Content uploaded by Peggy J Liu
Author content
All content in this area was uploaded by Peggy J Liu on Jul 13, 2022
Content may be subject to copyright.
THE SURPRISE OF REACHING OUT
1
In press, Journal of Personality and Social Psychology
The Surprise of Reaching Out: Appreciated More than We Think
Peggy J. Liu
University of Pittsburgh Katz Graduate School of Business
SoYon Rim
William Paterson University of New Jersey
Lauren Min
University of Kansas School of Business
Kate E. Min
Wheaton College
Correspondence concerning this article can be addressed to Peggy J. Liu, 3950 Roberto & Vera
Clemente Drive, Pittsburgh, PA 15260. Contact: peggy.liu@pitt.edu.
Author Note: We gratefully acknowledge financial funding for this research from the University
of Pittsburgh Joseph M. Katz Graduate School of Business, including through the Ben L. Fryrear
Faculty Fellowship and Chair, and from the University of Kansas School of Business. We also
thank Jessica Stenvig for data collection assistance with experiment 2; Vy Duong for data
collection assistance with experiment 3; the Debra Paget and Jeffrey Berg Business Simulation
Laboratory at Cornell University for data collection assistance with experiment 4b; Guadalupe
Arriola for data collection assistance with experiment 5a; and Josh Williams and the Katz
Business Research Center for data collection assistance with supplemental study S4.
All pre-registrations, survey materials, data, and syntax are available at ResearchBox
(https://researchbox.org/289).
© American Psychological Association, 2022. This paper is not the copy of record and may
not exactly replicate the authoritative document published in the APA journal. The final
article is available, upon publication, at: https://doi.org/10.1037/pspi0000402.
THE SURPRISE OF REACHING OUT
2
Abstract
People are fundamentally social beings and enjoy connecting with others, sometimes reaching
out to others—whether simply to say hello and to check in on how others are doing with a brief
message, or to send a small gift to show that one is thinking of the other person. Yet despite the
importance and enjoyment of social connection, do people accurately understand how much
other people value being reached out to by someone in their social circle? Across a series of pre-
registered experiments, we document a robust underestimation of how much other people
appreciate being reached out to. We find evidence compatible with an account wherein one
reason this underestimation of appreciation occurs is because responders (vs. initiators) are more
focused on their feelings of surprise at being reached out to; such a focus on feelings of surprise
in turn predicts greater appreciation. We further identify process-consistent moderators of the
underestimation of reach-out appreciation, finding that it is magnified when the reach-out
context is more surprising: when it occurs within a surprising (vs. unsurprising) context for the
recipient and when it occurs between more socially distant (vs. socially close) others. Altogether,
this research thus identifies when and why we underestimate how much other people appreciate
us reaching out to them, implicating a heightened focus on feelings of surprise as one underlying
explanation.
Keywords: appreciation, social relationships, surprise, gifts, prediction
THE SURPRISE OF REACHING OUT
3
The Surprise of Reaching Out: Appreciated More than We Think
Maintaining connections in one’s social circle is essential to well-being (Baumeister &
Leary, 1995; Cohen & Wills, 1985; Sandstrom & Dunn, 2014a, 2014b). Sometimes, life gives
people opportunities to interact with one another without much effort by either party. Work
colleagues may run into each other at the office coffee room, neighbors may pass each other as
they are walking around the neighborhood, or friends may spontaneously meet at their local café.
These brief interactions with a wide range of relationship partners are associated with positive
psychological outcomes, including increased social and emotional well-being (Sandstrom &
Dunn, 2014a, 2014b), protection from stressful events (Cohen & Wills, 1985) and anxiety
(Gump & Kulik, 1997), personal growth (Lee et al., 2018) and increased cognitive functioning
(Granovetter, 1973; Lee et al., 2018). However, beyond opportune moments of social connection
which are vital for well-being, people are often left disconnected from friends and acquaintances
with whom they typically have pleasant interactions, if neither party takes the initiative to
connect.
Although connecting with friends and acquaintances is typically a pleasant experience
(Baumeister & Leary, 1995), people often find themselves out of touch with such people in their
lives. In many parts of the United States, for example, many people’s lives are set up for
isolation rather than spontaneous social interactions—whether at the workplace, in the
neighborhood, or at a café—and therefore, there is a dearth of opportunities to maintain ties with
those we are already acquainted with in various settings (Putnam, 2000). As reflected in a steady
decline of social interactions in society (McPherson et al., 2006; Putnam, 2000), there may exist
various costs to taking the initiative to reach out to a friend or an acquaintance that outweigh the
THE SURPRISE OF REACHING OUT
4
benefits. If people perceive both the costs and the benefits accurately, then the resulting cost-
benefit analysis that they perform is appropriate and correct.
However, in this work, we examine whether people accurately perceive the benefits that
would accrue to others from their reaching out. We define “reaching out” broadly to involve a
minimum criterion consisting of a gesture to check-in with someone to show that one is thinking
about them, for instance by sending a short message (e.g., to say hi, to say “I’m thinking of you,”
to say “I hope you are well”) or a small gift. By this minimum criterion definition, more specific
instantiations of social exchange behavior that have been examined in prior research would
qualify as reach-outs but also involve additional elements that are not necessary to qualify as a
reach-out. For instance, contacting someone to ask for a favor (e.g., Zhang & Epley, 2009)
qualifies as a reach-out but also adds an element of asking for help that is not necessary to
qualify as a reach-out. Likewise, contacting someone to offer a compliment (e.g., as in Boothby
& Bohns, 2021; Zhao & Epley, 2021a, 2021b) qualifies as a reach-out but also adds an element
of saying something positive about a recipient’s traits or behaviors that is not necessary to
qualify as a reach-out. As a final example, contacting someone to express gratitude (e.g., as in
Grant & Gino, 2010) qualifies as a reach-out but also adds an element of thanking someone for
some positive behavior they have done for oneself or others in the past—which is not necessary
to qualify as a reach-out. Altogether then, in the present research, we offer a broad definition of
“reaching out” (one which is also consistent with its colloquial usage to refer to checking in on
someone; e.g., Pennell, 2021 on today.com) based on a minimum threshold. We then investigate
it empirically through testing reach-outs that meet this minimum threshold, but do not
necessarily exceed it with “added elements” (i.e., reach-outs can—but do not have to—include
help requests, compliments, or gratitude expressions).
THE SURPRISE OF REACHING OUT
5
We propose that people underestimate how much others appreciate being reached out to
by others. Examining whether people underestimate how much others value their attempts to
reach out is important, because underestimation may deter people from reaching out as often as
would benefit themselves and others. For instance, imagine that you thought of a friend this
morning while stopping by at a café you have been to together. You have not interacted with this
friend in a while. How much do you think your friend would appreciate it if you reached out and
sent a brief text message or even a small gift? Does your prediction match your friend’s actual
appreciation? We propose that you would likely underestimate how much your friend would
appreciate being reached out to.
Next, we discuss the theoretical framework for our predictions. We first discuss prior
research on mispredictions as a function of differences in perspective across a variety of
domains, ranging from the prosocial behavior and gratitude domains to the gift-giving domain.
We then discuss the rationale for our focus on responder appreciation of being reached out to—
including why accurately estimating responder appreciation is important. Finally, we discuss the
underlying role of focusing on responder surprise as one potential explanation for the
underestimation of reach-out appreciation.
Theoretical Framework
Mispredictions as a Function of Perspective Differences
Our prediction draws on and contributes to an extensive literature documenting the
difficulty of engaging in accurate perspective-taking and relatedly, the powerful effect of
perspective differences across various domains. This literature shows that people tend to have
difficulty making accurate judgments of others’ perspectives on a social situation in part due to
egocentric biases in perspective-taking (Epley et al., 2004; Kardas, Kumar, & Epley, 2022; Liu
THE SURPRISE OF REACHING OUT
6
& Kwon, 2022; Liu & Min, 2020; Ross & Sicoly, 1979). For instance, following social
interactions, people tend to be overly focused on their own actions (e.g., conversation
performance) and overestimate the salience of these actions to others (Gilovich et al., 2000),
leading to a systematic underestimation of how much their conversation partners liked them and
enjoyed their company (Boothby et al., 2018). They often focus on their own internal
monologues, which can be self-critical and negative after social interactions; however, others do
not perceive them with such a critical or negative lens (Boothby et al., 2018; Savitsky et al.,
2001; Savitsky & Gilovich, 2003). Additionally, people also underestimate how much others,
especially strangers, are interested in their intimate disclosures in conversations, which serves as
a barrier to deeper—yet more socially satisfying—conversations (Kardas et al., 2022).
In the prosocial behavior literature, perspective differences often occur as well. Potential
helpers and help-seekers are often misaligned in their understanding of each other’s thoughts and
behaviors, suggesting that people often act suboptimally when help is needed. Help-seekers
underestimate the extent to which potential helpers will comply with their direct requests for help
because help-seekers and potential helpers have a differential focus on the costs incurred by the
potential helper. Compared to the potential helper, help-seekers are more attentive to the
instrumental costs of accepting a request for help, and less attentive to the social costs of
rejecting a request for help (Flynn & Lake, 2008). By mis-predicting how much a potential
helper is willing to comply with their request, people often forego opportunities to request (and
receive) assistance. Help-seekers also tend to underestimate the amount of effort put forth by
potential helpers because they overlook the discomfort (e.g., guilt) that potential helpers
experience from being unhelpful (Newark et al., 2017). Just as help-seekers underestimate
THE SURPRISE OF REACHING OUT
7
compliance to requests for help, they also underestimate the quality of such help and by doing so,
they forego opportunities to request assistance.
On the flip side, potential helpers are also often inaccurate in predicting help-seekers’
willingness to ask for help: they overestimate help-seeking willingness because they overlook the
role of embarrassment in discouraging people from requesting help (Bohns & Flynn, 2010).
Potential helpers rely to some extent on their own perspective when judging how help-seekers
are thinking (Epley et al., 2004, 2006) and do not recognize that in situations involving exposure
of vulnerabilities and embarrassment, help-seekers often want to “save face” (Bohns & Flynn,
2010). This inaccuracy in perceiving others’ thoughts not only leads to missed opportunities for
helping others, but also leads to inefficient social interactions involving incorrect inferences that
people do not need help because they do not ask for help.
Once help is received, research on gratitude expressions to thank others for their prior
help has shown that people may not always accurately judge how much others value their
expressions of gratitude (Kumar & Epley, 2018; Liu et al., 2015). For instance, people
expressing gratitude underestimated how positive and surprising an expression of gratitude
would be for a receiver and overestimated how awkward it would be (Kumar & Epley, 2018).
Relatedly, research on compliments has shown that people offering compliments—which say
something positive about a recipient’s traits or behavior—underestimated the positive impact of
their compliments on recipients (Boothby & Bohns, 2021; Zhao & Epley, 2021a, 2021b). Such
misestimations are also explained by expressors’ failure to sufficiently adjust from their own
perspective to the recipient’s perspective (Epley et al., 2004, 2006).
Finally, research on gift-giving further demonstrates judgment discrepancies as a function
of a person’s perspective as either a giver or a recipient. The domain of gift-giving is relevant to
THE SURPRISE OF REACHING OUT
8
the present work because one way in which a reach-out can occur is through the giving of a gift.
However, in previous work, the focus has been on documenting and explaining discrepancies
between the types of gifts that givers prefer to give and that recipients prefer to receive (Baskin
et al., 2014; Cavanaugh et al., 2015; Galak et al., 2016; Gino & Flynn, 2011; Givi & Galak,
2022; Kupor et al., 2017; Liu, Dallas, et al., 2019; Liu & Baskin, 2021; Yang & Urminsky,
2018). For instance, givers tend to be more focused on what would make the recipient have the
most positive emotional reaction at the moment of gift exchange, whereas recipients are more
focused on what would be more satisfying to them overall beyond just the moment of gift
exchange (Yang & Urminsky, 2018). Other work documents various other mispredictions of
appreciation for different types of gifts, such as the tendency to underestimate appreciation of
requested gifts (Gino & Flynn, 2011) and the tendency to overestimate how much distant others
appreciate socially responsible gifts (Cavanaugh et al., 2015). By contrast, in the present work,
we focus on appreciation of the reach-out gesture itself. In other words, rather than focusing on
whether people accurately predict which type of gift is appreciated more, we examine more
generally whether people accurately predict how much others appreciate being reached out to in
the first place. We also examine the generalizability of the underestimation of appreciation of
reaching out across different reach-out modes (e.g., a message or a gift).
