Content uploaded by Lara Aknin
Author content
All content in this area was uploaded by Lara Aknin on Apr 12, 2020
Content may be subject to copyright.
1
Does Spending Money on Others Promote Happiness?: A Registered Replication Report
Lara B. Aknin1, Elizabeth W. Dunn2, Jason Proulx1, Iris Lok2 & Michael I. Norton3
1 Simon Fraser University, 2 University of British Columbia, 3 Harvard Business School
Corresponding author:
Lara Aknin
Department of Psychology
Simon Fraser University
8888 University Drive
Burnaby, B.C., Canada, V5A 1S6
Email: lara_aknin@sfu.ca
©American Psychological Association, 2020. This paper is not the copy of record and may
not exactly replicate the authoritative document published in the APA journal. Please do
not copy or cite without author's permission. The final article is available, upon
publication, at: https://doi.org/10.1037/pspa0000191
Acknowledgements
The authors would like to thank Armaghan Aliabadi, Jayna Bhindi, Angie Fan, Alyssa Greco,
George Guo, Jessamyn Hung, Burdett Kwon, Sejin Lee, Connor MacMillan, Cassandra Natura,
John Nweke, Jason Roh, Angela Starnaman, and Oksana Soychuke for their assistance.
2
Abstract
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 momentary
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
3
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, Aknin & Norton, 2008), we
presented a series of studies demonstrating that spending money on others (i.e., prosocial
spending) was associated with self-reported happiness. The strongest evidence for causality 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 five or twenty dollars 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 five or twenty dollars – 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 randomly assigned to think about a time they spent 20 dollars (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 dollars on
4
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 (1475 for Dunn et al., 2008; 516 for Aknin
et al., 2013 according to Google Scholar on December 11, 2019). These findings are described in
widely-used introductory textbooks (e.g., Aronson, Wilson, Akert & Fehr, 2013; Myers, 2010;
Smith, Mackie & Claypool, 2014) and have been discussed in media outlets worldwide (e.g., The
Guardian, The Wall Street Journal, 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 UK 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 pre-
registration (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
5
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 sample sizes, although these
studies were designed primarily to test novel hypotheses about prosocial spending rather than to
replicate earlier findings. In a pre-registered study (n=150 per condition), O’Brien and Kassirer
(2019, Study 2) examined whether the emotional benefits of giving are resistant to hedonic
adaptation. Mturk workers earned a series of five-cent bonus payments for solving ten 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 pre-registered, however, and thus 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
6
prosocial spending condition reported higher levels of happiness when controlling for baseline
well-being, but the effect sizes were very small (ds = .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 antisocial
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, pre-registered studies with
participants drawn from the broader population. The first experiment investigated the immediate
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 second experiment used a recollection procedure (as in Aknin et al.,
2013, Studies 2a-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 emotional 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 additional test of
the recollection paradigm in which participants were required to provide longer spending
descriptions to encourage vivid recollections and greater engagement. Materials, data,
7
hypotheses, and syntax for all studies are posted on the Open Science Framework (OSF;
https://osf.io/d6ymu/).
Experiment 1
Methods
Sample. We pre-registered 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 alpha = .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 practical
limitations surrounding subject pool size, cost, and time, we pre-registered our intention to
perform sequential analyses (Lakens, 2014). This strategy allowed us to examine our primary
pre-registered hypothesis using an alpha of 0.0387 after collecting data from 694 participants.
The pre-registration for Experiment 1 can be found on the OSF (https://osf.io/gz7a6/).
We ended up recruiting a sample of 730 individuals (Mage = 19.91, SD = 2.71; 73.6%
female, 25.6% male, 0.4% other, 0.4% missing). This sample is slightly higher than our interim
target of 694 because we replaced participants who were excluded (based on our pre-registered
criteria) with random assignment. As in the original study (Aknin et al., 2013, Study 3),
participants were run in small group sessions to facilitate timely data collection. Importantly, all
responses were provided in private behind desk dividers to minimize self-presentation concerns.
This study was approved by our institutional review boards.
Procedure. Baseline emotion was assessed using the same items as Study 3 in Aknin et
al (2013). Participants reported their baseline level of happiness on a state (“Do you feel happy
8
right now?”; from 1, not at all, to 5, extremely) and trait (“In general, I consider myself . . . ”;
from 1, not a very happy person, to 7, a very happy person) measure (Lyubomirsky & Lepper,
1999). As expected and specified in our pre-registered analysis plan, these scores were
correlated, r(710) = .423, p < .001, so we standardized and averaged them to create a baseline
measure of happiness. Baseline happiness items were presented among a few filler questions
(e.g., “How tired are you feeling right now?”) to disguise our interest in happiness.
Goody-Bag Paradigm. After completing the baseline measures, participants were
provided with a questionnaire informing them that they had earned $2.50 in addition to course
credit for their participation. Funds were presented in the form of a paper voucher and
participants were asked to sign a receipt to encourage feelings of ownership over the funds. The
questionnaire then invited participants to use their voucher to purchase a goody-bag filled with
either chocolate, juice, or both, valued at $3.00. Critically, participants were randomly assigned
to one of two spending conditions. In the personal spending condition, participants were told that
the goody-bag they purchased was for them and available for pickup at the conclusion of the
experiment. In the prosocial spending condition, participants were told that the goody-bag they
purchased would be donated to a sick child at a local children’s hospital.
Participants made their spending choice (two chocolates, two juice boxes, or one
chocolate and one juice box) in private by selecting their preferred option on a purchase card.
Once selected, participants took the purchase card to a research assistant in a private room. Here,
the research assistant packaged the requested items so the participant could see their purchase
was real. The packaged goody bag containing the purchased items was then marked with the
appropriate participant number and set-aside until the completion of the study. The research
assistant then handed each participant a pre-prepared thank you note reinforcing their condition
9
assignment. Specifically, participants in the personal condition received a note saying, “Thanks
for your purchase! Your items will be available for pickup at the end of the study!” Meanwhile,
participants in the prosocial condition received a note saying, “Thanks for your purchase! Your
items will be donated to a sick child at Children’s Hospital at the end of the study!” Critically,
the research assistant did not know what condition participants had been assigned to, and hence
the recipient of the goody bag, to ensure similar interactions with all participants. Thus, because
all study materials were identical, research assistants were blind to condition assignment during
the experiment. Research assistants only learned of a participant’s condition assignment at the
end of the experimental session so that they could return items to individuals in the personal
spending condition. Gifts purchased in the prosocial spending condition were donated to a local
charity for sick children and their families.
Opt-out. Past research has shown that a sense of volition is essential for experiencing the
emotional rewards of prosocial behavior (Weinstein & Ryan, 2010). Therefore, participants in
both conditions had the opportunity to opt-out of purchasing a goody-bag and take the cash value
($2.50) for themselves. This option ensured that participants in the prosocial spending condition
felt as though they had chosen to give a gift. To discourage participants from opting out, cash
collections were only available on one early morning at the end of the semester. As noted in the
pre-registration and consistent with past research, participants assigned to the prosocial spending
condition who chose to opt out of making a purchase (n = 18, < 2.5% of the sample) were
excluded from the primary analysis because they did not engage in a prosocial act. This meant
our final sample, after exclusions, included 712 individuals (Mage = 19.91, SD = 2.73; 74.2%
female, 25.1% male, 0.3% other, 0.4% missing). However, we also conducted additional
10
analyses to examine the consequences of prosocial spending among the full sample of
participants, including opt-outs (see below).
