Balancing Prosocial Effort Across Social Categories: Mental Accounting Heuristics in Helping Decisions

Abstract

Three studies examine whether individuals might use mental accounting heuristics in helping decisions, budgeting their prosocial effort in similar ways to how money is budgeted. In a hypothetical scenario study (N = 283), participants who imagined that they previously helped someone of a specific social category (e.g., “family,” “colleagues”) were less willing to help someone of that category again. Similarly, when reporting actual instances of day-to-day help in a diary study (N = 443), having helped more than usual in a social category yesterday was associated with less effort and less time spent on helping in the same category today. In contrast, helping more than usual in other social categories did not reduce helping today. Finally, a scenario study (N = 489) suggested that the mental accounting effect in helping decisions may, in part, be explained by perceived utility of help (helping others in the same social category is seen as less rewarding).

Keywords

prosocial behavior interpersonal relationships mental accounting

Prosocial behavior such as doing small and large favors for others, providing practical and emotional support, giving advice, and sharing in care for dependents is the “social glue” (Eisenberg & Miller, 1987, p. 92) of any society. On a more personal basis, prosocial behavior helps maintain or improve social relationships. Helping a colleague with the printer may solidify a budding cordiality and provide a sense of connectedness and positive emotions in the helper. However, prosocial behavior is also costly: Helping costs effort, time, and sometimes money (Vohs & Ciarocco, 2004). Given that a person likely encounters multiple potential helping situations every day, and motivation to engage in effortful activities is limited (Inzlicht et al., 2014), how do people decide who and how to help? We propose that individuals aim to balance prosocial efforts across their social network. Investing extensive prosocial effort in one particular social category might lead people to be more reluctant to keep investing effort in that same mental category. We examine such mental accounting patterns in hypothetical helping decisions as well as in reported spontaneous prosocial behavior reported day to day.

Managing Prosocial Behavior

For most people, any given day brings multiple opportunities to be helpful. From fixing lunch for one’s partner in the morning, covering for a colleague while they take a break, to encouraging one’s sister in her job search over dinner, a person might act prosocially in any number of ways. In an experience sampling study tracking social interactions, North American adults reported an average of 12 social interactions a day (Zhaoyang et al., 2018), which included interactions with family members, friends, and others (e.g., coworkers, acquaintances, strangers). These multiple affiliations likely provide multiple demands or opportunities for prosocial behavior.

Prosocial effort can be experienced as pleasant and rewarding (e.g., Dunn et al., 2014; Inagaki & Orehek, 2017) but it also requires effort (e.g., Cameron et al., 2019; Vohs & Ciarocco, 2004), such as the effort involved in identifying the help needed, the effort involved in overriding selfish impulses, and the effort involved in the helping act itself. Thus, if there are multiple opportunities to help but limited available time and willingness to exert effort, individuals have to decide when to help or not to help on a daily basis. A number of factors in decisions to help have been identified in past research: Factors about the help recipient (e.g., perceived need for social support, Lemay & Clark, 2008; whether help was requested previously, Fennis & Janssen, 2010), factors about the relationship between helper and recipient (e.g., commitment, Rusbult et al., 1991; type of relationship, Fiske, 1992), factors about the helper (see Thielmann et al., 2020, for a meta-analysis) and aspects of the situation itself (e.g., game theory, Camerer, 2011) might influence prosocial decisions. In the present studies, we examine whether previous help is another factor influencing prosocial decisions. Do people manage the effort and time involved in helping in a fashion similar to how they manage other resources such as money?

Mental Accounting Heuristics

Financial resources have been shown to be allocated according to the heuristic of “mental accounting” (Cheema & Soman, 2006; Heath & Soll, 1996; Rajagopal & Rha, 2009; Thaler, 1999; Thaler & Shefrin, 1981; Tversky & Kahneman, 1981). According to mental accounting principles, when faced with multiple demands on limited resources, these demands are aggregated into mental categories. Resources are then managed within these categories (Tversky & Kahneman, 1981): Spending more than usual in one category reduces willingness to spend further in that particular category but does not necessarily reduce willingness to spend in other categories. The mental accounting heuristic is mostly implicit—people are often not aware of the various mental accounts they keep. Instead, mental accounting heuristics are typically inferred from people’s behavior or intended behavior (Cheema & Soman, 2006; Heath & Soll, 1996; Thaler, 1999; Thaler & Shefrin, 1981; Tversky & Kahneman, 1981). For example, in a hypothetical scenario study (Heath & Soll, 1996), people who were told to imagine they had purchased a US$50 ticket to a basketball game earlier in the week were less willing to buy a ticket to a play (i.e., same consumption category: entertainment) than people who were told to imagine that earlier in the week they had received a US$50 parking ticket (i.e., a different consumption category). Mental accounting has been examined primarily for individuals’ management of money (Chatterjee et al., 2009; Cheema & Soman, 2006; Heath & Soll, 1996; Thaler, 1999; Thaler & Shefrin, 1981; Tversky & Kahneman, 1981). There is some limited evidence that the same principles apply to how individuals manage time (e.g., Rajagopal & Rha, 2009; but see Soman, 2001, for contradictory evidence). In the present research, we propose that mental accounting resources might also apply to an entirely different domain: to how individuals manage prosocial effort. We propose that helping decisions might be influenced by heuristics that budget “prosocial resources” (effort and time spent on helping others) in mental accounts similar to how monetary resources are budgeted.

Applying mental accounting principles to prosocial behavior presumes that individuals track their prosocial effort within mental accounts. One such mental account might be social categories, or more specifically the types of people in one’s social network. Individuals tend to organize their social network into distinct social categories (e.g., Bacev-Giles & Peetz, 2016; Fiske & Haslam, 1997; Milardo, 1988; Zhaoyang et al., 2018). If mental accounting heuristics apply to prosocial effort, then helping more than usual in one social category might reduce willingness to help further in that same category but not affect willingness to help people in other social categories. In the present studies, we examine whether individuals follow social accounting heuristics in hypothetical and actually reported day-to-day decisions to act prosocial.

Social Categories

As the popularity of social media tools organizing people’s virtual social networks (e.g., Google+, where social connections are sorted within circles) suggests, individuals’ social connections are likely organized and grouped in social categories. For instance, when asked to nominate a person and then substitute that person for someone similar, people find it easy to think of someone else who is similar on several dimensions (Fiske & Haslam, 1997). Cluster analyses of individual’s ratings of their social networks suggest groupings or clusters of similarly rated groups of individuals (Haslam & Fiske, 1999). The theory of social relationships suggests four relational models: communal sharing relationships (e.g., family, close friends), authority ranking relationships (e.g., employer–employee), equality matching relationships (e.g., colleagues, less close friends), and “market-pricing” relationships involving transactions with measurable, single unit investments (e.g., monetary transactions, Fiske, 1992).¹ Similarly, social network theory suggests that a person’s social connections can be divided into significant others (e.g., family and close friends), exchange relationships (e.g., colleagues, less close friends), and interactive networks (e.g., acquaintances; Milardo, 1988). When asked to organize their social network into distinct social groups, the majority of university and community samples used labels such as family, friends, and colleagues/coworkers in one set of studies (Bacev-Giles & Peetz, 2016); and family, friends, and colleagues/coworkers or acquaintances in another set of studies (Zhaoyang et al., 2018). For the purpose of this research, we approximated participants’ social networks by relying on four social categories that broadly capture these theoretical and empirical distinctions: family, friends, colleagues, and acquaintances/strangers.

Possible Reasons for Relational Mental Accounting

Previous help might affect subsequent help for the same social category via several avenues. First, previous prosocial behavior can reduce subsequent willingness to help due to feeling morally “licensed” to make more selfish decisions. Imagining a past charitable act or remembering prior instances in which one has been helpful has been shown to decrease helping if prosocial acts are seen as “credits” for one’s morality (Jordan et al., 2011; Miller & Effron, 2010; Sachdeva et al., 2009; Zhong et al., 2009; although there is also evidence that the moral licensing effect may be smaller and less reliable than previously thought; Blanken et al., 2015; Kuper & Bott, 2019; Simbrunner & Schlegelmilch, 2017, for meta-analyses; and Ebersole et al., 2016, for a large-scale replication). Applying this phenomenon to mental accounting heuristics, it may be that individuals feel they have earned prosocial “credits” for an entire social category by helping an individual belonging to that category. For example, they might feel like they no longer need to prove that they care about their friends because they have demonstrated that concern via yesterday’s help to their friend Cindy from high school.

