Reconsidering the roles of gratitude and indebtedness in social exchange

ABSTRACT

Receiving favors is often a mixed blessing and commonly triggers two emotions: the positive emotion gratitude and negative emotion indebtedness. In three studies, we examined the hypothesis that gratitude and indebtedness have distinct functions in social exchange. Contrary to current views, we believe that the function of gratitude does not primarily reside in facilitating social exchange. Instead, we propose that indebtedness motivates people to repay favours received, and thus accounts for most of the prosocial effects commonly attributed to gratitude. On the other hand consistent with current views, we believe that gratitude signals the potential for developing a relationship and fosters proximity seeking. Supporting these assumptions, in Study 1 we found that gratitude and indebtedness were associated with aspects of the favour that reflect the concern for relationship and the level of inequity. Studies 2 and 3 provided causal support for these relations, and revealed the unique associations between gratitude and the motivation of proximity seeking, and between indebtedness and the motivation to reciprocate. We argue that this functional distinction has escaped research attention as gratitude and indebtedness are naturally correlated because they stem from the same eliciting event. To appreciate this functional distinction, both emotions should be studied simultaneously in the context of social exchange.

KEYWORDS:

• Gratitude
• indebtedness
• social exchange
• reciprocity
• social relationship

Receiving favours is a mixed blessing. It can make people feel both grateful and indebted (Fisher, Nadler, & Whitcher-Alagna, 1982; Gleason, Lida, Shrout, & Bolger, 2008). Gratitude is a positive experience that is found to promote future helping by the receiver (Bartlett & DeSteno, 2006; McCullough, Kilpatrick, Emmons, & Larson, 2001; McCullough, Kimeldorf, & Cohen, 2008; Tsang, 2006b; for a meta-analytic review, see Ma, Tunney, & Ferguson, 2017) and thereby strengthens the social bond between people (Algoe, Fredrickson, & Gable, 2013; Bartlett, Condon, Cruz, Baumann, & DeSteno, 2012; Gordon, Impett, Kogan, Oveis, & Keltner, 2012; Kubacka, Finkenauer, Rusbult, & Keijsers, 2011). Indebtedness is a negative experience that produces aversive intrapsychic and interpersonal consequences (Greenberg & Shapiro, 1971), such as dampening self-esteem, compromising autonomy (Fisher et al., 1982; Greenberg & Westcott, 1983), and triggering negative evaluations of the benefactor (e.g. Greenberg, 1980). It is remarkable that one event can trigger, at the same time, a positive and a negative emotion. We believe that as yet, this simultaneous elicitation of gratitude and indebtedness has not received sufficient research attention, and these two emotions are too often studied in isolation. What’s more, research has focused much more on gratitude than on indebtedness, virtually ignoring the role of the latter. Consequentially, much scholarly attention has focused on how to shape helping behaviour so that the positive feeling of gratitude is maximised and the resulting feeling of indebtedness is minimal. For instance, if the help is not perceived as benevolent (e.g. conceals a hidden agenda on the part of the benefactor) or clearly entails a return expectation (Tsang, 2006a; Watkins, Scheer, Ovnicek, & Kolts, 2006), it is less likely that gratitude is experienced. We believe that this approach is interesting, but limited, because it overlooks the distinct roles of gratitude and indebtedness in social exchange.

As noted above, the simultaneous elicitation of gratitude and indebtedness implies something crucial for our understanding of these emotions.¹ According to appraisal theories (for an overview, see Moors, 2014), positive feelings arise when events are perceived as goal-congruent and negative emotions arise when events are perceived as goal-incongruent. We think that this implies that receiving a favour appeals to two distinct goals, and that both goals are served separately by gratitude and indebtedness. This general position is consistent with the basic tenets of functional perspectives of emotions, which hold that different emotions are elicited by distinct concerns and elicit qualitatively different behavioural responses aimed at attaining distinct goals (DeSteno, Petty, Wegener, & Rucker, 2000; Izard & Ackerman, 2000; Lerner & Keltner, 2000; Roseman, Wiest, & Swartz, 1994; Zeelenberg, Nelissen, Breugelmans, & Pieters, 2008). So what then would be the respective concerns and functions of gratitude and indebtedness in social exchange?

In short, the idea is that receiving help elicits an immediate concern to restore the inequity resulting from receiving benefits (Adams, 1965; Walster, Berscheid, & Walster, 1973), and a more distal concern related to the opportunity to establish, build, or strengthen a social relationship through exchange (Burgess & Huston, 1979). This perspective on the consequences of receiving help parallels what Molm (2010) refers to as the material and symbolic effects of social exchange (see also Molm, Collett, & Schaefer, 2007; Molm, Whitham, & Melamed, 2012). We contribute to this line of work by identifying the emotions of gratitude and indebtedness as independent psychological mechanisms underlying these two functions of social exchange. Corresponding to what Molm referred as the symbolic effect of helping, gratitude concerns for relational value and drives behaviour to seek proximity with the benefactor so as to build up social bond. Corresponding to what Molm refers to as the material effect of helping, indebtedness concerns for inequity after receiving favours and drives to repay the benefactor so as to restore inequity.