After the purchase, all participants were asked to report their current positive affect on
the Positive and Negative Affect Schedule (PANAS; Watson, Clark & Tellegen, 1988), which
included the additional word “happy” (consistent with our more recent work, e.g., Aknin, Dunn,
Sandstrom & Norton, 2013; Aknin, Dunn, Whillans, Grant & Norton, 2013; Aknin, Mayraz &
Helliwell, 2016; Whillans et al., 2019). As noted in the pre-registration, positive affect was
computed by taking the average of the 11 positive items (10 original positive affect items from
the PANAS and “happy”), which served as the primary dependent variable of interest. In
addition, participants were asked to report their positive and negative emotion on the Scale of
Positive and Negative Experience (SPANE; Diener et al., 2009), which assessed both general
and specific positive and negative states. Positive emotion on the SPANE served as a second and
exploratory dependent variable; positive emotion was computed by summing together responses
for all positive items. Finally, participants reported their demographic information (gender, age,
household income).
Hypotheses and Pre-registered Analyses
In line with past research, we predicted that participants who purchased a goody bag for a
sick child would report higher levels of positive affect than participants who purchased a goody
bag for themselves when controlling for baseline happiness. We tested this pre-registered
directional hypothesis with an ANCOVA in which spending condition (personal vs. prosocial
spending) was entered as the independent variable, average post-spending positive affect was
entered as the dependent variable, and baseline happiness was entered as a covariate. As
predicted, participants who purchased a goody bag for a sick child reported higher positive affect
11
(M = 2.972, SD = .78, n = 344) than participants who purchased a goody bag for themselves (M
= 2.732, SD = .69, n = 359), F(1,700) = 22.767, p < .001, d = .36, r = .18; see Table 2 for
summary of all experiment results. Results are similar when we included participants in the
prosocial spending condition who opted out of purchasing a goody bag; participants in the
prosocial spending condition reported higher positive affect (M = 2.945, SD = .79, n = 362) than
participants who purchased a goody bag for themselves (M = 2.732, SD = .69, n = 359), F(1,718)
= 19.627, p < .001, d = .33, r = .16.
We also conducted a secondary exploratory analysis to examine whether participants
randomly assigned to purchase a goody bag for others reported higher levels of positive emotion
on the SPANE than participants who purchase a goody bag for themselves when controlling for
baseline happiness. We tested this hypothesis with an ANCOVA in which spending condition
(personal vs. prosocial spending) was entered as the independent variable, average post-spending
positive emotion on the SPANE was entered as the dependent variable, and baseline happiness
was entered as a covariate. Analyses revealed that participants who purchased a goody bag for a
sick child reported higher positive emotion on the SPANE (M = 21.028, SD = 5.29, n = 341) than
participants who purchased a goody bag for themselves (M = 19.654, SD = 4.56, n = 358),
F(1,696) = 17.822, p < .001, d = .32, r = .16. Results are similar when we included participants
in the prosocial spending condition who opted out of purchasing a goody bag; participants in the
prosocial spending condition reported higher positive emotion (M = 20.784, SD = 5.40, n = 359)
than participants who purchased a goody bag for themselves (M = 19.654, SD = 4.56, n = 358),
F(1,714) = 13.430, p < .001, d = .27, r = .14.
Discussion
12
Experiment 1 provides clear evidence that participants felt happier after purchasing a
goody bag for a sick child than after purchasing a goody bag for themselves, consistent with the
hypothesis that spending money on others promotes happiness. We opted to use the goody bag
paradigm from Aknin et al (2013) because it offers numerous methodological advances over the
original paradigm used by Dunn and colleagues (2008) in which participants were assigned to
spend $5 or $20 on themselves or others. Specifically, the goody bag paradigm requires that all
participants purchase identical items (i.e. juice and/or treats). This consistency ensures that any
emotional differences observed across conditions are not a result of purchasing differential
content (e.g., experiences vs. material goods). In addition, because participants do no interact
with the recipient of their gift and their choice is made privately, the benefits of prosocial
spending cannot easily be explained by gratitude or praise.
Experiment 1 compared the immediate emotional consequences of personal and prosocial
spending. In Experiment 2, we investigated the long-term outcomes of personal and prosocial
spending by comparing how people felt when they reflected on a previous purchase made with
their own money.
Experiment 2
Methods
Sample. We recruited a final sample of 1950 participants (Mage = 47.72, SD = 15.19;
68.7% female, 31.0% male, 0.3% other, 0.1% prefer not to say) using Qualtrics’ online national
panel (see Table 3 for complete sample demographics of participants in Experiments 2 and 3).
This sample size was slightly higher than our target sample of 1926 (calculated using G*Power
with alpha =.05, one-tailed, power = .95, and an average effect size of d = .15) because we
oversampled slightly to exclude participants who did not recall a spending experience; this
13
exclusion criterion is explained in our pre-registration available on the OSF
(https://osf.io/x39cu/). As shown in Table 1, past research using the prosocial spending
recollection paradigm indicates an average effect size of d = .20, which requires a sample of
1084 with alpha =.05, one-tailed, and power = .95. However, given that published research may
have overestimated the true effect and the costs of online data collection are relatively low, we
were able to collect a larger sample enabling us to capture a true effect size of d = .15 with .95
power. This study was approved by our institutional review boards.
Procedure. Participants were asked to report their baseline well-being using the same
items and procedure described above. Specifically, participants reported their current level of
happiness on a single-item (“Do you feel happy right now?”; from 1, not at all, to 5, extremely)
and completed a measure of trait-level happiness (“In general, I consider myself . . . ”; from 1,
not a very happy person, to 7, a very happy person; Lyubomirsky & Lepper, 1999). As expected
and specified in our pre-registered analysis plan, these scores were highly correlated,
r(1948)=.600, p < .001, so we standardized and averaged them to create a baseline measure of
happiness. Baseline happiness items were presented among a few filler questions (e.g., “How
tired are you feeling right now?”) to disguise our interest in happiness.
After reporting their baseline happiness, participants were randomly assigned to recall
and describe a previous spending experience in which they used 20 dollars to benefit themselves
or someone else. Specifically, participants randomly assigned to the personal spending condition
saw the prompt:
Please think back to and describe as vividly and in as much detail as possible the last
time you spent approximately $20 on yourself. In the space provided below please
describe this event.
14
Participants randomly assigned to the prosocial spending condition saw the prompt:
Please think back to and describe as vividly and in as much detail as possible the last
time you spent approximately $20 on someone else. In the space provided below please
describe this event.
All participants then reported their current positive affect on the PANAS (Watson et al.,
1988) and the additional word “happy” (as in Experiment 1). Positive affect was computed by
taking the average of the 11 positive items (10 original positive affect items from the PANAS
and “happy”); this served as the primary dependent variable of interest. As in Experiment 1,
participants were also asked to report their positive and negative emotion on the SPANE (Diener
et al., 2009). Positive emotion on the SPANE served as a second and exploratory dependent
variable; positive emotion was computed by summing together responses for all positive items.