Mispredicting Appreciation for Reaching Out and a Differential Focus on Responder Surprise
We propose that people considering reaching out (“initiators”) underestimate how much
other people (“responders”) will appreciate their gesture. Note that we are not suggesting that
initiators are entirely unaware of responders’ appreciation. Most people understand that gestures
to socially connect are valued. Indeed, nearly all reach-outs are likely to be in the positive
THE SURPRISE OF REACHING OUT
9
domain of appreciation.
1
Instead, our prediction regarding underestimation is that initiators do
not understand the full extent to which their reach outs are appreciated. Understanding the full
extent to which their reach outs are appreciated is important because it would likely contribute to
people initiating social contact, to the benefit of themselves and others. Furthermore, examining
responder appreciation of reaching out is an important outcome because appreciation is an
important signal of relationship development (Gordon et al., 2012), and thus how much an
exchange is appreciated is also a focal outcome shared by prior work on gifting (Cavanaugh et
al., 2015; Flynn & Adams, 2009; Gino & Flynn, 2011; Liu, Lamberton, & Haws, 2015) and by
prior work on prosocial behaviors (Converse & Fishbach, 2012).
Why might people mis-predict the extent to which a friend would appreciate them simply
reaching out? Our prediction relies on prior theorizing on self-other discrepancies, including the
work discussed in the prior section, wherein people bring their own egocentric perspectives to
bear when predicting others’ mental states (Epley et al., 2004). In particular, we predict that one
reason this mismatch occurs is because initiators are less focused on the responder’s positive
feelings of surprise when predicting responder appreciation than responders are actually focused
on their own feelings of positive surprise. Specifically, by virtue of their differing perspectives,
we suggest that responders may be more focused on their own feelings of positive surprise when
they are reached out to, both because the unexpectedness of the event is especially salient for
them and also because they are especially attuned to cues to others’ warmth (Abele & Wojciszke,
2007; Zhao & Epley, 2021a). By contrast, initiators may be less focused on the responder’s
feelings of surprise both because the unexpectedness of the event is less salient for them, given
1
In fact, if we compare all participants’ appreciation ratings to the appreciation scale mid-point across studies, we
find that nearly all participants across our studies—whether in the initiator condition or in the responder condition—
indicate appreciation ratings greater than the midpoint.
THE SURPRISE OF REACHING OUT
10
that they themselves have not experienced it, and because they are more likely attuned to
alternative cues such as those of the competence of their reach-outs (Abele & Wojciszke, 2007;
Zhao & Epley, 2021a).
How might focusing on feelings of positive surprise shape the appreciation of being
reached out to by a friend or an acquaintance? Early work on surprise conceptualized it as one of
the basic emotions, along with happiness, sadness, anger, fear, and disgust (Ekman, 1982;
Ekman et al., 1972). Later work conceptualized surprise as a cognitive process that precedes an
emotional experience (Fiske, 1992; Teigen & Keren, 2003), detecting whether an event is
discrepant with one’s existing schemas and thus triggering motivational resources to analyze the
event (Meyer et al., 1997; Reisenzein et al., 2019). Of note, surprise can make a positive event
more positive or can make a negative event more negative. According to decision affect theory
(Mellers et al., 1997), when making decisions, people consider not only subjective expected
utility, but subjective expected emotions. Surprise amplifies the pleasantness or unpleasantness
of an experience, such that an unexpected gain of a smaller amount can feel more pleasurable
than an expected gain of a larger amount, and an unexpected loss of a smaller amount can feel
worse than an expected loss of a larger amount. For instance, participants felt more pleasure after
an unexpected win of $5.40 than an expected win of $9.70, and they felt worse after an
unexpected loss of $17.50 than after an expected loss of $31.50 (Mellers et al., 1997).
Subsequent work corroborated the finding that surprise enhances pleasurable experiences
(Oliver, 1997; Valenzuela et al., 2010). For example, participants felt more pleased, happy, and
excited after receiving a small, unexpected gift for participating in a study than when they
received no gift or the same gift but under circumstances where the gift was expected
(Valenzuela et al., 2010). In a related line of work, Kurtz, Wilson, and Gilbert (2007) found that
THE SURPRISE OF REACHING OUT
11
people who received a gift after a period of uncertainty about which one of two they would
receive were in a positive mood for longer than those who received a gift without a period of
uncertainty. In essence, the “surprise” of the gift increased the pleasure of the experience.
In the context of social exchange, surprise gifts may be positive when they signal warmth
and care on the part of the giver (cues to which responders may be especially attuned; Abele &
Wojciszke, 2007; Zhao & Epley, 2021a). The positive surprise elicited by such gifts can have a
positive impact on the relationship by strengthening and affirming it (Ruth et al., 1999). On the
other hand, surprise gifts can be negative when they violate relational appropriateness because
they are considered overkill for the current state of the relationship (e.g., an expensive bouquet
for someone one does not know very well) or they are insulting given the relationship (e.g., a
used frying pan as a wedding gift for close friends), and these negative surprise gifts could
damage the relationship (Ruth et al., 1999). In our research, we are specifically interested in the
context of reach-outs that we suggest would be considered positive surprises among
acquaintances. Neither a brief message nor a small gift—as we examine—are likely to be
perceived as a bad or an uncomfortable surprise (see also, footnote 1). Accordingly, a greater
focus on the responder’s positive surprise should amplify the focal outcome of appreciation,
which is an especially relevant outcome for positive (not negative) surprises.
Drawing on these prior findings, we suggest that being reached out to in a positive way
by someone with whom one has had positive interactions in the past is likely to be a pleasant
experience bolstering feelings of appreciation in the responder. When the reach-out occurs
unexpectedly, the element of positive surprise for the responder likely leads to an amplification
of the positive experience (Mellers et al., 1997; Oliver, 1997; Valenzuela et al., 2010) and thus, a
corresponding boost in feelings of appreciation, a fact not fully accounted for by initiators.
THE SURPRISE OF REACHING OUT
12
Therefore, our prediction is that initiators underestimate the extent to which responders will
appreciate their reach-out because initiators are less focused on the positive surprise that
responders would feel in such a situation.
The Present Research
Table 1 contains an overview of the thirteen pre-registered studies (nine main studies and
four supplemental studies). Experiments 1 to 4 examine our focal prediction that initiators will
underestimate the extent to which their reach out is appreciated by a responder. Experiment 1
tests this focal prediction using a recall paradigm in which participants considered actual
experiences in which they had reached out to someone (vs. had been reached out to). Experiment
2 is a field experiment on a college campus, involving actual reach-out experiences conducted
with dyads, wherein people wrote notes checking in on a college classmate with whom they had
not had contact in some time. Experiment 3 is another field experiment on a college campus, also
involving actual reach-out experiences conducted with dyads, wherein people wrote notes and
sent small gifts to college classmates with whom they had not had contact in some time.
Experiments 4a-4b then further test this focal prediction using a scenario paradigm, thereby
allowing us to control reach-out content and addressing other limitations of the recall and dyadic
paradigms.
Experiments 5 to 7 then examine the proposed role of a differential focus on the
responder’s feelings of surprise on being reached out to. Experiments 5a-5b use a mediation-
based approach to test for evidence consistent with our account that the underestimation of
reach-out appreciation is driven by a greater responder (vs. initiator) focus on the responder’s
feelings of surprise on being reached out to, both in a field experiment with actual reach-out
experiences in dyads (experiment 5a) and in a scenario paradigm with a larger sample size
THE SURPRISE OF REACHING OUT
13
(experiment 5b). Experiments 6 and 7 then identify process-consistent moderators of the
underestimation of reach-out appreciation by comparing the magnitude of the underestimation of
reach-out appreciation across reach-out contexts that differ in their surprisingness. Specifically,
experiment 6 tests whether the underestimation of reach-out appreciation exists when the reach-
out occurs in a surprising context for the responder but is eliminated when it occurs in an
unsurprising context. Experiment 7 tests whether the underestimation of reach-out appreciation is
magnified when the reach-out occurs within a relationship in which it is relatively more (vs. less)
surprising: when the initiator and responder are weak (vs. strong) ties. The four supplemental
studies are discussed briefly in the paper with full information in the web appendix.
All thirteen studies’ pre-registrations, materials, data, and syntax are available at
https://researchbox.org/289. All measures, manipulations, and exclusions in experiments are
disclosed. All sample sizes were determined in advance and pre-registered; sample sizes were
not determined based on intermittent data analysis. We pre-registered a target sample size of 100
participants per cell for our 2-cell independent-groups experiments focused on the basic effect
(i.e., experiments 1, 4a, and 4b), a target sample size of 50 dyads for our first dyadic field
experiment (experiment 2), an available sample size based on all potential initiators coming to
the in-person spring 2022 behavioral lab sessions available to one of our researchers for our
second dyadic field experiment (experiment 3), a target sample size of 100 valid initiators for our
third dyadic field experiment (experiment 5a), a target sample size of 200 participants per cell for
our 2-cell independent-groups mediation experiment (experiment 5b), and target sample sizes of
400 participants per cell for both 2´2 interaction experiments (experiments 6 and 7). For our
supplemental studies, we pre-registered a target sample size of 200 participants per cell for our
2´2 interaction experiment (supplemental study S1), a target sample size of 100 participants per
THE SURPRISE OF REACHING OUT
14
cell for our mediation experiment (supplemental study S2), a target sample size of 200
participants for our single-cell correlational study (supplemental study S3), and an available
sample size based on all potential initiators coming to an in-person spring 2022 behavioral lab
session over four days for our 2-cell intervention experiment on potential initiators (supplemental
study S4).
We also conducted sensitivity power analyses for all studies using G*Power (Faul et al.,
2007), indicating the minimum effect size detectable with statistical power of .80 given the final
sample sizes achieved. In general, the observed effect sizes were either larger than the minimum
effect size detectable at .80 for some experiments or slightly smaller than could be reliably
detected at .80 for others. See table 2 for a summary of the sensitivity power analyses.
Table 1. Summary of Main Findings
Studies
Design
Main Findings
1
2 (role: initiator,
responder)
When recalling a past experience in their lives, people who recalled a time
when they were the initiator predicted lower responder appreciation than
people who recalled a time when they were the responder and indicated how
much they appreciated being reached out to.
2
2 (role: initiator,
responder)
We replicate the underestimation of reach-out appreciation using a dyadic
paradigm in a field experiment involving actual reach-outs wherein college
students sent reach-out notes to a classmate they had not been in contact
with for a while.
3
2 (role: initiator,
responder)
We replicate the underestimation of reach-out appreciation using a dyadic
paradigm in a field experiment involving actual reach-outs wherein college
students sent reach-out notes with small gifts to a classmate they had not
been in contact with for a while.
4a-4b
2 (role: initiator,
responder)
We replicate the underestimation of reach-out appreciation using a scenario
paradigm, both with reach-outs consisting of brief notes (experiment 4a) and
with reach-outs consisting of small gifts (experiment 4b).
5a-5b
2 (role: initiator,
responder)
The underestimation of reach-out appreciation is mediated by greater
responder (vs. initiator) focus on the responder’s surprise when indicating
responder appreciation, both with a dyadic paradigm in a field experiment
involving actual reach-outs (experiment 5a) and with a scenario paradigm
(experiment 5b).
6
2 (role: initiator,
responder) ´ 2
(reach-out context:
surprising vs.
unsurprising)
The underestimation of reach-out appreciation exists when the reach-out
occurs in a context in which it is surprising to responders and is eliminated
when the reach-out occurs in a context in which it is unsurprising to
responders.