A second avenue by which previous help may reduce subsequent help in the same social category is fatigue through effort exerted when helping others. The motivated switching model of self-regulation (Inzlicht et al., 2014) posits that previous exertion of effort can reduce subsequent effort due to reduced motivation to exert further effort. If prosocial effort is managed in line with mental accounting heuristics, then feeling drained by previous help toward someone of one social category might lead people to shift their efforts to another category. Generally, helping others has been shown to affect feelings of fatigue or depletion (Bolino et al., 2015; Lam et al., 2016; Lanaj et al., 2016; Lin et al., 2016). For example, employees who reported having helped more of their coworkers on one day reported more subjective depletion (e.g., feeling drained) the next morning (Lanaj et al., 2016). In turn, there is evidence that feeling fatigued makes people less prosocial overall. Romantic relationship partners who felt drained were less likely to respond constructively in relationship conflicts (Finkel & Campbell, 2001), were less likely to offer support to romantic partners (Feeney & Collins, 2003; Iida et al., 2008), and were less willing to perform self-sacrifices for a partner (Findley et al., 2014).² Outside the context of romantic relationships, feeling drained has been linked to less prosocial behavior toward colleagues (Bolino et al., 2015; Trougakos et al., 2015). One study examining prosocial effort in the workplace (i.e., same social category) showed that supervisors who showed ethical leader behavior (e.g., discussing values with employees) on one day were more likely to act cruel toward employees (e.g., being rude, putting someone down in front of others) the next day (Lin et al., 2016). Extending the model of motivated self-regulation to mental accounting heuristics, it may be that fatigue due to helping others in a specific social category reduces further motivation to exert themselves particularly in that same social category.

A third avenue by which previous help may reduce subsequent help in the same social category may be via expectations for help given in return. Social exchange theory (e.g., Cropanzano & Mitchell, 2005; Thibaut & Kelley, 1959) suggests that individuals prefer balanced relationships where each person contributes equally over time. This is particularly the case in equity-matching types of relationships (Fiske, 1992) but can be a concern even in friendships (Xue & Silk, 2012). After exerting effort for a specific relationship (or a social category), individuals might reduce their effort toward further relationship maintenance of this social category until receiving some help in return.

Fourth, cognitive principles in economic decisions (e.g., Camerer, 2011; Kool & Botvinick, 2014) might play a role: Helping similar help recipients might have less utility than helping different help recipients. The rewards for prosocial behavior (e.g., positive affect, maintaining the relationship, signaling social standing) may be lower for repeated help for the same type of relationship, than for a different type of help recipient. For example, if one reward for helping one’s colleagues is to contribute to a better work climate, the first helping instance might bring greater perceived progress toward this goal than any subsequent act. Following the principles of labor/leisure trade-off (Kool & Botvinick, 2014), where decisions are motivated by finding an optimal balance between effortful engagement of executive control and leisure, perceiving diminishing returns for prosocial effort benefiting the same social category might explain less helping in that category.

Overview of Studies

The objective of this research was to examine how individuals manage prosocial behavior across multiple relationships in day-to-day life. We hypothesized that past prosocial behavior would affect subsequent prosocial behaviors differently depending on whether the help recipients are mentally organized into the same or different social categories. Specifically, we expected that individuals’ helping decisions would follow a pattern consistent with mental accounting heuristics: where more help given to someone in one social category reduces subsequent help in the same category but not in other categories.

In Study 1, we examined the role of social categories in hypothetical scenarios similar to scenario designs used in mental accounting research (e.g., Heath & Soll, 1996). We expected that people would report less willingness to exert effort for someone of the same social category as a recent help recipient compared with someone of a different social category.³ In Study 2, we examined self-reported actual prosocial behavior in a longitudinal study. Some people are genuinely more helpful than others across contexts (e.g., Cortes et al., 2014; Penner et al., 1995; Wilhelm & Bekkers, 2010), and longitudinal designs allow a separation of day-to-day variability in prosocial behavior from dispositional helpfulness. We expected that helping more than usual in one social category would lead to spending less effort, time, and money on helping in that same social category the next day. In Study 3, we examined four possible underlying processes (licensing, fatigue, exchange expectations, utility) for the mental accounting effect in helping decisions, using hypothetical scenarios. We determined sample sizes in advance, and all data were collected prior to analysis. All manipulations and exclusions in the study are disclosed in the main text. The unabridged surveys are available in the online methodology file in the Supplemental material. Surveys, data, syntax, and supplemental analyses are available at https://osf.io/3brhq/.

Study 1: Hypothetical Help

In an initial study, we assessed willingness to help via hypothetical helping scenarios (constructed similarly to financial scenarios in the mental accounting literature; Cheema & Soman, 2006; Heath & Soll, 1996). We examined social accounting patterns in people’s reported willingness to help people they actually know. Participants listed the names of actual people in their own lives. These names were then integrated in two helping scenarios: In one scenario, they were asked to imagine having helped someone of one social category yesterday and rated their willingness to exert time and effort today helping someone of the same category. In another scenario, participants were asked to imagine having helped someone of one social category yesterday and rated their willingness to exert time and effort today helping someone of a different social category. This study was preregistered before data collection (https://aspredicted.org/blind.php?x=xp5ya3).

Participants

We determined sample size in advance based on a power analysis, 80% power to detect a small (d = 0.20) within-participant effect requires 199 participants, and we overrecruited to account for potential exclusions. We posted 300 participation slots on Amazon Mechanical Turk (MTurk) and 308 U.S. MTurk workers completed the study. In line with preregistered exclusion criteria, 25 participants were excluded from analyses because they failed the attention check (final N = 283, M_age = 36.87 years, SD = 9.72 years, 44.4% female). All analyses were conducted after data collection was complete.

Method

Participants first thought of six people from their own social network. They listed two friends, two colleagues, and two family members by name or nickname. These names were later piped into the scenario text to make the hypothetical scenario more personally relevant. They then read two scenarios on hypothetical help that varied in the social category of the first help recipient and in the social category of the second help recipient:

Yesterday you helped your [friend<insert name1>/ colleague<insert name1>/ relative <insert name1>/ a stranger you met on your way home]. Helping them took some effort and took about 45 minutes. You felt good about being able to help [<insert name 1>/ this person].

Today, your [friend<insert name2>/ colleague<insert name2>/ relative <insert name2>/ a different stranger] is asking you for help. You don’t know yet how much time and effort this help would require.

One of the two scenarios specified two people of the same social category (e.g., “your colleague <name1>”; “your colleague <name 2>”; four variations) and one scenario specified two people of different social categories (e.g., “your colleague <name1>”; “your friend <name 2>”; 12 variations). Order of scenarios was counterbalanced. Participants rated the amount of effort they would be willing to invest in helping the second person in each scenario (“How much effort would you put into helping under these circumstances”) on a scale from 1 (no effort at all) to 7 (a lot of effort).⁴ They also explained their reasoning in open-ended format, which served as attention check.

Results

We conducted a within-subject analysis of variance (ANOVA) comparing willingness to exert effort for someone after reading the same category scenario with willingness to exert effort for someone after reading the other category scenario, controlling for order of scenario, and the social categories of the help recipients as covariates. There was a small but significant effect of the scenarios, with participants being more willing to exert effort for someone of a different social category (M = 5.41, 95% confidence interval [CI] = [5.25, 5.57]) than someone of the same social category (M = 5.39, 95% CI = [5.22, 5.56]), F(1, 277) = 4.62, p = .033, η² = .016, d = 0.26. Order did not have a significant effect, F(1, 277) = 1.64, p = .202, η² = .006, d = 0.16. The recipients of the help (stranger, colleague, friend, family) had a strong effect on willingness to exert effort on their behalf, for both the help recipient in the same category scenario, F(1, 277) = 20.32, p < .001, η² = .068, and the help recipient in the other category scenario, F(1, 277) = 11.53, p = .001, η² = .040. Participants were least willing to exert effort on behalf of strangers and most willing to exert effort on behalf of family members.

We also conducted this same analysis without those scenarios involving a family member (this additional analysis was preregistered and based on the assumption that helping norms might be different among family members). The effect of the scenarios was slightly stronger when leaving out family members, with participants being more willing to exert effort for someone of a different social category (M = 5.40, 95% CI = [5.14, 5.67]) than someone of the same social category (M = 5.18, 95% CI = [4.89, 5.46]), F(1, 102) = 5.41, p = .022, η² = .050.