Finally, participants reported their demographic information (gender, age, household income).
Although we originally used the Subjective Happiness Scale (Lyubomirsky & Lepper,
1999) – a relatively trait like measure of happiness – our more recent work and current best
practice suggests that measuring state happiness is preferred. As such, we measured current
positive affect and positive emotion as our dependent variables.
Hypotheses and Pre-registered Analyses
In line with past research, we predicted that participants randomly assigned to reflect
upon a previous instance of prosocial spending would report higher current levels of positive
affect than participants assigned to reflect upon a previous instance of personal spending. As
outlined in our pre-registered analysis plan, we tested this directional hypothesis with an
ANCOVA in which spending condition (personal vs. prosocial spending) was entered as the
independent variable, average levels of current positive affect reported on the PANAS were
15
entered as the dependent variable, and baseline happiness was entered as a covariate. Contrary to
our predictions and past work, participants in the prosocial spending condition did not report
higher levels of current positive affect (M = 3.104, SD = .90, n = 963) than participants in the
personal spending condition (M = 3.083, SD = .91, n = 983), F(1,1943) =.457, p = .499, d = .03,
r = .02.
As in Experiment 1, we conducted a second exploratory analysis to examine whether
participants randomly assigned to recall spending money on others reported higher levels of
current positive emotion on the SPANE than participants assigned to recall spending money on
something for themselves. Consistent with the pre-registered analysis plan, we tested this
hypothesis with an ANCOVA in which spending recall condition (personal vs. prosocial
spending) was entered as the independent variable, current post-recall positive emotion on the
SPANE was entered as the dependent variable, and baseline happiness was entered as a
covariate. Analyses revealed that participants in the prosocial spending condition did not report
higher levels of positive emotion on the SPANE (M = 19.925, SD = 6.01, n = 964) than
participants in the personal spending condition (M = 19.986, SD = 5.93, n = 983), F(1,1944)
=.118, p = .732, d = .02, r = .01.
Discussion
Experiment 2 did not replicate the long-term emotional rewards of prosocial (vs.
personal) spending with the recollection paradigm. Specifically, participants assigned to recall a
time they spent approximately 20 dollars on someone else reported similar happiness levels in
the current moment to participants assigned to recall a time they spent approximately 20 dollars
on themselves. One reason for the null effect may be that many participants did not provide
detailed spending recollections despite the request to describe their purchase “as vividly and in as
16
much detail as possible.” Indeed, participants in Experiment 2 wrote an average of 22 words in
response to the recollection prompts, which is less than half the length of responses captured in a
recent study demonstrating the emotional benefits of prosocial spending upon reflection (an
average of 45 words in Hanniball et al., 2019), signalling a possible lack of engagement in the
task. Whereas participants in our past studies have offered vivid descriptions of their
experiences, many participants in the present study wrote just a brief phrase, such as “as a
birthday gift” (in the prosocial spending condition) or “buy coffee” (in the personal spending
condition).
In interpreting the results of Experiment 2, it is also worth noting that we measured how
participants felt in the present after recalling a past spending experience. Theoretically, recalling
a positive past experience should only make people feel happy in the present if they vividly
recall it, akin to mentally re-living the experience (Strack, Schwarz & Gschneidinger, 1985).
Indeed, studies on buying experiences (vs. material things) typically ask participants to recall
how they felt at the time of the purchase, rather than asking about their current feelings, which
bypasses the problem of having to get participants to vividly re-live the past event.
Therefore, in Experiment 3 we conducted an additional test of the recollection paradigm,
using the same design as Experiment 2 with three small but important methodological
improvements. First, to ensure that participants engaged with the open-ended recollection task,
we recruited only Mturk participants who had provided high-quality responses in past studies.
Second, to encourage participants to engage in vivid reflection on this task, we required them to
write at least 150 characters when describing their spending experience. Finally, as well as
measuring participants’ current feelings (as in Experiment 2), participants also recalled their
feelings at the time of the purchase; this measurement strategy was not employed in the original
17
work and therefore does not represent a replication. However, we included this measure because
it allowed us to examine the emotional benefits of prosocial spending even for participants who
did not vividly re-live the past experience. To examine the potential value of these
methodological improvements, we conducted an exploratory study incorporating these changes
(materials and data for this exploratory study are available on the OSF at https://osf.io/d6ymu/).
After obtaining promising results in this exploratory study, we conducted a third pre-registered
study in Experiment 3.
Experiment 3
Methods
Sample. We recruited a final sample of 5199 participants (Mage = 39.00, SD = 11.77;
57.0% female, 42.4% male, 0.4% other, 0.04% prefer not to say) online using Amazon’s
Mechanical Turk (Mturk) system. We used Mturk as our recruitment platform because it gave us
the ability to reach respondents with a high-quality record (e.g., 97% approval rating in >5000
HITs), increasing the likelihood that participants would engage with the detailed recollection
task. Our target sample was 5050 participants. This target was based on G*Power calculations
using alpha =.05, one-tailed, power = .80, and an effect size of d = .07 (as observed in the
identical exploratory study mentioned above). However, we oversampled to 5300 participants so
we could exclude those who did not recall a spending experience; this exclusion criterion is
explained in our pre-registration available on the OSF at https://osf.io/y9s5f/. This study was
approved by our institutional review boards.
Procedure. Participants reported their current happiness and trait happiness on the same
single-item scales used in Experiments 1-2. Consistent with our earlier studies and our pre-
registered analysis plan, these scores were highly correlated, r(5197)=.688, p < .001, so we
18
standardized and averaged them to create a baseline measure of happiness. Happiness items were
presented among a few filler questions (e.g., “How tired are you feeling right now?”) to disguise
our interest in happiness.
Participants were randomly assigned to recall and describe a previous spending
experience in which they used 20 dollars to benefit themselves or someone else. Recollection
prompts were identical to those used in Experiment 2, however, participants were required to
write at least 150 characters (equivalent to approximately 25 words) when describing their
spending experience to encourage vivid recollection. Most participants wrote more than the
minimum requirement; the average spending recollection contained 60.52 words (SD = 33.23).
After describing their spending memory, all participants reported their current well-being
on the PANAS and SPANE (as before), as well their positive emotion at the time of purchase
using the SPANE. Consistent with the previous studies and our pre-registration, positive affect
was computed by taking the average of the 11 positive items (10 original positive affect items
from the PANAS and “happy”) and positive emotion on the SPANE was computed by summing
together responses for all positive items. Finally, participants reported their demographic
information (gender, age, household income).
Hypotheses and Pre-registered Analyses
We predicted that participants randomly assigned to reflect upon a previous instance of
prosocial spending would report higher current levels of positive affect than participants assigned
to reflect upon a previous instance of personal spending. As outlined in our pre-registered
analysis plan, we tested this directional hypothesis with an ANCOVA in which spending
condition (personal vs. prosocial spending) was entered as the independent variable, average
levels of current positive affect reported on the PANAS were entered as the dependent variable,
19
and baseline happiness was entered as a covariate. Consistent with predictions and past work,
participants in the prosocial spending condition reported higher levels of current positive affect
(M = 3.038, SD = .90, n = 2584) than participants in the personal spending condition (M = 3.006,
SD = .91, n = 2610), F(1,5191) = 5.341, p = .011, d = .06, r = .03.