THE SURPRISE OF REACHING OUT
15
7
2 (role: initiator,
responder) ´ 2
(relationship
closeness: weak tie
vs. strong tie)
The magnitude of the underestimation of reach-out appreciation is greater for
weak tie reach-outs (which may be more surprising to responders) than
strong tie reach-outs (which may be less surprising to responders).
S1
2 (role: initiator,
responder) ´ 2
(reach-out: brief
message, small gift)
This study is a replication of the findings in experiments 4a-4b, showing that
the underestimation of reach-out appreciation occurs for both brief message
reach-outs and for small gift reach-outs.
S2
2 (role: initiator,
responder)
This study is a replication of experiment 5b, which was conducted with a
smaller sample size prior to experiment 5b. This study found a trending but
non-significant mediation effect, which encouraged us to test for the
mediation effect with a larger sample size (i.e., experiment 5b).
S3
correlational
This correlational study shows that among initiators, being more focused on
the positive surprise of their reach-outs and believing that a responder would
appreciate a reach-out more were correlated with increased likelihood of
reaching out.
S4
2 (initiator control vs.
initiator intervention)
An intervention informing potential initiators about the research findings (i.e.,
that initiators often underestimate the positive feelings of surprise and thus
the appreciation of responders) found a trending but non-significant increase
in potential initiators’ interest in reaching out and found no significant
increase in potential initiators’ actual reach-out behavior during a brief
window of opportunity to reach-out during the experiment.
Table 2. Summary of Sensitivity Analyses
Studies
Design
N
Observed effect size
Minimum
detectable effect
size at 80% power
1
2-cell
200
d = .29
d = .40
2
Paired 2-cell
54 pairs
dz = .34
dz = .39
3
Paired 2-cell
67 pairs
dz = .56
dz = .35
4a
2-cell
201
d = .32
d = .40
4b
2-cell
196
d = .65
d = .40
5a
Paired 2-cell
85 pairs
dz = .22
dz = .31
5b
2-cell
402
d = .55
d = .28
6
2x2
interaction
1576
η!
" = 0.015 (interaction)
η2 = .019 (effect of role in surprising context)
η2 = .001 (effect of role in unsurprising context)
η!
" = .008
7
2x2
interaction
1602
η!
" = .007 (interaction)
η!
" = .046 (effect of role for weak ties)
η!
" = .010 (effect of role for strong ties)
η2 = .007
S1
2x2
interaction
(with a focus
on the main
effect)
801
η!
" = .045 (main effect)
η!
" = .007 (interaction)
η!
" = .009 (effect of role for brief message)
η!
" = .043 (effect of role for small gift)
η2 = .015
S2
2-cell
200
d = .49
d = .40
S3
correlational
199
r = .61
(appreciation predictions & reach-out interest)
r = .20
S4
2-cell
118
d = .41 (appreciation predictions)
d = .35 (reach-out interest)
d = .52
THE SURPRISE OF REACHING OUT
16
Note. All sample sizes were pre-registered. For any studies with pre-registered exclusions, the sample size denoted
in the table is the sample size after pre-registered exclusions. The minimum detectable effect size at 80% power was
calculated based on the observed sample size and G*Power settings (a = .05, two-tailed).
Experiment 1: Recalled Experiences
Experiment 1 tested our main hypothesis that people (“initiators”) underestimate the
extent to which others (“responders”) appreciate being reached out to. Participants were
randomly assigned to recall a time when they either reached out to someone in their social circle
or were reached out to by someone in their social circle. We hypothesized that initiators recalling
an instance of reaching out to someone in their social circle would estimate lower appreciation of
their reach-out compared to how much responders indicated having appreciated being reached
out to. We also collected control measures to gauge whether initiators and responders recalled
reach-out events that significantly differed on various dimensions: the extent of current (post-
reach-out) relationship closeness with the other person, temporal distance from the event, and
reach-out mode.
Methods
Participants and Design. This experiment was pre-registered at
https://aspredicted.org/kr8tk.pdf. Participants (N = 200; 46% female; Mage = 34.02) were U.S.-
based adults recruited from Prolific Academic. They were randomly assigned to one of two
roles: initiator versus responder.
Procedure. All participants were asked to recall an event from their personal history.
Based on random assignment, they were asked to recall the last time (initiator condition: they
reached out to someone in their social circle; responder condition: someone in their social circle
reached out to them) “just because” or “just to catch up” via email, text, or phone, after a
THE SURPRISE OF REACHING OUT
17
prolonged period of not interacting with them. Participants indicated the initials of the person
they thought about in response to the recall prompt.
Thinking back on this time, participants were then asked to rate a set of four items
measuring appreciation on a 7-point Likert scale (1 = not at all, 7 = to a great extent), adapted
from Flynn and Adams (2009). Initiators (responders) indicated the extent to which they thought
that this person (they) appreciated, felt grateful, felt thankful, and felt pleased that they (this
person) reached out to them. These four items were averaged to create an appreciation index (
𝛼
=
.95), serving as our dependent measure.
Participants then responded to three control measures to gauge whether initiators and
responders recalled events that were systematically different. Specifically, participants indicated
how long ago the reach-out occurred (1 = less than 1 month, 2 = 1-3 months, 3 = 3-6 months, 4 =
6-9 months, 5 = 9-12 months, 6 = more than 12 months) and the reach-out mode (email, via text,
via phone, other [please elaborate]). Participants also described their current—that is, post-
reach-out—closeness to the person they recalled using a single 8-point modified Inclusion of
Others in the Self (IOS) Scale item (Aron et al., 1992) (1 = [diagram of no overlap between self
and the other person], 8 = [diagram of most overlap between self and the other person]
2
).
Results and Discussion
Underestimation of Reach-out Appreciation. An independent-samples t-test showed
that initiators (M = 5.50, SD = 1.28) thought the gesture they recalled was appreciated
significantly less than responders (M = 5.87, SD = 1.27) appreciated the gestures they recalled,
Mdifference = -.37 (95% CI [-.73, -.01]), t(198) = -2.05, p = .041, Cohen’s d = .29.
2
We inadvertently indicated on the pre-registration for this experiment that participants would complete a 7-point
IOS scale; participants actually completed an 8-point IOS scale.
THE SURPRISE OF REACHING OUT
18
Control Measures. First, we checked whether there were systematic differences in the
nature of the reach-out events recalled by initiators versus responders. An independent-samples
t-test showed that the experience recalled did not differ between initiators and responders on
temporal distance (Minitiator = 1.70, SD = 1.19 vs. Mresponder = 1.65, SD = 1.02, Mdifference = .05
(95% CI [-.26, .36]), t(198) = .32, p = .753, d = .05). Further, a Chi-squared test showed that the
event recalled also did not differ between initiators and responders on reach-out mode (text
message: 70.0% vs. 70.1%; phone: 15.5% vs. 16.5%; email: 5.8% vs. 6.2%; other: 8.7% vs.
7.2%; χ2 (3, N = 200) = .184, p = .980, Cramer’s V = .030). Finally, the nature of the reach-out
events recalled did not differ between initiators and responders on current (i.e., post-reach-out)
relationship closeness to the person whom they recalled (Minitiator = 4.27, SD = 1.80 vs. Mresponder
= 4.44, SD = 1.84, Mdifference = -.17 (95% CI [-.68, .34]), t(198) = -.67, p = .506, d = .09).
We also conducted a one-way Analysis of Covariance (ANCOVA) to examine whether
initiators and responders differed on the appreciation index after controlling for temporal
distance from the reach-out event, the reach-out mode (coded -1 = less effortful [text message]; 1
= more effortful [phone or email]; 0 = other)
3
, and current (post-reach-out) closeness to the
person whom they recalled. The underestimation of reach-out appreciation remained significant
controlling for these three measures (F(1, 195) = 4.16, p = .043,
η!
"
= .021).
Discussion. Experiment 1 shows initial evidence for the underestimation of reach-out
appreciation using a recall paradigm. Specifically, initiators recalling the last time that they
reached out to someone in their social circle estimated lower responder appreciation, compared
to responders recalling the last time someone in their social circle reached out to them. These
3
We also conducted a similar ANCOVA where “other” was coded as “1” (thus included in the analysis as an
effortful reach-out mode), and all other aspects of the model remained unchanged. The underestimation of reach-out
appreciation in this model was similar (F(1, 195) = 4.14, p = .043,
η!
"
= .021).
THE SURPRISE OF REACHING OUT
19
findings provide initial evidence of the underestimation of reach-out appreciation with
experiences that participants recalled from their own lives.
Using a recall paradigm has the strength of allowing us to tap into participants’ real-life
experiences. In addition, we confirmed that the events recalled did not differ between initiators
and responders on some key dimensions—temporal distance from the event, reach-out mode, and
the current (i.e., post-reach-out) closeness to the person recalled—thus addressing some potential
alternative explanations based on systematic differences between initiators and responders in
aspects of the reach-out event recalled. Nonetheless, a recall paradigm is still susceptible to some
additional alternative explanations. For instance, although our control measures would argue
against it, it is possible that people in the responder role more selectively recalled instances of
being reached out to that were highly positive and memorable due to being highly appreciated. In
the next experiment, we thus move away from the recall paradigm to provide convergent support
for our predictions in a field experiment on a college campus, involving actual reach-out
experiences conducted with dyads.
Experiment 2: Underestimation of Reach-Out Appreciation in a Field Experiment with
Actual Reach-Outs Consisting of a Note
In experiment 2, we further tested our main hypothesis that initiators underestimate how
much responders appreciate being reached out to in the context of a field experiment examining
actual reach-outs, thereby addressing potential alternative explanations associated with the recall
paradigm. For this field experiment, we adapted a dyadic design and procedure from Zhao and
Epley's (2021a) experiment 2. Specifically, we recruited undergraduate students on a college
THE SURPRISE OF REACHING OUT
20
campus to write a reach-out message to an actual friend whom they had not connected with in a
while and to predict how much their friend would appreciate receiving this message. We then
delivered these reach-out messages to participants’ friends on their behalf and asked their friends
to read the message and to indicate how much they appreciated receiving this message. We
predicted that participants would underestimate the extent to which their friends appreciated their
reach-out gesture.
Methods
Participants and Design. This experiment was pre-registered at
https://aspredicted.org/zc7iu.pdf. We set a pre-registered target to collect data from 50 unique
dyads. We thus purposely recruited a larger number of initiators in our first stage, anticipating
that not all initiator reach-outs would lead to completed surveys by responders. In this first stage,
participants were recruited in-person on a college campus in the U.S. and participated in
exchange for a snack item. All participants recruited in-person in this first stage were assigned to
the role of the initiator and provided the email address of the person whom they would like to
reach out to; these people then served as the corresponding paired responders if they agreed to
take part in the study. A total of 109 students participated as initiators (Mage = 19.73, 56.0%
female) across two days of recruitment: Day 1 (n = 72) and Day 2 (n = 37). We sent emails to the
corresponding responder on behalf of 105 of these initiators.
4
From the 105 responders who
received our email regarding the reach-out note and a request to participate in our survey, 59
responders responded to our survey. We then excluded five of these dyads because the responder
4
One email was not delivered because the initiator did not indicate the responder’s email address. Two emails were
not delivered because one initiator named two different responders. Another email was not delivered to the
responder until after data collection was completed because the responder was addressed in notes by two different
initiators. We delivered the email on behalf of the first initiator who completed the survey during the data collection
period and delivered the email on behalf of the second initiator after data collection closed.
THE SURPRISE OF REACHING OUT
21
had already participated in the same experiment in the role of an initiator, meaning that there
were 100 valid initiators in this experiment and 54 valid responders
5
. Thus, 54 unique dyads
were included in the final analyses (i.e., a 54% response rate when considering 54 responders for
100 valid initiators; Mage = 19.80; 68.6% female; n = 3 did not provide age and gender). We
employed a 2-group (initiator, responder) between-subjects design with matched pairs of
initiators and responders.