Discussion

The social category of help recipients in a hypothetical help scenario mattered for decisions to invest resources to help. Participants were more willing to help someone in a different social category than someone in the same social category as a previous help recipient. However, although highly controlled, hypothetical helping scenarios do not capture the complexities of helping decisions in individuals’ day-to-day lives. In the next study, we examined self-reported helping in participants’ own actual social networks over several days.

Study 2

Participants in Study 2 recorded their daily prosocial behavior for 7 days. We expected that participants who invested more resources (effort, time) in helping someone of a specific social category on one day would be less likely to help people in that same social category the next day. The longitudinal design of this study allowed us to account for dispositional effects of average helping over the course of the week separately from day-to-day variability in prosocial behavior. By isolating variance at dispositional and situational levels (Curran & Bauer, 2011; Fleeson, 2004; Molenaar & Campbell, 2009), we are able to test whether helping more than usual in one social category affects helping in this category or in other social categories the next day, over and above any dispositional differences in helpfulness.

We hypothesized the following:

Hypothesis 1: Helping more than usual on one day would reduce helping the next day in the same type of social category as yesterday’s help.

Hypothesis 2: We expect a between-participant effect in which more dispositionally helpful people would tend to help more on any given day.

These hypotheses, data collection, and analysis plan were preregistered (https://osf.io/4ev2y/) on the last diary day of the study, before data were analyzed. We also registered an additional hypothesis (we expected that mental accounting principles would apply to type of help categories as well), however, we decided to limit the scope of this research report to the analysis of social categories rather than other possible categories.

Method

Participants

Our a priori recruitment goal was to collect 500 participants. This sample size was determined by effects in our previous studies, attrition concerns, as well as funds available to researchers (see preregistration for more detail on the data collection plan). We collected 511 U.S. participants through MTurk on one Sunday for the intake (January 20). Of these, 443 participants (M_age = 39.57 years, SD = 12.52 years, 54.3% female) completed at least one diary (N = 2,380 total diary reports) and were used in the analyses reported below. All analyses were conducted after data collection was complete.

Procedure

At intake, participants first completed a demographic survey and a number of dispositional measures that might be linked to helpfulness (see the Supplemental material). Participants were also trained on the mental categories that were going to be used in the diary. Specifically, they were given a definition of the five social categories (family members, romantic partner,⁵ friends, colleagues/coworkers, stranger/slight acquaintance), and were then asked to sort examples of specific relationships (“cousin,” “someone you are close to but not romantically involved with”) into these categories. All participants did these training tasks error free (i.e., they identified the correct category for each example).

Participants who indicated willingness to participate in the diary part of the study were asked to complete a report on their daily interactions every evening for the 7 days following the intake survey. They were sent a reminder email every evening at 8 p.m. and were asked to complete the survey at the end of the day before going to bed. Each survey was only available for that day. Two-hundred four participants (46%) completed all seven diaries, 121 participants completed five or six diaries (27.4%), 55 participants completed three to four diaries (11.2%), and 63 participants completed one to two diaries (12.9%).

In each diary entry, participants were first asked to “think back to what happened today and think about who you interacted with. Particularly, we would like you to think about who you helped or did a favor for today.” They were then asked “How many times did you help someone today?” If participants indicated that they had helped a person, they were asked to describe briefly how they had helped.

For each prosocial act, participants first described the helping instance, and selected the social category of the person they helped from drop-down menus. They then rated the effort involved in helping (“How much effort was involved in helping?” on a scale from hardly any effort [1] to a lot of effort [7]), and reported the time involved in helping (“Approximately how long did it take to help?”) in minutes. Participants also reported the money spent while helping in dollars. However, participants spent very little money on helping, and for sake of simplicity, the analyses reported below focus on effort and time. All analyses using money as an alternative helping resource are reported in the Supplemental material: https://osf.io/qasju/.

Data coding

For each diary report, the prosocial resource indices were aggregated by social category. For example, if a participant helped two friends on one day, the effort for both prosocial acts were summed to capture the total effort on helping friends that day. Because we aimed to compare next-day helping for the same social category with helping for other social categories, we also computed a complementary index of prosocial resources invested in helping nonfriends (i.e., family members + significant other + coworkers + strangers + others), nonfamily members, nonsignificant others, noncoworkers, and nonstrangers, by summing all other categories, respectively.

Analysis strategy

Analyses for the current and subsequent studies were conducted as multilevel structural equation models (SEM) in Mplus software (v.8.3; Muthén & Muthén, 1998–2017). We fit models to predict the prosocial resources invested (time, effort) in helping someone of each of the social categories. Models estimated lagged effects of yesterday’s resource investment on today’s resource investment in helping. A critical advantage of using multilevel SEM is its latent person mean-centering capability (e.g., Lüdtke et al., 2008). This method ensures that effects of yesterday’s helping on today’s helping reflect strictly within-person processes. Effects represent, for a given person, whether helping more than usual yesterday is associated with more or less helping today relative to their latent mean level of helpfulness. For models of helping within specific social categories, we estimate effects of yesterday’s resource investment in a specific social category (e.g., friends) as well as yesterday’s investment in all other categories (e.g., nonfriends). A between-person estimate of helpfulness in other relational domains is also included as a latent person mean predicting helping within a specific social category.

To take advantage of latent person mean centering, we used Bayesian estimation, implemented in Mplus as two-chain Markov chain Monte Carlo searches of the posterior distributions of each model parameter, given uninformative prior distributions. Results presented are the medians and standard deviations of each posterior parameter distribution, along with 95% credible intervals. These are conceptually similar to frequentist CIs, but instead reflect uncertainty in the estimate itself (rather than in reference to a null hypothesis) and can be interpreted as a 95% chance that the true population effect falls between the lower and upper bounds of the interval.

We had originally planned (and preregistered) to analyze the data with generalized multilevel linear modeling. However, upon further consideration, we were concerned about negative bias inherent in analyses involving lagged variables predicting their own later values in autoregressive models (Nickell, 1981). The multilevel SEM analysis we used instead that makes use of latent person mean centering eliminates this bias (McNeish & Hamaker, 2018), but the trade-off is that we are unable to make use of a response distribution suitable for highly skewed data. Below, we report the analysis we believe is the most conservative and best test of our hypothesis. The preregistered analysis is reported only briefly here and in more detail in the Supplemental material: https://osf.io/qasju/.

Results

Descriptive analyses of reported helping instances

For those reports that included at least one instance of help that day (n = 1,812 of 2,380 reports), the average number of helping instances per day was 1.97 (SD = 1.23, range = 1–6). The average prosocial effort invested per day was 2.74 (SD = 2.13) on a 7-point scale, and average time spent on helping was 31.62 min (SD = 96.23 min). As shown in Table 1, prosocial acts for family, friends, and significant other were comparably effortful and time consuming. Prosocial acts for colleagues were similarly effortful but required less time. Prosocial acts toward strangers were least effortful and time consuming.

Table 1.

Daily Help by Relationship Category; Number of Such Instances Across the Sample (n) and Average Resources Invested for These Instances of Helping (Effort, Time).

Social Category	n	Average effort per instance(7-point scale)		Average time per instance(in minutes)
Social Category	n	M	SD	M	SD
Friends	439	3.63	1.67	44.68	103.81
Family members	1,317	3.71	1.64	48.46	102.47
Significant other	534	3.91	1.78	39.75	47.97
Colleague	563	3.75	1.68	29.89	45.57
Stranger	500	2.70	1.87	16.98	47.93
Other category	91	2.35	1.34	13.90	18.39

Day-to-day effects across categories

In the initial analyses, we examined whether yesterday’s effort and time spent on helping affected today’s effort and time spent on helping, aggregated across social categories. Participants reported fewer instances of helping, B = −.16*, SD = 0.025, 95% CI = [−.21, −.11], less effort invested in helping, B = −.16*, SD = 0.028, 95% CI = [−.21, −.11], and less time spent on helping, B = −.10*, SD = 0.025, 95% CI = [−.15, −.05] on days after they helped more than their personal average the day before.

Day-to-day effects by social categories

In our main analyses, we examined day-to-day effects within social categories. Table 2 shows results of 10 models predicting effort and time expended on helping friends, family members, colleagues, strangers, and significant others today from yesterday’s effort and time expended in the same and in other categories, and from average effort and time expended over the week. We did not include helping instances in these analyses because when split across social categories, there were not enough instances to make such an analysis meaningful.