We also examined whether participants randomly assigned to recall spending money on
others reported higher levels of current positive emotion on the SPANE than participants
assigned to recall spending money on something for themselves. Following our pre-registered
analysis plan, we tested this directional hypothesis using an ANCOVA in which spending recall
condition (personal vs. prosocial spending) was entered as the independent variable, current post-
recall positive emotion on the SPANE was entered as the dependent variable, and baseline
happiness was entered as a covariate. As predicted, participants in the prosocial spending
condition reported higher levels of positive emotion on the SPANE (M = 19.638, SD = 6.20, n =
2584) than participants in the personal spending condition (M = 19.491, SD = 6.19, n = 2613),
F(1,5194) = 4.571, p = .017, d = .06, r = .03.
Finally, we examined whether participants randomly assigned to recall spending money
on others remember having felt happier after spending than participants assigned to recall
spending money on themselves. As stated in the pre-registration, we tested this directional
hypothesis using an ANCOVA in which spending recall condition (personal vs. prosocial
spending) was entered as the independent variable, ratings of positive emotion on the SPANE at
the time of purchase were entered as the dependent variable, and baseline happiness was entered
as a covariate. As predicted, participants recalled experiencing higher levels of positive emotion
(on the SPANE) after spending money on others (M = 22.696, SD = 5.69, n = 2584) than after
20
spending money on themselves (M = 21.901, SD = 5.85, n = 2613), F(1,5194) = 37.926, p <
.001, d = .17, r = .09.
Discussion
Using a modified version of our recollection paradigm that was designed to ensure
participants exhibited at least a moderate level of engagement with the task, we replicated the
emotional rewards of prosocial (vs. personal) spending. On both measures (the PANAS and the
SPANE), participants reported more positive feelings when they reflected on a time they spent
their own money on others versus themselves. However, this effect was very small (d =.06). We
also asked participants to recall how they had felt at the time of their purchase, which revealed a
larger positive effect of prosocial spending (d = .17). The greater difference between conditions
reported at the time of purchase may reflect the more immediate outcomes of personal and
prosocial spending choices, but may also stem from recollection biases or demand
characteristics. That is, asking participants to recall their past emotions may have led them to
rely on their naïve theories of how giving should make them feel, or their guesses of how the
experimenters expected them to feel.
General Discussion
This Registered Replication Report offers evidence that spending money on others
promotes happiness. Experiment 1 showed that university students randomly assigned to
purchase treats for a sick child reported feeling happier afterward than participants assigned to
purchase treats for themselves. In Experiment 2, however, American adults recruited through
Qualtrics did not report greater current happiness when they thought about spending their own
money on others (vs. themselves) in the past. We suspected that these null results might have
stemmed from a lack of participant engagement, and thus we made several modifications to
21
enhance engagement with the task in Experiment 3. Using this improved recollection paradigm,
we found that participants in the prosocial (vs. personal) spending condition reported slightly
higher happiness after recalling a time they spent money on others versus themselves. We
observed a similar, but larger difference between conditions when participants recalled their
feelings immediately after making the purchase. Given that these studies were pre-registered and
well-powered, they provide the most conclusive evidence to date for the emotional benefits of
prosocial spending.
The present studies also offer the best available estimates of effect sizes to guide future
experimental research on prosocial spending. In Experiment 1, we observed an effect size (d =
.36; r = .18) that was similar in magnitude to other well-established effects in social psychology
– such as the foot-in-the-door effect (average r = .16) and the effect of self-disclosure on liking
(average r = .16; Richard, Bond, & Stokes-Zoota, 2003) – as well as the typical effect reported in
the published social psychology literature more broadly (r = .21; Fraley & Marks, 2007). While
we observed null effects using a recollection paradigm in Experiment 2, the modified
recollection paradigm used in Experiment 3 yielded significant, but small effects on participants’
happiness in the moment after recalling a past spending experience (d = .06, r = .03), and a larger
effect on participants’ recall of their happiness immediately after making the purchase (d = .17, r
= .09). The latter effect is similar in magnitude to other well-known effects, such as the tendency
to attribute failure to external factors (average r = .09) and the tendency for people to be more
aggressive toward men than women (average r =.06; Richard et al., 2003).
As this discussion underscores, there is no one effect size of prosocial spending on
happiness. Instead, the magnitude of the effect critically depends on methodological choices,
including the experimental paradigm employed and the manner in which happiness is assessed.
22
More generally, the search for a single effect size to capture any phenomenon is challenging for
at least two reasons. First, the simple fact that moderators increase and decrease the magnitude of
an effect suggests that any phenomenon is inherently variable. In the context of prosocial
spending, for example, factors such as perceived prosocial impact, a sense of volition, and
connection with recipients critically influence the extent to which spending money on others
promotes happiness (Aknin, Dunn, Sandstrom, & Norton, 2013; Aknin, Dunn, Whillans, Grant
& Norton, 2013; Dunn, Aknin, & Norton, 2014; Lok & Dunn, 2019; Weinstein & Ryan, 2010).
Second, broader contextual factors are also likely to influence the observed effect size of a
psychological phenomenon. Our correlational research using the Gallup World Poll data
provides a clear example; while the effect of prosocial spending on life satisfaction is positive in
over 88% of countries surveyed, the correlation ranges from 1.33 (in Montenegro) to -.63 (in
Tunisia; Aknin et al., 2013).
Perhaps most importantly, our results suggest that at least a moderate level of participant
engagement may be necessary in order to detect the beneficial effects of prosocial spending.
Traditionally, psychologists from Milgram (1975) to Batson (e.g., Toi & Batson, 1982) to Latané
and Darley (1968) conducted elegant, high-impact experiments that immersed participants in the
phenomenon under study. More recently, in a well-justified effort to collect much larger samples
in less time, social psychologists have increasingly conducted studies online. Our findings point
to the conclusion that slow, costly lab studies still have an important place in social psychology;
using a highly-involved lab paradigm in Experiment 1, we were able to detect a robust effect of
prosocial spending on happiness, which required enormous samples to detect using a more pallid
online procedure. Of course, it is possible to design highly engaging experiments that can be
completed online (e.g., O’Brien & Kassirer, 2019, Exp. 2), and thus an important—and perhaps
23
under-recognized—challenge for psychologists lies in importing the strengths of traditional high-
impact labs studies into modern online data collection platforms.
It is also important to consider how these effects might shift over time. Taken together,
our studies suggest that people experience a clear happiness benefit immediately after spending
on others. When people look back on a past prosocial spending experience, they remember
feeling happy—and just thinking about this past experience produces a detectable, though very
small boost to happiness in the present.
1
Once again, it is worth noting, that original research on
prosocial spending did not ask participants to report how they felt immediately after their
purchase (we imported this paradigm from the experiential vs. material spending literature; e.g.,
Carter & Gilovich, 2010), so this finding should not be considered a direct replication of past
work.
