Why a Simple Act of Kindness Is Not as Simple as It Seems: Underestimating the Positive Impact of Our Compliments on Others

Abstract

A simple compliment can make someone’s day, start a new friendship, or just make the world a better, kinder place. So, why don’t people give more compliments? Perhaps people misforecast the effect their compliment will have. Five studies explored this possibility. In Studies 1a and 1b, compliment givers underestimated how positively the person receiving their compliment would feel, with consequences for their likelihood of giving a compliment. Compliment givers also overestimated how bothered and uncomfortable the recipient would feel (Study 2)—and did so even in hindsight (Study 3). Compliment givers’ own anxiety and concern about their competence led to their misprediction, whereas third-party forecasters were accurate (Study 4). Finally, despite compliment givers’ anxiety at the prospect of giving compliments across our studies, they felt better after having done so (Study 4). Our studies suggest that people misestimate their compliments’ value to others, and so they refrain from engaging in this prosocial behavior.

Full text

https://doi.org/10.1177/0146167220949003
Personality and Social
Psychology Bulletin
2021, Vol. 47(5) 826 –840
© 2020 by the Society for Personality
and Social Psychology, Inc
Article reuse guidelines:
sagepub.com/journals-permissions
DOI: 10.1177/0146167220949003
journals.sagepub.com/home/pspb
Article
“Whenever there is a human being, there is an opportunity for a
kindness.”
—Lucius Annaeus Seneca
In the time it takes to tie your shoes, you could compli-
ment a passerby on their style, you could approach someone
at a conference and tell them you loved their talk, or you
could tell the security guard in your building that you are ter-
ribly fond of his choice of socks. Any of these compliments
could make that person’s day. And you would likely walk
away feeling good having done so. It is difficult to contend
with the proposition that the world would be a better, kinder
place if people took the time to say nice things to one another
more often.
And yet, for an act so simple and easy that would increase
the well-being of everyone involved, why do people refrain
from doing it? Why don’t people give more compliments,
particularly to those whom they do not know well? There are
several possible reasons. Perhaps people fail to notice oppor-
tunities to do so. Or maybe they think they do not have time
or do not want to put in the effort. Maybe noticing someone
else’s positive qualities evokes jealousy and competitive-
ness, so compliments are withheld. However, we propose a
different explanation: People misforecast how their compli-
ments make recipients feel.
Although research on prosocial behavior has grown rap-
idly in the past several decades, it has focused largely on the
affective consequences of giving gifts (Dunn et al., 2008;
Kupor et al., 2017), giving money (Frey & Meier, 2004;
Weinstein & Ryan, 2010; Zaki & Mitchell, 2011), and giving
one’s time to help others (Freeman, 1997; Thoits & Hewitt,
2001). Recently, researchers have begun to look at more
intangible sorts of prosocial behavior such as giving thanks
(Kumar & Epley, 2018). However, despite its important role
in everyday life—for example, in facilitating social connec-
tions and increasing interpersonal liking (Jones, 1964; Jones
& Wortman, 1973)—little is known about the prosocial act
of giving a compliment, and the answers to such fundamen-
tal questions as whether people are aware of the effects their
compliments have on others.
949003PSPXXX10.1177/0146167220949003Personality and Social Psychology BulletinBoothby and Bohns
research-article2020
1University of Pennsylvania, Philadelphia, USA
2Cornell University, Ithaca, NY, USA
Corresponding Author:
Erica J. Boothby, The Wharton School, University of Pennsylvania,
Philadelphia, PA 19104, USA.
Email: ericajboothby@gmail.com
Why a Simple Act of Kindness Is Not
as Simple as It Seems: Underestimating
the Positive Impact of Our Compliments
on Others
Erica J. Boothby1 and Vanessa K. Bohns2
Abstract
A simple compliment can make someone’s day, start a new friendship, or just make the world a better, kinder place. So,
why don’t people give more compliments? Perhaps people misforecast the effect their compliment will have. Five studies
explored this possibility. In Studies 1a and 1b, compliment givers underestimated how positively the person receiving their
compliment would feel, with consequences for their likelihood of giving a compliment. Compliment givers also overestimated
how bothered and uncomfortable the recipient would feel (Study 2)—and did so even in hindsight (Study 3). Compliment
givers’ own anxiety and concern about their competence led to their misprediction, whereas third-party forecasters were
accurate (Study 4). Finally, despite compliment givers’ anxiety at the prospect of giving compliments across our studies, they
felt better after having done so (Study 4). Our studies suggest that people misestimate their compliments’ value to others,
and so they refrain from engaging in this prosocial behavior.
Keywords
compliment, social influence, prosocial behavior, well-being, conversation
Received December 24, 2019; revision accepted July 15, 2020

Boothby and Bohns 827
Why would people misforecast the impact of their com-
pliment on its recipient? For one, people are often anxious
about interacting with people whom they do not know well
(Boothby et al., 2018; Epley & Schroeder, 2014; Sandstrom
& Boothby, 2020; Sandstrom et al., manuscript under
review). However, in reality talking to strangers leaves peo-
ple feeling better than they did before (McIntyre et al., 1991;
Sandstrom & Dunn, 2014a, 2014b; Vittengl & Holt, 2000)
and better than they anticipated feeling (Epley & Schroeder,
2014). If anticipating an interaction with a stranger provokes
anxiety, those feelings might impact their estimate of the
effect their compliment will have on its recipient, either
because they are using their own affective state as informa-
tion when making their estimate (Clore et al., 2001) or
because they are egocentrically projecting their own affec-
tive state onto the recipient of their compliment (e.g., Ross
et al., 1977; Van Boven & Loewenstein, 2003).
People may also misforecast the impact of their compli-
ment in part because they are overly concerned about their
ability to deliver a compliment competently. While people
are indeed evaluated on the basis of their competence, assess-
ments of warmth (e.g., sociability) are usually more impor-
tant to perceivers (Fiske et al., 2007; Wojciszke, 1994). For
example, when expressing gratitude, people are overly con-
cerned about their ability to express themselves just right,
leading them to underestimate how the recipient of their
gratitude will feel (Kumar & Epley, 2018). We suspect that
people are similarly concerned about their ability to skill-
fully express a compliment to someone, which might inhibit
them from complimenting people they might otherwise.
In sum, previous research has found that people feel anx-
ious about interacting with new people and concerned about
their ability to express their feelings competently. We pre-
dicted that the combination of these factors would lead people
to misforecast how good their compliments would make peo-
ple on the receiving end feel, and that this error would result
in a lower likelihood of giving compliments. We explored this
hypothesis in four studies that sought to establish the exis-
tence of this forecasting error and identify its causes.
Study 1a: Do People Misestimate the
Value of a Compliment?
Method
Overview. Participants were recruited to the laboratory and
sent out onto campus to compliment a stranger. Beforehand,
they predicted how their compliment would make the person
feel. After giving their compliment, they asked the receiver
of their compliment to complete a brief survey (the contents
of which were unknown to the compliment giver until they
were debriefed at the end of the session) on how the interac-
tion made them feel.
Pre-registration and data accessibility. This study was pre-
registered at AsPredicted.org (#6286). Data and materials for
all studies are available at Open Science Foundation (OSF;
https://tinyurl.com/ydbrn4b7).
Participants. We pre-specified a sample of 100 compliment
givers, who would be paired with 100 compliment receivers.
A power analysis (GPower 3; Faul et al., 2007) indicated that
this sample size had 99.99% statistical power to detect a mini-
mum effect size of 0.40. We were able to recruit a total of 190
participants (70.50% female and missing gender information
for one participant; Mage = 21.15 years, SD = 4.54; 38.9%
White, 37.4% Asian, 6.8% Black/African American, 6.8%
“mixed,” 4.7% Hispanic/Latino, 2.6% Indian, 0.5% Native
American, 1.1% “other,” and missing information for two
participants) for cash compensation.
Procedure. After arriving at the laboratory, participants (our
compliment givers) were told they would go to an assigned
campus location (e.g., dining hall, building lobby), and give
a simple, straightforward compliment to a matched gender
stranger. Following the procedure of Roghanizad & Bohns
(2017) to reduce selection bias, participants were instructed
to compliment the fourth male or female (depending on their
own gender) they saw (compliment receivers) once they
arrived at their designated location. They were instructed to
say, “I like your shirt” (or jacket or dress, if no shirt was
visible).
Next, they were instructed to hand the person a sealed
envelope and say, “I’m supposed to give this to you as part of
an experiment I’m participating in. I don’t know what’s in it,
but I need to wait for you to fill it out and give it back to me.”
This envelope contained a short survey along with an empty
unsealed envelope so that, after completing the survey, com-
pliment receivers could submit their survey confidentially.
Once the person handed back the sealed envelope, compli-
ment givers returned to the lab.
This methodology maximized the ecological validity of
the study in two ways. First, the interaction took place out-
side of the lab, in people’s everyday lives. If a stranger was
complimented in the lab, it could seem artificial and raise
suspicion. Second, it minimized the unnaturalness of the
interaction to have compliment givers, rather than research
assistants, collect compliment receivers’ sealed envelopes.
It would have been more awkward, unexpected, and far less
natural if a brand-new person who was “in on” the event
appeared after the compliment was delivered to collect the
envelope.
Before giving their compliment, compliment givers
answered several questions asking how they thought the
other person would feel as a result of their compliment.
These questions were posed similarly to Epley & Schroeder
(2014), who asked participants to indicate how stimulating
they imagined a conversation with an unknown stranger
during their commute would be. Our studies focused instead
on participants’ expectations of an unknown stranger’s reac-
tion to their compliment. Participants were asked how much
the other person would enjoy the interaction, and would feel