This perspective readily accommodates recent findings in research on people’s responses to receiving help. First, it explains why the same event results in the simultaneous elicitation of a positive and a negative emotion: receiving help is congruent with the goal to establish or promote a relationship as it suggests an opportunity to do so (unless the help offered is guided by instrumental motives on behalf of the benefactor). As a positive feeling, gratitude will thus broaden a person’s thought action repertoire (cf. Fredrickson, 2004) and gear the receiver to identify opportunities for relationship enhancement. Receiving help is at the same time incongruent with the goal to maintain equity in social relationships. Equity restoration is important to ensure mutual cooperation in social groups (e.g. Trivers, 1971), and its importance to the functioning of human societies is evidenced by the fact that it is perceived to be a universal social norm (Gouldner, 1960). As a negative feeling, indebtedness will focus a person’s thoughts and actions towards opportunities to restore equity. Second, the dual functions perspective that we propose underscores the importance of benevolent intentions in eliciting gratitude (Tsang, 2006a; Watkins et al., 2006). Perceiving the help as intentional and benevolent is crucial for eliciting gratitude, as non-intentional and instrumental helping (i.e. helping with the sole objective to secure future help in return) does not clearly suggest that the helper seeks to develop a social relationship. On the other hand, the feeling of indebtedness should be affected neither by the intentionality nor by the benevolence with which a favour is extended, as neither ultimately affects the amount of inequity being generated.

There are different, albeit complementary explanations as to why the functional role of indebtedness has been overlooked. First, previous research simply studied indebtedness as a negative outcome of receiving favours. As a result, the positive role of indebtedness in promoting the restoration of equity has been neglected. To be more specific, feeling indebted would prompt the receiver to repay the benefactor so as to restore equity (Greenberg, 1980). This positive effect of indebtedness can only be revealed by studying the responses of people after receiving help. Therefore, to understand the function of indebtedness, we should study the responses to indebtedness in social exchange process (Cropanzano & Mitchell, 2005; Emerson, 1976).

A second reason as to why the functional role of indebtedness has been overlooked, one that is more methodological, is that many studies on the behavioural effects of gratitude simply did not include measures of indebtedness. Considering that gratitude and indebtedness are correlated by nature, as they are both emotional responses to the same event of receiving help, some of the relations between gratitude and reciprocal behaviour might be spurious and could in fact be caused by indebtedness. We believe that by neglecting indebtedness, researchers have ascribed both the goals to restore equity and to further enhance relationships to gratitude, stating that gratitude as a moral affect facilitates costly reciprocal behaviours (McCullough et al., 2001, 2008) and social bonding (Kubacka et al., 2011). Indeed, gratitude has found to be associated with both costly prosocial behaviour in return to receiving favours (e.g. Bartlett & DeSteno, 2006; Tsang, 2006b) even to third-parties (Bartlett & DeSteno, 2006; Nowak & Roch, 2007), and to prompt social bonding with the benefactor (Bartlett et al., 2012). In line with the latter, expressing gratitude is beneficial for relationship growth on both sides (Algoe et al., 2013), not only enhancing the receiver’s perceived bond with the benefactor (Lambert, Clark, Durtschi, Fincham, & Graham, 2010), but also consolidating the benefactor’s responsiveness to the receiver (Grant & Gino, 2010). Yet, to truly understand the dual functions of both emotions, they need to be studied in combination.

In sum, we acknowledge the role of gratitude in fostering social relationships, but we believe that it is indebtedness rather than gratitude that serves the function of restoring imbalance in the social exchange. In the present paper, we present three studies that were specifically designed to illustrate the dual functions of gratitude and indebtedness. Generally, we derived predictions concerning factors that should affect predominantly the experience of gratitude or predominantly indebtedness. If our ideas about the different functions of gratitude and indebtedness are correct, we should be able to find emotion-specific effects. More specifically, we expect to find that:

Gratitude and indebtedness should have distinct cognitive antecedents. Corresponding to the goal of promoting relationship development, gratitude should be affected by factors related to the relational value. First, the favour should be intentionally benevolent for the receiver to treat it as an opportunity for relationship promotion. An offered favour entails the most intentionality while an imposed favour (favour that is not initiated by the benefactor himself/herself) entails the least, and a requested favour lies in between (Broll, Gross, & Piliavin, 1974; Greenberg & Saxe, 1975). Second, the favour, and by extension the relationship with the benefactor should be perceived as instrumentally beneficial, so that it would indeed be worthwhile for the receiver to further proliferate the social bond with the benefactor. In brief, the benefits and intentionality of the favour signal the value of fostering a relationship, which results in gratitude. Corresponding to the goal of maintaining equity, indebtedness is elicited by the concern for equity, and should be affected by factors related to the amount of inequity. This inequity mainly comes from accounting for the benefactor’s costs in the exchange. Therefore, the cost of the favour signals the degree of inequity, which results in indebtedness.

Gratitude and indebtedness should have distinct behavioural consequences. As we argued earlier, gratitude is the response to the potential opportunity of building up a relationship, thus gratitude should drive the motivation to seek proximity with the benefactor. On the other hand, indebtedness is the response to inequity. To compensate the benefactor, indebtedness should drive the obligation to reciprocate.

Gratitude and indebtedness should serve as the mediators between their antecedents and consequences separately. Because their effects are goal-directed, gratitude and indebtedness serve functions through dealing with their causes. Thus, gratitude should mediate the effect of benefit and intentionality on the motivation to seek proximity, while indebtedness should mediate the effect of cost on the perceived obligation to reciprocate.

We conducted three studies to test these predictions. In Study 1, we surveyed people and asked for autobiographic recalls of situations in which they received help. In Study 2, we manipulated the cost and benefit of the favour. We expected that manipulations of cost would affect indebtedness and its motivational tendency of repaying the benefactor, while manipulations of benefit would affect gratitude and its motivational tendency of seeking proximity with the benefactor. In Study 3, we manipulated the intentionality of the favour, expecting that it only affected gratitude and its motivational tendency.

