensity to engage in gift giving (e.g., Garner & Wagner, 1991). However, excluding these demographic variables did not change the main findings.

3

It should be noted that Tsang (2006) did measure various reasons for donating money to the experimental partner, including obligation.

4

r(1422) = .28, p < .001.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are openly available in The Emotional Predictors of a Thank‐you Gift at https://osf.io/rfbwg/?view_only=655483c6c09546a4b2de55c4b74fb82d.

REFERENCES

  1. Aknin, L. B. , & Human, L. J. (2015). Give a piece of you: Gifts that reflect givers promote closeness. Journal of Experimental Social Psychology, 60(8), 8–16. 10.1016/j.jesp.2015.04.006 [DOI] [Google Scholar]
  2. Algoe, S. B. (2012). Find, remind, and bind: The functions of gratitude in everyday relationships. Social and Personality Psychology Compass, 6(6), 455–469. 10.1111/j.1751-9004.2012.00439.x [DOI] [Google Scholar]
  3. Algoe, S. B. , Haidt, J. , & Gable, S. L. (2008). Beyond reciprocity: Gratitude and relationships in everyday life. Emotion, 8(3), 425–529. 10.1037/1528-3542.8.3.425 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Aronson, E. , & Worchel, P. (1966). Similarity versus liking as determinants of interpersonal attractiveness. Psychonomic Science, 5, 157–158. 10.3758/BF03328329 [DOI] [Google Scholar]
  5. Bartlett, M. Y. , & DeSteno, D. (2006). Gratitude and prosocial behavior: Helping when it costs you. Psychological Science, 17, 319–325. 10.1111/j.1467-9280.2006.01705.x [DOI] [PubMed] [Google Scholar]
  6. Belk, R. W. (1996). The perfect gift. In Otnes C. & Beltramini R. F. (Eds.), Gift giving: A research anthology (pp. 59–84). Popular Press. [Google Scholar]
  7. Belk, R. W. , & Coon, G. S. (1993). Gift giving as Agapic love: An alternative to the exchange paradigm based on dating experiences. Journal of Consumer Research, 20(3), 393–417. 10.1086/209357 [DOI] [Google Scholar]
  8. Berger, F. R. (1975). Gratitude. Ethics, 85(4), 298–309. 10.1086/291969 [DOI] [Google Scholar]
  9. Caplow, T. (1984). Rule enforcement without visible means: Christmas gift giving in Middletown. American Journal of Sociology, 89, 1306–1323. 10.1086/228017 [DOI] [Google Scholar]
  10. Chang, Y.‐P. , & Algoe, S. B. (2020). On thanksgiving: Cultural variation in gratitude demonstrations and perceptions between the United States and Taiwan. Emotion, 20(7), 1185–1205. 10.1037/emo0000662 [DOI] [PubMed] [Google Scholar]
  11. Cohen, D. (2003). The American national conversation about (everything but) shame. Social Research, 70(4), 1073–1108. [Google Scholar]
  12. Collins, N. L. , & Miller, L. C. (1994). Self‐disclosure and liking: A meta‐analytic review. Psychological Bulletin, 116, 457–475. 10.1037/0033-2909.116.3.457 [DOI] [PubMed] [Google Scholar]
  13. Converse, B. A. , & Fishbach, A. (2012). Instrumentality boosts gratitude: Helpers are more appreciated while they are useful. Psychological Science, 23, 560–566. 10.1177/0956797611433334 [DOI] [PubMed] [Google Scholar]
  14. Corona, K. , Senft, N. , Campos, B. , Chen, C. , Shiota, M. , & Chentsova‐Dutton, Y. E. (2020). Ethnic variation in gratitude and well‐being. Emotion, 20(3), 518–524. 10.1037/emo0000582 [DOI] [PubMed] [Google Scholar]
  15. DeSteno, D. , Duong, F. , Lim, D. , & Kates, S. (2019). The grateful don't cheat: Gratitude as a fount of virtue. Psychological Science, 30, 979–988. 10.1177/0956797619848351 [DOI] [PubMed] [Google Scholar]