828 Personality and Social Psychology Bulletin 47(5)
good, pleased, and flattered as a result of the interaction
using 7-point Likert-type scales (1 = not at all, 7 = very/
very much).
The compliment receiver survey mirrored the compliment
giver survey, measuring how people felt as a result of receiv-
ing the compliment. They answered the same four questions
about how much they enjoyed the interaction, and felt good,
pleased, and flattered as a result of the interaction.1
Upon returning to the lab, compliment givers responded
to some additional questions and were debriefed and
compensated.2
Results
Data were excluded from four dyads in which at least one
person failed to complete the survey.3 Including data from
these dyads does not change any of the results.
Our primary research question was whether participants
accurately predicted how good their compliment would
make its recipient feel. Our four measures of how good com-
pliment givers thought the other person would feel were
highly correlated (α = .93). Following our pre-registration,
we averaged these measures into a single measure of per-
ceived positivity. Likewise, our four measures of how good
people actually felt as a result of receiving a compliment
were also highly correlated (α = .91), and they were aver-
aged to create a single measure of actual positivity. These
measures, collectively referred to as a positivity index, served
as our primary dependent variable.
Because participants were nested within dyads, for this
and all subsequent studies, as pre-registered, we fit mixed
linear models to the data in R using the lme4 package (Bates
et al., 2015) with role (giver or receiver) as the independent
variable, and our positivity index as the dependent variable.
Our model included our independent variable as a fixed
effect, and an intercept for each dyad as random effects. We
used the lmerTest package (Kuznetsova et al., 2014) to derive
p values and degrees of freedom. Note that the reported
means are predicted marginal means.
Predicted versus actual value of compliment. Did compliment
givers underestimate the value of their compliment to its
recipient? As predicted, there was a significant effect of role
(compliment giver vs. compliment receiver) on positivity,
b = .72, SE = .16, t(92) = 4.48, p < .001, 95% confidence
interval (CI) = [0.40, 1.05], with receivers feeling signifi-
cantly better, M = 4.66, 95% CI = [4.42, 4.90], than givers
thought they would, M = 3.93, 95% CI = [3.69, 4.17].
Discussion
These results indicate that people underestimate how good
their compliment will make someone feel. Participants who
were instructed to compliment a specific stranger believed
that their interaction would make that person feel less happy,
good, pleased, and flattered than it actually did.
Study 1b: Do People Refrain From
Giving Compliments Because They
Underestimate the Value of Their
Compliment?
Method
Overview. This study was designed to achieve four goals (a)
to provide a direct test of our thesis that people refrain from
giving compliments because they misestimate their value to
others, (b) to rule out an alternative explanation for why peo-
ple may avoid complimenting others (social comparison), (c)
to demonstrate that people misestimate the value of their
compliment even when their interaction does not involve
asking the compliment receiver to fill out a survey, which
may independently contribute to people’s pessimistic assess-
ments of how the compliment receiver will react (Flynn &
Lake, 2008), and (d) to find out what kinds of compliments
people normally give strangers in their daily lives to compare
these naturally occurring compliments to the compliments
used in our first and subsequent studies.
Pre-registration. This study was pre-registered at AsPredicted.
org (#41694).
Participants. We pre-specified a sample of at least 200 par-
ticipants and were able to recruit 237 people (51.1% female
and 2.1% preferred not to indicate their gender; Mage = 30.73
years, SD = 11.66; 68.8% White, 12.2% Asian, 7.2% Black/
African American, 5.5% Hispanic/Latino, 4.6% “more than
1 of the above,” 0.8% “prefer not to answer,” 0.4% Hawai-
ian/Pacific Islander, 0.4% “other”) on Prolific Academic.
Procedure. Participants were assigned to one of two roles:
compliment giver or compliment receiver.
Compliment giver. Compliment givers were told, “Imag-
ine you notice something you really like about someone you
don’t know (who’s the same gender as you). You approach
them and give them a genuine, straightforward compliment
(e.g., ‘I really like your shirt’).” Participants indicated how
happy, good, pleased, and flattered they thought this stranger
would feel as a result of this interaction using 7-point Lik-
ert-type scales (1 = not at all, 7 = very/very much). Next,
they reported how likely they would be to compliment the
stranger on something they genuinely liked about them using
a 7-point Likert-type scale (1 = not at all likely, 7 = very
likely).
To rule out an alternative explanation noted earlier—
namely, that people may refrain from giving compliments
out of jealousy and competitiveness due to upward social
comparison (Aspinwall & Taylor, 1993; Mendes et al., 2001;
Wills, 1981)–compliment givers also reported how bad they
would feel about themselves having noticed something
they liked about the stranger using a 7-point Likert-type
scale (1 = not at all bad, 7 = very bad).

