Aknin, Dunn, Proulx, Lok, Norton- 2020

Journal of Personality and Social Psychology Does Spending Money on Others Promote Happiness?: A Registered Replication Report Lara B. Aknin, Elizabeth W. Dunn, Jason Proulx, Iris Lok, and Michael I. Norton Online First Publication, April 6, 2020. http://dx.doi.org/10.1037/pspa0000191 CITATION Aknin, L. B., Dunn, E. W., Proulx, J., Lok, I., & Norton, M. I. (2020, April 6). Does Spending Money on Others Promote Happiness?: A Registered Replication Report. Journal of Personality and Social Psychology . Advance online publication. http://dx.doi.org/10.1037/pspa0000191

Does Spending Money on Others Promote Happiness?: A Registered Replication Report Lara B. Aknin Simon Fraser University Elizabeth W. Dunn University of British Columbia Jason Proulx Simon Fraser University Iris Lok University of British Columbia Michael I. Norton Harvard Business School Research indicates that spending money on others— prosocial spending —leads to greater happiness than spending money on oneself (e.g., Dunn, Aknin, & Norton, 2008, 2014). These findings have received widespread attention because they offer insight into why people engage in costly prosocial behavior, and what constitutes happier spending more broadly. However, most studies on prosocial spending (like most research on the emotional benefits of generosity) utilized small sample sizes ( n 100/cell). In light of new, improved standards for evidentiary value, we conducted high-powered registered replications of the central paradigms used in prosocial spending research. In Experiment 1, 712 students were randomly assigned to make a purchase for themselves or a stranger in need and then reported their happiness. As predicted, participants assigned to engage in prosocial (vs. personal) spending reported greater momen- tary happiness. In Experiment 2, 1950 adults recalled a time they spent money on themselves or someone else and then reported their current happiness; contrary to predictions, participants in the prosocial spending condition did not report greater happiness than those in the personal spending condition. Because low levels of task engagement may have produced these null results, we conducted a replication with minor changes designed to increase engagement; in this Experiment 3 ( N 5,199), participants who recalled a prosocial (vs. personal) spending memory reported greater happiness but differences were small. Taken together, these studies support the hypothesis that spending money on others does promote happiness, but demonstrate that the magnitude of the effect depends on several methodological features. Keywords: prosocial spending, generosity, happiness, well-being, replication People face countless spending choices in everyday life. Do some purchases offer greater happiness returns than others? In “Spending Money on Others Promotes Happiness” (Dunn et al., 2008), we presented a series of studies demonstrating that spending money on others (i.e., prosocial spending ) was associ- ated with self-reported happiness. The strongest evidence for cau- sality came from Study 3, in which 46 undergraduate students at the University of British Columbia were randomly assigned to spend a small windfall of either $5 or $20 on themselves or someone else by the end of the day. In the evening, all participants were called on the phone and asked to report their happiness. Participants randomly assigned to spend money on others— whether $5 or $20—were happier. Subsequent research has offered converging evidence for the emotional rewards of prosocial (vs. personal) spending. Aknin and colleagues (2013) examined the hedonic consequences of acts of prosocial spending in rich and poor nations, assessing whether the link between prosocial spending and happiness was limited to relatively wealthy nations. In one study, a sample of 207 students from Canada ( n 86) and South Africa ( n 121) earned a small monetary sum that they could use to purchase edible treats. They were randomly assigned either to a personal spending condition in which they purchased the treats for themselves, or a prosocial spending condition in which they purchased treats for a sick child at a local children’s hospital. Afterward, all