Procedure. We first asked initiators to name a fellow college student in their social circle
with whom they typically have pleasant encounters but have not interacted (either virtually or in-
person) in a while. Initiators were also instructed that this person should be someone whom they
know to some extent and who knows them to some extent—someone with whom they are
friendly. Initiators provided the email address and first and last name of such a person in their
lives, as well as their own email address and first and last name. On a separate page, initiators
were then asked to write a short note to the responder whom they named earlier to “check in and
say hello.” On the same page, initiators were informed that the responder would receive a survey
link that allows them to read their note and answer a question about how they feel. Finally,
participants predicted the extent to which the responder would appreciate their note reaching out
to them (1 = not at all, 7 = a great extent). Participants were informed that their appreciation
predictions would not be shared with the responders (only the notes would be shared).
5
There is a difference between the pre-registrations of experiments 2 and 3, with regard to sending the reach-outs of
initiators to responders who already participated as initiators. We realized after conducting experiment 2 that these
five responders should not have been reached out to as they had already participated as initiators (thus effectively,
the response rate should have been 54 responders out of 100 valid initiators, instead of 59 out of 105). Thus, we
modified our pre-registration for the next dyadic experiment, experiment 3, to clarify that if initiators listed a
responder who already participated as an initiator, they would not be reached out to (our response rate in experiment
3 was thus 67 responders out of 91 valid initiators). Experiment 5a, which is also a dyadic experiment, uses the same
pre-registered criteria for determining valid participants as experiment 3 does (our response rate in experiment 5a
was 85 responders out of 109 valid initiators).
THE SURPRISE OF REACHING OUT
22
We emailed responders later on the same day on which their corresponding initiator
completed the study. In the email, we informed them that one of their classmates (with first and
last name indicated) wrote a note reaching out to them as part of a research experiment
conducted on campus, and then we requested that the responder read the note and complete a
brief survey. Responders who agreed then read their note and reported the extent to which they
appreciated the initiator’s note reaching out to them, on the same 7-point appreciation scale that
initiators used.
Results and Discussion
Results. A paired-samples t-test was conducted to compare initiators’ predictions of
responder appreciation and respondents’ actual appreciation. Initiators (M = 5.57, SD = 1.35)
significantly underestimated how much their reach-outs were appreciated, compared to how
much responders (M = 6.17, SD = 1.24) appreciated them, Mdifference = -.59 (95% CI [-1.07, -
.12]), t(53) = -2.51, p = .015, dz = .34.
6
,
7
Discussion. In experiment 2, we found evidence of the underestimation of reach-out
appreciation effect using a dyadic paradigm on a college campus involving reach-outs to actual
friends, further generalizing our findings while addressing limitations with the recall paradigm.
Specifically, undergraduate students wrote a reach-out message to an actual friend that they had
6
Following Kenny et al. (2006), we also analyzed the data with linear mixed effects modeling, nesting initiators and
responders within friend dyads. Given that our paradigm requires complete dyads, the results from the linear mixed
effects model analysis were the same as those from the pre-registered paired t-test (i.e., B = -.59, SE = .24, t(53) = -
2.51, p = .015).
7
We also examined the content of participants’ messages. Nearly all participants followed directions in reaching out
to someone with whom they have had positive interactions in the past, except one participant whose message
indicated that they had reached out to someone with whom they had a recent falling out. Excluding this one
participant, the underestimation of reach-out appreciation effect remained significant (Minitiator = 5.60, SDinitiator = 1.35
vs. Mresponder = 6.26, SDresponder = 1.02; Mdifference = -.66 (95% CI [-1.12, -.20]), t(52) = -2.87, p = .006, dz = .39. In the
General Discussion, we discuss future research directions, including expanding to examine reach-outs in the context
of prior negative interactions and how they may differ from reaching out within mostly positive interaction
relationships.
THE SURPRISE OF REACHING OUT
23
not connected with in a while and then predicted how much they thought their friend would
appreciate receiving this message; their friends then read this reach-out message and indicated
how much they appreciated receiving this message. The results replicated our key finding,
wherein initiators underestimated the extent to which responders appreciated their act of reaching
out.
However, experiment 2 had the limitation that there was a relatively high attrition rate.
To address this issue, we conducted two additional dyadic field experiments (experiments 3 and
5a) aimed at mitigating the attrition rate through three modifications. First, we modified the
reach-outs to consist of a note with a small gift, which we expected would incentivize responders
to participate. Second, we modified the email subject line to include the initiator’s first and last
names to reduce the chances of being misidentified as spam email. Third, we used the Qualtrics
e-mail distribution platform to identify responders who did not open our initial email and
followed up with these responders with one reminder email. See also, the web appendix, where
we conducted exploratory analyses across experiments 2, 3, and 5a, examining whether initiators
whose reach-out recipients did (vs. did not) participate as responders differed systematically in
any observable characteristics. There were no systematic differences across the three
experiments, suggesting that selective responding is unlikely to account for the underestimation
of appreciation effect.
Experiment 3: Underestimation of Reach-Out Appreciation in a Field Experiment with
Actual Reach-Outs Consisting of a Note with a Small Gift
THE SURPRISE OF REACHING OUT
24
In experiment 3, we again further tested our hypothesis that initiators underestimate how
much responders appreciate being reached out to in the context of a field experiment examining
actual reach-outs, using a similar design and procedure as in experiment 2, but with reach-outs
consisting of a note with a small gift. We again predicted that participants would underestimate
the extent to which their friends appreciated their reach-out gesture.
Methods
Participants and Design. This experiment was pre-registered at
https://aspredicted.org/8ys68.pdf. In the first stage of experiment 3, participants were recruited
through a U.S. school’s behavioral lab for course credit. As in experiment 2, all participants in
this first stage were assigned to the role of the initiator and provided the email address of the
person whom they would like to reach out to; these people then served as the corresponding
paired responders if they agreed to take part in the study. A total of 109 students participated as
initiators (Mage = 20.55, 68.8% female) across three weeks of recruitment: Week 1 (n = 58),
Week 2 (n = 19), and Week 3 (n = 32). We sent emails to the corresponding responders on behalf
of 91 valid initiators. In accordance with our pre-registration, we did not consider someone to be
a valid initiator if they had already participated as an initiator in the same experiment (n = 6). We
also did not consider someone to be a valid initiator if they named a responder who was either (a)
already named by another participant (n = 1) or (b) had already participated in the same
experiment in the role of an initiator (n = 7). Additionally, we did not use reach-outs initiated by
participants who wrote to themselves (n = 2) or left the note field blank (n = 2). From the 91
responders to whom we sent an email regarding the reach-out note with a small gift and a request
to participate in our survey, 67 responders completed the survey (i.e., a 74% response rate; Mage
= 20.45; 65.7% female; n = 2 indicated “Other” gender), resulting in 67 unique dyads for the
THE SURPRISE OF REACHING OUT
25
final analyses. We employed a 2-group (initiator, responder) between-subjects design with
matched pairs of initiators and responders.
Procedure. We first asked initiators to name a fellow college student in their social circle
with whom they typically have pleasant encounters but have not interacted (either virtually or in-
person) in a while. Initiators were also instructed that this person should be someone whom they
know to some extent and who knows them to some extent—someone with whom they are
friendly. Initiators provided the email address and first and last name of such a person in their
lives, as well as their own email address and first and last name. On a separate page, initiators
were then asked to write a short note to the responder whom they named earlier to “check in and
say hello.” On the same page, initiators were informed that the responder would receive a survey
link that allows them to read their note, receive a gift voucher redeemable for a bag of assorted
candy/snacks, and answer a question about how they feel. Finally, participants predicted the
extent to which the responder would appreciate that they chose them to reach out with their note
and gift voucher (1 = not at all, 2 = slightly, 3 = moderately, 4 = very, 5 = extremely).
Participants were informed that their appreciation predictions would not be shared with the
responders.
We emailed responders later during the same week in which their corresponding initiator
completed the study. In the email, we informed them that one of their classmates (with first and
last name indicated) wrote a note and gifted them a voucher for a bag of assorted candy/snacks
reaching out to them as part of a research experiment conducted on campus, and then we
requested that the responder read the note and complete a brief survey. Responders who agreed
then read their note and reported the extent to which they appreciated the initiator’s note and gift
THE SURPRISE OF REACHING OUT
26
voucher reaching out to them, on the same 5-point appreciation scale that initiators used. The
final page contained instructions for redeeming the gift voucher on campus.
Results and Discussion
Results. A paired-samples t-test was conducted to compare initiators’ predictions of
responder appreciation and respondents’ actual appreciation. Initiators (M = 4.25, SD = .86)
significantly underestimated how much their reach-outs were appreciated, compared to how
much responders (M = 4.72, SD = .62) appreciated them, Mdifference = -.46 (95% CI [-.66, -.26]),
t(66) = -4.60, p < .001, dz = .56.
8
Discussion. Having provided converging evidence across both a recall study and two
field experiments involving different kinds of reach-outs, we next used a scenario experimental
paradigm in experiments 4a and 4b, which provided greater experimental control regarding
reach-out content and also removed attrition concerns. Specifically, in the subsequent scenario
studies, we had participants think of someone in their social circle first, before we randomly
assigned them to consider reaching out to this person or being reached out to by this person. We
then controlled the content of initiators’ reach-out by having initiators and responders imagine
the same reach-out made by initiators.
Experiments 4a-4b: Underestimation of Reach-out Appreciation
Experiments 4a and 4b further test our key hypothesis that initiators underestimate the
extent to which responders appreciate being reached out to, using a scenario paradigm. To
further test the generalizability of the underestimation of reach-out appreciation, there were
8
As in experiment 2, a dyadic analysis using linear mixed effects modeling, nesting initiators and responders within
friend dyads, led to the same results as those from the pre-registered paired t-test (i.e., B = -.46 , SE = .10, t(66) = -
4.60, p < .001).
THE SURPRISE OF REACHING OUT
27
several differences between experiments 4a and 4b. First, we examined two different reach-out
modes—brief messages (experiment 4a) and small gifts (experiment 4b). We also varied whether
participants considered a weak tie (experiment 4a) or more generally someone whom they have
not interacted with in a while (experiment 4b). Finally, we also used two different participant
pools—a U.S. adult Prolific sample (experiment 4a) and a U.S. college student sample
(experiment 4b).
Methods
Participants and Design. In experiment 4a, U.S.-based adult participants (N = 201;
54.7% female; Mage = 34.83) were recruited from Prolific Academic. In experiment 4b,
participants (N = 196; 52.8% female; Mage = 21.19, n = 1 did not provide age and gender) were
recruited from a U.S. university’s behavioral lab. All participants were randomly assigned to one
of two roles: initiator versus responder. Both experiments were pre-registered (experiment 4a:
https://aspredicted.org/ns5km.pdf; experiment 4b: https://aspredicted.org/a8i5q.pdf).
Procedure. All participants were first asked to think about someone in their social circle
with whom they typically have pleasant encounters. In experiment 4a, participants were further
instructed to think of someone whom they would consider a weak tie—someone whom they are
not very close to, whom they know to some extent and who knows them to some extent, with
whom they are friendly but would be unlikely to confide in (adapted from Sandstrom & Dunn,
2014b). In experiment 4b, participants read similar instructions but without specification that the
person should be a weak tie; they were asked to think of someone with whom they have not
interacted (either virtually or in-person) in a while and were told that this person should be
someone whom they know to some extent and who knows them to some extent, with whom they
THE SURPRISE OF REACHING OUT
28
are friendly. Participants then indicated the initials of the person they thought about in response
to the prompt presented to them.
Then, initiators imagined a situation in which they happened to think about this person
and are considering reaching out to them after a while of not spending any time with them.
Responders imagined a scenario in which this person happened to be thinking about them and
that this person is considering reaching out to them after a while of not spending any time with
them. In experiment 4a, the reach-out was through a brief message, whereas in experiment 4b,
the reach-out was with a small gift of a bag of cookies, a box of tea bags, a bag of coffee, or a
mini plant.
Finally, all participants rated a similar set of four items measuring appreciation from
experiment 1 but adapted for the scenario paradigm, again on a 7-point Likert scale (1 = not at
all, 7 = to a great extent) to create an appreciation index (experiment 4a:
𝛼
= .94; experiment 4b:
𝛼
= .89).