Study 1: recollections of real-life helping behaviour

We conducted a survey about the recipient’s feelings and thoughts after receiving help, using an autobiographic recall method.² This allowed us to collect data from a wide variety of daily events, ensuring that our results were robust across different situations in which people receive favours.

Method

Participants and procedure

A total of 145 participants, 84 males and 61 females ranging from 19 to 68 years (M_age = 35.50, SD = 10.77, U.S.A. residents only) were recruited from Amazon Mechanical Turk (Mturk, see Buhrmester, Kwang, & Gosling, 2011; Paolacci & Chandler, 2014). As this was an exploratory study, we aimed for 150 participants. Because 5 participants did not finish the questionnaire, we ended up with 145 participants. In addition, for all studies we have reported all measures, conditions, data exclusions, and how we determined our sample sizes. The data for all three studies can be found on the Open Science Framework page (https://osf.io/t8p2b). Participants received $0.80 as a compensation for completing our questionnaire that took on average eight minutes.

Participants were asked to describe an occasion of being helped in an interpersonal context. This requirement excluded helping cases such as organisational level charity activities and volunteer work that were not categorised as social exchange. The questionnaire started with several questions intended to characterise the type of help and the nature of the relationship between the helper and the receiver. Next, participants answered questions regarding to their feelings and perceived obligation to reciprocate after receiving help.

Materials

Autobiographic recall

Participants received the following instructions:

Please recall a recent occasion in which you personally received help from someone else. Please describe the situation with as many details as possible, paying attention to the information requested below. In your description, please take a first-person perspective (e.g. “I received help … ”), and try to be as clear as possible, so that somebody who was not there appreciates what the other person did to help you and why.

Participants were also asked to describe (1) who the benefactor was, (2) why they needed help, (3) what the benefactor did to help them, and (4) when this event happened. Finally, participants indicated whether the help was offered by the benefactor or requested by the receiver, to establish the intentionality of the favour. We further measured the following:

Relationship with the benefactor

In order to assess the relation with the benefactor in more detail, we measured: perceived closeness with the single-item Inclusion of Others in the Self scale (IOS; Aron, Aron, & Smollan, 1992) in which two circles representing the Self and the Other were presented with seven different degrees of overlap; similarity with one item asking: How similar or different are you to the other person on a seven-point scale from 1 (very dissimilar) to 7 (very similar); and status with one item asking: Please rate the extent to which you consider your social status to be higher or lower in comparison to the other person on a seven-point scale from 1 (my status is much lower) to 7 (my status is much higher).

Costs and benefits of the favour

We measured the costs of the favour to the benefactor with six items asking what the other person gave was: Effortful; Difficult; Expensive; Time-consuming; Inconvenient; Risky (α = .75) on a five-point scale from 1 (not at all) to 5 (very much), and benefits to the receiver with three items asking what participants received was: Beneficial; Helpful; Useful (α = .96) on the same five-point scale.

Emotional reactions to receiving help

We measured participants’ emotional reactions to receiving help. Specifically, we measured participants’ feelings of gratitude and indebtedness, along with four fillers: happiness, embarrassment, pride, and frustration that were presented in a random order. Participants responded to the question: Please rate the extent to which you experienced each of the following emotions after receiving help on five-point scales from 1 (not at all) to 5 (very much).

Obligation to reciprocate

We measured participants’ perceived obligation to reciprocate in the future with three items asking to what extent participants agree or disagree: I believe that I now owe the other person a favour; I have to do something for the other person someday; If I do not return the other person’s favour I will feel uneasy (α = .88) on seven-point Likert scales from 1 (strongly disagree) to 7 (strongly agree).

Results

Gratitude and indebtedness³

We predicted that factors related to relational value affect the feeling of gratitude while factors related to equity affect the feeling of indebtedness. Particularly, the benefits and the intentionality of the favour (Requested = 45%, Offered = 55%) should affect gratitude while the costs of the favour should affect indebtedness. Table 1 presents the descriptive statistics and bivariate correlations. Indeed, gratitude and indebtedness are positively correlated. Moreover, gratitude correlated positively with both the benefit and the intentionality of the favour. Indebtedness correlated positively with both the costs and the benefits of the favour, and did not correlate with the intentionality of the favour. Finally, we observed an unexpected, positive correlation between indebtedness and closeness with the benefactor.

Table 1. Descriptive statistics and bivariate correlations for all measures in Study 1 (N = 145).

Variables (range)	M (SD)	1	2	3	4	5	6	7	8
1. Intentionality^a	n.a.	–
2. Closeness (1–7)	4.25 (1.91)	−.03	–
3. Similarity (1–7)	5.00 (1.50)	−.05	.56**	–
4. Status (1–7)	3.72 (1.00)	.04	−.02	−.08	–
5. Cost (1–5)	2.42 (0.78)	.12	.24**	−.01	−.11	–
6. Benefit (1–5)	4.57 (0.68)	.07	.14	.03	−.09	.15	–
7. Gratitude (1–5)	4.45 (0.85)	.17*	.07	.07	−.03	.16	.56**	–
8. Indebtedness (1–5)	3.32 (1.35)	−.01	.18*	.04	−.10	.47**	.28**	.36**	–
9. Obligation to reciprocate (1–7)	4.65 (1.69)	.02	.10	.09	−.05	.38**	.16	.15	.57**

^aIntentionality: 0 = requested, 1 = offered.