participants reported their happiness. In both Canada and South Africa, students who engaged in prosocial spending were happier. In another study, people from Canada ( n 140) and Uganda ( n 680) who were X Lara B. Aknin, Department of Psychology, Simon Fraser University; Elizabeth W. Dunn, Department of Psychology, University of British Columbia; Jason Proulx, Department of Psychology, Simon Fraser Uni- versity; Iris Lok, Department of Psychology, University of British Colum- bia; Michael I. Norton, Marketing Unit, Harvard Business School. We thank Armaghan Aliabadi, Jayna Bhindi, Angie Fan, Alyssa Greco, George Guo, Jessamyn Hung, Burdett Kwon, Sejin Lee, Connor Mac- Millan, Cassandra Natura, John Nweke, Jason Roh, Angela Starnaman, and Oksana Soychuke for their assistance. Correspondence concerning this article should be addressed to Lara B. Aknin, Department of Psychology, Simon Fraser University, 8888 Univer- sity Drive, Burnaby, BC V5A 1S6, Canada. E-mail: lara_aknin@sfu.ca This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. Journal of Personality and Social Psychology: Attitudes and Social Cognition © 2020 American Psychological Association 2020, Vol. 2, No. 999, 000 ISSN: 0022-3514 http://dx.doi.org/10.1037/pspa0000191 1

randomly assigned to think about a time they spent $20 (or its equivalent in Ugandan shillings) of their own money on someone else reported greater happiness than those randomly assigned to think about a time they spent $20 on themselves. Similar results were observed in a separate study conducted with 101 participants from India in which respondents recalled a time they spent money on either themselves, spent money on others, or did not recall a spending memory. Research examining the well-being consequences of prosocial spending has received a large amount of attention in the academic literature and beyond. The two papers mentioned above have been cited nearly 2,000 times collectively (1,475 for Dunn et al., 2008; 516 for Aknin et al., 2013 according to Google Scholar on De- cember 11, 2019). These findings are described in widely used introductory textbooks (e.g., Aronson, Wilson, Fehr, & Akert, 2013; Myers, 2010; Smith, Mackie, & Claypool, 2014) and have been discussed in media outlets worldwide (e.g., The Guardian , The Wall Street Journal , and The New York Times ). These findings have also been featured in popular science books (e.g., “Give and Take,” Grant, 2014; “Drive,” Pink, 2011) and utilized in providing behavioral insights for policymakers (e.g., the U.K. Cabinet’s Charitable Giving Assessment ). Given the wide dissemination of research on the happiness benefits of prosocial spending, replicating the foundational studies is likely to be of interest to a wide range of social scientists. In the years since these studies were conducted, our field has undergone substantial change, and there is a growing recognition of the importance of using larger samples and preregistration (Button et al., 2013; Ioannidis, 2005; Nosek, Ebersole, DeHaven, & Mellor, 2018). According to a recent meta-analysis of existing studies, generosity (broadly defined) has a small to medium causal effect on happiness (Curry et al., 2018) and, thus, sample sizes of at least 200 participants per condition are needed to detect a main effect of prosocial behavior. Unfortunately, very few studies examining the hedonic consequences of prosocial spending meet this threshold (including our own). In fact, most experiments on this topic have used cell sizes of 100 or less (see Table 1) and, thus, should be interpreted with caution, given that underpowered studies carry an increased risk of false positives (Button et al., 2013; Fraley & Vazire, 2014). Very recently, several relevant studies have utilized larger sam- ple sizes, although these studies were designed primarily to test novel hypotheses about prosocial spending rather than to replicate earlier findings. In a preregistered study ( n 150 per condition), O’Brien and Kassirer (2019, Study 2) examined whether the emo- tional benefits of giving are resistant to hedonic adaptation. Am- azon Mechanical Turk (MTurk) workers earned a series of five- cent bonus payments for solving 10 puzzles, and