Results and Discussion
Results. Independent-samples t-tests with role as the independent variable and the
appreciation index as the dependent variable were conducted in both experiments.
9
In both
experiments, initiators significantly underestimated how much responders would appreciate their
reaching out (experiment 4a: Minitiator = 4.36, SD = 1.31 vs. Mresponder = 4.81, SD = 1.53, Mdifference
= -.45 (95% CI [-.84, -.05]), t(199) = -2.23, p = .027, d = .32, experiment 4b: Minitiator = 5.72, SD
= 1.03 vs. Mresponder = 6.32, SD = .82, Mdifference = -.60 (95% CI [-.86, -.34]), t(194) = -4.52, p <
.001, d = .65).
9
We also conducted an independent-samples Welch’s t-test in experiment 4b due to heterogeneity of variances
between initiators and responders. The results remain the same.
THE SURPRISE OF REACHING OUT
29
Discussion. In sum, experiments 4a-4b provide converging evidence that people
consistently underestimate the extent to which responders would appreciate being reached out to,
both with a message (experiment 4a) and with a gift (experiment 4b), with specifying weak ties
(experiment 4a) and more generally without specifying weak ties (experiment 4b), and in both an
online panel of U.S. adults (experiment 4a) and in college undergraduates (experiment 4b). We
also replicated these findings, showing that the underestimation of appreciation occurs for both a
message and a gift reach-out, in a supplemental experiment, reported in the web appendix (see
supplemental study S1).
Experiments 5a-5b: Mediation by a Differential Focus on the Responder’s Surprise
Experiments 5a-5b both test for the underlying role of a differential focus on the
responder’s surprise between initiators and responders. Specifically, using a mediation approach,
we tested whether the responder (vs. initiator) role is more focused on the responder’s surprise
when predicting how much the responder would appreciate being reached out to, and whether a
greater focus on the responder’s surprise is linked to predictions of higher appreciation.
Experiment 5a was conducted with a field experiment paradigm involving actual reach-outs
within dyads (similar to experiment 3). Experiment 5b was conducted using a scenario paradigm
with a much larger sample size and higher statistical power (similar to experiment 4b). Given the
differing paradigms and the slightly different wording of the measures in each experiment, we
discuss the methods and results separately for each experiment next.
Experiment 5a Methods
THE SURPRISE OF REACHING OUT
30
Participants and Design. This experiment was pre-registered at
https://aspredicted.org/8rd52.pdf. Participants were recruited on a U.S. college campus. As in
experiments 2 and 3, all participants in this first stage were assigned to the role of the initiator
and provided the email address of the person whom they would like to reach out to; these people
then served as the corresponding paired responders if they agreed to take part in the study. A
total of 131 students participated as initiators (Mage = 20.08, 61.8% female) spanning five weeks
of recruitment: Week 1 (n = 32), Week 2 (n = 18), Week 3 (n = 7), Week 4 (n = 26), and Week 5
(n = 48). We sent emails to the corresponding responders on behalf of 109 valid initiators. In
accordance with our pre-registration, we did not consider someone to be a valid initiator if they
had already participated as an initiator in the same experiment (n = 6). We also did not consider
someone to be a valid initiator if they named a responder who was either (a) already named by
another participant (n = 3) or (b) had already participated in the same experiment in the role of an
initiator (n = 12).
10
Additionally, we did not use reach-outs initiated by participants who wrote to
themselves (n = 2). Following our pre-registration, we stopped recruiting new initiators at the
end of the day on which we reached 100 valid initiators. From the 109 potential responders to
whom we sent an email with the reach-out note and a small gift voucher and a request to
participate in our survey, 85 responders completed the survey (i.e., a 78% response rate; Mage =
20.04; 65.9% female; n = 1 indicated “Other” gender), resulting in 85 unique dyads for the final
analyses. We employed a 2-group (initiator, responder) between-subjects design with matched
pairs of initiators and responders.
Procedure. Initiators first completed the same procedure from experiment 3, following
the same set of instructions. After writing their short note to the responder whom they named
10
One dyad was excluded for two reasons: both the initiator had already participated as an initiator in the same
study, and the responder had already participated as an initiator.
THE SURPRISE OF REACHING OUT
31
earlier to “check in and say hello,” initiators responded to two questions on the next page. First,
participants predicted the extent to which the responder would appreciate that they chose them to
reach out to with their note and gift voucher. Second, participants indicated how much they were
thinking about how pleasantly surprised the responder would feel by their reach-out. For both
questions, participants responded on a 5-point Likert scale (1 = not at all, 2 = slightly, 3 =
moderately, 4 = very, 5 = extremely).
Following the same procedure as in experiment 3, we emailed responders with our
request to complete a brief survey later during the same week in which their corresponding
initiator completed the study, and any responders who did not open our initial email were sent
one reminder email. The content of the email was identical to that of experiment 3, except for a
change in the name of the university as this experiment was conducted with students from a
different university. Responders who agreed then read their note and reported the extent to which
they appreciated the initiator’s note and gift voucher reaching out to them and the extent to
which they were thinking about how pleasantly surprised they feel by the reach-out, both on the
same 5-point scales that initiators used.
Experiment 5a Results
Underestimation of Appreciation. A paired-samples t-test was conducted to compare
initiators’ predictions of responder appreciation and responders’ actual appreciation. Initiators
(M = 3.91, SD = .95) significantly underestimated how much their note was appreciated,
compared to how much responders (M = 4.19, SD = .87) appreciated it, Mdifference = -.28 (95% CI
-.564, -.001]), t(84) = -1.99, p = .049, dz = .22.
11
11
As in experiments 2 and 3, a dyadic analysis using linear mixed effects modeling, nesting initiators and
responders within friend dyads, led to the same results as those from the pre-registered paired t-test (i.e., B = -.28 ,
SE = .14, t(84) = -1.99, p = .049).
THE SURPRISE OF REACHING OUT
32
Mediation. We then used methods for testing mediation in within-subject designs (5,000
bootstrapped samples; MEMORE v2.1, Model 1; Montoya & Hayes, 2017) to examine whether
the underestimation of reach-out appreciation was mediated by a focus on the responder’s
surprise. First, a paired-samples t-test showed that responders (M = 4.00, SD = 1.07) were more
focused on their pleasant surprise at the reach-out than initiators were (M = 3.62, SD = 1.08),
Mdifference = .38 (95% CI [.04, .72]), t(84) = 2.21, p = .030, dz = .24.
12
Next, a regression of the
difference between the responder’s actual appreciation and the initiator’s predicted appreciation
within each dyad (i.e., responder appreciation – initiator appreciation) on two predictors—(1) the
difference in the focus on pleasant surprise between responder and initiator within each dyad
(i.e., responder surprise – initiator surprise) and (2) the mean-centered average of the dyad’s
focus on pleasant surprise (i.e., [responder surprise + initiator surprise] / 2) showed that the
regression coefficient for the difference in the focus on pleasant surprise predictor was
significant (B = .49, SE = .07, 95% CI [.35, .64] t(82) = 6.70, p < .001). The indirect effect was
significant as the 95% CI excluded zero (B = .18, SE = .08, 95% CI [.02, .36]), indicating that
initiators’ underestimation of responders’ appreciation of their reach-out is mediated by a
differential focus on the responder’s surprise between initiators and responders.
Experiment 5b Methods
Participants and Design. This experiment was pre-registered at
https://aspredicted.org/m8w7f.pdf. U.S.-based adult participants (N = 402; 79.9% female; Mage =
28.62) recruited from Prolific Academic were randomly assigned to one of two roles: initiator
versus responder.
12
A dyadic analysis using linear mixed effects modeling, nesting initiators and responders within friend dyads, led
to the same results as those from the paired t-test (i.e., B = .38 , SE = .17, t(84) = 2.21, p = .030).
THE SURPRISE OF REACHING OUT
33
Procedure. All participants were asked to think about someone in their social circle after
being presented with the same prompt as in experiment 4b. Participants then indicated the initials
of the person they thought about in response to the prompt. Initiators were then asked to consider
a situation in which they happened to be thinking about this person and that they are considering
reaching out to this person, whereas responders were asked to think of a situation in which this
person happened to be thinking about them and was considering reaching out to them. In both
conditions, reaching out involved sending a small gift of cookies, a box of tea bags, a bag of
coffee, or a mini plant. Both initiators and responders then rated the same four appreciation items
from experiment 4a (1 = not at all, 7 = to a great extent), which were averaged to form an
appreciation index (
α
= .94).
Next, participants completed two items assessing the proposed mediator: a differential
focus on the responder’s surprise: “How much were you thinking about how surprised you
([initials]) would feel by this gift?” and “How much were you thinking about how unexpected
this gift feels (would feel for [initials])?” (1 = not at all, 7 = very much). We averaged the items
to create a focus on responder’s surprise index (r = .68, p < .001).
Experiment 5b Results
Underestimation of appreciation. We conducted an independent-samples t-test with
role as the independent variable and the appreciation index as the dependent variable.
13
Initiators
(M = 6.22, SD = 1.00) significantly underestimated responder appreciation (M = 6.67, SD = .57)
of being reached out to, Mdifference = -.45 (95% CI [-.61, -.29]), t(400) = -5.53, p < .001, d = .55,
exhibiting the underestimation of reach-out appreciation.
13
We also conducted an independent-samples Welch’s t-test. The results were very similar, t(319.65) = -5.53, p <
.001.
THE SURPRISE OF REACHING OUT
34
Mediation. We then examined whether the underestimation of reach-out appreciation
was mediated by a focus on the responder’s surprise. Mediation analysis (5,000 bootstrapped
samples; PROCESS v3.5, Model 4; Hayes, 2013) with role as the predictor, appreciation as the
outcome, and surprise focus as the mediator showed that the mediation pathway through the
focus on the responder’s surprise was significant as the 95% CI excluded zero (B = .06, SE =
.03, 95% CI [.02, .13]). See figure 1 for information on each mediation pathway.
Figure 1. Mediation Model (Experiment 5b)
Note. Regression coefficients are unstandardized; SE in parentheses; 95% confidence intervals of
the regression coefficients in brackets.
***p < .001.
Discussion
In both experiments 5a and 5b, we found mediation evidence compatible with our
proposed account involving a differential responder (vs. initiator) focus on the surprise
anticipated from the reach-out. Specifically, we find that responders are more focused than
initiators on the surprise of being reached out to when predicting how much they would
appreciate the reach out, and this differential focus on surprise is one reason why initiators
underestimate the extent to which their gestures of reaching out are appreciated. We also tested
THE SURPRISE OF REACHING OUT
35
this mediation effect in another supplemental study, reported in the web appendix (supplemental
study S2).
To test for further evidence for our proposed explanatory framework, we next examine
process-consistent moderators of the underestimation of reach-out appreciation in experiments 6-
7. We conjecture that when a reach-out occurs unexpectedly, the element of surprise for the
responder is likely to lead to an amplification of the positive experience (Mellers et al., 1997;
Oliver, 1997; Valenzuela et al., 2010) and thus, a corresponding boost in feelings of
appreciation. Therefore, our prediction is that the more surprising a reach-out is, the more
misaligned initiators and responders will be.
Experiment 6: The Underestimation of Reach-Out Appreciation for Surprising versus
Unsurprising Reach-Outs
In experiment 6, we tested whether the underestimation of reach-out appreciation occurs
when the reach-out is within a context that would be surprising to the responder (i.e., as in our
experiments thus far), whereas it is eliminated when the reach-out is within a context that would
not be surprising to the responder. In particular, we suggest that the experiments thus far have
examined reach-out situations in which the initiator reaches out without the responder expecting
it. However, there are also some reach-out situations in which initiators may reach out with the
responder expecting it (e.g., support programs or mentorship programs that involve people paired
together to reach out to each other). We thus randomly assigned participants both to a role
(initiator vs. responder) and to a reach-out context (surprising vs. unsurprising). We tested
whether the underestimation of reach-out appreciation occurs when the context of the reach-out
THE SURPRISE OF REACHING OUT
36
is surprising for the responder, whereas the underestimation of reach-out appreciation is
eliminated when the context of the reach-out is unsurprising for the responder. In doing so, we
are able to further test our framework based on differential focus on the responder’s surprise.