Table 2.

Between and Within-Person Effects of Helping Someone in the Same or in Other Social Categories Across Diary Days.

	Effort (on 7-point scales)Est (SD), CI [LL, UL]	Time (in minutes)Est (SD), CI [LL, UL]
Predicting daily help for friends
Help for friends yesterday	−.10* (.028), [−.16, −.05]	−.26* (.028), [−.32;−.20]
Help for other categories yesterday	−.04* (.013), [−.06, −.01]	−.00 (.010), [−.02, −.01]
Mean help for other categories	.04 (.019), [.00, .07]	.03 (.030), [−.04, .08]
Predicting daily help for family
Help for family yesterday	−.11* (.028), [−.17, −.06]	−.08* (.025), [−.13, −.03]
Help for other categories yesterday	−.01 (.018), [−.05, −.02]	−.02 (.056), [−.13, .09]
Mean help for other categories	−.05 (.080), [−.20, −.10]	.07 (.199), [−.33, .46]
Predicting daily help for colleagues
Help for colleagues yesterday	−.05 (.029), [−.11, .01]	−.22* (.030), [−.28, −.16]
Help for other categories yesterday	−.01 (.016), [−.0, .02]	−.00 (.005), [−.01, .01]
Mean help for other categories	.00 (.0024), [−.05, .05]	−.00 (.020), [−.04, .04]
Predicting daily help for strangers
Help for strangers yesterday	−.07* (.035), [−.14, −.00]	−.00 (.032), [−.07, .06]
Help for other categories yesterday	.02 (.014), [−.00, .05]	.00 (.004), [−.01, .01]
Mean help for other categories	−.03 (.017), [−.06, .01]	−.01 (.018), [−.04, .03]
Predicting daily help for significant other
Help for significant other yesterday	.01 (.029), [−.05, .06]	.00 (.033), [−.07, .06]
Help for other categories yesterday	.00 (.015), [−.03, .03]	.00 (.006), [−.01, .01]
Mean help for other categories	.02 (.033), [−.04, .08]	−.01 (.028), [−.06, .05]

Note. Est = unstandardized regression coefficient calculated as the mode of the posterior distribution for the given parameter; SD = standard deviation of the posterior distribution; CI = Bayesian credible interval, LL = lower limit; UL = upper limit.

Credible intervals that do not include zero (corresponds to p < .05 under a frequentist statistical model).

Overall, prosocial effort and time expended in one social category yesterday predicted fewer resources expended in the corresponding category today (i.e., a mental accounting effect, Hypothesis 1). These mental accounting effects are described below. For effort spent on helping, spending more effort than usual on friends, family, and strangers yesterday significantly reduced today’s effort in the same social categories (but yesterday’s effort did not reduce today’s effort in the colleague or significant-other category). In contrast, spending more effort than usual on other social categories yesterday reduced effort in the friend category only. For time spent on helping, spending more time than usual on friends, family, and colleagues yesterday significantly reduced time spent today in the same social categories (but yesterday’s time did not reduce today’s time in the stranger or significant-other category). In contrast, spending more time than usual in other social categories yesterday did not reduce time spent on any of the categories. In additional analyses, we also included in each model a difference test of the regression paths from help in the same category versus help in the other category (see the Supplemental material: https://osf.io/qasju/).

In additional exploratory analyses, we also examined these models for data from weekdays only (n = 443; Table 3). It is possible that interactions (especially with colleagues) change during the weekend. Overall, prosocial effort and time expended in one social category yesterday predicted fewer resources expended in the corresponding category today (i.e., a mental accounting effect, Hypothesis 1). For effort spent on helping, spending more effort than usual on friends, family, colleagues, strangers, and even significant others yesterday significantly reduced today’s effort in the same social categories. In contrast, spending more effort than usual on other social categories yesterday reduced effort in the friend category only and increased effort in the stranger category. For time spent on helping, spending more time than usual on friends, colleagues, and significant others yesterday significantly reduced time spent today in the same social categories. In contrast, spending more time than usual in other social categories yesterday reduced time spent in the significant-other category only and did not affect helping in the other categories. Notably, yesterday’s time spent on helping increased today’s time in the family and stranger category, despite also reporting that they spent less effort today. These effects may represent helping projects that spanned several days or scheduled weekday time commitments such as babysitting for family members or volunteering.

Table 3.

Between and Within-Person Effects of Helping Someone in the Same or in Other Social Categories Across Diary Days—Weekdays Only.

	Effort (on 7-point scales)Est (SD), CI [LL, UL]	Time (in minutes)Est (SD), CI [LL, UL]
Predicting daily help for friends
Help for friends yesterday	−.22* (.033), [−.28, −.15]	−.12* (.038), [−.19, −.04]
Help for other categories yesterday	−.03* (.015), [−.06; −.00]	−.00 (.024), [−.04, .05]
Mean help for other categories	.01 (.024), [−.04, .06]	.06* (.032), [.00, .13]
Predicting daily help for family
Help for family yesterday	−.16* (.032), [−.22, −.10]	.23* (.037), [.16, .30]
Help for other categories yesterday	−.00 (.023), [−.05, .04]	.00 (.029), [−.05, .06]
Mean help for other categories	−.13 (.078), [−.30, .02]	.23 (.208), [−.17, .67]
Predicting daily help for colleagues
Help for colleagues yesterday	−.29* (.032), [−.35, −.23]	−.37* (.038), [−.45, −.30]
Help for other categories yesterday	.01 (.019), [−.03, .04]	.01 (.013), [−.02, .03]
Mean help for other categories	−.09* (.037), [−.16, −.01]	−.02 (.042), [−.10, .06]
Predicting daily help for strangers
Help for strangers yesterday	−.16* (.045), [−.25, −.07]	.14* (.043), [.06, .23]
Help for other categories yesterday	.06* (.017), [.03, .09]	.02 (.009), [−.00, .04]
Mean help for other categories	−.04 (.023), [−.08, .01]	−.01 (.017), [−.04, .03]
Predicting daily help for significant other
Help for significant other yesterday	−.11* (.039), [−.18, −.03]	−.37* (.049), [−.47, −.28]
Help for other categories yesterday	−.02 (.017), [−.05, .02]	−.04* (.015), [−.06, −.01]
Mean help for other categories	.09* (.033), [.02, .14]	.04 (.030), [−.02, .10]

Credible intervals that do not include zero (corresponds to p < .05 under a frequentist statistical model).

Alternative analysis

We had originally planned to analyze the data with generalized multilevel linear modeling. Results for these alternative analyses are available in the Supplemental material. Using generalized multilevel linear modeling, prosocial resources expended in one social category yesterday also predicted less effort and less time expended in the same category today in all five social categories, with overall stronger effects than shown in the multilevel SEM analysis. However, we suspect these stronger effects are attributable to negative bias in lagged prediction described earlier (Nickell, 1981).

It is also important to note that the present effects do not reflect regression to the mean. Regression to the mean describes the tendencies of extreme values to be less extreme and closer to the mean in repeated assessment, a problem that commonly occurs when difference scores between assessments are used as outcomes without adjusting for initial status.

Discussion

This study provided evidence for mental accounting principles in day-to-day prosocial decisions. The overall effect of less helping on days after helping more than usual appears to be driven by a pronounced drop off in helping within the same social categories, but less of a drop off in helping in other social categories. Investing effort and time in one social category previously reduced these prosocial resources invested in the same category the next day. This pattern of effects was apparent across several types of social categories (family members, friends, colleagues, and strangers) but did not extend to the significant-other category (at least when considered across the entire week). This pattern suggests that help for very close others such as romantic partners is not tracked in the same way as it is for other social categories (this finding is in line with relational models theory, which suggests tracking should not occur in communal sharing relationships; Fiske, 1992). However, exploratory analyses considering only weekdays showed that the next-day drop in helping within the romantic partner category was significant. This finding might suggest that help within communal relationships may be prioritized when ample resources are available (e.g., when there are few other demands on time and effort in the weekend) but are budgeted similar to other relationships when resources are scarce. Indeed, exploratory analyses showed that participants reported spending more time (∆M = 3.51 min more, SE = 0.14, p = .015) and more effort (∆M = 0.17 more on a 7-point scale, SE = 0.08, p = .045) helping their romantic partners during the weekend than during weekdays.