*p < .05.

**p < .01.

We applied multivariate regression analyses to further test our predictions, in which we specifically took intentionality, cost, and benefit as predictors of gratitude and indebtedness. The regression results are summarised in Table 2. Consistent with our predictions, gratitude was predicted by the perceived benefits but not by the perceived costs of the favour. So, favours with higher benefits trigger more gratitude. Furthermore, we found a marginally significant effect of intentionality, suggesting that offered favours trigger more gratitude than requested ones. Also consistently, indebtedness was predominantly predicted by the perceived costs of the favour but not by the intentionality. Furthermore, the perceived benefits of the favour had a smaller effect on indebtedness.

Table 2. Summary of the multivariate regression analyses of Study 1 for variables predicting gratitude, indebtedness, and the obligation to reciprocate (N = 145).

Predictors	Gratitude				Indebtedness
	B	SE B	β	p	B	SE B	β	p
Intentionality^a	0.21	0.12	.13	.07	−0.20	0.20	−.07	.31
Cost	0.07	0.08	.06	.37	0.77	0.13	.44	<.001
Benefit	0.68	0.09	.55	<.001	0.43	0.14	.22	.003
Model R^2 = .34, F = 24.15***					Model R^2 = .27, F = 17.16***

^aIntentionality: 0 = requested, 1 = offered.

***p < .001.

Reciprocation

Finally, the results also confirmed that the perceived obligation to reciprocate correlated positively with indebtedness, whereas the correlation with gratitude was only marginally significant (see Table 1). When entered simultaneously in a multivariate regression analysis, only indebtedness predicted the obligation to reciprocate, β = .59, t(142) = 8.02, p < .001, whereas gratitude did not have a unique effect, β = −.07, t(142) = −0.89, p = .38.

The feeling of indebtedness also mediated the effects of the perceived costs of the favour on the obligation to reciprocate (see Figure 1). We used the Preacher and Hayes (2008, model 4) PROCESS procedure and the corresponding SPSS macro to test for an indirect effect of cost on perceived obligation to reciprocate through indebtedness. A 5000 bootstrap resample supported the indirect effect, b = 0.50, SE = 0.11, 95% CI [.31, .76], since the CI did not include zero. Furthermore, the direct effect of cost to the perceived obligation to reciprocate became marginally significant after controlling for the effect through indebtedness, b = 0.33, SE = 0.17, p = .051. This pattern of findings confirmed that indebtedness mediated the effect of cost on the perceived obligation to reciprocate.

Figure 1. Indebtedness as the mediator of the effect of the cost on perceived obligation to reciprocate in Study 1. The numbers are unstandardised coefficients in the multivariate regression models. ***p < .001.

Discussion

In summary, the data provided support to the hypothesis that gratitude and indebtedness have distinct functions in social exchange. Specifically, gratitude was associated with factors related to the concern for relational value (that is by the intentionality and the perceived benefits of the favour), while indebtedness was associated with factors related to the concern for equity (i.e. the perceived costs of the favour). Note that these results are collected from a wide variety of daily events, which suggests that our hypotheses hold across different social situations.

Moreover, our findings suggested that indebtedness rather than gratitude serves the function of restoring equity after receiving favours, which further supports that gratitude and indebtedness serve distinct functions. We found a marginally significant correlation between gratitude and the behavioural tendency to reciprocate, a relation previous research (Bartlett & DeSteno, 2006; McCullough et al., 2001, 2008; Tsang, 2006b) also observed. However, gratitude did not have an independent effect when the feeling of indebtedness was also taken into account. Therefore, the relation between gratitude and the motivation for reciprocation might be spurious and could perhaps be attributed to indebtedness.

For the autobiographic recall method, a potential concern is that participants may have already reciprocated the favour they received. As a result, the feeling of indebtedness as well as perceived obligation to reciprocate have been reduced. However, we believe that if this was the case, the real effect should be even stronger. Because the independent variable of interest (i.e. perceived costs of the favour) should not be affected by whether the favour is reciprocated or not, the observed associations between perceived costs and indebtedness should have been underestimated.

We believe using vignette methodology can circumvent the concern regarding recall methodology, and provide further evidence beyond correlational investigations. In the next two studies, we experimentally manipulate factors related to gratitude (benefits and intentionality of the favour) and indebtedness (costs of the favour). If those factors impact gratitude or indebtedness correspondingly, this experimental evidence would further support the dual function model.

Study 2: cost and benefit manipulations

In Study 2, we applied a scenario that was specifically created to manipulate the costs and the benefits of the favour separately. To further examine their hypothesised unique behavioural implications, we not only assessed response tendencies related to the feeling of indebtedness (i.e. perceived obligations to reciprocate) but also assessed response tendencies related to the feeling of gratitude (i.e. motivation to seek proximity with the benefactor).

Method

Participants and procedure

We aimed for 300 participants (75 per condition). A power analysis indicated that we needed 128 participants to find a medium effect (f = 0.25) with 80% power. A total of 300 participants, 172 males and 128 females ranging from 18 to 68 years (M_age = 33.76, SD = 9.45, U.S.A. residents only) were recruited from Amazon Mturk. Participants were randomly assigned to one out of four conditions in a 2 (Cost: low vs. high) × 2 (Benefit: low vs. high) between-subjects design. Participants received $0.40 as a compensation for completing our questionnaire, which took on average four minutes.