Methods
Participants and Design. This experiment was pre-registered at
https://aspredicted.org/s3988.pdf. U.S.-based adult participants (N = 1602) were recruited from
Prolific Academic and completed this experiment. As pre-registered, we excluded participants
with dual IP addresses with another participant in the dataset (n = 4), and we also excluded
participants who failed an attention check included at the end of the study (n = 22), resulting in
1576 participants (64.2% female, 34.1% male, 1.7% other; Mage = 38.02) for analysis. Results
remained similar (i.e., all significant findings remain significant) if these participants are not
excluded, however. Participants were randomly assigned to one of four conditions in a 2 (role:
initiator vs. responder) ´ 2 (reach-out context: surprising vs. unsurprising) between-subjects
design.
Procedure. Participants were randomly assigned to one of the four conditions. All
participants were first asked to think about someone in their circle with whom they typically
have pleasant encounters but haven’t interacted in a while and to write the person’s first initial to
personalize the scenario for them. The surprising reach-out context condition then involved
stimuli similar to those in the prior studies. Specifically, initiators in the surprising reach-out
context condition were asked to imagine that they happened to be thinking about this person and
that they are considering reaching out by sending a brief note to say hi with a small gift of a bag
of cookies. Responders in the surprising reach-out context condition were asked to imagine that
THE SURPRISE OF REACHING OUT
37
this person happened to be thinking about them and is reaching out with a brief note to say hi
with a small gift of a bag of cookies.
By contrast, the unsurprising reach-out context condition was different from the contexts
in the prior studies. Specifically, initiators in the unsurprising reach-out context condition were
asked to imagine that they happen to be thinking about this person because they are paired in a
program where it is expected that they will reach out to this person sometime this month to check
in on them and send a small gift, such that it should not be surprising to this person when they
do. They were then asked to imagine that they are considering reaching out by sending a brief
note to say hi with a small gift of a bag of cookies, as they are expected to do in this program.
Responders in the unsurprising reach-out context condition were asked to imagine that this
person happened to be thinking of them because they are paired in a program where it is
expected that this person will reach out to them sometime this month to check in on them and
send a small gift, such that it should not be surprising to them when they do. They were then
asked to imagine that the other person is reaching out by sending a brief note to say hi with a
small gift of a bag of cookies, as they are expected to do in this program.
Finally, initiators and responders were asked to complete a similar four-item appreciation
index (
α
= .95) as in experiments 4a-4b but on a 5-point Likert scale (1 = not at all, 2 = slightly,
3 = moderately, 4 = very, 5 = extremely).
Results and Discussion
Results. A 2 (role: initiator vs. responder) ´ 2 (reach-out context: surprising vs.
unsurprising) ANOVA
14
was conducted on the appreciation index, revealing a significant
14
We pre-registered a linear regression interaction analysis in experiment 6 but as the results are the same regardless
of whether we do a linear regression interaction analysis or a 2´2 ANOVA analysis, we report a 2´2 ANOVA
analysis here for consistency with experiment 7, which uses a similar 2´2 design and reports a 2´2 ANOVA
THE SURPRISE OF REACHING OUT
38
interaction (F(1, 1572) = 23.30, p < .001,
$η!
"
= .015). Given that the interaction was significant,
we conducted simple-effects tests within each reach-out context. In the surprising reach-out
context, we replicated the underestimation of appreciation effect: Initiators significantly
underestimated responders’ appreciation of being reached out to (Minitiator = 4.14, SD = .81 vs.
Mresponder = 4.46, SD = .67; Mdifference = -.32 (95% CI [-.44, -.21]); F(1, 1572) = 30.83, p < .001,
η!
"
= .019). By contrast, in the unsurprising reach-out context, the underestimation of
appreciation effect was eliminated (Minitiator = 3.93, SD = .85 vs. Mresponder = 3.86, SD = .90;
Mdifference = .07 (95% CI [-.04, .19]); F(1, 1572) = 1.63, p = .202,
η!
"
= .001). See figure 2.
Figure 2. Underestimation of Reach-Out Appreciation by Surprising versus Unsurprising
Context (Experiment 6)
Note. The figure depicts means and standard errors of the mean.
analysis. The pre-registered linear regression interaction analysis for experiment 6 is presented in the web appendix
for completeness of reporting.
1
2
3
4
5
Surprising Context Unsurprising Context
Responder Appreciation
Initiator Responder
THE SURPRISE OF REACHING OUT
39
Discussion. Experiment 6 provides further support for our surprise account via a
moderation approach. In particular, experiment 6 shows that the underestimation of reach-out
appreciation occurred when the reach-out was in a context that was surprising to the responder,
whereas it was eliminated when the reach-out was in a context that was not surprising to the
responder. We suggest that these moderation findings are in line with our surprise account
involving a differential focus on responder surprise. The next experiment proposes another
moderator that draws on similar rationale that the more surprising a reach-out is, the more
misaligned initiators and responders will be.
Experiment 7: The Underestimation of Reach-Out Appreciation for Weak versus Strong
Ties
In experiment 7, we examined another potential moderator to the underestimation of
reach-out appreciation: the degree of relationship closeness between the initiator and responder.
In experiment 4a, we observed the underestimation of reach-out appreciation when we asked
participants to imagine a weak tie. How might strong tie partners differ, if at all, in their
estimations of responder appreciation? We surmised that a reach-out from a weak tie would be
more surprising for a responder than a strong tie, thus leading to a magnification of the
underestimation of reach-out appreciation. We first report a pretest confirming that weak tie
reach-outs are indeed more surprising for responders than strong tie reach-outs. Then, in our
main experiment, we randomly assigned participants both to a role (initiator vs. responder) and
to relationship closeness (weak tie vs. strong tie). We tested whether the underestimation of
reach-out appreciation will be even greater between weak (vs. strong) tie partners, as a reach-out
THE SURPRISE OF REACHING OUT
40
will be more surprising in the context of a weak tie relationship than in the context of a strong tie
relationship.
Pretest
A pretest was conducted to establish whether reach-outs are indeed more surprising when
received from weak ties (vs. strong ties). Fifty participants recruited from Prolific Academic
(68% female; Mage = 34.02) indicated the initials of two different individuals within their social
circle: one who is a strong tie and one who is a weak tie, after being presented with the same
prompts as in the main experiment. Participants then indicated which person’s reach-out,
described as in the main experiment, would surprise them more on a 7-point scale (-3 =
Definitely a reach-out from my strong tie, 3 = Definitely a reach-out from my weak tie). The
presentation order of the two relationship types as well as the scale labels were counterbalanced
and accordingly re-coded. A one-sample t-test compared to the scale midpoint of zero showed
that a reach-out from a weak tie partner was perceived as more surprising than the same reach-
out from a strong tie partner (M = 1.94, SD = 1.53), t(49) = 8.96, p < .001, d = 1.27. In the main
experiment, we thus examined both strong tie relationships and weak tie relationships, enabling
us to directly compare the magnitude of the underestimation of reach-out appreciation for each
kind of relationship tie.
Methods
Participants and Design. This experiment was pre-registered at
https://aspredicted.org/9qi6g.pdf. U.S. and U.K.-based adult participants (N = 1,602; 65.9%
female; Mage = 36.81) recruited from Prolific Academic were randomly assigned to one of four
conditions in a 2 (role: initiator vs. responder) ´ 2 (relationship closeness: strong tie vs. weak tie)
between-subjects design.
THE SURPRISE OF REACHING OUT
41
Procedure. Participants were randomly assigned to one of the four conditions. They were
first asked to think about someone in their social circle with whom they typically have pleasant
encounters and with whom they do not currently live. In the weak tie condition, participants were
further asked to think about a weak tie—someone they are not very close to, whom they know to
some extent and who knows them to some extent, with whom they are friendly but would be
unlikely to confide in or talk to about themselves or their problems. In the strong tie condition,
participants were further asked to think about a strong tie–someone who they are very close to,
whom they know very well and who knows them very well, in whom they confide or talk to
about themselves or their problems (adapted from Sandstrom & Dunn, 2014b). Participants then
indicated the initials of the person they thought about in response to the prompt.
Next, participants in the initiator role were asked to imagine that they happened to be
thinking about this person and that they are considering reaching out with a small gift. By
contrast, participants in the responder role were asked to imagine that this person happened to be
thinking about them and is considering reaching out with a small gift.
Finally, both initiators and responders completed a similar 4-item appreciation index (
α
=
.95) as in experiments 4a-4b on a 7-point Likert scale (1 = not at all, 7 = to a great extent).
Results and Discussion
Results. A 2 (role: initiator vs. responder) ´ 2 (relationship closeness: strong tie vs. weak
tie) ANOVA was conducted on the appreciation index, revealing a significant interaction (F(1,
1598) = 11.82, p < .001,
$η!
"
= .007). Given that the interaction was significant, we conducted
simple-effects tests to examine the magnitude of the underestimation effect for each level of
relationship closeness. Initiators significantly underestimated responders’ appreciation of being
reached out to, both in the case of weak tie reach-outs (Minitiator = 5.24, SD = 1.43 vs. Mresponder =
THE SURPRISE OF REACHING OUT
42
5.92, SD = 1.20; Mdifference = -.67 (95% CI [-.82, -.52]); F(1, 1598) = 77.32, p < .001,
η!
"=
$.046
), and as evidenced by the significant interaction, to a lesser extent in the case of strong tie
reach-outs (Minitiator = 6.35, SD = .87 vs. Mresponder = 6.65, SD = .64; Mdifference = -.30 (95% CI [-
.45, -.15]); F(1, 1598) = 15.76, p < .001,
η!
"= $ .010
). See figure 3.
Figure 3. Underestimation of Reach-Out Appreciation by Relationship Closeness (Experiment 7)
Note. The figure depicts means and standard errors of the mean.
Discussion. In experiment 7, we identified another process-consistent moderator of the
underestimation of reach-out appreciation: degree of relationship closeness. Experiment 7 shows
that the underestimation of reach-out appreciation occurs both within strong and weak tie
relationships but is even stronger within weak tie relationships. We suggest that these moderation
findings are also in line with our surprise account: a weak tie reach-out is more surprising to a
responder than a strong tie reach-out and thus leads to even greater misalignment between
initiator and responder.
1
2
3
4
5
6
7
Weak Tie Strong Tie
Responder Appreciation
Initiator Responder
THE SURPRISE OF REACHING OUT
43
General Discussion
Social connections with people in one’s life are essential to happiness and well-being
(Baumeister & Leary, 1995; Cohen & Wills, 1985; Sandstrom & Dunn, 2014a, 2014b). Yet,
staying connected can be challenging. Modern life is often not set up to facilitate serendipitous
social connection with one’s acquaintances and thus, staying connected often requires at least
one person to take the initiative to reach out. Although there are certainly multiple benefits and
costs of taking the initiative to reach out, we examined whether people accurately perceive the
benefits that would accrue to others when they take the initiative to reach out.
Across a series of preregistered studies, initiators underestimated the extent to which
responders appreciate their act of reaching out. This effect occurred when considering events that
participants recalled from their personal history (experiment 1), in three dyadic field experiments
in which people reached out to their actual acquaintances (experiments 2-3, 5a), and in scenarios
in which we controlled for reach-out mode and content (experiments 4a-4b, 5b, 6, 7,
supplemental studies S1-S2). Additionally, this effect persisted across both brief message and
small gift reach-outs, weak tie and strong tie relationships, and across both undergraduate and
online adult samples.