Study 3

Several processes potentially underlie the use of mental accounting heuristics for helping decisions. In Study 3, we conducted an exploration into these possible reasons for the mental accounting effect. Specifically, we examined whether the effect might be stronger among those who feel they have earned prosocial credits for past help (licensing), those who feel more fatigued after past help, and those who expect help in return after past help (exchange expectations). If these factors moderate the strength of the mental accounting effect, this moderation would provide initial evidence for the importance of these cognitions. To examine the role of utility of help in mental accounting heuristics, we compared participants’ judgments of how rewarding helping would be (affective utility) in each social category. Note that the design of this study provides initial insights into the role of four possible underlying processes of the mental accounting effect in helping decisions, but cannot establish true mechanisms (Bullock et al., 2010).

This study employed hypothetical scenarios, which were based on common helping descriptions in Study 2. Participants were asked to imagine two helping situations in which friends or colleagues had required their help “yesterday,” and one helping situation in which a friend or colleague required their help “today.” The latter helping recipient was either of the same or of a different social category as the helping recipients in the first two scenarios. We expected that participants would report less hypothetical effort and time spent helping in the different social category condition than in the same social category condition. We also examined whether this link would be stronger if people reported greater feelings of licensing, greater fatigue, and greater expectations for receiving help in return after “yesterday’s” helping scenarios, and when perceiving “today’s” help as more rewarding in one condition over the other (utility). This study was preregistered before data collection (https://aspredicted.org/blind.php?x=zp9ji4).

Method

Participants

We determined sample size in advance, aiming for 80% power to detect a condition effect of the size found for hypothetical scenarios in Study 1 (d = 0.26), and overrecruiting to account for exclusions. We posted 500 participation slots on MTurk, and 506 U.S. MTurk workers completed the study. In line with preregistered exclusion criteria, 17 participants were excluded from analyses because they failed the attention check (final N = 489, M_age = 39.32 years, SD = 12.21 years, 42.1% female). All analyses were conducted after data collection was complete.

Procedure

After reporting only age and gender, participants read three scenarios that described an opportunity to help. These scenarios were taken from real helping descriptions in Study 2 (helping someone move, helping someone program a new phone, helping someone drop off a parcel at the post office). Two of the scenarios were described as happening on the same day (i.e., similar to “yesterday’s helping” in the diary study), and the third scenario was described as happening the day after that (i.e., similar to “today’s helping” in the diary study). For the third scenario, participants were randomly assigned to read about a help recipient of the same social category (n = 242) or a different social category (n = 247) than the first two scenarios. The social categories of the initial and the third scenarios (friends, colleagues) were counterbalanced, and the sex of the help recipient was matched to the participant’s sex. Controlling for these variables did not affect results and they will not be discussed further.

After each scenario, participants rated the amount of effort they would be willing to invest (“How much effort would you put into helping in this situation?”) on a scale from 1 (no effort at all) to 7 (a lot of effort). We measured how much time they would be willing to invest (“How much time would you spend helping?”) on a 7-point scale (1 = 0 min, 2 = 1–5 min, 3 = 6–10 min, 4 = 11–15 min, 5 = 16–30 min, 6 = half to 1 hr, 7 = more than 1 hr).

Participants also rated their expected affective utility of providing help in this situation on two items (e.g., “How rewarding would it be to help your friend/colleague?”) from not at all happy (1) to very happy (7), which correlated (r = .79) and were averaged into a measure of utility. To assess the other three potential processes, participants rated how they would feel at the end of the day after the first two scenarios detailing “yesterday’s” help. Two questions assessed licensing (e.g., “I would feel like I’ve earned credit with my friendships/colleagues,” r = .52), two questions assessed fatigue (e.g., “I would feel mentally or physically tired.,” r = .72), and two questions assessed exchange expectations (e.g., “I would feel that it would be my friends’/colleagues’ turn to help me,” r = .78). All questions were answered on scales from not at all (1) to very much (7). These six items loaded on three components that matched our proposed factors (see the Supplemental material for factor analysis: https://osf.io/qasju/). Table 4 presents means and correlations. Finally, participants explained their reasoning about the helping decisions they had made for the scenarios in open-ended format, which served as attention check.

Table 4.

Descriptive Statistics and Correlations for Cognitions About Licensing, Fatigue, Exchange Expectations, and Utility (Study 3).

	Condition
	Same category		Other category		Correlations
	M	SD	M	SD	1.	2.	3.	4.
1. Feeling they earned credits with “yesterday’s” help	4.00	0.80	4.08	0.71	1	−.04	.06	.37**
2. Feeling fatigued due to “yesterday’s” help	2.92	1.11	2.96	1.12	−.05	1	.34**	−.11
3. Expecting help in return due to “yesterday’s” help	2.90	1.09	2.94	1.10	.20**	.32**	1	−.01
4. Utility of “today’s” help	4.48	1.74	4.85	1.66	.20**	.03	−.06	1

Note. Correlations above the diagonal for same category condition, correlations below the diagonal for other category condition.

p < .05. **p < .01.

Results

Independent t tests revealed no significant effect of condition on hypothetical effort, t(486) = 1.29, p = .199, d = 0.12. Participants reported that they would invest slightly more effort in the different social category condition (M = 5.05 [4.95, 5.15]) than in the same social category condition (M = 4.85 [4.74, 4.96]), but this difference was not significant. There was, however, a significant effect of condition on the time participants reported being willing to invest in hypothetical helping, t(486) = 2.10, p = .037, d = 0.19, with participants reporting that they would invest more time in the different social category condition (M = 4.56 [4.46, 4.66]) than in the same social category condition (M = 4.25 [4.14, 4.36]).

Next, we examined how the condition effect may be moderated by feelings of having earned credits, fatigue, and exchange expectations. We focused on participants’ reports of the time they would invest in helping because condition effects were significant only for this variable (effort analyses are reported in the Supplemental material: https://osf.io/qasju/). In multiple regression analyses, we entered condition, the three potential moderators, and their interaction terms with condition as predictor variables, and time as outcome variable.⁶ All predictor variables were centered. The regression analysis revealed a significant main effect of condition, B = .29, SE = 0.14, β = .09, t(480) = 2.01, p = .045, a significant main effect of licensing, B = .34, SE = 0.10, β = .16, t(480) = 3.50, p = .001, and exchange expectations, B = −.16, SE = 0.07, β = −.11, t(480) = −2.24, p = .025, but not fatigue, B = −.04, SE = 0.07, β = −.03, t(480) = −0.55, p = .581. The Licensing × Condition interaction was significant, B = −.45, SE = 0.19, β = −.11, t(480) = −2.32, p = .021; the Fatigue × Condition interaction was not significant, B = .14, SE = 0.14, β = .05, t(480) = 1.02, p = .310; and the Exchange Expectations × Condition interaction was not significant, B = .12, SE = 0.14, β = .04, t(480) = 0.86, p = .390. The Licensing × Condition interaction is depicted in Figure 1. The condition effect was particularly strong for those who did not feel like they had earned credit with that particular social category after the first two helping scenarios, B = .63 [0.22, 1.03], t(480) = 3.05, p = .002, but was not significant for those who felt they had earned credits, B = −.05 [−0.45, 0.35], t(480) = −0.26, p = .795. This is the opposite pattern one might expect if licensing is a process underlying the mental accounting pattern in helping intentions. Instead, this pattern suggests that if previous help did not provide a sense of having earned prosocial credits with that social category, people are more likely to switch to helping a different social category.

Figure 1.

Estimated time invested in hypothetical help for a person of the same or a different social category than previous help recipients at high and low levels of feeling they earned credit with that social category.

To examine the role of utility, we regressed the expected affective utility on condition, controlling intended effort and time spent on helping, as preregistered. Condition marginally affected perceived utility, B = .22, SE = 0.12, t(482) = 1.88, p = .061, suggesting that participants thought they would feel somewhat happier and more rewarded by helping someone of a different social category than someone of the same social category. In exploratory (not preregistered) analyses, we also examined the indirect effect of condition on time intended for helping, with perceived utility, licensing, fatigue, and exchange expectations as simultaneous mediators (PROCESS V3.8; Hayes, 2018, Model 4). Analyses showed a significant indirect effect for perceived utility, B = .11 [0.02, 0.21], but no significant indirect effect for licensing, B = .002 [−0.01, 0.02], fatigue, B = −.001 [−0.01, 0.01], and exchange expectations, B = −.002 [−0.02, 0.01]. This result suggests that the condition effect on investing more time in helping for others of a different social category than others of the same social category as previous help recipients could be partly explained by expecting this help to be more rewarding (Figure 2).