Boothby and Bohns 829
Compliment receiver. Compliment receivers were told,
“Imagine someone you don’t know (who’s the same gender
as you) notices something they really like about you. This
person approaches you and gives you a genuine, straight-
forward compliment (e.g., ‘I really like your shirt’).” Using
7-point Likert-type scales (1 = not at all, 7 = very/very
much), they indicated how happy, good, pleased, and flat-
tered they would feel as a result of this interaction.
All participants were asked to report on their desire for
strangers to compliment one another more often in two ques-
tions using 7-point scales with endpoints labeled “Strongly
disagree” and “Strongly agree.” Finally, participants were
asked to report on the kinds of compliments they have given
to strangers using a text box.
Results
Data were excluded from 15 participants who failed to cor-
rectly answer attention checks. Including data from these
participants does not change any of the results.
We first examined whether we replicated our effect from
Study 1a. Specifically, we tested whether compliment receiv-
ers reported that they would feel better than compliment giv-
ers believed they would. To do so, we created composite
variables from the four dependent measures of perceived
positivity (α = .92) and actual positivity (α = .95). These
measures, collectively referred to as a positivity index, served
as our primary dependent variable for this analysis. We fit a
linear model to the data with role (giver or receiver) as the
independent variable, and our positivity index as the depen-
dent variable. As predicted, the analysis revealed a significant
effect of role on positivity, b = .60, SE = .15, t(220) = 3.92,
p < .001, 95% CI = [0.30, 0.90], with receivers reporting
that they would feel significantly better, M = 5.65, 95%
CI = [5.45, 5.86], than givers thought receivers would,
M = 5.06, 95% CI = [4.84, 5.28].
Did compliment givers’ misestimation affect their likeli-
hood of complimenting a stranger? A linear model with
compliment givers’ perceived positivity as the independent
variable and their likelihood of giving a compliment as the
dependent variable revealed a significant effect, b = .55,
SE = .11, t(102) = 4.91, p < .001, 95% CI = [0.33, 0.78].
People were more likely to compliment a stranger when
they believed their compliment would have a positive effect.
We next fit a linear model with compliment givers’ antici-
pated negative feelings about themselves as a result of social
comparison as the independent variable, and their likelihood
of giving a compliment as the dependent variable. This anal-
ysis revealed a non-significant effect, p = .38.
Our two measures of whether people believe strangers
ought to give each other more compliments were highly cor-
related (α = .88), and we averaged them into a single mea-
sure. A one-sample t-test comparing participants’ responses
to the scale midpoint (“4—Neither agree nor disagree”)
revealed that people wished strangers complimented one
another more often, M = 5.57, SD = 1.02, t(221) = 22.87,
p < .001, 95% CI = [5.43, 5.70].
Finally, 81% of participants in our sample reported com-
plimenting strangers on physical characteristics (e.g., clothes,
hairstyle, tattoo).4,5
Discussion
In Study 1b, people reported being more likely to compli-
ment a stranger to the extent that they anticipated their com-
pliment would be received positively, whereas people’s
social comparison concerns were unrelated to their likeli-
hood of complimenting a stranger. Furthermore, participants
underestimated the value of their compliment even when
they did not have to ask the compliment receiver to fill out a
survey as part of the study design. The vast majority of com-
pliments people reported giving strangers in their everyday
lives were for physical characteristics, in line with the kind
of compliment participants gave strangers in Study 1a, sug-
gesting that clothing and other aspects of people’s appear-
ance are common subjects of compliments among strangers.
Study 2: Do People Underestimate the
Value of a Genuine Compliment?
In Study 1a, people gave a scripted compliment, that is, par-
ticipants were instructed to compliment someone on their
shirt. Although this enabled us to standardize compliments
across dyads, it may also have contributed to compliment
givers’ pessimism about how good their compliment would
make its recipient feel, that is, compliment givers may have
been concerned that their compliment would not seem genu-
ine. To eliminate this possibility, in Study 2 compliment giv-
ers complimented the stranger on whatever they genuinely
liked about them.
Furthermore, compliment givers in both Studies 1a and
1b underestimated the value of their compliment to the recip-
ient. But might people realize its value after having compli-
mented the person and observed their reaction? In Study 2,
we tested the possibility that people are unable to adequately
update their perception of their compliment’s value, which
may contribute to the persistence of their misestimation.
Pre-Registration
This study was pre-registered at AsPredicted.org (#8690).
Participants
We aimed to recruit 50 compliment givers for a total of 100
participants. We were able to recruit 100 participants (72%
female and missing gender information for three participants;
Mage = 21.59 years, SD = 5.53; 48% White, 28% Asian, 14%
Black/African American, 3% Hispanic/Latino, 2% Indian,
2% “mixed,” 1% “other,” and missing information from two

830 Personality and Social Psychology Bulletin 47(5)
participants). A power analysis indicated that this sample size
had 97.72% statistical power to detect a minimum effect size
of 0.40 for our primary analyses.
Procedure
The procedures were identical to those of Study 1a with two
exceptions. First, compliment givers were instructed to give
the stranger a compliment of their choice. Participants were
told, “We recommend finding something about them that
you like and complimenting them on that. It’s up to you
what, specifically, you want to compliment the person on.”
Since complimenting someone new on something about
them that one likes is a common way to break the ice, one
benefit of this design is its ecological validity and relevance
to people’s everyday lives. Second, upon returning to the lab,
compliment givers were asked how they thought the person
they complimented felt as a result of their interaction on the
same 7-point Likert-type scales (enjoyed, good, pleased,
flattered).
Results
Data were excluded from four dyads in which at least one
person failed to complete the survey or follow the instructions.6
Including data from these dyads does not change any of the
results. We created composite variables from the four depen-
dent measures of perceived positivity (α = .91) and actual
positivity (α = .85).
Predicted versus actual value of compliment. Our primary
research question was whether compliment givers underes-
timate the value of their compliment to its recipient. Indeed,
we found a significant effect of role (compliment giver vs.
compliment receiver) on positivity, b = .90, SE = .19,
t(46) = 4.63, p < .001, 95% CI = [0.51, 1.28], with
receivers feeling significantly better, M = 5.08, 95% CI =
[4.77, 5.39], than givers thought they would, M = 4.18,
95% CI = [3.87, 4.50] (see Figure 1).
Compliment givers’ pre- versus post-interaction perceptions of the
value of their compliment. It is clear that compliment givers
underestimate the value of their compliment to its recipient
prior to giving their compliment. But do they continue to
undervalue their compliment afterward? If so, this mistaken
belief might be one reason for people’s hesitance to compli-
ment strangers in everyday life. Our analysis revealed that
although compliment givers believed that the recipient of
their compliment felt better after they gave their compliment,
M = 4.43, 95% CI = [4.10, 4.77], than they had forecasted,
M = 4.18, 95% CI = [3.85, 4.52], this difference did not
reach significance, p = .12, that is, compliment givers failed
to sufficiently update their mistaken beliefs after having
given their compliment, indicating a failure to realize just
how positively their compliment affected the stranger after
the fact.
Discussion
In Study 1a, participants were constrained to complimenting
someone on their shirt, which because of the artificiality of
the compliment might have contributed to their undervaluing
of their compliment. In Study 2, compliment givers could
compliment someone on anything they genuinely liked about
them. Even when people were free to compliment someone
on something they genuinely liked about them, compliment
givers continued to underestimate how good their compli-
ment would make a recipient feel. We also found that partici-
pants complimented people on a variety of features—primarily
clothing, accessories, and physical features (see Table 1).
Although such compliments may seem superficial, they often
reflect affirmation of people’s personal choices, identity,
Figure 1. Results of Study 2: Means of actual and perceived effect of compliment. Error bars show the 95% confidence intervals around
the means.

Boothby and Bohns 831
culture, and so on, and so may be more meaningful than they
first appear.
Furthermore, even after giving their compliment, people
failed to realize how positively their compliment made its
recipient feel. This is not terribly surprising. After all, polite-
ness norms dictate that people be courteous and positive in
most social interactions (Jones & Wortman, 1973). Thus, it
can be difficult to discern whether someone’s positive
response to your compliment indicates that they truly
enjoyed receiving your compliment or they are simply being
polite. This inability to realize just how positively strangers
actually feel after receiving a compliment likely contributes
to people’s tendency to refrain from giving compliments
when they have the opportunity to do so.
Study 3: Do People Overestimate the
Cost of a Compliment?
Compliment givers in Studies 1a, 1b, and 2 underestimated
the positive impact their compliment had on its recipient.
Study 3 tested whether people additionally overestimate the
negative impact they will have on the stranger they compli-
ment. We hypothesized that because people tend to feel
awkward and uncomfortable about interacting with new
people (Boothby et al., 2018; Epley & Schroeder, 2014),
they would assume the converse to also be true—that is,
that by approaching a stranger they would make that person
feel awkward and uncomfortable. Thus, because of peo-
ple’s anxiety about interacting with a stranger, we predicted
that participants would overestimate how annoyed and
bothered by the interaction the other person would be.
Pre-Registration
This study was pre-registered at AsPredicted.org (#8961).
Participants
We aimed to recruit 50 compliment givers, for a total of 100
participants. We recruited 100 participants (76% female,
Mage = 20.61 years, SD = 4.48; 35% White, 29% Asian,
12% Black/African American, 10% Hispanic/Latino, 14%
“other”). A power analysis indicated that this sample size had
97.72% statistical power to detect a minimum effect size of
0.40.
Procedure
The procedures were identical to those of Study 2, with the
exception that in addition to asking about the predicted and
actual positive impact of their compliment, compliment giv-
ers and receivers were asked about the predicted and actual
negative impact of their compliment. Compliment givers
were asked how annoyed, bothered, and uncomfortable the
other person would feel as a result of the interaction. These
items were combined into a perceived negativity (α = .82).
Likewise, for compliment receivers, actual negativity
(α = .89).
Results
Predicted versus actual positive effect of compliment. As in Stud-
ies 1 and 2, our analysis revealed a significant effect of role
(compliment giver vs. compliment receiver) on our positivity
index, b = .72, SE = .17, t(50) = 4.20, p < .001, 95% CI =
[0.38, 1.05], with compliment receivers feeling significantly
better, M = 4.87, 95% CI = [4.58, 5.16], than compliment
givers thought they would, M = 4.16, 95% CI = [3.86, 4.45].
Predicted versus actual negative effect of compliment. We have
repeatedly found that compliment givers underestimate the
value of their compliment to its recipient. Here we also tested
whether compliment givers also overestimate the negative
impact their compliment has on its recipient. Our analysis
revealed a significant effect of role (compliment giver vs.
compliment receiver) on negativity, b = −1.75, SE = .23,
t(50) = −7.76, p < .001, 95% CI = [−2.20, −1.30], with
receivers feeling significantly less annoyed, bothered, and
uncomfortable, M = 2.42, 95% CI = [2.04, 2.80], than giv-
ers anticipated, M = 4.17, 95% CI = [3.79, 4.54], as a result
of being complimented by them.
Compliment givers’ pre- versus post-interaction perceptions of the
value of their compliment. Replicating the results of Study 2,
there was no significant effect of time (pre- vs. post-interac-
tion) on compliment givers’ beliefs about how positively
their compliment would affect its recipient, p = .59. How-
ever, compliment givers believed their compliment had a sig-
nificantly less negative effect on the compliment receiver,
M = 3.39, 95% CI = [3.03, 3.75], post-interaction than they
had initially forecasted, M = 4.17, 95% CI = [3.81, 4.53];
b = −.77, SE = .19, t(50) = −4.13, p < .001, 95% CI =
[−1.15, −0.40] (see Figure 2), indicating that compliment
givers did update their beliefs about the negative impact of
their compliments somewhat. However, this adjustment was
insufficient; compliment givers’ perceptions of how nega-
tively their compliment made the stranger feel was still
Table 1. Content of Compliments Given in Study 2.
Percent of
compliment givers Content of compliment
20 Physical feature (e.g., hair, eyebrows)
42 Top (e.g., shirt, sweater, sweatshirt, jacket)
24 Accessory (e.g., scarf, jewelry, backpack)
6 Outfit
8 Shoes