Moreover, we highlight a differential focus on responder surprise as one underlying
explanation for why initiators underestimate how much responders appreciate being reached out
to. Specifically, we find that responders place greater focus than initiators on the surprise of
being reached out to, and this heightened focus on surprise predicts higher appreciation
(experiments 5a-5b). Further supporting this process based on differential focus on surprise, we
also find that the underestimation of reach-out appreciation is magnified under circumstances in
which the reach-out is more surprising to responders: when it occurs unexpectedly instead of as
THE SURPRISE OF REACHING OUT
44
part of an expected reach-out program (experiment 6) and for reach-outs from weak tie partners
compared to strong tie partners (in line with the notion that reach-outs from weak tie partners are
more surprising than reach-outs from strong tie partners; experiment 7).
Finally, the present research was focused on documenting when and why the
underestimation of reach-out appreciation occurs. However, it is also interesting to consider
whether and if so, how, this underestimation could be mitigated. We thus explored this issue in
two supplemental studies. First, we found correlational evidence that potential initiators who
may be more well-calibrated in terms of focusing on responders’ positive surprise and on how
much their reach-outs would be appreciated are more interested in reaching out (supplemental
study S3). Then, given that people typically use their own perspectives as a default and either
insufficiently or inaccurately adjust from this egocentric perspective (for reviews, see Epley &
Caruso, 2009; Keysar & Barr, 2002; Nickerson, 1999), we conducted supplemental study S4,
which explored a potential intervention. This intervention consisted of informing potential
initiators about our research findings—and in particular the tendency to not fully appreciate the
large extent to which their reach-outs are surprising and appreciated—and then gauging potential
initiators’ interest in reaching out and their actual reach-outs during a brief window of
opportunity during the experiment. We did not find much evidence in favor of this intervention,
at least when using this brief reach-out window paradigm, possibly because multiple additional
factors likely shaped the interest in reaching out and actual reach-out behavior during the study.
Implications
Initiating social contact after a prolonged period of disconnect can feel daunting for
various reasons, one of which is uncertainty about how such a gesture might be received. When
people take the initiative to reach out, they risk being rejected, and this worry could keep them
THE SURPRISE OF REACHING OUT
45
from reaching out in the first place. Indeed, social rejection is a highly negative and painful
social experience (Eisenberger, 2012; Leary, 2010; Williams et al., 2000) that people are
motivated to avoid given the fundamental need to belong and to feel socially connected with
others (Baumeister & Leary, 1995). Our findings take some of this challenge out by
demonstrating that small acts of reaching out are highly appreciated by responders and that the
extent of this appreciation is in fact systematically underestimated by initiators.
We add to this existing literature and contribute beyond it by broadening the scope of the
contexts to which a phenomenon of underestimation of appreciation of a social exchange applies.
In particular, we defined “reaching out” as involving a minimum criterion consisting of a gesture
to check-in with someone to show that one is thinking about them. Such a reach-out could for
instance entail sending a short message (e.g., “I’m thinking of you”) or sending a small gift.
Importantly, reaching out by this broad definition could, but need not, include expressing
gratitude (Kumar & Epley, 2018) or giving compliments (Zhao & Epley, 2021a). Indeed, while
gratitude and compliments can sometimes be warranted (Kumar & Epley, 2018; Zhao & Epley,
2021a), not all relationships involve a prior exchange that elicits feelings of gratitude that one
wants to reach out and express (as in Kumar & Epley, 2018), nor is giving a compliment always
natural or appropriate (as in Zhao & Epley, 2021a). This point is especially true with weak ties,
with whom one may have lost touch recently. Accordingly, our studies considered reaching out
without expressing gratitude or giving compliments, instead focusing on reach-outs that met the
minimum criterion of checking in with someone to show that one is thinking about them. In this
way, we hope that our findings will encourage people to reach out to their social contacts more
often “just because.” Such small gestures are likely to be appreciated more than people predict.
THE SURPRISE OF REACHING OUT
46
Moreover, by highlighting the role of a differential focus on the responder’s surprise and
identifying process-consistent moderators for this account, we also offer additional contributions.
First, we contribute to the literature demonstrating that interaction partners are often focused on
different aspects of an experience, which can subsequently lead to misalignment in judgment and
suboptimal decision-making (Boothby et al., 2018; Kupor et al., 2017; Liu & Min, 2020; Zhang
& Epley, 2012). We also note that past related work, which did not find a mediating role for
surprise, measured surprise in terms of whether receiving the message and the specific content of
the message itself was differentially surprising in magnitude based on perspective (i.e.,
consistent with the curse of knowledge; (Kumar & Epley, 2018; Zhao & Epley, 2021a). By
contrast, our theorized mediator is a differential focus on the responder’s surprise. We suggest
that initiators and responders are misaligned in the extent to which they are focusing on the
responder’s feelings of surprise when predicting responder appreciation. Our prediction thus
relies on prior theorizing on self-other discrepancies, including on egocentric perspectives that
can lead to misalignment in judgments (Epley et al., 2004) and on differential foci on one party’s
reactions (Kupor et al., 2017; Liu & Min, 2020). We reasoned that by virtue of their differing
perspectives, responders may be more focused on their own feelings of surprise when they are
reached out to, both because the unexpectedness of the event is more salient for them and also
because they are especially attuned to cues to others’ warmth (Abele & Wojciszke, 2007; Zhao
& Epley, 2021a). On the other hand, initiators may be less focused on the responder’s feelings of
surprise, both because the surprise aspect is less salient for them, given that they are not
themselves experiencing (or imagining experiencing) the event, and because they may be more
attuned to alternative cues such as those regarding the competence of their reach-outs (Abele &
Wojciszke, 2007; Zhao & Epley, 2021a). Consistent with this account, and further adding
THE SURPRISE OF REACHING OUT
47
contribution by understanding the boundaries of the tendency to underestimate others’
appreciation, we showed that when reach-outs occur within a context in which they are
unsurprising to responders (e.g., as part of a program in which one is expecting to be reached out
to by someone; experiment 6), then the underestimation of reach-out appreciation is eliminated.
Of note, it is not that unsurprising reach-outs are unpleasant. In fact, responders still indicate that
they would appreciate them to a substantial extent (i.e., a mean responder appreciation of 3.86 on
a 1-5 appreciation index, where 3 was labeled “moderately,” 4 was labeled “very,” and 5 was
labeled “extremely”)—but this appreciation is now aligned with what initiators predict.
Finally, we examined responders’ feelings of appreciation (and initiators’ predictions
about those feelings) as our outcome measure in experiments 1-7 and supplemental studies S1
and S2, consistent with past research on social exchange (Cavanaugh et al., 2015; Gino & Flynn,
2011; Kumar & Epley, 2018; Liu et al., 2015). Appreciation is a meaningful outcome variable
for its relational component, as it predicts not only general increases in life satisfaction and well-
being (Adler & Fagley, 2005) but is an important factor in building social bonds. Partner
appreciation within dyadic relationships has been shown to predict greater relationship
satisfaction and relationship commitment (Gordon et al., 2012) . Thus, another implication of our
findings is that people may underestimate the extent to which simple reach-outs may serve not
just to maintain relationships but to strengthen them as well.
Limitations and Future Directions
Our findings come with several limitations that present opportunities for future research.
First, we focus on reach-outs that carry positive surprise, as people presumably have initiated the
social contact to check in with someone and signal that they care. However, future work might
examine whether the underestimation of reach-out appreciation differs for reach-outs that carry
THE SURPRISE OF REACHING OUT
48
with them more neutral or even slightly negative feelings of surprise, such as reach-outs that are
inappropriate for the relationship (e.g., an expensive bouquet of flowers for someone one does
not know very well or a cheap, thoughtless gift for one’s partner on one’s ten-year wedding
anniversary). Such reach-outs could be perceived by the responder as inappropriate for the
relationship (Liu et al., 2015; Liu, Lamberton, et al., 2019), which may mitigate how
appreciative they are of such gifts and ultimately lead to negative consequences for the
relationship (Ruth et al., 1999). Relatedly, future research might expand to examine reach-outs in
the context of prior negative interactions and how they may differ from reaching out within
mostly positive interaction relationships. One possibility is that responders may feel more
negative feelings of surprise, which initiators might be less focused on as they predict responder
appreciation, such that initiators might even overestimate appreciation.
Second, future research may identify additional barriers to reaching out and relatedly, test
additional interventions to increasing reaching out. Our research suggests that underestimating
reach-out appreciation is one barrier that decreases interest in reaching out (see supplemental
study S3), which if addressed, may increase interest in reaching out. Although the intervention
that we tested in supplemental study S4 did not successfully increase interest in reaching out and
actual reach-out behavior during a brief time window during the study, identifying successful
interventions remains an important topic for future work. Of note, there are likely additional
types of barriers to reaching out. For instance, people may sometimes not think about reaching
out at all, or they may think about reaching out but then procrastinate and forget to do it later.
Additionally, people may overestimate the costs of reaching out. Future work could thus
examine these various potential barriers and identify ways of addressing them.
THE SURPRISE OF REACHING OUT
49
Third and finally, people initiate social contact for many reasons, but one major
overarching purpose is likely the motive to maintain or increase closeness. Thus, another
potential direction for future research is to examine how reaching out affects relationship
closeness, both from the initiator perspective and from the responder perspective (Chan &
Mogilner, 2017; Min et al., 2018; Rim et al., 2019). In the present research, we did not examine
how relationship closeness changes (we measured current post-reach-out closeness in experiment
1, which had a recall paradigm). Future work might compare pre-reach-out and post-reach-out
closeness to examine how reaching out affects the closeness felt by each party.
Final Remarks
Altogether, this research identifies a robust tendency to underestimate how much others
appreciate being reached out to and shows through both mediation and moderation evidence that
a differential focus on feelings of responder surprise is one underlying explanation. Although this
work is not explicitly related to the Covid-19 pandemic, it is not lost on us that the findings offer
particular relevance during a global pandemic that has separated millions of people from their
social contacts for a prolonged period of time. We would thus be remiss if we did not
acknowledge the implications for the millions who are now starting to reconnect or are
contemplating reconnecting with both close and distant contacts as efforts to mitigate the
pandemic progress. For those treading back into the social milieu with caution and trepidation,
feeling woefully out of practice and unsure, our work provides robust evidence and an
encouraging green light to go ahead and surprise someone by reaching out. Such reach-outs are
likely to be appreciated more than one thinks.
Open Practices
THE SURPRISE OF REACHING OUT
50
The experiments in this article adhere to Pre-registration, Open Materials, and Open Data. Pre-
registrations, materials, and data for all experiments are available at https://researchbox.org/289.
THE SURPRISE OF REACHING OUT
51
References
Abele, A. E., & Wojciszke, B. (2007). Agency and communion from the perspective of self
versus others. Journal of Personality and Social Psychology, 93(5), 751–763.
https://doi.org/10.1037/0022-3514.93.5.751
Adler, M. G., & Fagley, N. S. (2005). Appreciation: Individual differences in finding value and
meaning as a unique predictor of subjective well-being. Journal of Personality, 73(1),
79–114. https://doi.org/10.1111/j.1467-6494.2004.00305.x
Aron, A., Aron, E. N., & Smollan, D. (1992). Inclusion of Other in the Self Scale and the
structure of interpersonal closeness. Journal of Personality and Social Psychology, 63(4),
596–612. https://doi.org/10.1037/0022-3514.63.4.596
Baskin, E., Wakslak, C. J., Trope, Y., & Novemsky, N. (2014). Why feasibility matters more to
gift receivers than to givers: A construal-level approach to gift giving. Journal of
Consumer Research, 41(1), 169–182. https://doi.org/10.1086/675737
Baumeister, R. F., & Leary, M. R. (1995). The need to belong: Desire for interpersonal
attachments as a fundamental human motivation. Psychological Bulletin, 117(3), 497–
529. https://doi.org/10.1037/0033-2909.117.3.497
Bohns, V. K., & Flynn, F. J. (2010). “Why didn’t you just ask?” Underestimating the discomfort
of help-seeking. Journal of Experimental Social Psychology, 46(2), 402–409.
https://doi.org/10.1016/j.jesp.2009.12.015
Boothby, E. J., & Bohns, V. K. (2021). Why a simple act of kindness is not as simple as it seems:
Underestimating the positive impact of our compliments on others. Personality and
Social Psychology Bulletin, 47(5), 826–840. https://doi.org/10.1177/0146167220949003
THE SURPRISE OF REACHING OUT
52
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(11),
1742–1756. https://doi.org/10.1177/0956797618783714
Cavanaugh, L. A., Gino, F., & Fitzsimons, G. J. (2015). When doing good is bad in gift giving:
Mis-predicting appreciation of socially responsible gifts. Organizational Behavior and
Human Decision Processes, 131, 178–189. https://doi.org/10.1016/j.obhdp.2015.07.002
Chan, C., & Mogilner, C. (2017). Experiential gifts foster stronger social relationships than
material gifts. Journal of Consumer Research, 43(6), 913–931.
https://doi.org/10.1093/jcr/ucw067
Cohen, S., & Wills, T. A. (1985). Stress, social support, and the buffering hypothesis.