Participants were instructed to read a scenario in which they received help from a colleague. Next, participants answered questions about their feelings, perceived obligation to reciprocate and motivation to seek proximity after receiving help. The manipulation check on the costs and benefits of the favour was presented in the end.

Materials

The scenario

Participants read the following scenario:

Imagine that you started a new job at a company just one month ago. One day last week, you worked so late that you missed the last bus home. Your home is 10 miles away from where you work, which takes around 20 minutes to drive. Your colleague Catherine from the same department happened to work late that day too, and she offered to drive you home.

Subsequently, to manipulate the costs of the favour, the scenario further specified that the colleague lived either in the same direction, so the favour was not a big deal for her (low cost) or in the opposite direction, so she had to take a 40-minute detour (high cost). To manipulate the benefits of the favour, the situation was either described as urgent because without a ride the receiver would definitely miss the reservation for celebrating his/her partner’s birthday that night (high benefit). In the low-benefit condition, there was no particular urgency to get a ride home.

After reading the scenario, participants completed the following measures:

Emotional reactions to receiving help

We measured participants’ feeling of gratitude with two items: Grateful; Thankful, feeling of indebtedness with two items: Indebted; Obligated, along with four fillers: happiness, embarrassment, pride, and frustration that were presented in a random order. Participants responded to the question: Please indicate on the scales below the extent to which you experience these feelings after receiving help on scales from 1 (not at all) to 10 (extremely).

Obligation to reciprocate

We measured participants’ perceived obligation of returning the favour in the future, with the same three items used in Study 1 (α = .83).

Proximity seeking

We measured participants’ motivation to seek proximity with the benefactor with three items: I would like to work in the same team with the other person; I would like to share an office with the other person; I would like to get to know the other person better(α = .84), on seven-point Likert scales from 1 (strongly disagree) to 7 (strongly agree).

Manipulation check

The final questions asked for participants’ perception of the costs and the benefits of the favour. Cost was measured with three items asking what the other person did was: Effortful; Time-consuming; Inconvenient (α = .90) on a seven-point scale from 1 (not at all) to 7 (extremely). Benefit was measured with three items asking what participants received was: Beneficial; Helpful; Useful (α = .85) on the same seven-point scale.

Results

Manipulation checks

The descriptive statistics for all dependent variables are shown in Table 3. The cost manipulation was successful. A 2 (Cost: low vs. high) × 2 (Benefit: low vs. high) analysis of variance (ANOVA) on the perceived costs of the favour yielded only a main effect of the cost manipulation, F(1, 296) = 295.59, p < .001, η² = .50. As expected, participants in the High Cost condition reported higher perceived costs of the favour compared with participants in the Low Cost condition. There was neither a main effect of the benefit manipulation, F(1, 296) = 2.33, p = .13, η² = .01, nor an interaction effect, F(1, 296) = 0.13, p = .72, η² = .000.

Table 3. Means (standard deviations within parentheses) by conditions manipulating cost and benefit in Study 2 (N = 300).

	Low cost		High cost
	Low benefit	High benefit	Low benefit	High benefit
Perceived cost	3.13a (1.35)	3.30a (1.24)	5.50b (1.15)	5.77b (1.13)
Perceived benefit	6.26a (0.87)	6.27a (0.87)	6.12a (0.94)	6.36a (0.78)
Indebtedness	5.45a (2.42)	5.48a (2.20)	6.27b (2.31)	6.43b (2.60)
Gratitude	8.92a (1.39)	8.87a (1.40)	8.74a (1.82)	8.97a (1.52)
Obligation to reciprocate	5.29a (1.08)	5.57a,b (1.16)	5.79b (1.01)	5.86b (1.14)
Proximity seeking	5.68a (0.72)	5.50a (0.99)	5.62a (1.10)	5.67a (0.96)

Note: Within each row, means with different subscripts are significantly different by p < .05 in a pairwise comparison.

The benefit manipulation failed. The same ANOVA on the perceived benefits of the favour yielded neither main effect of the benefit manipulation, F(1, 296) = 1.52, p = .22, η² = .01, nor main effect of the cost manipulation, F(1, 296) = 0.04, p = .84, η² = .000, and interaction effect, F(1, 296) = 1.41, p = .24, η² = .01.

Hypothesis testing⁴

We predicted that the cost manipulation would have an effect on the feeling of indebtedness but not on the feeling of gratitude. As Table 3 shows, consistent with this prediction, only the cost manipulation had an effect on the feeling of indebtedness, F(1, 296) = 10.31, p = .001, η² = .03. There was neither a main effect of the benefit manipulation, F(1, 296) = 0.12, p = .73, η² = .000, nor an interaction effect, F(1, 296) = 0.05, p = .82, η² = .000. Again consistent with our predictions, only the cost manipulation had an effect on the perceived obligation to reciprocate, F(1, 296) = 9.54, p = .002, η² = .03. There was neither a main effect of the benefit manipulation, F(1, 296) = 1.92, p =.17, η² = .01, nor an interaction effect, F(1, 296) = 0.68, p = .41, η² = .000. Also as expected, the cost manipulation affected neither the feeling of gratitude, F(1, 296) = 0.06, p = .81 η² = .00, nor the motivation to seek proximity, F(1, 296) = 0.26, p = .61, η² = .00. There were no differences between conditions manipulating high and low cost in the ratings of gratitude or motivation to seek proximity.