That said, while the warm glow of any one generous purchase appears to fade over time,
daily life offers numerous opportunities to spend money on others. Funder and Ozer (2019)
argue that rather than apologizing for small effects, researchers should consider how these small
effects may accumulate over time. For example, someone who spends small amounts of money
helping others each day for a week may get a small happiness boost each time, which may
compound into a larger benefit for well-being. Although it is also possible that people may
habituate to repeated behaviors, recent research suggests that people adapt more slowly to
treating others than to treating themselves (O’Brien & Kassirer, 2019). Consistent with the
notion that small boosts in happiness from prosocial spending may aggregate into greater well-
1
It is worth noting that another recent experiment found that simply recalling acts of kindness
led to well-being benefits that were comparable to the effects of actually engaging in acts of
kindness (Ko, Margolis, Revord, & Lyubomirsky, 2019). This online study was conducted with
university students, who were required to spend at least 5 minutes on the recall task, further
underscoring the potential importance of promoting participants’ engagement.
24
being, recent analyses of correlational data from over a million people around the globe show
that prosocial spending—in the form of charitable giving—is one of the top six predictors of life
satisfaction around the world (Helliwell, Huang & Wang, 2019).
Finally, our hope is that this research – and our learning along the way about the
importance of key factors for consideration such as engagement – serves as a model for other
researchers interested in examining the robustness of their previously published findings. While
daunting, we found it immensely valuable to reflect upon our past research in selecting the
paradigms we felt were most important to replicate, while being transparent about our selection
criteria and our methodological choices. New norms of transparency and rigor are already
improving psychological science; our results suggest that revisiting old paradigms with higher
research standards are one such tool to ensure that our science rests on a solid foundation.
25
References
Aknin, L. B., Barrington-Leigh, C. P., Dunn, E. W., Helliwell, J. F., Burns, J., Biswas-Diener,
R., . . . Norton, M. I. (2013). Prosocial spending and well-being: Cross-cultural evidence
for a psychological universal. Journal of Personality and Social Psychology, 104(4), 635-
652. doi:10.1037/a0031578
Aknin, L. B., Broesch, T., Hamlin, J. K., & Van de Vondervoort, J. W. (2015). Prosocial
behavior leads to happiness in a small-scale rural society. Journal of Experimental
Psychology. General, 144(4), 788-795. doi:10.1037/xge0000082
Aknin, L. B., Dunn, E. W., & Norton, M. I. (2012). Happiness runs in a circular motion:
Evidence for a positive feedback loop between prosocial spending and happiness. Journal
of Happiness Studies, 13(2), 347-355. doi:10.1007/s10902-011-9267-5
Aknin, L. B., Dunn, E. W., Sandstrom, G. M., & Norton, M. I. (2013). Does social connection
turn good deeds into good feelings? On the value of putting the ‘social’ in prosocial
spending. International Journal of Happiness and Development, 1(2), 155–171.
http://dx.doi.org/10.1504/ijhd.2013.055643.
Aknin, L. B., Dunn, E. W., Whillans, A. V., Grant, A. M., & Norton, M. I. (2013). Making a
difference matters: Impact unlocks the emotional benefits of prosocial spending. Journal
of Economic Behavior and Organization, 88, 90-95. doi:10.1016/j.jebo.2013.01.008.
Aknin, L. B., Fleerackers, A. L., & Hamlin, J. K. (2014). Can third-party observers detect the
emotional rewards of generous spending? Journal of Positive Psychology, 9(3), 198-203.
doi:10.1080/17439760.2014.888578
Aknin, L. B., Hamlin, J. K., & Dunn, E. W. (2012). Giving leads to happiness in young children.
PLoS One, 7(6), e39211. doi:10.1371/journal.pone.0039211
26
Aknin, L. B., Mayraz, G., & Helliwell, J. F. (2017). The emotional consequences of donation
opportunities. Journal of Positive Psychology, 12(2), 169-177.
doi:10.1080/17439760.2016.1163409
Alden, L. E., & Trew, J. L. (2013). If it makes you happy: Engaging in kind acts increases
positive affect in socially anxious individuals. Emotion, 13(1), 64-75.
doi:10.1037/a0027761
Anik, L., Aknin, L. B., Norton, M. I., Dunn, E. W., & Quoidbach, J. (2013). Prosocial bonuses
increase employee satisfaction and team performance. PLoS One, 8(9), e75509.
doi:10.1371/journal.pone.0075509
Aronson, E., Wilson, T. D., Fehr, B., Akert, R. M. (2013). Social Psychology (5th ed.). Upper
Saddle River, NJ: Pearson Education, Inc.
Buchanan, K. E., & Bardi, A. (2010). Acts of kindness and acts of novelty affect life
satisfaction. The Journal of Social Psychology, 150(3), 235-237.
doi:10.1080/00224540903365554
Button, K. S., Ioannidis, J. P. A., Mokrysz, C., Nosek, B. A., Flint, J., Robinson, E. S. J., &
Munafò, M. R. (2013). Power failure: Why small sample size undermines the reliability
of neuroscience. Nature Reviews. Neuroscience, 14(5), 365-376. doi:10.1038/nrn3475
Carter, T. J. & Gilovich, T. (2010). The relative relativity of material and experiential purchases.
Journal of Personality and Social Psychology, 98(1), 146-159. doi:10.1037/a0017145
Chancellor, J., Margolis, S., Jacobs Bao, K., & Lyubomirsky, S. (2018). Everyday prosociality in
the workplace: The reinforcing benefits of giving, getting, and glimpsing. Emotion, 18(4),
507-517. doi:10.1037/emo0000321
27
Curry, O. S., Rowland, L. A., Van Lissa, C. J., Zlotowitz, S., McAlaney, J., & Whitehouse, H.
(2018). Happy to help? A systematic review and meta-analysis of the effects of
performing acts of kindness on the well-being of the actor. Journal of Experimental
Social Psychology, 76, 320-329. doi:10.1016/j.jesp.2018.02.014
Diener, E., Wirtz, D., Tov, W., Kim-Prieto, C., Choi, D., Oishi, S., & Biswas-Diener, R. (2009).
New measures of well-being: Flourishing and positive and negative feelings. Social
Indicators Research, 39, 247-266. doi:10.1007/978-90-481-3354-4_12
Donnelly, G. E., Lamberton, C., Reczek, R. W., & Norton, M. I. (2017). Social recycling
transforms unwanted goods into happiness. Journal of the Association for Consumer
Research, 2(1), 48-63. doi:10.1086/689866
Dunn, E. W., Aknin, L. B., & Norton, M. I. (2008). Spending money on others promotes
happiness. Science, 319(5870), 1687-1688. doi:10.1126/science.1150952
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, 13(2),
347-355.
Fraley, R. C., & Marks, M. J. (2007). The null hypothesis significance-testing debate and its
implications for personality research. In R. W. Robins, R. C. Fraley, & R. F. Krueger
(Eds.), Handbook of research methods in personality psychology (p. 149–169). The
Guilford Press.