  16. Dickens, L. , & DeSteno, D. (2016). The grateful are patient: Heightened daily gratitude is associated with attenuated temporal discounting. Emotion, 16, 421–425. 10.1037/emo0000176 [DOI] [PubMed] [Google Scholar]
  17. Dunn, E. W. , Huntsinger, J. , Lun, J. , & Sinclair, S. (2008). The gift of similarity: How good and bad gifts influence relationships. Social Cognition, 26, 469–481. 10.1521/soco.2008.26.4.469 [DOI] [Google Scholar]
  18. Emmons, R. A. (2007). Thanks! How the new science of gratitude can make you happier. Houghton‐Mifflin. [Google Scholar]
  19. Faul, F. , Erdfelder, E. , Lang, A. G. , & Buchner, A. (2007). G* power 3: A flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behavior Research Methods, 39(2), 175–191. 10.3758/BF03193146 [DOI] [PubMed] [Google Scholar]
  20. Fehr, B. (1988). Prototype analysis of the concepts of love and commitment. Journal of Personality and Social Psychology, 55, 557–579. 10.1037/0022-3514.55.4.557 [DOI] [Google Scholar]
  21. Flynn, F. J. , & Adams, G. S. (2009). Money can't buy love: Asymmetric beliefs about gift price and feelings of appreciation. Journal of Experimental Social Psychology, 45, 404–409. 10.1016/j.jesp.2008.11.003 [DOI] [Google Scholar]
  22. Frei, J. R. , & Shaver, P. R. (2002). Respect in close relationships: Prototype definition, self‐report assessment, and initial correlates. Personal Relationships, 9, 121–139. [Google Scholar]
  23. Galak, J. , Givi, J. , & Williams, E. F. (2016). Why certain gifts are great to give but not to get: A framework for understanding errors in gift giving. Current Directions in Psychological Science, 25, 380–385. 10.1177/0963721416656937 [DOI] [Google Scholar]
  24. Garner, T. I. , & Wagner, J. (1991). Economic dimensions of household gift giving. Journal of Consumer Research, 18, 368–379. 10.1086/209266 [DOI] [Google Scholar]
  25. Gignac, G. E. , & Szodorai, E. T. (2016). Effect size guidelines for individual differences researchers. Personality and Individual Differences, 102, 74–78. 10.1016/j.paid.2016.06.069 [DOI] [Google Scholar]
  26. Givi, J. , & Galak, J. (2017). Sentimental value and gift giving: Givers' fears of getting it wrong prevents them from getting it right. Journal of Consumer Psychology, 27, 473–479. 10.1016/j.jcps.2017.06.002 [DOI] [Google Scholar]
  27. Givi, J. , & Galak, J. (2019). Keeping the joneses from getting ahead in the first place: Envy's influence on gift giving behavior. Journal of Business Research, 101, 375–388. 10.1016/j.jbusres.2019.04.046 [DOI] [Google Scholar]
  28. Goranson, R. E. , & Berkowitz, L. (1966). Reciprocity and responsibility reactions to prior help. Journal of Personality and Social Psychology, 3(2), 227–232. 10.1037/h0022895 [DOI] [PubMed] [Google Scholar]
  29. Goulder, A. W. (1960). The norm of reciprocity: A Preliminary statement. American Sociological Review, 25, 161–178. [Google Scholar]
  30. Greenberg, M. S. , & Westcott, D. R. (1983). Indebtedness as a mediator of reactions to aid. In Fisher J. D., Nadler A., & DePaulo B. (Eds.), New directions in helping (Vol. 1). Recipient Reactions to Aid. (pp. 85–112). Academic Press. [Google Scholar]
  31. Hitokoto, H. (2016). Indebtedness in cultural context: The role of culture in the felt obligation to reciprocate. Asian Journal of Social Psychology, 19, 16–25. 10.1111/ajsp.12122 [DOI] [Google Scholar]
  32. Hitokoto, H. , Niiya, Y. , & Tanaka‐Matsumi, J. (2008). Own benefit and other's cost: Cross‐cultural comparison of “indebtedness” among Americans and Japanese. Japanese Journal of Research on Emotions, 16, 3–24. 10.4092/jsre.16.3 [DOI] [Google Scholar]