Again, indebtedness mediated the relation between the cost manipulation and the perceived obligation to reciprocate. As Figure 2 shows, the cost manipulation predicted indebtedness and the perceived obligation to reciprocate. Indebtedness also predicted the perceived obligation to reciprocate. By applying PROCESS procedure, we found support for the indirect effect through indebtedness, b = 0.23, SE = 0.07, 95% CI [.08, .39] which excluded zero. Furthermore, the direct effect of cost manipulation became insignificant after controlling for the effect through indebtedness, p = .13. This pattern of findings confirmed again that indebtedness mediated the effect of cost on the perceived obligation to reciprocate.

Figure 2. Indebtedness as the mediator of perceived obligation to reciprocate in Study 2. Cost condition was coded as follows: low cost = 0, high cost = 1. The numbers are unstandardised coefficients in the multivariate regression models. **p < .01, ***p < .001.

Discussion

The data confirmed that the costs of the favour are the central concern of indebtedness and its corresponding motivation to reciprocate. This again indicates that indebtedness responds to inequity and drives to restore it. The present results build upon the results from Study 1 by establishing the causal nature of the relations between cost, indebtedness, and the obligation to reciprocate. The manipulation of the benefits of the favour did not work. Hence, we did not find evidence to support or reject the effect of benefit on the feeling of gratitude and its motivation to seek proximity. The benefit manipulation appeared to be too strong. Even in the low-benefit conditions participants reported an average score of benefit higher than six on a seven-point scale. Therefore, in order to also find support for the proposed function of gratitude in response to receiving help, we manipulated the intentionality of the favour rather than its perceived benefits in Study 3.

Study 3: intentionality manipulation

To manipulate the intentionality of the favour, we again used a scenario in which the help was either offered by the benefactor or imposed to the benefactor by a third-party.⁵ Considering that an offered favour signals more benevolence from the benefactor and thus suggests more relational value than an imposed one, we expected that an offered favour would elicit more gratitude and in turn a higher motivation for proximity seeking towards the benefactor, compared with an imposed one. According to our hypotheses, the intentionality of the favour should not affect indebtedness and resulting obligations to reciprocate, however.

Method

Participants and procedure

We aimed for 150 participants (75 per condition). A power analysis indicated that we needed 102 participants to find a medium effect (d = 0.5) with 80% power. A total of 151 participants, 78 males and 73 females ranging from 19 to 72 years (M_age = 34.81, SD = 10.12, U.S.A. residents only) were recruited from Amazon Mturk. Participants were randomly assigned either to the offered help condition (N = 76) or to the imposed help condition (N = 75). Participants received $0.40 as a compensation for an average of four minutes.

Materials and measures

Participants were instructed to read a scenario of being helped by a colleague. In both conditions, participants read the following scenario:

Imagine that you started a new job at a company just one month ago. You are still in the process of getting familiar with the work and with people there. One day last week, you felt a bit overwhelmed and you were completely troubled as to how to finish your current assignment.

In the offered help condition, the scenario continued as follows:

Your colleague David who works at the same department saw what was happening. David offered to help you out without you asking for it. Although David was busy himself and he also had a number of unfinished tasks, he volunteered to stay late to help you, by showing you how to do the tasks you were assigned and how to organize your results.

In the imposed help condition, the final part of the scenario was:

Your manager saw what was happening, so he told David - a colleague who works at the same department – to assist you. Although David was busy himself and he also had a number of unfinished tasks, he was told to stay late to help you, by showing you how to do the tasks you were assigned and how to organize your results.

After reading the scenario, participants completed the same questions as in Study 2 concerning their emotional reactions, obligation to reciprocate and proximity seeking. Finally, as a manipulation check, participants indicated the extent to which they thought the favour was initiated by the benefactor himself or by someone else on a scale from 1 (by himself) to 7 (by someone else).

Results

The results are shown in Table 4. The manipulation was successful. Participants in the imposed help condition tended to report that the favour was initiated by someone else whereas participants in the offered help condition tended to report that the favour was initiated by the benefactor himself, t(120) = 20.03, p < .001.

Table 4. Means (standard deviations within parentheses) by conditions manipulating intentionality in Study 3 (N = 151).

	Imposed	Offered
Initiation of the help^a	5.95a (1.59)	1.67b (0.94)
Indebtedness	6.01a (2.26)	5.74a (2.53)
Gratitude	8.51a (1.55)	8.96b (1.24)
Obligation to reciprocate	5.49a (1.20)	5.20a (1.28)
Proximity seeking	5.24a (0.99)	5.80b (1.03)

Note: Within each row, means with different subscripts are significantly different by p < .05 in a pairwise comparison.

^aInitiation of the Help: 1 = by himself, 7 = by someone else.

We predicted that an offered favour would cause more gratitude and proximity seeking than imposed favour. Consistent with our prediction, participants in the offered help condition reported higher feeling of gratitude, t(141) = −1.99, p = .049, d = 0.32, and a stronger motivation to seek proximity with the benefactor, t(149) = −3.40, p = .001, d = 0.55, than participants in the imposed help condition. Also as expected there were no differences in the feeling of indebtedness, t(149) = 0.71, p = .48, d = 0.11, and the perceived obligation to reciprocate, t(149) = 1.44, p = .15, d = 0.23 between the two conditions.

To test whether gratitude mediated the relation between the intentionality of the favour and the motivation to seek proximity, we applied the same regression procedure as used in Studies 1 and 2. After controlling for gratitude, the effect of the intentionality manipulation on proximity seeking motivation decreased from b = 0.56, SE = 0.16, p < .001 to b = 0.46, SE = 0.16, p < .01. The size of the indirect effect through gratitude was b = 0.10, 95% CI [.01, .21] which excluded zero (see Figure 3). Therefore, gratitude partially mediated the effect of intentionality of the favour on the perceived motivation to seek proximity.