832 Personality and Social Psychology Bulletin 47(5)
significantly worse than the effect it actually had, M = 2.42,
95% CI = [2.02, 2.82]; b = −.97, SE = .22, t(50) = −4.50,
p < .001, 95% CI = [−1.41, −0.54].
Discussion
Study 3 provides evidence that people not only fail to realize
how positive another person will feel as a result of receiving
their compliment; they also drastically overestimate how
bothered, uncomfortable, and annoyed the other person will
feel. People’s overly pessimistic assessments of how uncom-
fortable and bothered a stranger will feel as a result of receiv-
ing their compliment may hinder them from complimenting
strangers in their everyday lives. This seems especially likely
given that participants in Study 3 were unable to adequately
update their initial forecasts of how negatively their compli-
ment impacted its recipient. Although they did update their
beliefs somewhat in the right direction, they continued to be
anchored on their pre-compliment forecasts, believing that
their interaction had a more negative impact on the stranger
they complimented than it actually did. This miscalibration
likely reduces people’s likelihood of complimenting strang-
ers of their own accord.
Study 4: Causes of Underestimating the
Value of a Compliment
The results of Studies 1 to 3 clearly demonstrate a tendency
to underestimate the value of one’s compliment. But why?
We predicted two major reasons why people underestimate
the value of their compliment. First, we hypothesized that
people are anxious about complimenting a stranger, which
contributes to their pessimistic beliefs about the effect of
their compliment. There are several reasons to believe that
approaching a stranger to give them a compliment would
cause people anxiety. For one, research has demonstrated
that people by and large underestimate strangers’ interest in
interacting with them (Epley & Schroeder, 2014; Shelton &
Richeson, 2005). And when people anticipate being socially
evaluated, they tend to expect to be judged more harshly by
others than they actually are Savitsky et al. (2001), Savitsky
& Gilovich (2003). Indeed, consistent with prior research,
the results of Study 3 reveal that people believed that by giv-
ing a stranger a compliment they were “bothering” the per-
son to a greater extent than they actually were. Furthermore,
when people are socially anxious they exhibit an attenuated
self-serving attribution bias (Mezulis et al., 2004), presenting
their abilities and attributes significantly more modestly than
do individuals low in social anxiety (Arkin et al., 1980). In
short, people may feel anxious leading up to giving a stranger
a compliment, resulting in overly pessimistic forecasts about
the effect their compliment will have on its recipient.
A second reason people may underestimate the value of
their compliment is that people are concerned about their
ability to give a compliment competently. That is, one result
of people’s tendency to underestimate strangers’ interest in
interacting with them, as discussed above, is that people
likely feel some pressure to perform well in giving their
compliment, to make the interaction worth the stranger’s
while. Research on people’s “willingness to communicate”
(McCroskey, 1992) reveals that people report being less will-
ing to talk to strangers to the extent that they are concerned
about their own communication competence (McCroskey &
Richmond, 1990). Indeed, when people’s communication
skills increase as a result of training, people become more
willing to communicate, underscoring the importance of
feeling competent. People’s doubts about their own social
competence have also been found in more general beliefs
that oneself has fewer friends than one’s peers, goes to fewer
parties, and has less rich social networks (Deri et al., 2017),
Figure 2. Results of Study 3: Means of actual and perceived positive and negative effects of compliment. Error bars show the 95%
confidence intervals around the means.

Boothby and Bohns 833
and in specific beliefs about their ability to adequately
express gratitude to friends and acquaintances (Kumar &
Epley, 2018). Thus, in Study 4, we tested whether people’s
anxiety and doubts about their own competence in giving a
compliment predict their pessimistic beliefs about the effect
of their compliment in Study 4.
In addition, Study 4 included a group of third-party fore-
casters who predicted the value a stranger’s compliment
would have on its recipient. This allowed us to address a
couple of questions. First, this design provided an additional
test of the primary aim of this study—to examine how anxi-
eties related to actually giving a compliment shapes compli-
ment givers’ misforecasts about how it will be received. We
hypothesized that third-party predictors’ estimates would be
more accurate than those of compliment givers because they
have no reason to feel anxious, as they are not about to give
someone a compliment. Second, this design allowed us to
address concerns regarding the potential demand effects of
having compliment givers hand the recipients of their com-
pliments a survey. If compliment receivers felt pressure to
inflate their responses, we would expect differences between
compliment givers’ and compliment recipients’ ratings to be
caused solely by the recipients’ responses, rather than being
characteristics of the compliment givers’ pessimism. Thus, if
outside forecasters’ predictions look different from compli-
ment givers’ predictions, this would suggest that our results
must be explained, at least in part, by the attributes of com-
pliment givers, such as their anxieties—not just attributes of
the compliment recipients.
Finally, Study 4 tested the prediction that giving a com-
pliment confers benefits not only to the receiver but also to
the giver, and that people would feel more inclined to give
a compliment in the future after having done so before—
suggesting a way we might be able to increase the number
of compliments people are willing to give.
Pre-Registration
This study was pre-registered at AsPredicted.org (#20640).
Participants
We aimed to recruit as many participants as possible by the
end of the semester. We were able to recruit 50 compliment
givers and 50 third-party predictors to the laboratory, for a
total of 150 participants (57.3% female; Mage = 19.62 years,
SD = 2.28; 46.7% White, 26% Asian, 5.3% Black/African
American, 5.3% Indian, 6.7% Hispanic/Latino, 0.7% Native
American, 1.3% “other”). A power analysis indicated that
this sample size had 99.81% statistical power to detect a
minimum effect size of 0.40 for our primary analyses.
Procedure
Participants who arrived in the laboratory were randomly
assigned to one of two conditions, namely compliment giver
or third-party predictor. Compliment givers were given
instructions identical to those in Study 3. Before leaving the
lab to give their compliment, compliment givers answered
the same positivity and negativity questions as in Study 3,
which were again combined into a perceived positivity index
(α = .86) and a perceived negativity index (α = .78). In addi-
tion, compliment givers were asked several questions on
7-point Likert-type scales about how good and anxious they
currently felt, how competently and skillfully they would be
able to deliver their compliment to a stranger, and how likely
they would be to give a stranger a compliment if it was not
part of a study. Upon returning to the lab, compliment givers
were once again asked how good they felt and how likely
they would be to compliment a stranger in future.
Participants assigned to be third-party predictors were
told about a study being run by researchers at their university
and instructed to answer a few questions about it:
In that study, participants are told they will be asked to give a
stranger a compliment. The stranger they compliment must
match their own gender, and they should give the person a
simple, straightforward compliment (e.g., “I like your shirt”).
It’s up to the participant what, specifically, they want to
compliment the person on. Now think of one participant giving
a stranger a compliment. Answer the following questions about
how you think the stranger receiving the compliment will feel as
a result of this interaction.
They were asked the same questions about a receiver’s
positive and negative reactions as were the participants
assigned to the compliment giver condition. These responses
were combined into a third-party predicted positivity
index (α = .88) and a third-party predicted negativity index
(α = .83).
Compliment receivers were asked the same positive and
negative questions as in Study 2, which were again com-
bined into an actual positivity index (α = .91) and an actual
negativity index (α = .82).
Results
Data from four dyads in which at least one person did not com-
plete the survey or directions were excluded from analyses.7
Including data from these dyads does not change any of the
results.
Predicted versus actual positive effect of compliment. As pre-
dicted, we again found a significant effect of role (compli-
ment giver vs. compliment receiver) on our positivity index,
b = .71, SE = .20, t(46) = 3.55, p = .001, 95% CI = [0.31,
1.11], with compliment receivers feeling significantly better,
M = 5.16, 95% CI = [4.84, 5.47], than compliment givers
thought they would, M = 4.45, 95% CI = [4.13, 4.76]. Fur-
thermore, descriptively the CI of the third-party predictors’
mean positivity index rating, M = 5.26, 95% CI = [4.99,
5.53], does not overlap with the CI of the compliment giv-
ers’ mean positivity index rating. However, the CI of the