Figure 2.

Coefficients presented are unstandardized regression coefficients (B) with standard error (SE) in parentheses.

Discussion

When making helping decisions in hypothetical scenarios, participants showed a preference for investing more time (but not significantly more effort) for people of the same rather than for people of a different social category as previous hypothetical help recipients. There was little evidence that feeling “morally licensed,” feeling fatigued, or expecting help in return increased this preference. On the contrary, the preference for helping people of a different social category was stronger among those who felt the previous help had not earned them credit with the previous help recipients’ group. It may be that the feeling of having earned credits with one social group is a form of utility and people tend to prioritize social categories that provide more utility. In line with this potential importance of perceived utility, participants expected that helping others of a different social category than the previous help instances would feel more rewarding than helping others of the same social category. In other words, the expected affective utility of help for the helper appeared to contribute to the mental accounting effect. It is important to note that we focused on utility in terms of affective rewards (how rewarding would helping be, how happy would it make you), whereas there are a number of other utility aspects that might play a role in determining decisions to help (e.g., maintaining the relationship, signaling social standing). More research is needed to examine the different aspects of utility of help in helping decisions across social categories.

General Discussion

How do people manage the multiple opportunities for prosocial behavior they encounter in day-to-day life? Two scenario studies of hypothetical helping decisions and one longitudinal diary study of actual helping in daily life showed that the social category of the help recipients matters. Imagining or actually helping in one social category predicted less helping in the same social category the next day, suggesting that individuals balance their prosocial resources across the social categories of people in their life using heuristics similar to heuristics that have been shown to influence budgeting of financial decisions (e.g., Thaler, 1999/2008). This mental accounting effect may, in part, be driven by a greater perceived utility of helping others of diverse social categories than the same social category (Study 3).

Social behavior such as decisions to help others differs from financial decisions to buy or invest money in several ways: Prosocial effort is subjective (e.g., the same task might feel more effortful at some times than others), there is more uncertainty about how much helping might “cost” the helper (e.g., the helping tasks might require more effort than anticipated), as well as uncertainty about returns for effort (paying US$1 for chocolate guarantees a chocolate; investing more effort in helping others does not guarantee positive outcome for the other person or for the self). However, the present studies suggest that the same mental principles appear to apply when managing effort across multiple helping occasions as when managing money across multiple purchase occasions. Mental categories might be a spontaneous cognitive strategy to organize a large set of variables into meaningful structures not just in the financial domain but in many other domains as well.

Limitations

Self-report

This study relies on self-reports of helping behavior. One issue of self-report is that we relied on participants’ own definition of what helping meant to them. This context not only provides ecological validity but also led to a widely varied set of helping behaviors. In Study 2, participants gave examples varying from letting someone cut in line in a traffic jam to minding a baby for several hours or preparing an elaborate birthday party. Note, however, that such inclusiveness reflects the wide range of prosocial behavior studied in past research: Even acts that were not committed (e.g., racist actions not performed) can be regarded as prosocial behavior (Effron et al., 2012).

Different mental accounts

We focused on four social categories as mental accounts (family, friends, colleagues, strangers). There may well be other mental accounts that are used to organize individuals’ social network beyond the distinctions we made. Social categories might vary in level of abstraction with some individuals making finer distinctions than others (e.g., separating or including “in-laws” with “family”), and vary with regard to the criteria used for grouping relationships (e.g., closeness, kinship, goal instrumentality). Mental accounting tendencies can differ across cultures (Banerjee et al., 2019) and culture might also influence how relational categories are shaped and the tendency to manage social effort across them. In addition to variations in social categories, there might also be different mental accounts that influence how effort is distributed across helping instances: for example, the type of help might matter. Individuals might also invest more effort and time in a different type of help than previous help (e.g., running errands vs. cooking dinner).

Time course of accounting

We conceptualized subsequent helping instances as helping the next day (Study 1-3). This time interval was chosen for methodological reasons (diary assessments at the end of the day make it more likely that participants have some helping to report, whereas helping assessed throughout the same day might result in a lot of null helping reports) rather than conceptual reasons. It is at least equally likely that having helped earlier on one day affects helping decisions later that same day. Future research might examine more immediate effects (e.g., within the same day) or more lagged effects (e.g., several days later) of prosocial efforts. The time course of mental accounting patterns might also help determine the specifics of how accounts are balanced. Willingness to help might “reset” after a certain amount of time has passed or after a perceived new temporal category boundary.

Future Directions

Consistency as an alternative heuristic

Contrasting the mental accounting pattern of prosocial behavior, there are also reasons to expect the opposite pattern: Sometimes prior help might increase helping in the same social category rather than a different social category. The moral behavior literature also provides evidence for a possible consistency effect, where recalling a previous prosocial behavior leads people to view themselves as having a prosocial identity, and therefore increases their willingness for further prosocial behaviors such as volunteering and donations to a local food bank (Aquino & Reed, 2002), donations to charities (Conway & Peetz, 2012; Reed & Aquino, 2003), and buying presents for others (Gneezy et al., 2012). In the context of mental categories, prior prosocial behavior might signal the importance of helping in general, helping a particular person, or the importance of helping one social category (e.g., “family is important to me”). If the latter consideration is salient, people may prefer helping the same social category again rather than switching to another social category. Future research could examine the boundaries of the mental accounting pattern and examine whether and how this pattern can be flipped to a consistency pattern for helping people of the same mental category.

Account input

In interpersonal relationships, prosocial resources are not only expended but also received. First, receiving relational support by others can help to conserve energy (Fitzsimons et al., 2015; Fitzsimons & Finkel, 2011) and might free up (or add) available resources for later helping behaviors. Second, giving relational support can energize or otherwise benefit the support giver directly (Brown et al., 2003; Inagaki & Orehek, 2017). Future studies may examine whether receiving help (i.e., incoming prosocial resources) increases willingness to help others across all social categories or specifically within the social category that increased regulatory resources. How received prosocial effort or prosocial support affects individuals’ own helping behavior might also depend on other variables such as whether the relationships are communal (Clark et al., 1989; Rusbult et al., 1991) or whether the helping individual tends toward a communal or exchange orientation.

Potential advantages and pitfalls of mental accounting heuristics

Mental accounting heuristics for prosocial behavior might provide benefits such as distributing maintenance resources across different types of relationships, ensuring that people can sustain a variety of relationships. This could be an advantage because we reap different rewards from different types of relationships (e.g., Constant et al., 1996), and because it minimizes the risk of losing an entire social network over one “falling out.” Second, such heuristics might lead to prioritizing rewarding types of relationships that fulfill more than one function. For instance, people may end up investing more prosocial resources in relationships that fall in multiple social categories.

There may also be potential pitfalls of using social categories to help determine where relationship maintenance efforts are invested. One problem may be the underresponding in categories that have been draining one’s resources. Because mental accounting heuristics track expenses not on the individual level, but on the category level, individual relationships might be neglected because other individual relationships within the same category are particularly draining. Another potential problem might be overresponding in a social category—paying attention and providing support where none is needed. Such behavior might backfire (Gleason et al., 2008; Maisel & Gable, 2009) and make prorelational investments less efficient. In sum, more research is needed to examine the functionality of using mental accounting heuristics in managing relationship maintenance efforts.

Conclusion

This research examined a novel aspect about interpersonal cognition: Individuals’ helping decisions followed mental accounting principles that take into account the mental categories of the help recipients. More generally, our studies suggest that individuals may balance between the different social categories in their social network when making helping decisions. This research contributes to the emerging literature examining judgment and decision processes in relationships (Joel et al., 2013) and suggests that mental accounting heuristics (Thaler, 1998/2008) can apply to cognition well beyond financial decision making. Of course, any prosocial behavior will be determined by multiple dispositional and situational factors. However, the present studies suggest that beyond the current situation, it might also matter what individuals have previously done and for whom they have done it.