Figure 3. Gratitude as the mediator of motivation of proximity seeking in Study 3. Intentionality condition was coded as follows: imposed help = 0, offered help = 1. The numbers are unstandardised coefficients in the multivariate regression models. *p < .05, **p < .01, ***p < .001.

Discussion

The data confirmed that the intentionality of the favour affected gratitude and its corresponding motivation to seek proximity, but not indebtedness and its corresponding motivation to reciprocate. This suggests that gratitude is triggered by different aspects of the favour than indebtedness. Specifically, gratitude is affected by factors that reflect genuine benevolence and thus reflects a concern for relational value, and functions to foster the relationship.

General discussion

We studied the emotional responses to receiving favours in the context of social exchange. In order to explain why receiving favours simultaneously elicits two emotions of opposite valence: the positive emotion gratitude and the negative emotion indebtedness, we proposed that receiving a favour appeals to two different yet coinciding goals: First, receiving a favour triggers a need for equity restoration and consequently a goal to compensate one’s benefactor. Second, receiving a favour signals an opportunity for relationship promotion and a goal to seek proximity with one’s benefactor. Based on these different functions, we hypothesised that favours should elicit more gratitude to the extent that they are intentional, benevolent, and beneficial. On the other hand, favours should elicit more indebtedness to the extent that they increase the level of inequity. We found support for these predictions in three studies. Study 1 employed an autobiographic recall method to sample various real-life instances of receiving favours and supported our core predictions at a correlational level. Gratitude was associated with factors related to relational value (i.e. benefits and the intentionality of the favour), while indebtedness was associated with factors related to the concern for equity (i.e. costs of the favour). By experimentally manipulating costs and benefits in Study 2, and intentionality in Study 3, we were able to test and confirm the causal nature of the predicted relations. Moreover, our studies confirmed that gratitude and indebtedness are associated with distinct behavioural consequences. That is, gratitude was associated with the motivation of seeking proximity with the benefactor, while indebtedness was associated with the obligation to repay the benefactor. Indeed, these associations were uniquely mediated by feelings of gratitude and indebtedness, respectively. Combining those results, we suggest that gratitude and indebtedness have distinct functions in governing the relational and reciprocal aspects of social exchange (cf. Molm, 2010).

The prediction that a more beneficial favour would elicit more gratitude but not more indebtedness was not supported by the data. Instead, we found that perceived benefits affected both gratitude and indebtedness in Study 1. This is altogether not very surprising as in most real-life situations the benefits and costs of favours are strongly correlated. It was unfortunate, however, that we did not succeed in manipulating the benefits of the favour in Study 2. We observed a ceiling effect to the extent that benefits were perceived to be very high, even in the Low-Benefit condition. As such, we did not observe differences in experiences of gratitude. We do not know whether this lack of differences is the result of people feeling gratitude already over small favours, or whether it is a social desirability bias, preventing participants from reporting lower levels of appreciation out of a concern for appearing ungrateful. Future studies may be more successful in designing more sensitive manipulations of benefits.

Our results and conclusions were built upon self-report measures, which can be biased by social desirability and recall biases. However, if social desirability biases do exist, the real effects might tend to be stronger. It seems plausible that the social desirability bias would be particularly higher when the favours entailed lower relational value (i.e. less beneficial and/or less intentional), under which circumstances the concern for appearing ungrateful is likely to be stronger. Participants in these circumstances were thus relatively more motivated to report higher gratitude than the others. As a result, the real effects on gratitude would have probably been underestimated. Also, recall biases should have not substantially affected the outcomes in Study 1. We did find consistent results with recall (Study 1) and vignette methodologies (Studies 2 and 3), in the latter of which recall biases are irrelevant.

The present research has functionally disentangled gratitude and indebtedness at the level of cognitive antecedents and in terms of their ensuing motivational tendencies, by adopting the functional analysis of specific emotions proposed by Roseman et al. (1994). Future studies may further disentangle them at the level of behavioural responses, which would definitely add to the validity of our findings. Future research may take into consideration that the goals of restoring equity and enhancing proximity may be (and usually are) achieved by the same behaviour (reciprocating a favour). For example, Tsang (2007) suggested that both gratitude and indebtedness are predictors of the same social exchange behaviour (i.e. distributing resources after receiving help), of which gratitude predicts better. Therefore, to tease apart the functions of gratitude and indebtedness respectively at the behavioural level, it would be promising to find behavioural measures that can separate these two goals. Still, we hope evidence from behavioural measurement in the future will in turn complement our work.

We also believe that future studies may lend further support to the proposed dual functions of gratitude and indebtedness by studying other relevant aspects of social exchange. For instance, the experience of gratitude and indebtedness may vary with the type of social relationship (communal relationship vs. exchange relationship) in which favours are extended (e.g. Clark & Mills, 1979, 1993; Fiske, 1992). Compared with favours in an exchange relationship, favours in a communal relationship should elicit less indebtedness, because in a communal relationship one does not keep score (favours are exchanged on the basis of needs).

Moreover, the type of favour may also have an impact on the experience of gratitude and indebtedness (e.g. Foa & Foa, 1980, 2012). For example, financial favours should elicit less gratitude and more indebtedness. This is due to the fact that money signals less relational value to the receiver but causes more salient inequity situations compared with other resources.