834 Personality and Social Psychology Bulletin 47(5)
third-party predictors’ mean positivity index rating does
overlap with the compliment receivers’ mean positivity
index rating. These results suggest that third-party predictors
are better able to forecast the positive effect a compliment
will have on its recipient, compared with the person giving
the compliment (see Figure 3).
Predicted versus actual negative effect of compliment. There
was a significant effect of role (compliment giver vs. compli-
ment receiver) on our negativity index, b = −1.93, SE = .23,
t(92) = −8.34, p < .001, 95% CI = [−2.39, −1.47], with
compliment receivers feeling significantly less negatively,
M = 1.96, 95% CI = [1.63, 2.28], than compliment givers
anticipated, M = 3.88, 95% CI = [3.56, 4.21]. Furthermore,
descriptively the CI of the third-party predictors’ mean nega-
tivity index rating, M = 2.33, 95% CI = [2.06, 2.59], does
not overlap with the CI of the compliment givers’ mean neg-
ativity index rating. However, the CI of the third-party pre-
dictors’ mean negativity index rating does overlap with the
compliment receivers’ mean negativity index rating. Consid-
ered together, this suggests that compliment givers’ forecasts
about how negatively their compliment will impact its recip-
ient are less accurate than are those of third-party predictors
(see Figure 3).
Benefits to compliment givers. A paired t-test revealed that
compliment givers felt significantly better after compliment-
ing someone, M = 5.26, 95% CI = [4.84, 5.68], than they
had beforehand, M = 4.13, 95% CI = [3.71, 4.55]; t(45) =
−4.36, p < .001, 95% CI = [−1.65, −0.61] (see Figure 4),
suggesting that there may be benefits not only of receiving a
compliment but also of giving one.
Causes of underestimating how good a compliment makes its
recipient feel. We hypothesized that compliment givers’ pre-
dictions about how their compliment would make its recipi-
ent feel are inaccurate for two primary reasons (a) people
feel anxious about delivering a compliment to someone they
do not know and (b) people doubt their competence in giving
a compliment to someone they do not know.
Anxiety. We first tested whether anxious feelings were a better
predictor of people’s forecasts about how the compliment
recipient would feel for those assigned to the compliment giver
condition than those assigned to the third-party predictor con-
dition, that is, do people who are going to give a compliment
feel more anxious than third-party predictors, and is their anxi-
ety significantly correlated with the effect they expect their
compliment to have on its recipient? Our analysis revealed that
compliment givers, M = 4.17, 95% CI = [3.71, 4.64], felt
significantly more anxious than did third-party predictors,
M = 2.86, 95% CI = [2.41, 3.31]; b = −1.31, SE = .32,
t(94) = −4.05, p < .001, 95% CI = [−1.96, −0.67]. Moreover,
as predicted, compliment givers’ feelings of anxiety signifi-
cantly predicted their forecasts of how negatively the stranger
would be impacted by their interaction, b = .29, SE = .09,
t(44) = 3.16, p = .003, 95% CI = [0.11, 0.48], and a mediation
model revealed a significant indirect effect of anxiety on their
forecasts (p < .01). Third-party predictors’ feelings of anxiety,
on the contrary, were not significantly related to their forecasts
of how negatively the stranger would feel as a result of the
interaction, p = .10. Neither compliment givers’ (p = .22), nor
third-party predictors’ (p = .93), feelings of anxiety were sig-
nificantly related to their beliefs about how positively the
stranger would feel as a result of the interaction.
Figure 3. Results of Study 4: Means of actual and perceived negative and positive effects of compliment. Error bars show the 95%
confidence intervals around the means.

Boothby and Bohns 835
Competence. In addition, as predicted, our pre-registered
analysis testing compliment givers’ beliefs about how compe-
tently they would be able to deliver their compliment were
significantly correlated with how negatively, b = −.39,
SE = .10, t(44) = −3.97, p < .001, 95% CI = [−0.58, −0.19],
and how positively, b = .36, SE = .09, t(44) = 3.80, p < .001,
95% CI = [0.17, 0.54], they thought the stranger would feel.
The less competently participants believed they would be
able to deliver their compliment, the less positively and the
more negatively they believed the recipient of their compli-
ment would feel as a result of receiving their compliment.
Likelihood of giving a compliment to a stranger. We also pre-
dicted that people would be more inclined to compliment a
stranger of their own accord after having given a compli-
ment as a participant in our study compared with before-
hand, because they would be more cognizant of the fact that
by doing so they are not actually as bothersome to the
stranger they compliment as they originally believed.
Indeed, a pre-registered paired t-test revealed that compli-
ment givers reported being significantly more likely to com-
pliment a stranger in the future after participating in our
study, M = 4.59, 95% CI = [4.08, 5.10], than they were
prior to participating, M = 3.00, 95% CI = [2.49, 3.51];
t(45) = 6.91, p < .001, 95% CI = [1.12, 2.05] (see Figure 4).
Exploratory analyses revealed that participants’ reported
likelihood of complimenting a stranger in their everyday life
(apart from their participation in this study) prior to giving
their compliment was significantly negatively correlated
with the extent to which they believed their compliment
would impact its recipient negatively, b = −.67, SE = .22,
t(45) = −3.06, p = .004, 95% CI = [−1.12, −0.23], that is,
to the extent people believed their compliment would not
pose an annoyance to its recipient or cause them discomfort,
they indicated being more likely to compliment a stranger.
Although participants reported being more likely to compli-
ment a stranger to the extent they believed their compli-
ment would affect its recipient positively, this effect was
only marginally significant, b = .45, SE = .25, t(44) = 1.85,
p = .07, 95% CI = [−0.04, 0.95].
Furthermore, an exploratory analysis revealed that the
extent to which compliment givers’ likelihood of giving a
stranger a compliment increased after giving a compliment
in the study compared with beforehand was predicted by the
improvement in their mood from pre-interaction to post-
interaction, b = .31, SE = .13, t(45) = −2.50, p = .016.
Consistent with past research showing that positive moods
promote sociality (Isen & Levin, 1972), compliment givers’
mood change may have contributed to their greater willing-
ness to give compliments after participating in our study.
Discussion
Just as in Studies 1 to 3, participants in Study 4 misfore-
casted the impact their compliment would have on its recipi-
ent. Building on the previous studies, Study 4 additionally
provided reasons why people tend to be overly pessimistic
about the effect their compliment will have on its recipient.
First, compliment givers felt anxious prior to giving a com-
pliment, and their anxiety predicted how negatively they
thought their compliment would affect its recipient. Second,
compliment givers were concerned about their own compe-
tence in giving a compliment to a stranger. Both of these con-
cerns predicted their expectations about how their compliment
would make its recipient feel. Importantly, estimates made
by third-party predictors, who by and large did not report
feeling anxious, as they were not about to interact with a
stranger, were significantly less pessimistic than those of
compliment givers. Third-party predictors expected the
receiver of a compliment to feel significantly better (more
flattered, pleased, and good) and significantly less bad
(uncomfortable, annoyed, and bothered) than did people who
were about to give someone a compliment, that is, third-
party predictors’ estimates of the effect a compliment would
have on a stranger were accurate, reflecting the way the
receivers of compliments actually felt as a result of being
Figure 4. Results of Study 4: Compliment givers’ mood and likelihood of giving a stranger a compliment before and after interaction.
Error bars show the 95% confidence intervals around the means.