Supplemental Material

sj-docx-1-psp-10.1177_0146167220976683 – Supplemental material for Balancing Prosocial Effort Across Social Categories: Mental Accounting Heuristics in Helping Decisions

Supplemental material, sj-docx-1-psp-10.1177_0146167220976683 for Balancing Prosocial Effort Across Social Categories: Mental Accounting Heuristics in Helping Decisions by Johanna Peetz and Andrea Howard in Personality and Social Psychology Bulletin

Footnotes

Acknowledgements

We thank Mariya Davydenko, Iuliia Kolotylo, Marta Kolbuszewska, and Aaron Maccosham for research assistance.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The research was funded by a Grant of the Social Science and Humanities Research Council of Canada (#435-2012-1211) to the first author.

ORCID iD

Johanna Peetz

Supplemental Material

Supplemental material is available online with this article.

Notes

References

Aquino

Reed

A . (2002). The self-importance of moral identity. Journal of Personality and Social Psychology, 83, 1423–1440. https://doi.org/10.1037//0022-3514.83.6.1423

Bacev-Giles

Peetz

(2016). Thinking big or small: Does mental abstraction affect social network organization? PLOS ONE, 11(7), Article e0160047. https://doi.org/10.1371/journal.pone.0147325

Banerjee

Chatterjee

Mishra

A. A.

(2019). Loss is a loss, why categorize it? Mental accounting across cultures. Journal of Consumer Behaviour, 18(2), 77–88.

Blanken

van de Ven

Zeelenberg

(2015). A meta-analytic review of moral licensing. Personality and Social Psychology Bulletin, 41(4), 540–558. https://doi.org/10.1177/0146167215572134

Bolino

M. C.

Hsiung

H. H.

Harvey

LePine

J. A.

(2015). “Well, I’m tired of tryin’!” Organizational citizenship behavior and citizenship fatigue. Journal of Applied Psychology, 100, 56–74. https://doi.org/10.1037/a0037583

Brown

S. L.

Nesse

R. M.

Vinokur

A. D.

Smith

D. M.

(2003). Providing social support may be more beneficial than receiving it: Results from a prospective study of mortality. Psychological Science, 14(4), 320–327. https://doi.org/10.1111/1467-9280.14461

Bullock

J. G.

Green

D. P.

S. E.

(2010). Yes, but what’s the mechanism? (Don’t expect an easy answer). Journal of Personality and Social Psychology, 98(4), 550–558.

Burkley

Anderson

Curtis

(2011). You wore me down: Self-control strength and social influence. Social and Personality Psychology Compass, 5(7), 487–499. https://doi.org/10.1111/j.1751-9004.2011.00367.x

Camerer

C. F.

(2011). Behavioral game theory: Experiments in strategic interaction. Princeton University Press.

10.

Cameron

C. D.

Hutcherson

C. A.

Ferguson

A. M.

Scheffer

J. A.

Hadjiandreou

Inzlicht

(2019). Empathy is hard work: People choose to avoid empathy because of its cognitive costs. Journal of Experimental Psychology: General, 148(6), 962–976. https://doi.org/10.1037/xge0000595

11.

Chatterjee

Heath

T. B.

Min

(2009). The susceptibility of mental accounting principles to evaluation mode effects. Journal of Behavioral Decision Making, 22(2), 120–137.

12.

Cheema

Soman

(2006). Malleable mental accounting: The effect of flexibility on the justification of attractive spending and consumption decisions. Journal of Consumer Psychology, 16, 33–44. https://doi.org/10.1207/s15327663jcp1601_6

13.

Clark

M. S.

Mills

J. R.

Corcoran

D. M.

(1989). Keeping track of needs and inputs of friends and strangers. Personality and Social Psychology Bulletin, 15, 533–542. https://doi.org/10.1177/0146167289154007

14.

Constant

Sproull

Kiesler

(1996). The kindness of strangers: The usefulness of electronic weak ties for technical advice. Organization Science, 7, 119–135. https://doi.org/10.1287/orsc.7.2.119

15.

Conway

Peetz

(2012). When does feeling moral actually make you a better person? Conceptual abstraction moderates whether past moral deeds motivate consistency or compensatory behavior. Personality and Social Psychology Bulletin, 38, 907–919. https://doi.org/10.1177/0146167212442394

16.

Cortes

Kammrath

L. K.

Scholer

A. A.

Peetz

(2014). Self-regulating the effortful “social dos.” Journal of Personality and Social Psychology, 106(3), 380–397. https://doi.org/10.1037/a0035188

17.

Cropanzano

Mitchell

M. S.

(2005). Social exchange theory: An interdisciplinary review. Journal of Management, 31(6), 874–90.

18.

Curran

P. J.

Bauer

D. J.

(2011). The disaggregation of within-person and between-person effects in longitudinal models of change. Annual Review of Psychology, 62, 583–619. https://doi.org/10.1146/annurev.psych.093008.100356

19.

Dunn

E. W.

Aknin

L. B.

Norton

M. I.

(2014). Prosocial spending and happiness: Using money to benefit others pays off. Current Directions in Psychological Science, 23(1), 41–47.

20.

Ebersole

C. R.

Atherton

O. E.

Belanger

A. L.

Skulborstad

H. M.

Allen

J. M.

Banks

J. B.

. . . Brown

E. R.

(2016). Many Labs 3: Evaluating participant pool quality across the academic semester via replication. Journal of Experimental Social Psychology, 67, 68–82.

21.

Effron

D. A.

Miller

D. T.

Monin

(2012). Inventing racist roads not taken: The licensing effect of immoral counterfactual behaviors. Journal of Personality and Social Psychology, 103(6), 916–932. https://doi.org/10.1037/a0030008

22.

Eisenberg

Miller

P. A.

(1987). The relation of empathy to prosocial and related behaviors. Psychological Bulletin, 101(1), 91–119. https://doi.org/10.1037/0033-2909.101.1.91

23.

Feeney

B. C.

Collins

N. L.

(2003). Motivations for caregiving in adult intimate relationships: Influences on caregiving behavior and relationship functioning. Personality and Social Psychology Bulletin, 29(8), 950–968. https://doi.org/10.1177/0146167203252807

24.

Fennis

B. M.

Janssen

(2010). Mindlessness revisited: Sequential request techniques foster compliance by draining self-control resources. Current Psychology, 29, 235–246. https://doi.org/10.1007/s12144-010-9082-x

25.

Findley

M. B.

Carvallo

Bartak

C. P.

(2014). The effect of self-control on willingness to sacrifice in close relationships. Self and Identity, 13, 334–344. https://doi.org/10.1080/15298868.2013.826595

26.

Finkel

E. J.

Campbell

W. K.

(2001). Self-control and accommodation in close relationships: An interdependence analysis. Journal of Personality and Social Psychology, 81, 263–277. https://doi.org/10.1037/0022-3514.81.2.263

27.

Fiske

A. P.

(1992). The four elementary forms of sociality: Framework for a unified theory of social relations. Psychological Review, 99(4), 689–723.

28.

Fiske

A. P.

Haslam

(1997). The structure of social substitutions: A test of relational models theory. European Journal of Social Psychology, 27(6), 725–729.

29.

Fitzsimons

G. M.

Finkel

E. J.

(2011). Outsourcing self-regulation. Psychological Science, 22, 369–375. https://doi.org/10.1177/0956797610397955

30.

Fitzsimons

G. M.

Finkel

E. J.

vanDellen

M. R.

(2015). Transactive goal dynamics. Psychological Review, 122, 648–674. https://doi.org/10.1037/a0039654

31.

Fleeson

(2004). Moving personality beyond the person-situation debate. Current Directions in Psychological Science, 13(2), 83–87. https://doi.org/10.1111/j.0963-7214.2004.00280.x

32.

Gleason

M. E. J.

Iida

Shrout

P. E.

Bolger

(2008). Receiving support as a mixed blessing: Evidence for dual effects of support on psychological outcomes. Journal of Personality and Social Psychology, 94, 824–838. https://doi.org/10.1037/0022-3514.94.5.824

33.

Gneezy

Imas

Brown

Nelson

L. D.

Norton

M. I.

(2012). Paying to be nice: Consistency and costly pro-social behavior. Management Science, 58, 179–187. https://doi.org/10.1287/mnsc.1110.1437

34.

Haslam

Fiske

A. P.

(1999). Relational models theory: A confirmatory factor analysis. Personal Relationships, 6(2), 241–250.

35.

Hayes

A. F.

(2018). The PROCESS macro for SPSS and SAS (version 3).

36.

Heath

Soll

J. B.

(1996). Mental budgeting and consumer decisions. Journal of Consumer Research, 23, 40–52. https://doi.org/10.1086/209465

37.