Fraley, R. C., & Vazire, S. (2014). The N-pact factor: Evaluating the quality of empirical
journals with respect to sample size and statistical power. Plos One, 9(10), e109019.
doi:10.1371/journal.pone.0109019
28
Funder, D. C., & Ozer, D. J. (2019). Evaluating effect size in psychological research: Sense and
nonsense. Advances in Methods and Practices in Psychological Science, 2(2), 156-168.
doi:10.1177/2515245919847202
Grant, A. M. (2014). Give and take: Why helping others drives our success. New York, NY:
Penguin.
Geenen, N. Y. R., Hohelüchter, M., Langholf, V., & Walther, E. (2014). The beneficial effects of
prosocial spending on happiness: Work hard, make money, and spend it on others? The
Journal of Positive Psychology, 9(3), 204-208. doi:10.1080/17439760.2014.891154
Hanniball, K. B., & Aknin, L. B. (2016). Does helping others foster happiness in newcomer
populations? (unpublished data).
Hanniball, K. B., Aknin, L. B., Douglas, K. S., & Viljoen, J. L. (2019). Does helping promote
well-being in at-risk youth and ex-offender samples? Journal of Experimental Social
Psychology, 82, 307-317. doi:10.1016/j.jesp.2018.11.001
Helliwell, J. F., Huang, H. & Wang, S. (2019). Changing world happiness. In J. F. Helliwell, R.
Layard & J. Sachs (Eds.), World Happiness Report 2019 (pp. 10-45). New York:
Sustainable Development Solutions Network
Ioannidis, J. P. A. (2005). Why most published research findings are false. PLoS Medicine, 2(8),
e124. doi:10.1371/journal.pmed.0020124
Ko, K., Margolis, S., Revord, J., & Lyubomirsky, S. (2019). Comparing the effects of
performing and recalling acts of kindness. Journal of Positive Psychology, 1-9.
doi:10.1080/17439760.2019.1663252
Lakens, D. (2014). Performing high-powered studies efficiently with sequential analyses:
Sequential analyses. European Journal of Social Psychology, 44(7), 701–710.
29
https://doi.org/10.1002/ejsp.2023
Latane, B., & Darley, J. M. (1968). Group inhibition of bystander intervention in
emergencies. Journal of Personality and Social Psychology, 10(3), 215-221.
Layous, K., Kurtz, J., Margolis, S., Chancellor, J., & Lyubomirsky, S. (2017). Make someone
happy and you will be happy too: An other-oriented path to well-being. Manuscript
submitted for publication.
Layous, K., Lee, H., Choi, I., & Lyubomirsky, S. (2013). Culture matters when designing a
successful happiness-increasing activity: A comparison of the united states and south
korea. Journal of Cross-Cultural Psychology, 44(8), 1294-1303.
doi:10.1177/0022022113487591
Layous, K., Nelson, S. K., Oberle, E., Schonert-Reichl, K., & Lyubomirsky, S. (2012). Kindness
counts: Prompting prosocial behavior in preadolescents boosts peer acceptance and well-
being. PLoS One, 7(12), 1-3. doi:10.1371.0051380
Lok, I., & Dunn, E. W. (2019, April 12). Under what conditions does prosocial spending
promote happiness? https://doi.org/10.31234/osf.io/9bpg5
Lyubomirsky, S., & Lepper, H. S. (1999). A measure of subjective happiness: Preliminary
reliability and construct validation. Social Indicators Research, 46(2), 137-155.
doi:10.1023/A:1006824100041
Martela, F., & Ryan, R. M. (2016). Prosocial behavior increases well-being and vitality even
without contact with the beneficiary: Causal and behavioral evidence. Motivation and
Emotion, 40(3), 351-357. doi:10.1007/s11031-016-9552-z
Milgram, S. (1975). Obedience to authority: An experimental view. New York, NY: Harper.
30
Mongrain, M., Chin, J. M., & Shapira, L. B. (2011). Practicing compassion increases happiness
and self-esteem. Journal of Happiness Studies, 12(6), 963-981. doi:10.1007/s10902-010-
9239-1
Myers, D. G. (2010). Social psychology (10th ed.). New York, NY: McGraw-Hill.
Nelson, S. K., Della Porta, M. D., Jacobs Bao, K., Lee, H. C., Choi, I., & Lyubomirsky, S.
(2015). 'it's up to you': Experimentally manipulated autonomy support for prosocial
behavior improves well-being in two cultures over six weeks. Journal of Positive
Psychology, 10(5), 463. doi:10.1080/17439760.2014.983959
Nelson, S. K., Layous, K., Cole, S. W., & Lyubomirsky, S. (2016). Do unto others or treat
yourself? the effects of prosocial and self-focused behavior on psychological flourishing.
Emotion, 16(6), 850-861. doi:10.1037/emo0000178.supp (Supplemental)
Nosek, B. A., Ebersole, C. R., DeHaven, A. C., & Mellor, D. T. (2018). The preregistration
revolution. Proceedings of the National Academy of Sciences of the United States of
America, 115(11), 2600. doi:10.1073/pnas.1708274114
O’Brien, E., & Kassirer, S. (2019). People are slow to adapt to the warm glow of
giving. Psychological Science, 30(2), 193-204. doi:10.1177/0956797618814145
O'Connell, B. H., O'Shea, D., & Gallagher, S. (2016). Enhancing social relationships through
positive psychology activities: A randomised controlled trial. Journal of Positive
Psychology, 11(2), 149-162. doi:10.1080/17439760.2015.1037860
Ouweneel, A., Blanc, L., PM Pascale, & Schaufeli, W. (2014). On being grateful and kind:
Results of two randomized controlled trials on study-related emotions and academic
engagement. Journal of Psychology, 148(1), 37-60. doi:10.1080/00223980.2012.742854
31
Pink, D. H. (2011). Drive: The surprising truth about what motivates us. New York, NY:
Penguin.
Richard, F. D., Bond, C. F., & Stokes-Zoota, J. J. (2003). One hundred years of social
psychology quantitatively described. Review of General Psychology, 7(4), 331-363.
doi:10.1037/1089-2680.7.4.331
Schreier, H. M., Schonert-Reichl, K. A., & Chen, E. (2013). Effect of volunteering on risk
factors for cardiovascular disease in adolescents: A randomized controlled trial. JAMA
Pediatrics, 167(4), 327-332. doi:10.1001/jamapediatrics.2013.1100
Smith, E. R., Mackie, D. M., & Claypool, H. M. (2014). Social psychology: Fourth
edition (Fourth ed.). London, ENG: Psychology Press. doi:10.4324/9780203833698
Strack, F., Schwarz, N., & Gschneidinger, E. (1985). Happiness and reminiscing: the role of time
perspective, affect, and mode of thinking. Journal of Personality and Social
Psychology, 49(6), 1460-1469.
Toi, M., & Batson, C. D. (1982). More evidence that empathy is a source of altruistic
motivation. Journal of Personality and Social Psychology, 43(2), 281-292.