836 Personality and Social Psychology Bulletin 47(5)
complimented. And recipients of compliments felt rather
good—indeed, one compliment receiver wrote on their sur-
vey, “Thanks for making my day more human!”
It is possible that compliment givers’ forecasts about the
effect their compliment would have on a stranger were overly
pessimistic in part because they did not make their estimates
about a specific person (it would simply be the fourth same-
gender person they saw once they arrived at their designated
location). Perhaps they imagined someone who would reject
them or respond particularly poorly to their compliment.
However, it is worth noting that third-party predictors made
their estimates under the same conditions, that is, third-party
predictors also did not have a specific person in mind when
making their forecasts. Nevertheless, their forecasts were
accurate. It is therefore unlikely that there is simply some-
thing about making a forecast about an unknown other that
result in overly pessimistic beliefs. The only difference
between compliment givers and third-party predictors is that
the former group would be giving a stranger a compliment.
And a major difference between these two groups is that
compliment givers felt anxious about giving their compli-
ment and doing so competently, whereas third-party predic-
tors were unencumbered by such feelings and therefore
better able to more accurately foresee how compliment
recipients would feel. We additionally note that the fact that
third-party forecasters’ estimates of the impact compliments
had on recipients were aligned with the reactions of compli-
ment recipients speaks against the likelihood that our find-
ings can be attributed solely to demand effects on the part of
recipients.
Study 4 also demonstrated that there may additionally be
benefits to complimenting strangers for the person giving the
compliment. Compliment givers were in a better mood and
expressed a greater likelihood of giving a compliment to a
stranger in the future after complimenting someone compared
with beforehand. Just as practice talking to strangers improves
people’s initially pessimistic expectations of such conversa-
tions (Sandstrom et al., manuscript under review), it appears
people may be more inclined to give more compliments to
strangers in the future after having done so before, suggesting
a way we might be able to increase the number of compli-
ments people are willing to give in their everyday lives.
General Discussion
Expressions of admiration, endorsement, and general posi-
tive regard are an important component of social life, in large
part because they satisfy people’s fundamental need to
belong (Baumeister & Leary, 1995). Giving compliments
therefore has significance in everyday social life because
doing so fosters positive self-regard and social acceptance,
which are critical for people’s self-esteem (Leary & Downs,
1995; Leary et al., 1995) and health and well-being
(Baumeister & Leary, 1995; House et al., 1988). However,
compliment giving is an understudied social psychological
phenomenon in contrast to other forms of prosocial behavior
(e.g., prosocial spending, gratitude, helping) which have
received ample attention (for a review, see Dunn et al., 2008;
Helliwell & Aknin, 2018), and so compliment giving is not
well understood. We hope the present research sets the stage
for more work in this area, as compliment giving is a rela-
tively easy and low-cost behavior that can make the world a
friendlier place. Although previous work has to our knowl-
edge studied compliment giving only as a form of strategic
ingratiation (e.g., Jones, 1964), perhaps people would be
more inclined to compliment strangers if they viewed it as an
act of kindness rather than an instrumental act designed to
curry favor with someone.
People in our studies systematically underestimated the
value of their compliment to its recipient, and this reduced
their likelihood of giving compliments (Study 1b). Not only
did people underestimate how positive their compliment
would make someone feel (Studies 1–4), they also overesti-
mated how bothered and annoyed the person would feel as a
result of being approached (Studies 3 and 4). Even after giv-
ing their compliment, people failed to adequately update
their beliefs about the effect they had on the person they
complimented (Studies 2 and 3). Compliment givers’ fore-
casts were inaccurate in part because they felt anxious and
concerned about their ability to compliment a stranger com-
petently (Study 4). In Study 4, compliment givers felt signifi-
cantly more anxious than did third-party predictors, and
compliment givers’ feelings of anxiety were related to their
pessimistic beliefs about the effect their compliment would
have on its recipient, whereas third-party predictors’ fore-
casts were more accurate and unrelated to their feelings of
anxiety. Despite the difficulty of realizing the true impact
they had on the person they complimented, after compli-
menting a stranger people were in a better mood and more
likely to compliment someone they do not know in the future
(Study 4), likely in part because they realized their compli-
ment did not affect the person quite as negatively as they had
initially anticipated (Study 3).
We all know it feels good to receive a compliment, and
that it feels good to make others feel good. So, why are people
so anxious about giving compliments? Research suggests
several reasons. First, social norms often dictate that strangers
in public places give one another privacy, perhaps acknowl-
edging one another but rarely engaging in full-blown interac-
tion (Goffman, 1963). Violating this social norm may cause
some discomfort, in part because people believe strangers
would prefer not to interact with them at all (Epley &
Schroeder, 2014; Sandstrom et al., manuscript under review).
But people of course do interact with strangers quite regu-
larly, whether they are buying coffee, being introduced to a
new colleague, or meeting someone brand new who may
become a friend or romantic partner—and to be sure, many
such first conversations even start with a compliment.
Second, people fear social judgment, a fear which is often
overblown because people overestimate how harshly others

Boothby and Bohns 837
judge them during social interactions (Savitsky et al., 2001).
When interacting with someone new, people believe their
awkwardness is on display and that people are noticing—and
judging—them for their many flaws and faux pas (Boothby
et al., 2018; Gilovich et al., 2000; Schegloff et al., 1977; Van
Boven et al., 2005). Despite its ubiquity, we know that inter-
acting with new people is a social activity that people find
particularly challenging and stressful (Duronto et al., 2005).
As demonstrated in Study 4, focusing on one’s own inability
to deliver a compliment competently prevented people from
realizing just how positive an impact their compliment actu-
ally had on its recipient. In people’s own minds, they are
stammering and nervous and searching for the right words,
but in the eyes of the recipient of their compliment, they are
simply nice, friendly folks.
Interestingly, when people are not bogged down by their
own anxieties and insecurities, they are able to anticipate the
positive impact of a compliment. In Study 4, forecasters
were able to discern how someone else’s compliment would
make a person feel. Although it is clear from the outside that
a compliment will make someone feel good, this fact is lost
on potential compliment givers, who are blinded by their
own feelings of anxiety and incompetence (DePaulo & Tang,
1994).
It is noteworthy that even in hindsight people continue
to mistakenly believe their compliment had a more nega-
tive impact than it actually did. This may be because social
interactions, especially with new people, are characterized
by politeness norms which prescribe people respond to one
another civilly. Thus, even when someone responds warmly
to a compliment, it may be difficult to tell whether one’s
compliment really had a positive effect or if the recipient is
just being polite (Brown & Levinson, 1987; Tesser &
Rosen, 1975). Thus, the difficulty of taking at face value
someone’s positive reactions to being complimented likely
contributes to people’s continued devaluation of their
compliment.
The results of these studies are noteworthy in part because
they differ from a large, well-established body of research on
self-serving biases showing that people tend to be overly
optimistic about their own outcomes (Sharot, 2011; Sweeny
et al., 2006; Weinstein, 1980). People believe they are less
likely than others to fall victim to diseases such as cancer,
heart attack, and alcoholism (Kirscht et al., 1966; Perloff,
1983; Weinstein, 1987), and that they are less susceptible to
negative life events such as injury due to car accidents and
divorce (Perloff & Fetzer, 1986). However, in stark contrast
to these pervasive optimism biases, and in line with recent
work showing that people are sometimes remarkably pessi-
mistic about their social abilities (e.g., Boothby et al., 2018;
Deri et al., 2017; Srivastava & Beer, 2005; Whillans et al.,
2017), we found that people are unduly pessimistic when it
comes to the effect their compliments have on others. The
present research suggests that this is a domain of social life
in which people are not comfortably buffered by optimism in
their abilities and outcomes.
Limits on Generality
We do not expect people always underestimate how positively
their compliments are received; there are likely to be instances
in which people instead overestimate how positively and
underestimate how negatively their compliments affect recipi-
ents. This is likely to be the case for a specific subset of com-
pliments that are largely of a romantic or sexual nature, as in
the cases of “cat calling” or of inappropriate remarks directed
at work colleagues. In such cases, compliment givers may in
fact overestimate the extent to which their compliments are
welcomed and underestimate the extent to which they make
their targets uncomfortable (e.g., Abbey, 1982; Bohns &
DeVincent, 2019; Farris et al., 2008; Jacques-Tiura et al.,
2007). Such instances are problematic, for sure. However,
they are also more limited in scope and less surprising, in the
light of the extensive literature on overconfidence (Moore &
Healy, 2008), than the findings we report here.
We additionally note that people may overestimate the
value of their compliments to strangers, or be more accurate
in their estimates, in cultures in which there is a stronger
modesty norm, or in contexts when giving a compliment is
not as anxiety-provoking (e.g., complimenting a subordinate,
or a close friend; although see Zhao & Epley, 2020a & Zhao
& Epley, 2020b for an alternative perspective), such as social
media platforms in which people can compliment strangers
anonymously with little risk of anxiety or embarrassment.
That is, we do not claim that undervaluing one’s compliment
is a universal bias; the bias we have documented and repli-
cated several times is robust, at least in some circumstances,
and future research ought to investigate more extensively the
factors that affect people’s beliefs about the effects their
compliments have on others.
Coda
Compliments are an important part of social life, and the
world would be a better place if there were more of them.
Compliments make us feel good, and they can provide an
easy entrée to conversation with someone new. But when
people are anxious about giving compliments and concerned
about their ability to do successfully, people may be inhibited
from giving as many compliments as they might otherwise.
And so, they lamentably forgo opportunities to increase peo-
ple’s well-being, leaving everyone—themselves included—
worse off.
Acknowledgments
The authors thank Meredith Anderer, Ava Barnett, Lauren Ann
DeVincent, David Navadeh, Courtney Noll, Megan Rodriguez,
Kendra Sober, Carlie Stewart, and Charlotte Walden.