Our findings clearly demonstrate that studying gratitude and indebtedness in combination can help clarify their respective functions. Gratitude and indebtedness have the same antecedent since they are caused by the same event. As a result, gratitude and indebtedness are highly correlated, and in order to understand their separate effects they must be studied in combination, so that the influence of the other emotion can be controlled for statistically. When studied in isolation, researchers may for example find that gratitude is correlated with some behaviours that are actually caused by the experience of indebtedness. Unfortunately, this studying of gratitude and indebtedness in isolation is common practice among researchers. For example, Bartlett and DeSteno (2006) concluded that gratitude promotes prosocial behaviours (operationalised as reciprocating help). In their studies, they measured only gratitude after receiving help, and reported the mediating effect of gratitude on the relation between receiving help and reciprocating help. This effect of gratitude might be spurious and could perhaps be attributed to indebtedness. If that would be the case, a totally different conclusion can be drawn instead: indebtedness promotes reciprocity. Our studies measured both gratitude and indebtedness after receiving favours. We did find an effect of gratitude on the perceived obligation to reciprocate, but this effect disappeared after taking indebtedness into account. Indebtedness rather than gratitude fostered the motivation to reciprocate.

Our findings further highlight the importance of studying gratitude and indebtedness in the context of (continued) social exchange, rather than as the mere outcome of receiving favours or help. There are two reasons why using the framework of social exchange theory provides new insight into understanding the roles of gratitude and indebtedness. First, it allows for a more comprehensive view on the role of gratitude. Rather than being just a psychological mechanism that upholds reciprocity, many researchers have proposed that gratitude plays a pivotal role in the development of social relationships (Algoe et al., 2013; Bartlett et al., 2012; Kubacka et al., 2011). Still, this proposition is primarily based on the observed effects of gratitude on reciprocity (Bartlett & DeSteno, 2006; Tsang, 2006b) and on the associations between the experience of gratitude and positive relational outcomes (closeness, conflict resolution, maintenance, goal contagion, mimicry) in social relationship (Algoe et al., 2013; Gordon et al., 2012; Grant & Gino, 2010; Jia, Lee, & Tong, 2015; Jia, Tong, & Lee, 2014; Kubacka et al., 2011; Lambert et al., 2010). Nevertheless, an actual demonstration of the relationship promoting effects of gratitude would require monitoring exchange processes over a long period, to investigate the ongoing dynamics between the reception of favours, the experience of gratitude, and the development of relational ties.

A second reason for using the framework of social exchange theory is that it helps reveal the positive role of indebtedness. The positive role of indebtedness cannot be revealed when simply assessed as an outcome of receiving a favour (for then it is just a negative experience that creates disutility), but is revealed only when studying the recipients’ response to a favour. It has been acclaimed that gratitude is “an adaptation for altruism” (McCullough et al., 2008), but so is indebtedness, as maintaining equity is believed to be the basis of sustainable social cooperation (e.g. Adams, 1965; Trivers, 1971). In many enduring social relationships such as between colleagues, friends, and neighbours, where people give favours back and forth that leads to multiple exchanges, reciprocity is essential for maintaining equity. Reciprocity in social exchange is often not instant, but is possibly implemented when the benefactor needs a favour someday in the future (e.g. Blau, 1964; Molm, Takahashi, & Peterson, 2000). This temporal delay on reciprocity requires long-term commitment, which is not guaranteed through any formal contracts or negotiations, but is guaranteed by the receiver’s feeling of indebtedness. When receiving a favour produces inequity, indebtedness is an internal reminder that keeps the receiver obligated until there is a chance that the benefits can be reciprocated. Without such a mechanism of indebtedness, social exchange might not be possible because people have to turn to formal contracts and negotiations to guarantee equity. So, we should embed helping event in the social exchange process to understand the adaptive function of indebtedness.

To conclude, we contend that gratitude and indebtedness serve distinct yet complementary social functions. Gratitude promotes social bonding, while indebtedness promotes the restoration of equity. To appreciate that, both emotions should be studied simultaneously in the context of social exchange. Finally, it deserves mentioning that even though we suggest that gratitude and indebtedness have distinct functions, they are also in some way co-dependent. The promotion of a relationship is unlikely without some level of reciprocation, and reciprocation is unlikely to be continued without the development of a relationship.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This work was supported by China Scholarship Council: [Grant Number 201507720023].

Notes

1. Although some researchers consider indebtedness a feature of gratitude (e.g. Lambert, Graham, & Fincham, 2009; Tesser, Gatewood, & Driver, 1968), the consensual view is that gratitude and indebtedness are distinct emotions (e.g. Goei & Boster, 2005; Tsang, 2006a; Watkins et al., 2006).

2. We only report here the part of survey that is relevant to our present research question. The remaining part of this survey can be found in the supplementary materials at https://osf.io/t8p2b, where all the datasets can also be accessed.

3. There were no gender effects on gratitude and indebtedness in Study 1.

4. Females reported overall more gratitude than males in both Study 2 (benefactor being female) and Study 3 (benefactor being male). This is consistent with findings of Kashdan, Mishra, Breen, and Froh (2009). There were no interaction effects between gender and cost manipulation in Study 2, and between gender and intentionality manipulation in Study 3.

5. We did not include a requested condition because the fact that requested help is initiated by the receiver may introduce confounding factors related to the receiver’s motivations (e.g. the receiver may request help from someone s/he already feels close to). Given that both offered and imposed favor are initiated by others, comparing these two conditions can avoid such confounding factors, and provide a comparison on the motivation of proximity seeking.