  33. Hudson, S. D. (1980). The nature of respect. Social Theory and Practice, 6(1), 69–90. 10.5840/soctheorpract19806112 [DOI] [Google Scholar]
  34. Joy, A. (2001). Gift giving in Hong Kong and the continuum of social ties. Journal of Consumer Research, 28(2), 239–256. 10.1086/322900 [DOI] [Google Scholar]
  35. Kates, S. , & DeSteno, D. (2021). Gratitude reduces consumption of depleting resources. Emotion, 21(5), 1119–1123. 10.1037/emo0000936 [DOI] [PubMed] [Google Scholar]
  36. Kim, H. S. , Sherman, D. K. , & Taylor, S. E. (2008). Culture and social support. American Psychologist, 63(6), 518–526. [DOI] [PubMed] [Google Scholar]
  37. Kupor, D. , Flynn, F. , & Norton, M. I. (2017). Half a gift is not half‐hearted: A giver‐receiver asymmetry in the thoughtfulness of partial gifts. Personality and Social Psychology Bulletin, 43(12), 1686–1695. 10.1177/0146167217727003 [DOI] [PubMed] [Google Scholar]
  38. Lambert, N. M. , Clark, M. S. , Durtschi, J. , & Fincham, F. D. (2010). Benefits of expressing gratitude to a partner changes one's view of the relationship. Psychological Science, 21, 574–580. 10.1177/0956797610364003 [DOI] [PubMed] [Google Scholar]
  39. Lee, H. W. , Bradburn, J. , Johnson, R. E. , Lin, S.‐H. , & Chang, C. H. (2019). The benefits of receiving gratitude for helpers: A daily investigation of proactive and reactive helping at work. Journal of Applied Psychology, 104(2), 197–213. 10.1037/apl0000346 [DOI] [PubMed] [Google Scholar]
  40. Markus, H. R. , & Kitayama, S. (1991). Culture and the self: Implications for cognition, emotion, and motivation. Psychological Review, 98, 224–253. 10.1037/0033-295X.98.2.224 [DOI] [Google Scholar]
  41. Mauss, M. (1925/1990). The gift: The form and reason for exchange in archaic societies. Routledge. [Google Scholar]
  42. McCullough, M. E. , Tsang, J. , & Emmons, R. A. (2004). Gratitude in intermediate affective terrain: Links of grateful moods to individual differences and daily emotional experience. Journal of Personality and Social Psychology, 86, 295–309. 10.1037/0022-3514.86.2.295 [DOI] [PubMed] [Google Scholar]
  43. Menard, S. (1995). Applied logistic regression analysis: Sage University quantitative applications in the social sciences. Sage. [Google Scholar]
  44. Mesquita, B. , & Leu, J. (2007). The cultural psychology of emotion. In Kitayama S. & Cohen D. (Eds.), Handbook of cultural psychology (pp. 734–759). The Guilford Press. [Google Scholar]
  45. Mullin, A. (2016). Dependent children, gratitude, and respect. Journal of Moral Philosophy, 13, 720–738. 10.1163/17455243-01306004 [DOI] [Google Scholar]
  46. O'Brien, R. M. (2007). A caution regarding rules of thumb for variance inflation factors. Quality & Quantity, 41, 673–690. 10.1007/s11135-006-9018-6 [DOI] [Google Scholar]
  47. Oishi, S. , Koo, M. , Lim, N. , & Suh, E. M. (2019). When gratitude evokes indebtedness. Applied Psychology. Health and Well‐Being, 11(2), 286–303. 10.1111/aphw.12155 [DOI] [PubMed] [Google Scholar]
  48. Ortony, A. , Clore, G. L. , & Collins, A. (1988). The cognitive structure of emotion. New York: University Press. [Google Scholar]
  49. Peng, C. , Nelissen, R. M. A. , & Zeelenberg, M. (2018). Reconsidering the roles of gratitude and indebtedness in social exchange. Cognition and Emotion, 32, 760–772. 10.1080/02699931.2017.1353484 [DOI] [PubMed] [Google Scholar]
  50. Raudenbush, S. W. , & Bryk, A. S. (2002). Hierarchical linear models. Sage. [Google Scholar]