838 Personality and Social Psychology Bulletin 47(5)
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) received no financial support for the research, author-
ship, and/or publication of this article.
ORCID iD
Erica J. Boothby https://orcid.org/0000-0001-6171-3563
Supplemental Material
Supplemental material is available online with this article.
Notes
1. In this first study, we also included several additional items
asking about the perceived and actual impact of the interaction
on the compliment receiver. However, as our understanding of
the phenomenon evolved, we no longer included these items in
subsequent studies. These items and results can be found in the
Supplemental Material.
2. All questions for this and all subsequent studies are available in
the Supplemental Material.
3. Two compliment givers failed to complete the survey prior to
giving a compliment, one compliment giver did not actually
give a compliment, and one compliment receiver did not hear
the compliment.
4. The full qualitative data with information about what people
compliment strangers on in their everyday lives are available on
OSF (https://tinyurl.com/ydbrn4b7).
5. We note that we found no differences in the pattern of results for
male and female participants, for this or subsequent studies.
6. One compliment giver had already participated in the study as
a compliment receiver, two participants failed to complete the
survey before giving a compliment, and one participant acciden-
tally complimented someone who did not match their gender.
7. One participant failed to complete the prediction survey prior
to giving a compliment, two participants misunderstood the
instructions and complimented someone who did not match
their gender, and one compliment giver returned to the lab after
giving their compliment with an unsealed envelope from the
compliment receiver.
References
Abbey, A. (1982). Sex differences in attributions for friendly
behavior: Do males misperceive females’ friendliness? Journal
of Personality and Social Psychology, 42, 830–838.
Arkin, R. M., Appelman, A. J., & Burger, J. M. (1980). Social
anxiety, self-presentation, and the self-serving bias in causal
attribution. Journal of Personality and Social Psychology, 38,
23–35.
Aspinwall, L. G., & Taylor, S. E. (1993). Effects of social
comparison direction, threat, and self-esteem on affect, self-
evaluation, and expected success. Journal of Personality and
Social Psychology, 64, 708–722.
Bates, D., Maechler, M., Bolker, B., & Walker, S. (2015). Fitting
linear mixed-effects models using lme4. Journal of Statistical
Software, 67, 1–48.
Baumeister, R. F., & Leary, M. R. (1995). The need to belong:
Desire for interpersonal attachments as a fundamental human
motivation. Psychological Bulletin, 117, 497–529.
Bohns, V. K., & DeVincent, L. A. (2019). Rejecting unwanted
romantic advances is more difficult than suitors realize. Social
Psychological and Personality Science, 10, 1102–1110.
Boothby, E. J., Cooney, G., Sandstrom, G. M., & Clark, M. S.
(2018). The liking gap in conversations: Do people like us
more than we think? Psychological Science, 29, 1742–1756.
Brown, P., & Levinson, S. C. (1987). Politeness: Some universals
in language usage (Vol. 4). Cambridge University Press.
Clore, G. L., Gasper, K., & Garvin, E. (2001). Affect as informa-
tion. In J. P. Forgas (Ed.), Handbook of affect and social
cognition (pp. 121–144). Lawrence Erlbaum Associates
Publishers.
DePaulo, B. M., & Tang, J. (1994). Social anxiety and social judg-
ment: The example of detecting deception. Journal of Research
in Personality, 28, 142–153.
Deri, S., Davidai, S., & Gilovich, T. (2017). Home alone: Why
people believe others’ social lives are richer than their own.
Journal of Personality and Social Psychology, 113, 858–877.
Dunn, E. W., Aknin, L. B., & Norton, M. I. (2008). Spending money
on others promotes happiness. Science, 319, 1687–1688.
Duronto, P. M., Nishida, T., & Nakayama, S. I. (2005).
Uncertainty, anxiety, and avoidance in communication with
strangers. Inter-national Journal of Intercultural Relations,
29, 549–560.
Epley, N., & Schroeder, J. (2014). Mistakenly seeking solitude.
Journal of Experimental Psychology, 143, 1980–1999.
Farris, C., Treat, T. A., Viken, R. J., & McFall, R. M. (2008). Sexual
coercion and the misperception of sexual intent. Clinical
Psychology Review, 28, 48–66.
Faul, F., Erdfelder, E., Lang, A. G., & Buchner, A. (2007). G*
Power 3: A flexible statistical power analysis program for the
social, behavioral, and biomedical sciences. Behavior Research
Methods, 39, 175–191.
Fiske, S. T., Cuddy, A. J., & Glick, P. (2007). Universal dimen-
sions of social cognition: Warmth and competence. Trends in
Cognitive Sciences, 11, 77–83.
Flynn, F. J., & Lake, V. K. B. (2008). If you need help, just ask:
Underestimating compliance with direct requests for help.
Journal of Personality and Social Psychology, 95, 128–143.
Freeman, R. B. (1997). Working for nothing: The supply of volun-
teer labor. Journal of Labor Economics, 15, 140–166.
Frey, B. S., & Meier, S. (2004). Social comparisons and pro-social
behavior: Testing “conditional cooperation” in a field experi-
ment. American Economic Review, 94, 1717–1722.
Gilovich, T., Medvec, V. H., & Savitsky, K. (2000). The spotlight
effect in social judgment: An egocentric bias in estimates of
the salience of one’s own actions and appearance. Journal of
Personality and Social Psychology, 78, 211–222.
Gilovich, T., Savitsky, K., & Medvec, V. H. (1998). The illusion
of transparency: Biased assessments of others’ ability to read
one’s emotional states. Journal of Personality and Social
Psychology, 75, 332–346.
Goffman, E. (1963). Behavior in public places. Free Press.