Trew, J. L., & Alden, L. E. (2015). Kindness reduces avoidance goals in socially anxious
individuals. Motivation and Emotion, 39(6), 892-907. doi:10.1007/s11031-015-9499-5
Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of brief measures
of positive and negative affect: The PANAS scales. Journal of Personality and Social
Psychology, 54(6), 1063-1070. doi:10.1037/0022-3514.54.6.1063
Weinstein, N., & Ryan, R. M. (2010). When helping helps: Autonomous motivation for prosocial
behavior and its influence on well-being for the helper and recipient. Journal of
Personality and Social Psychology, 98(2), 222-244. doi:10.1037/a0016984
32
Whillans, A. V., Aknin, L. B., Ross, C. J., Chen, L., & Chen, F. S. (2019). Common variants of
the oxytocin receptor gene do not predict the positive mood benefits of prosocial
spending. Emotion, doi:10.1037/emo0000589
Whillans, A. V., Dunn, E. W., Sandstrom, G. M., Dickerson, S. S., & Madden, K. M. (2016). Is
spending money on others good for your heart? Health Psychology, 35(6), 574-583.
doi:10.1037/hea0000332
Whillans, A. V., Seider, S. C., Chen, L., Dwyer, R. J., Novick, S., Gramigna, K. J., ... & Dunn,
E. W. (2016). Does volunteering improve well-being? Comprehensive Results in Social
Psychology, 1(1-3), 35-50.
33
Table 1. Observed effect sizes in prosocial spending experiments and other prosocial intervention experiments on happiness.
Source
Study
DV
Experimental (E)
Control (C)
nE
nC
Sample (location)
Cohen’s d [CI95]
Baseline
Happiness
Controlled
Aknin,
Barrington
-Leigh et
al., 2013
2a†
2b
3*
3*
H
H
PA
SWLS
Recall prosocial spending
Recall prosocial spending
Prosocial spending
Prosocial spending
Recall personal spending
Recall personal spending
Personal spending
Personal spending
410
22
104
104
410
18
103
103
Student (Canada; Uganda)
Community (India)
Student (Canada; South Africa)
Student (Canada; South Africa)
.20 [.06, .34]
.55 [-.10, 1.19]
.46 [.18, .74]
.13 [-.14, .40]
No
No
Yes
No
Aknin et
al., 2015
1
2
PA
Smiling
Prosocial spending
Donate sweets (own)
Personal spending
Donate sweets (other)
13
20
13
20
Community (Vanuatu)
Children (Vanuatu)
.87 [.01, 1.70]
.30 [-.33, .92]
Yes
No
Aknin,
Dunn et
al., 2013
WB
Prosocial spending
Personal spending
25
25
Community (Canada)
.24 [-.32, .80]
No
Aknin et
al., 2014
PA
ORH
Prosocial spending
Prosocial spending
Personal spending
Personal spending
60
60
59
59
Student (Canada)
Research Assistant (Canada)
.38 [.01, .74]
.44 [.07, .81]
Yes
Yes
Aknin,
Dunn, &
Norton,
2012
H
Recall prosocial spending
Recall personal spending
26
25
Students (Canada)
.61 [.03, 1.18]
No
Aknin,
Hamlin, &
Dunn,
2012
Smiling
Donate sweets (own)
Donate sweets (other)
20
20
Children (Canada)
.46 [-.18, 1.09]
No
Alden &
Trew,
2013
PA
PA
Prosocial behaviour
Prosocial behaviour
Safety behaviour experiment
Life details tracking
43
43
40
43
Socially anxious student (Canada)
Socially anxious student (Canada)
.59 [.14, 1.04]
.54 [.10, .98]
No
No
Anik et al.,
2013
1
1
PA
PA
Prosocial spending ($25)
Prosocial spending ($50)
None
None
41
41
48
48
Community (Australia)
Community (Australia)
-.15 [-.57, .27]
.49 [.06, .92]
Yes
No
Buchanan
& Bardi,
2010
1
1
SWLS
SWLS
Prosocial behaviour
Prosocial behaviour
New behaviour
No behaviour
28
28
28
28
Community (UK)
Community (UK)
.41 [-.13, .94]
.62 [.07, 1.16]
No
No
Chancellor
et al., 2018
SHS
SWLS
SHS
SWLS
Prosocial behaviour
Prosocial behaviour
Prosocial behaviour
Prosocial behaviour
Prosocial receiver
Prosocial receiver
None
None
16
16
16
16
34
34
33
33
Community (Spain)
Community (Spain)
Community (Spain)
Community (Spain)
a
a
a
a
-
-
-
-
(continued on next page)
34
Table 1. (continued)
Source
Study
DV
Experimental (E)
Control (C)
nE
nC
Sample (location)
Cohen’s d [CI95]
Baseline
Happiness
Controlled
Donnelly
et al., 2017
1
1
2b
H
H
PA
Social recycling
Social recycling
Social recycling
Trash/recycling
Take item
Trash
59
59
107
56
59
108
Student (USA)
Student (USA)
Community (USA)
.77 [.38, 1.16]
.85 [.45, 1.24]
1.25 [.93, 1.56]
No
No
No
Dunn et
al., 2008
3
H
Prosocial spending
Personal spending
23
23
Student (Canada)
.67 [.05, 1.27]
Yes
Geenan et
al., 2014
H
Prosocial spending
Personal spending
34
34
Student (Germany)
.70 [.19, 1.20]
Yes
Hanniball
& Aknin,
2016
PA
Prosocial behaviour
Self-helping behaviour
51
56
Students (Canada)
-.46 [-.84, -.07]
No
Hanniball
et al., 2019
1†
2
3*
4*
PA
PA
PA
PA
Recall prosocial spending
Prosocial spending
Prosocial spending
Prosocial spending
Recall personal spending
Personal spending
Personal spending
Personal spending
250
31
357
588
251
33
420
707
Ex-offender adults (USA)
Delinquent youth (Canada)
Ex-offender adults (USA)
Ex-offender adults (USA)
.20 [.02, .38]
.70 [.17, 1.22]
.16 [.02, .30]
.11 [.00, .22]
Yes
Yes
Yes
Yes
Layous et
al., 2017
1
1
1
2
2
SHS
WB
EWB
SHS
WB
Prosocial behaviour
Prosocial behaviour
Prosocial behaviour
Prosocial behaviour
Prosocial behaviour
Track daily activity
Track daily activity
Track daily activity
Make self happier
Make self happier
70
70
70
178
178
69
69
69
81
81
Student (USA)
Student (USA)
Student (USA)
Student (USA)
Student (USA)
.08 [-.25, .41]
.20 [-.13, .53]
.26 [-.08, .59]
.30 [.04, .56]
.12 [-.14, .38]
No
No
No
No
No
Layous et
al., 2013
WB
Prosocial behaviour
Track locations
213
104
Student (USA/Korea)
.18 [-.06, .41]
No