Iida

Seidman

Shrout

P. E.

Fujita

Bolger

(2008). Modeling support provision in intimate relationships. Journal of Personality and Social Psychology, 94, 460–478. https://doi.org/10.1037/0022-3514.94.3.460

38.

Inagaki

T. K.

Orehek

(2017). On the benefits of giving social support: When, why, and how support providers gain by caring for others. Current Directions in Psychological Science, 26(2), 109–113. https://doi.org/10.1177/0963721416686212

39.

Inzlicht

Schmeichel

B. J.

Macrae

C. N.

(2014). Why self-control seems (but may not be) limited. Trends in Cognitive Science, 18, 127–133. https://doi.org/10.1016/j.tics.2013.12.009

40.

Joel

MacDonald

Plaks

J. E.

(2013). Romantic relationships conceptualized as a judgment and decision-making domain. Current Directions in Psychological Science, 22(6), 461–465. https://doi.org/10.1177/0963721413498892

41.

Jordan

Mullen

Murnighan

J. K.

(2011). Striving for the moral self: The effects of recalling past moral actions on future moral behavior. Personality and Social Psychology Bulletin, 37, 701–713. https://doi.org/10.1177/0146167211400208

42.

Kammrath

L. K.

Peetz

Hara

Demarco

Wood

Meirovich

Allen

(2015). It’s a matter of time: The effect of depletion on helpful behavior in romantic relationships is moderated by relationship length. Journal of Personality and Social Psychology, 109, 276–291. https://doi.org/10.1037/pspi0000023

43.

Kool

Botvinick

(2014). A labor/leisure tradeoff in cognitive control. Journal of Experimental Psychology: General, 143(1), 131–146.

44.

Kuper

Bott

(2019). Has the evidence for moral licensing been inflated by publication bias? Meta-Psychology, 3.https://doi.org/10.15626/mp.2018.878

45.

Lam

C. F.

Wan

W. H.

Roussin

C. J.

(2016). Going the extra mile and feeling energized: An enrichment perspective of organizational citizenship behaviors. Journal of Applied Psychology, 101, 379–391. https://doi.org/10.1037/apl0000071

46.

Lanaj

Johnson

R. E.

Wang

(2016). When lending a hand depletes the will: The daily costs and benefits of helping. Journal of Applied Psychology, 101, 1097–1110. https://doi.org/10.1037/apl0000118

47.

Lemay

E. P.

Jr. Clark

M. S.

(2008). How the head liberates the heart: Projection of communal responsiveness guides relationship promotion. Journal of Personality and Social Psychology, 94(4), 647–671. https://doi.org/10.1037/0022-3514.94.4.647

48.

Lin

S. H. J.

Johnson

R. E.

(2016). When ethical leader behavior breaks bad: How ethical leader behavior can turn abusive via ego depletion and moral licensing. Journal of Applied Psychology, 101(6), 815–830. https://doi.org/10.1037/apl0000098

49.

Lüdtke

Marsh

H. W.

Robitzsch

Trautwein

Asparouhov

Muthén

(2008). The multilevel latent covariate model: A new, more reliable approach to group-level effects in contextual studies. Psychological Methods, 13(3), 203–229. https://doi.org/10.1037/a0012869

50.

Maisel

N. C.

Gable

S. L.

(2009). The paradox of received social support: The importance of responsiveness. Psychological Science, 20, 928–932. https://doi.org/10.1111/j.1467-9280.2009.02388.x

51.

McNeish

Hamaker

E. L.

(2018, November 28). A primer on two-level dynamic structural equation models for intensive longitudinal data. https://doi.org/10.31234/osf.io/j56bm

52.

Milardo

R. M.

(1988). Families and social networks. SAGE.

53.

Miller

D. T.

Effron

D. A.

(2010). Psychological license: When it is needed and how it functions. In Zanna

M. P.

Olson

J. M.

(Eds.), Advances in experimental social psychology (Vol. 43, pp. 115–155). Academic Press. https://doi.org/10.1016/S0065-2601(10)43003-8

54.

Molenaar

P. C.

Campbell

C. G.

(2009). The new person-specific paradigm in psychology. Current Directions in Psychological Science, 18(2), 112–117. https://doi.org/10.1111/j.1467-8721.2009.01619.x

55.

Muthén

L. K.

Muthén

B. O.

(2017). 1998–2017. Mplus user’s guide. Muthén & Muthén.

56.

Nickell

(1981). Biases in dynamic models with fixed effects. Econometrica: Journal of the Econometric Society, 49(6), 1417–1426.

57.

Penner

L. A.

Fritzsche

B. A.

Craiger

J. P.

Freifeld

T. R.

(1995). Measuring the prosocial personality. Advances in Personality Assessment, 10, 147–163.

58.

Rajagopal

Rha

(2009). The mental accounting of time. Journal of Economic Psychology, 30, 772–781. https://doi.org/10.1016/j.joep.2009.06.008

59.

Reed

II.

A Aquino

K. F

. (2003). Moral identity and the expanding circle of moral regard toward out-groups. Journal of Personality and Social Psychology, 84, 1270–1286. https://doi.org/10.1037/0022-3514.84.6.1270

60.

Righetti

Finkenauer

Finkel

E. J.

(2013). Low self-control promotes the willingness to sacrifice in close relationships. Psychological Science, 24(8), 1533–1540. https://doi.org/10.1177/0956797613475457

61.

Rusbult

C. E.

Verette

Whitney

G. A.

Slovik

L. F.

Lipkus

(1991). Accommodation processes in close relationships: Theory and preliminary empirical evidence. Journal of Personality and Social Psychology, 60(1), 53–78. https://doi.org/10.1037/0022-3514.60.1.53

62.

Sachdeva

Iliev

Medin

D. L.

(2009). Sinning saints and saintly sinners: The paradox of moral self-regulation. Psychological Science, 20(4), 523–528. https://doi.org/10.1111/j.1467-9280.2009.02326.x

63.

Simbrunner

Schlegelmilch

B. B.

(2017). Moral licensing: A culture-moderated meta-analysis. Management Review Quarterly, 67, 201–225.

64.

Soman

(2001). The mental accounting of sunk time costs: Why time is not like money. Journal of Behavioral Decision Making, 14(3), 169–18.

65.

Thaler

R. H.

(1999/2008). Mental accounting matters. Journal of Behavioral Decision Making, 12, 183–206. https://doi.org/10.1002/(SICI)1099-0771(199909)12:3<183::AID-BDM318>3.0.CO;2-F

66.

Thaler

R. H.

Shefrin

H. M.

(1981). An economic theory of self-control. Journal of Political Economy, 89, 392–406. https://doi.org/10.1086/260971

67.

Thibaut

J. W.

Kelley

H. H.

(1959). The social psychology of groups. John Wiley.

68.

Thielmann

Spadaro

Balliet

(2020). Personality and prosocial behavior: A theoretical framework and meta-analysis. Psychological Bulletin, 146(1), 30–90. https://doi.org/10.1037/bul0000217

69.

Trougakos

J. P.

Beal

D. J.

Cheng

B. H.

Hideg

Zweig

(2015). Too drained to help: A resource depletion perspective on daily interpersonal citizenship behaviors. Journal of Applied Psychology, 100(1), 227–236. https://doi.org/10.1037/a0038082

70.

Tversky

Kahneman

(1981). The framing of decisions and the psychology of choice. Science, 211, 453–458. https://doi.org/10.1126/science.7455683

71.

Vohs

K. D.

Ciarocco

N. J.

(2004). Interpersonal functioning requires self-regulation. Guilford Press.

72.

Wilhelm

M. O.

Bekkers

(2010). Helping behavior, dispositional empathic concern, and the principle of care. Social Psychology Quarterly, 73(1), 11–32. https://doi.org/10.1177/0190272510361435

73.

Xue

Silk

J. B.

(2012). The role of tracking and tolerance in relationship among friends. Evolution and Human Behavior, 33(1), 17–25.

74.

Zhaoyang

Sliwinski

M. J.

Martire

L. M.

Smyth

J. M.

(2018). Age differences in adults’ daily social interactions: An ecological momentary assessment study. Psychology and Aging, 33(4), 607–618. https://doi.org/10.1037/pag0000242

75.

Zhong

Liljenquist

Cain

D. M.

(2009). Moral self-regulation: Licensing and compensation. In De Cremer

(Ed.), Psychological perspectives on ethical behavior and decision making (pp. 75–89). Information Age Publishing.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.03 MB