depending on condition, all the payments went to participants themselves or to a charity of their choice. The pleasure of winning money declined at a lower rate when participants got money for charity versus for themselves; in fact, participants in the charity condition showed no evidence of hedonic adaptation at all. Another large study ( n 210 per condition) investigated whether individual differences in oxytocin receptor genes predict the emotional rewards derived from prosocial spending (Whillans, Aknin, Ross, Chen, & Chen, 2019). Students were randomly assigned to purchase treats for either themselves or a sick child at a local hospital before reporting their happiness and providing a saliva sample to collect genetic information. Although this research revealed no effects of oxytocin receptor genes, individuals randomly assigned to buy treats for a sick child reported greater happiness than those who bought for themselves. This study was not preregistered, however, and it is appropriate to treat this replication with caution. Finally, Hanniball and colleagues (2019) conducted three large experiments with cell sizes ranging from 250 to 588 among ex- offender samples reporting elevated levels of antisocial tendencies. Participants were asked to either recall or engage in an act of personal or prosocial spending before reporting their momentary well-being. In each study, participants assigned to the prosocial spending condition reported higher levels of happiness when con- trolling for baseline well-being, but the effect sizes were very small ( d s .11 to .16). In light of these findings, a plausible conclusion (drawn by some reviewers of that paper) would be that the happiness benefits of prosocial spending border onto trivial, and that past research greatly overestimated the size of this effect. However, another plausible interpretation is that people with an- tisocial tendencies should be least likely to exhibit the warm glow of giving, and observing even a small effect with this special population underscores the robustness of the hedonic benefits of giving. To resolve this debate and enable an accurate estimate of effect size, we replicated the key paradigms in this research stream using well-powered, preregistered studies with participants drawn from the broader population. The first experiment investigated the im- mediate emotional rewards of prosocial (vs. personal) spending by providing participants the opportunity to spend a small monetary windfall on edible treats for themselves or an unknown sick child at a local children’s hospital (see Aknin et al., 2013, Study 3). Afterward, participants reported their current well-being. The sec- ond experiment used a recollection procedure (as in Aknin et al., 2013, Studies 2a and b) in which participants were randomly assigned to reflect upon a previous purchase made for themselves or others before reporting their well-being. In both experiments, we predicted that participants randomly assigned to engage in or to reflect upon prosocial (vs. personal) spending would report greater momentary happiness when controlling for baseline well-being. Experiment 1 provided robust evidence for the immediate emo- tional rewards of generous spending, but Experiment 2 revealed a null result, possibly because participants did not fully engage with the task. Therefore, in a third experiment, we conducted an addi- tional test of the recollection paradigm in which participants were required to provide longer spending descriptions to encourage vivid recollections and greater engagement. Materials, data, hy- potheses, and syntax for all studies are posted on the Open Science Framework (OSF; https://osf.io/d6ymu/). Experiment 1 Method Sample. We preregistered recruiting a maximum sample of 896 individuals across two Canadian universities in exchange for course credit. According to G Power, this sample size allowed us to detect an effect size of d .22 with 95% power using .05, one-tailed. The effect size estimate of d .22 was based on the average effect observed in past prosocial spending studies with cell sizes of n 100 ( d .22; see Table 1). However, given the This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 2 AKNIN, DUNN, PROULX, LOK, AND NORTON