Layous et
al., 2012
SHS
PA
SWLS
Prosocial behaviour
Prosocial behaviour
Prosocial behaviour
Whereabouts
Whereabouts
Whereabouts
208
208
208
208
208
208
Youth (Canada)
Youth (Canada)
Youth (Canada)
-.05 [-.24, .14]
-.12 [-.31, .07]
.07 [-.12, .26]
No
No
No
Martela &
Ryan,
2016
PA
Benevolence
Neutral activity
34
42
Students (USA)
.55 [.08, 1.02]
No
(continued on next page)
35
Table 1. (continued)
Source
Study
DV
Experimental (E)
Control (C)
nE
nC
Sample (location)
Cohen’s d [CI95]
Baseline
Happiness
Controlled
Mongrain
et al., 2011
SHI
Prosocial behaviour
Memory
237
237
Community (Canada)
.01 [-.17, 19]
No
Nelson et
al., 2015
SHS
SWLS
PE
Prosocial behaviour
Prosocial behaviour
Prosocial behaviour
Work activity
Work activity
Work activity
54.5
54.5
55
54.5
54.5
55
Students (USA; Korea)
Students (USA; Korea)
Students (USA; Korea)
.23 [-.15, .61]
.27 [-.11, .65]
.09 [-.28, .46]
No
No
No
Nelson et
al.,2016
PE
PE
Prosocial behaviour
Prosocial behaviour
Track activities
Self
238
238
116
116
Community/Student (USA)
Community/Student (USA)
.30 [.08, .52]
20 [-.02, .42]
No
No
O’Brien &
Kassirer,
2019
1
1
1
1
2*
H
H
H
H
H
Prosocial spending
Prosocial spending
Prosocial spending
Prosocial spending
Prosocial spending
Personal spending
Personal spending
Personal spending
Personal spending
Personal spending
59
59
59
59
249
54
54
54
54
253
Student (USA)
Student (USA)
Student (USA)
Student (USA)
Community (USA)
.35 [-.03, .73]
.46 [.08, .84]
.35 [-.03, .72]
.09 [-.28, .46]
.20 [.02, .38]
No
No
No
No
No
O’Connell
et al., 2016
SHS
SHS
Prosocial behaviour
Prosocial behaviour
List activities
Self
28
28
12
31
Community (USA)
Community (USA)
.02 [-.66, .70]
.12 [-.39, .63]
No
No
Ouweneel
et al., 2014
2
PE
Prosocial behaviour
Neutral activity
25
24
Student (Netherlands)
.27 [-.30, .83]
No
Trew &
Alden,
2015
PA
PA
Prosocial behaviour
Prosocial behaviour
Social exposure
List activities
38
36
41
41
Socially anxious student (Canada)
Socially anxious student (Canada)
-.05 [-.49, .39]
-.33 [-.78, .13]
Yes
Yes
Whillans et
al., 2016
WB
Prosocial spending
Personal spending
36
37
Hypertense older adults (Canada)
.19 [-.27, .65]
No
Whillans et
al., 2019
*
PA
Prosocial spending
Personal spending
218
219
Students (Canada)
.23 [.04, .42]
Yes
Average Prosocial Spending Effect Size (ns>100)
d = .22
Average Recall Prosocial Spending Effect Size (ns>100)
d = .20
Note: EWB = Eudaimonic Well-Being; H = Happiness; ORH = Other Rated Happiness; PA = Positive Affect; PE = Positive Emotion; SHS = Subjective Happiness
Scale; SHI = Steen Happiness Index; SWLS = Satisfaction With Life Scale; WB = Well-Being. * = Prosocial spending intervention with ns>100; entered in
estimate of Average Prosocial Spending Effect Size (ns>100). † = Recall prosocial spending intervention with ns>100; entered in estimate of Average Recall
Prosocial Spending Effect Size (ns>100). a = Statistics needed to calculate effect size were not reported in the paper, nor available from the authors.
36
Table 2. Summary of experimental results
Experiment 1 (N = 712)
Outcome
npersonal
nprosocial
Mpersonal (SD)
Mprosocial (SD)
F
p
ηp2
Cohen’s d
r
Current PANAS
359
344
2.73 (.69)
2.97 (.78)
22.77
< .001
.03
.36
.18
Current SPANE
358
341
19.65 (4.56)
21.03 (5.29)
17.82
< .001
.02
.32
.16
Experiment 2 (N = 1,950)
Outcome
npersonal
nprosocial
Mpersonal (SD)
Mprosocial (SD)
F
p
ηp2
Cohen’s d
r
Current PANAS
983
963
3.08 (.91)
3.10 (.90)
.457
.499
.00
.03
.02
Current SPANE
983
964
19.99 (5.93)
19.93 (6.01)
.118
.732
.00
.02
.01
Experiment 3 (N = 5,199)
Outcome
npersonal
nprosocial
Mpersonal (SD)
Mprosocial (SD)
F
p
ηp2
Cohen’s d
r
Current PANAS
2610
2584
3.01 (.91)
3.04 (.90)
5.34
.010
.00
.06
.03
Current SPANE
2613
2584
19.49 (6.19)
19.64 (6.20)
4.57
.016
.00
.06
.03
Retrospective SPANE
2613
2584
21.90 (5.85)
22.69 (5.69)
37.93
< .001
.01
.17
.09
Note. PANAS: Positive and Negative Affect Schedule (Watson, Clark & Tellegen, 1988). SPANE: Scale of Positive and Negative Experience
(Diener et al., 2009)
37
Table 3. Summary of sample demographics (recall studies)
Experiment 2
(N = 1,950)
Experiment 3
(N = 5,199)
Experiment 2
(N = 1,950)
Experiment 3
(N = 5,199)
Race/Ethnicity
N
%
N
%
Age
M
(SD)
Mdn
(Range)
M
(SD)
Mdn
(Range)
First Nation/Native
American
21
1.1%
34
0.7%
42.72
(15.19)
40.00
(14–90)
39.00
(11.77)
36.00
(18–90)
African
American/Black
236
12.1%
443
8.5%
Annual Pre-tax
Household Income
N
%
N
%
Hispanic
140
7.2%
252
4.8%
Less than $10,000
124
6.4%
223
4.3%
Caucasian/White
1419
72.8%
3998
76.9%
$10,000 - $19,999
167
8.6%
341
6.6%
Asian
77
3.9%
322
6.2%
$20,000 - $29,999
245
12.6%
526
10.1%
Middle Eastern
6
0.3%
11
0.2%
$30,000 - $39,999
222
11.4%
650
12.5%
Multi-racial
36
1.8%
103
2.0%
$40,000 - $49,999
196
10.1%
529
10.2%
Other
10
0.5%
16
0.3%
$50,000 - $59,999
186
9.5%
595
11.4%
Prefer not to answer
5
0.3%
18
0.3%
$60,000 - $69,999
148
7.6%
443
8.5%
Missing Data
0
0.0%
2
0.0%
$70,000 - $79,999
128
6.6%
423
8.1%
Gender
N
%
N
%
$80,000 - $89,999
82
4.2%
281
5.4%
Male
605
31.0%
2205
42.4%
$90,000 - $99,999
105
5.4%
280
5.4%
Female
1339
68.7%
2962
57.0%
$100,000 - $149,999
219
11.2%
589
11.3%
Other
5
0.3%
22
0.4%
More than $150,000
113
5.8%
222
4.3%
Prefer not to answer
1
0.1%
8
0.2%
Prefer not to answer
15
0.8%
95
1.8%
Missing Data
0
0.0%
2
0.0%
Missing Data
0
0.0%
2
0.0%