  51. Regan, D. T. (1971). Effects of a favor and liking on compliance. Journal of Experimental Social Psychology, 7(6), 627–639. 10.1016/0022-1031(71)90025-4 [DOI] [Google Scholar]
  52. Richard, F. D. , Bond, C. F. Jr. , & Stokes‐Zoota, J. (2003). One hundred years of social psychology quantitatively described. Review of General Psychology, 7(4), 331–363. 10.1037/1089-2680.7.4.331 [DOI] [Google Scholar]
  53. Ruffle, B. J. (1999). Gift giving with emotions. Journal of Economic Behavior & Organization, 39, 399–420. 10.1016/S0167-2681(99)00048-7 [DOI] [Google Scholar]
  54. Schwartz, B. (1967). The social psychology of the gift. American Journal of Sociology, 73, 1–11. 10.1086/224432 [DOI] [PubMed] [Google Scholar]
  55. Septianto, F. , Nallaperuma, K. , Bandyopadhyay, A. , & Dolan, R. (2020). Proud powerful, grateful powerless: The interactive effect of power and emotion on gift giving. European Journal of Marketing, 54, 1703–1729. 10.1108/EJM-10-2018-0727 [DOI] [Google Scholar]
  56. Shen, H. , Wan, F. , & Wyer, R. S. Jr. (2011). Cross‐cultural differences in the refusal to accept a small gift: The differential influence of reciprocity norms on Asians and north Americans. Journal of Personality and Social Psychology, 100, 271–281. 10.1037/a0021201 [DOI] [PubMed] [Google Scholar]
  57. Thibaut, J. W. , & Kelley, H. H. (1959). The social psychology of groups. Wiley. [Google Scholar]
  58. Tsai, J. L. , & Clobert, M. (2019). Cultural influences on emotion: Established patterns and emerging trends. In Cohen D. & Kitayama S. (Eds.), Handbook of cultural psychology (pp. 292–318). The Guilford Press. [Google Scholar]
  59. Tsang, J. (2006). Gratitude and prosocial behavior: An experimental test of gratitude. Cognition & Emotion, 20, 138–148. 10.1080/02699930500172341 [DOI] [Google Scholar]
  60. Watkins, P. C. , Emmons, R. A. , Greaves, M. R. , & Bell, J. (2018). Joy is a distinct positive emotion: Assessment of joy and relationship to gratitude and well‐being. Journal of Positive Psychology, 13, 522–539. 10.1080/17439760.2017.1414298 [DOI] [Google Scholar]
  61. Watkins, P. C. , Scheer, J. , Ovnicek, M. , & Kolts, R. (2006). The debt of gratitude: Dissociating gratitude and indebtedness. Cognition and Emotion, 20, 217–241. 10.1080/02699930500172291 [DOI] [Google Scholar]
  62. Watkins, P. C. , Woodward, K. , Stone, T. , & Kolts, R. (2003). Gratitude and happiness: Development of a measure of gratitude, and relationships with subjective well‐being. Social Behavior and Personality, 31, 431–452. 10.2224/sbp.2003.31.5.431 [DOI] [Google Scholar]
  63. West, S. G. , Finch, J. F. , & Curran, P. J. (1995). Structural equation models with nonnormal variables: Problems and remedies. In Hoyle R. H. (Ed.), Structural equation modeling: Concepts, issues and applications (pp. 56–75). Sage. [Google Scholar]
  64. Wooten, D. B. (2000). Qualitative steps toward an expanded model of anxiety in gift giving. Journal of Consumer Research, 27, 84–95. 10.1086/314310 [DOI] [Google Scholar]
  65. Zhang, Y. , & Epley, N. (2012). Exaggerated, mispredicted, and misplaced; when “it's the thought that counts” in gift exchanges. Journal of Experimental Psychology: General, 141, 667–681. 10.1037/a0029223 [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Availability Statement

The data that support the findings of this study are openly available in The Emotional Predictors of a Thank‐you Gift at https://osf.io/rfbwg/?view_only=655483c6c09546a4b2de55c4b74fb82d.

Articles from Applied Psychology. Health and Well-Being are provided here courtesy of Wiley

RESOURCES