Boothby and Bohns 839
Helliwell, J. F., & Aknin, L. B. (2018). Expanding the social sci-
ence of happiness. Nature Human Behavior, 2, 248–252.
House, J. S., Landis, K. R., & Umberson, D. (1988). Social relation-
ships and health. Science, 241, 540–545.
Isen, A. M., & Levin, P. F. (1972). Effect of feeling good on help-
ing: Cookies and kindness. Journal of Personality and Social
Psychology, 21(3), 384–388.
Jacques-Tiura, A. J., Abbey, A., Parkhill, M. R., & Zawacki, T.
(2007). Why do some men misperceive women’s sexual inten-
tions more frequently than others do? An application of the
confluence model. Personality and Social Psychology Bulletin,
33, 1467–1480.
Jones, E. E. (1964). Integration: A social-psychological analysis.
Appleton-Century-Crofts.
Jones, E. E., & Wortman, C. (1973). Integration: An attributional
approach. General Learning Corp.
Kirscht, J. P., Haefner, D. P., Kegeles, S. S., & Rosenstock, I. M.
(1966). A national study of health beliefs. Journal of Health
and Human Behavior, 7, 248–254.
Kumar, A., & Epley, N. (2018). Undervaluing gratitude: Expressers
misunderstand the consequences of showing appreciation.
Psychological Science, 29, 1423–1435.
Kupor, D., Flynn, F., & Norton, M. I. (2017). Half a gift is not half-
hearted: A giver–receiver asymmetry in the thoughtfulness of
partial gifts. Personality and Social Psychology Bulletin, 43,
1686–1695.
Kuznetsova, A., Brockhoff, P. B., & Christensen, R. H. B. (2014).
lmerTest: Tests for random and fixed effects for linear mixed
effect models (lmer objects of lme4 package). R package ver-
sion 2.0-6. R-project.org/package=lmerTest.
Leary, M. R., & Downs, D. L. (1995). Interpersonal functions of the
self-esteem motive. In M. H. Kernis (Eds.), Efficacy, agency,
and self-esteem (pp. 123–144). Springer.
Leary, M. R., Tambor, E. S., Terdal, S. K., & Downs, D. L. (1995).
Self-esteem as an interpersonal monitor: The sociometer
hypothesis. Journal of Personality and Social Psychology, 68,
518–530.
McCroskey, J. C. (1992). Reliability and validity of the willingness
to communicate scale. Communication Quarterly, 40, 16–25.
McCroskey, J. C., & Richmond, V. P. (1990). Willingness to com-
municate: A cognitive view. Journal of Social Behavior and
Personality, 5, 19–37.
McIntyre, C. W., Watson, D., Clark, L. A., & Cross, S. A. (1991).
The effect of induced social interaction on positive and nega-
tive affect. Bulletin of the Psychonomic Society, 29, 67–70.
Mendes, W. B., Blascovich, J., Major, B., & Seery, M. (2001).
Challenge and threat responses during downward and upward
social comparisons. European Journal of Social Psychology,
31(5), 477–497.
Mezulis, A. H., Abramson, L. Y., Hyde, J. S., & Hankin, B. L.
(2004). Is there a universal positivity bias in attributions? A
meta-analytic review of individual, developmental, and cultural
differences in the self-serving attributional bias. Psychological
Bulletin, 130, 711–747.
Moore, D. A., & Healy, P. J. (2008). The trouble with overconfi-
dence. Psychological Review, 115(2), 502–517.
Perloff, L. S. (1983). Perceptions of vulnerability to victimization.
Journal of Social Issues, 39, 41–61.
Perloff, L. S., & Fetzer, B. K. (1986). Self–other judgments and
perceived vulnerability to victimization. Journal of Personality
and Social Psychology, 50, 502–510.
Roghanizad, M. M., & Bohns, V. K. (2017). Ask in person:
You’re less persuasive than you think over email. Journal of
Experimental Social Psychology, 69, 223–226.
Ross, L., Greene, D., & House, P. (1977). The “false consensus
effect”: An egocentric bias in social perception and attribution
processes. Journal of Experimental Social Psychology, 13,
279–301.
Sandstrom, G. M. & Boothby, E. J. (manuscript accepted). Why
do people avoid talking to strangers? A mini meta-analysis of
predictions and experiences. Self and Identity.
Sandstrom, G. M., Boothby, E. J., & Cooney, G. Talking to strang-
ers: A week-long intervention reduces barriers to social con-
nection. [Unpublished manuscript]
Sandstrom, G. M., & Dunn, E. W. (2014a). Social interactions and
well-being: The surprising power of weak ties. Personality and
Social Psychology Bulletin, 40, 910–922.
Sandstrom, G. M., & Dunn, E. W. (2014b). Is efficiency over-
rated? Minimal social interactions lead to belonging and posi-
tive affect. Social Psychological and Personality Science, 5,
437–442.
Savitsky, K., Epley, N., & Gilovich, T. (2001). Do others judge us
as harshly as we think? Overestimating the impact of our fail-
ures, shortcomings, and mishaps. Journal of Personality and
Social Psychology, 81, 44–56.
Savitsky, K., & Gilovich, T. (2003). The illusion of transparency
and the alleviation of speech anxiety. Journal of Experimental
Social Psychology, 39, 618–625.
Schegloff, E. A., Jefferson, G., & Sacks, H. (1977). The preference
for self-correction in the organization of repair in conversation.
Language, 53, 361–382.
Sharot, T. (2011). The optimism bias. Current Biology, 21(23),
R941–R945.
Shelton, J. N., & Richeson, J. A. (2005). Intergroup contact and
pluralistic ignorance. Journal of Personality and Social
Psychology, 88, 91–107.
Sweeny, K., Carroll, P. J., & Shepperd, J. A. (2006). Is optimism
always best? Future outlooks and preparedness. Current
Directions in Psychological Science, 15, 302–306.
Tesser, A., & Rosen, S. (1975). The reluctance to transmit bad
news. In L. Berkowitz (Ed.), Advances in experimental social
psychology (Vol. 8, pp. 193–232). Academic Press
Thoits, P. A., & Hewitt, L. N. (2001). Volunteer work and well-
being. Journal of Health and Social Behavior, 42, 115–131.
Van Boven, L., & Loewenstein, G. (2003). Social projection of
transient drive states. Personality and Social Psychology
Bulletin, 29, 1159–1168.
Van Boven, L., Loewenstein, G., & Dunning, D. (2005). The
illusion of courage in social predictions: Underestimating
the impact of fear of embarrassment on other people.
Organizational Behavior and Human Decision Processes,
96(2), 130–141.
Vittengl, J. R., & Holt, C. S. (2000). Getting acquainted: The
relationship of self-disclosure and social attraction to posi-
tive affect. Journal of Social and Personal Relationships, 17,
53–66.

840 Personality and Social Psychology Bulletin 47(5)
Weinstein, N., & Ryan, R. M. (2010). When helping helps:
Autonomous motivation for prosocial behavior and its influ-
ence on well-being for the helper and recipient. Journal of
Personality and Social Psychology, 98, 222–224.
Weinstein, N. D. (1980). Unrealistic optimism about future life
events. Journal of Personality and Social Psychology, 39,
806–820.
Weinstein, N. D. (1987). Unrealistic optimism about susceptibility
to health problems: Conclusions from a community-wide
sample. Journal of Behavioral Medicine, 10, 481–500.
Whillans, A. V., Christie, C. D., Cheung, S., Jordan, A. H., &
Chen, F. S. (2017). From misperception to social connection:
Correlates and consequences of overestimating others’ social
connectedness. Personality and Social Psychology Bulletin,
43(12), 1696–1711.
Wills, T. A. (1981). Downward comparison principles in social
psychology. Psychological Bulletin, 90, 245–271.
Wojciszke, B. (1994). Multiple meanings of behavior: Construing
actions in terms of competence or morality. Journal of
Personality and Social Psychology, 67, 222–232.
Zaki, J., & Mitchell, J. P. (2011). Equitable decision making is
associated with neural markers of intrinsic value. Proceedings
of the National Academy of Sciences, 108, 19761–19766.
Zhao, X., & Epley, N. (2020a). Kind words do not become tired
words: Undervaluing the positive impact of frequent compli-
ments. Self and Identity, 1–22. https://doi.org/10.1080/15298
868.2020.1761438
Zhao, X., & Epley, N. (2020b). Insufficiently complimentary?
Underestimating the positive impact of compliments cre-
ates a barrier to expressing them [Unpublished manuscript].
University of Chicago.