Table 1 Observed Effect Sizes in Prosocial Spending Experiments and Other Prosocial Intervention Experiments on Happiness Source Study Dependent variable Experimental (E) Control (C) n E n C Sample (location) Cohen’s d [95% CI] Baseline happiness controlled Aknin, Barrington-Leigh, et al. (2013) 2a b H Recall prosocial spending Recall personal spending 410 410 Student (Canada; Uganda) 0.20 [0.06, 0.34] No 2b H Recall prosocial spending Recall personal spending 22 18 Community (India) 0.55 [ 0.10, 1.19] No 3 a PA Prosocial spending Personal spending 104 103 Student (Canada; South Africa) 0.46 [0.18, 0.74] Yes 3 a SWLS Prosocial spending Personal spending 104 103 Student (Canada; South Africa) 0.13 [ 0.14, 0.40] No Aknin, Broesch, Hamlin, and Van de Vondervoort (2015) 1 PA Prosocial spending Personal spending 13 13 Community (Vanuatu) 0.87 [0.01, 1.70] Yes 2 Smiling Donate sweets (own) Donate sweets (other) 20 20 Children (Vanuatu) 0.30 [ 0.33, 0.92] No Aknin, Dunn, et al. (2013) WB Prosocial spending Personal spending 25 25 Community (Canada) 0.24 [ 0.32, 0.80] No Aknin, Fleerackers, and Hamlin (2014) PA Prosocial spending Personal spending 60 59 Student (Canada) 0.38 [0.01, 0.74] Yes ORH Prosocial spending Personal spending 60 59 Research Assistant (Canada) 0.44 [0.07, 0.81] Yes Aknin, Dunn, and Norton (2012) H Recall prosocial spending Recall personal spending 26 25 Students (Canada) 0.61 [0.03, 1.18] No Aknin, Hamlin, and Dunn (2012) Smiling Donate sweets (own) Donate sweets (other) 20 20 Children (Canada) 0.46 [ 0.18, 1.09] No Alden and Trew (2013) PA Prosocial behavior Safety behavior experiment 43 40 Socially anxious student (Canada) 0.59 [0.14, 1.04] No PA Prosocial behavior Life details tracking 43 43 Socially anxious student (Canada) 0.54 [0.10, 0.98] No Anik, Aknin, Norton, Dunn, and Quoidbach (2013) 1 PA Prosocial spending ($25) None 41 48 Community (Australia) 0.15 [ 0.57, 0.27] Yes 1 PA Prosocial spending ($50) None 41 48 Community (Australia) 0.49 [0.06, 0.92] No Buchanan and Bardi (2010) 1 SWLS Prosocial behavior New behavior 28 28 Community (U.K.) 0.41 [ 0.13, 0.94] No 1 SWLS Prosocial behavior No behavior 28 28 Community (U.K.) 0.62 [0.07, 1.16] No Chancellor, Margolis, Jacobs Bao, and Lyubomirsky (2018) SHS Prosocial behavior Prosocial receiver 16 34 Community (Spain) c — SWLS Prosocial behavior Prosocial receiver 16 34 Community (Spain) c — SHS Prosocial behavior None 16 33 Community (Spain) c — SWLS Prosocial behavior None 16 33 Community (Spain) c — Donnelly, Lamberton, Reczek, and Norton (2017) 1 H Social recycling Trash/recycling 59 56 Student (U.S.) 0.77 [0.38, 1.16] No 1 H Social recycling Take item 59 59 Student (U.S.) 0.85 [0.45, 1.24] No 2b PA Social recycling Trash 107 108 Community (U.S.) 1.25 [0.93, 1.56] No Dunn, Aknin, and Norton (2008) 3 H Prosocial spending Personal spending 23 23 Student (Canada) 0.67 [0.05, 1.27] Yes Geenen et al. (2014) H Prosocial spending Personal spending 34 34 Student (Germany) 0.70 [0.19, 1.20] Yes Hanniball and Aknin (2016) PA Prosocial behavior Self-helping behavior 51 56 Students (Canada) 0.46 [ 0.84, 0.07] No Hanniball, Aknin, Douglas, and Viljoen (2019) 1 b PA Recall prosocial spending Recall personal spending 250 251 Ex-offender adults (U.S.) 0.20 [0.02, 0.38] Yes 2 PA Prosocial spending Personal spending 31 33 Delinquent youth (Canada) 0.70 [0.17, 1.22] Yes 3 a PA Prosocial spending Personal spending 357 420 Ex-offender adults (U.S.) 0.16 [0.02, 0.30] Yes 4 a PA Prosocial spending Personal spending 588 707 Ex-offender adults (U.S.) 0.11 [0.00, 0.22] Yes Layous, Kurtz, Margolis, Chancellor, and Lyubomirsky (2017) 1 SHS Prosocial behavior Track daily activity 70 69 Student (U.S.) 0.08 [ 0.25, 0.41] No 1 WB Prosocial behavior Track daily activity 70 69 Student (U.S.) 0.20 [ 0.13, 0.53] No 1 EWB Prosocial behavior Track daily activity 70 69 Student (U.S.) 0.26 [ 0.08, 0.59] No 2 SHS Prosocial behavior Make self happier 178 81 Student (U.S.) 0.30 [0.04, 0.56] No 2 WB Prosocial behavior Make self happier 178 81 Student (U.S.) 0.12 [ 0.14, 0.38] No ( table continues ) This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 3 SPENDING MONEY ON OTHERS PROMOTES HAPPINESS