Psychological Bulletin, 98(2), 310–357. https://doi.org/10.1037/0033-2909.98.2.310
Converse, B. A., & Fishbach, A. (2012). Instrumentality boosts appreciation: Helpers are more
appreciated while they are useful. Psychological Science, 23(6), 560–566.
https://doi.org/10.1177/0956797611433334
Eisenberger, N. I. (2012). The pain of social disconnection: Examining the shared neural
underpinnings of physical and social pain. Nature Reviews Neuroscience, 13, 421–434.
https://doi.org/10.1038/nrn3231
Ekman, P. (1982). Methods for measuring facial action. In K. R. Scherer & P. Ekman (Eds.),
Handbook of methods in nonverbal behavior research (pp. 45–135). Cambridge
University Press.
Ekman, P., Friesen, W. V., & Ellsworth, P. (1972). Emotion in the human face: Guidelines for
research and an integration of findings. Pergamon Press.
THE SURPRISE OF REACHING OUT
53
Epley, N., & Caruso, E. M. (2009). Perspective taking: Misstepping into others’ shoes. In K. D.
Markman, W. M. P. Klein, & J. A. Suhr (Eds.), Handbook of imagination and mental
simulation (pp. 295–309). Psychology Press.
Epley, N., Caruso, E. M., & Bazerman, M. H. (2006). When perspective taking increases taking:
Reactive egoism in social interaction. Journal of Personality and Social Psychology,
91(5), 872–889. https://doi.org/10.1037/0022-3514.91.5.872
Epley, N., Keysar, B., Van Boven, L., & Gilovich, T. (2004). Perspective taking as egocentric
anchoring and adjustment. Journal of Personality and Social Psychology, 87(3), 327–
339. https://doi.org/10.1037/0022-3514.87.3.327
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. https://doi.org/10.3758/bf03193146
Fiedler, K., Harris, C., & Schott, M. (2018). Unwarranted inferences from statistical mediation
tests–An analysis of articles published in 2015. Journal of Experimental Social
Psychology, 75, 95-102. https://doi.org/10.1016/j.jesp.2017.11.008
Fiedler, K., Schott, M., & Meiser, T. (2011). What mediation analysis can (not) do. Journal of
Experimental Social Psychology, 47(6), 1231-1236.
https://doi.org/10.1016/j.jesp.2011.05.007
Fiske, S. T. (1992). Thinking is for doing: Portraits of social cognition from Daguerreotype to
laserphoto. Journal of Personality and Social Psychology, 63(6), 877–889.
https://doi.org/10.1037/0022-3514.63.6.877
THE SURPRISE OF REACHING OUT
54
Flynn, F. J., & Adams, G. S. (2009). Money can’t buy love: Asymmetric beliefs about gift price
and feelings of appreciation. Journal of Experimental Social Psychology, 45(2), 404–409.
https://doi.org/10.1016/j.jesp.2008.11.003
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(1), 128–
143. https://doi.org/10.1037/0022-3514.95.1.128
Galak, J., Givi, J., & Williams, E. F. (2016). Why certain gifts are great to give but not to get: A
framework for understanding errors in gift giving. Current Directions in Psychological
Science, 25(6), 380–385. https://doi.org/10.1177/0963721416656937
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(2), 211–222. https://doi.org/10.1037/0022-
3514.78.2.211
Gino, F., & Flynn, F. J. (2011). Give them what they want: The benefits of explicitness in gift
exchange. Journal of Experimental Social Psychology, 47(5), 915–922.
https://doi.org/10.1016/j.jesp.2011.03.015
Givi, J., & Galak, J. (2022). Gift recipients’ beliefs about occasion‐based and nonoccasion‐based
gifts: The importance of signaling care and meeting expectations in gift giving. Journal
of Consumer Psychology. https://doi.org/10.1002/jcpy.1239
Gordon, A. M., Impett, E. A., Kogan, A., Oveis, C., & Keltner, D. (2012). To have and to hold:
Gratitude promotes relationship maintenance in intimate bonds. Journal of Personality
and Social Psychology, 103(2), 257–274. https://doi.org/10.1037/a0028723
THE SURPRISE OF REACHING OUT
55
Granovetter, M. S. (1973). The strength of weak ties. American Journal of Sociology, 78(6),
1360–1380. https://doi.org/10.1086/225469
Grant, A. M., & Gino, F. (2010). A little thanks goes a long way: Explaining why gratitude
expressions motivate prosocial behavior. Journal of Personality and Social Psychology,
98(6), 946–955. https://doi.org/10.1037/a0017935
Gump, B. B., & Kulik, J. A. (1997). Stress, affiliation, and emotional contagion. Journal of
Personality and Social Psychology, 72(2), 305–319. https://doi.org/10.1037/0022-
3514.72.2.305
Hayes, A. F. (2013). Methodology in the social sciences. Introduction to mediation, moderation,
and conditional process analysis: A regression-based approach. Guilford Press.
Kardas, M., Kumar, A., & Epley, N. (2022). Overly shallow?: Miscalibrated expectations create
a barrier to deeper conversation. Journal of Personality and Social Psychology, 122(3),
367–398. https://doi.org/10.1037/pspa0000281
Kenny, D. A., Kashy, D. A., & Cook, W. L. (2006). The analysis of dyadic data. Guilford Press.
Keysar, B., & Barr, D. J. (2002). Self-anchoring in conversation: Why language users do not do
what they “should.” In T. Gilovich, D. Griffin, & D. Kahneman (Eds.), Heuristics and
Biases (1st ed., pp. 150–166). Cambridge University Press.
https://doi.org/10.1017/CBO9780511808098.010
Kumar, A., & Epley, N. (2018). Undervaluing gratitude: Expressers misunderstand the
sonsequences of showing appreciation. Psychological Science, 29(9), 1423–1435.
https://doi.org/10.1177/0956797618772506
THE SURPRISE OF REACHING OUT
56
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(12), 1686–1695. https://doi.org/10.1177/0146167217727003
Kurtz, J., Wilson, T., & Gilbert, D. T. (2007). Quantity versus uncertainty: When winning one
prize is better than winning two. Journal of Experimental Social Psychology, 6(43), 979–
985. https://doi.org/10.1016/j.jesp.2006.10.020
Leary, M. R. (2010). Affiliation, acceptance, and belonging. In S. T. Fiske, D. T. Gilbert, & G.
Lindzey (Eds.), Handbook of Social Psychology (5th ed., Vol. 2, pp. 864–897). Wiley.
Lee, D. S., Ybarra, O., Gonzalez, R., & Ellsworth, P. (2018). I-Through-We: How supportive
social relationships facilitate personal growth. Personality and Social Psychology
Bulletin, 44(1), 37–48. https://doi.org/10.1177/0146167217730371
Liu, P. J., & Baskin, E. (2021). The quality versus quantity tradeoff: Why and when choices for
self versus others differ. Personality and Social Psychology Bulletin, 47(5), 728–740.
https://doi.org/10.1177/0146167220941677
Liu, P. J., Dallas, S. K., & Fitzsimons, G. J. (2019). A framework for understanding consumer
choices for others. Journal of Consumer Research, 46(3), 407–434.
https://doi.org/10.1093/jcr/ucz009
Liu, P. J., & Kwon, T. A. (2022). Predicting others’ social interaction preferences: What to do,
for how long, and how often. Current Opinion in Psychology, 43, 139–145.
https://doi.org/10.1016/j.copsyc.2021.07.001
Liu, P. J., Lamberton, C., Bettman, J., & Fitzsimons, G. J. (2019). Delicate snowflakes and
broken bonds: A conceptualization of consumption-based offense. Journal of Consumer
Research, 45(6), 1164–1193. https://doi.org/10.1093/jcr/ucy051
THE SURPRISE OF REACHING OUT
57
Liu, P. J., Lamberton, C., & Haws, K. L. (2015). Should firms use small financial benefits to
express appreciation to consumers? Understanding and avoiding trivialization effects.
Journal of Marketing, 79(3), 74–90. https://doi.org/10.1509/jm.14.0091
Liu, P. J., & Min, K. E. (2020). Where do you want to go for dinner? A preference expression
asymmetry in joint consumption. Journal of Marketing Research, 57(6), 1037–1054.
https://doi.org/10.1177/0022243720949497
McPherson, M., Smith-Lovin, L., & Brashears, M. E. (2006). Social isolation in America:
Changes in core discussion networks over two decades. American Sociological Review,
71(3), 353–375. https://doi.org/10.1177/000312240607100301
Mellers, B. A., Schwartz, A., Ho, K., & Ritov, I. (1997). Decision affect theory: Emotional
reactions to the outcomes of risky options. Psychological Science, 8(6), 423–429.
https://doi.org/10.1111/j.1467-9280.1997.tb00455.x
Meyer, W., Reisenzein, R., & Schützwohl, A. (1997). Toward an analysis of emotions: The case
of surprise. Motivation and Emotion, 21(3), 251–274.
https://doi.org/10.1023/A:1024422330338
Min, K. E., Liu, P. J., & Kim, S. (2018). Sharing Extraordinary Experiences Fosters Feelings of
Closeness. Personality and Social Psychology Bulletin, 44(1), 107–121.
https://doi.org/10.1177/0146167217733077
Montoya, A. K., & Hayes, A. F. (2017). Two-condition within-participant statistical mediation
analysis: A path-analytic framework. Psychological Methods, 22(1), 6–27.
https://doi.org/10.1037/met0000086
THE SURPRISE OF REACHING OUT
58
Newark, D. A., Bohns, V. K., & Flynn, F. J. (2017). A helping hand is hard at work: Help-
seekers’ underestimation of helpers’ effort. Organizational Behavior and Human
Decision Processes, 139, 18–29. https://doi.org/10.1016/j.obhdp.2017.01.001
Nickerson, R. S. (1999). How we know—and sometimes misjudge—what others know:
Imputing one’s own knowledge to others. Psychological Bulletin, 125(6), 737–759.
https://doi.org/10.1037/0033-2909.125.6.737
Oliver, R. (1997). Customer delight: Foundations, findings, and managerial insight. Journal of
Retailing, 73(3), 311–336. https://doi.org/10.1016/S0022-4359(97)90021-X
Pennell, J. (2021, February 4). 5 things to say when reaching out to a friend right now.
Today.Com. https://www.today.com/tmrw/5-things-say-when-reaching-out-friend-right-
now-t207962
Putnam, R. D. (2000). Bowling alone: The collapse and revival of American community. Simon
& Schuster.
Reisenzein, R., Horstmann, G., & Schützwohl, A. (2019). The cognitive-evolutionary model of
surprise: A review of the evidence. Topics in Cognitive Science, 11(1), 50–74.
https://doi.org/10.1111/tops.12292
Rim, S., Min, K. E., Liu, P. J., Chartrand, T. L., & Trope, Y. (2019). The gift of psychological
closeness: How feasible versus desirable gifts reduce psychological distance to the giver.
Personality and Social Psychology Bulletin, 45(3), 360–371.
https://doi.org/10.1177/0146167218784899
Ross, M., & Sicoly, F. (1979). Egocentric biases in availability and attribution. Journal of
Personality and Social Psychology, 37(3), 322–336. https://doi.org/10.1037/0022-
3514.37.3.322
THE SURPRISE OF REACHING OUT
59
Ruth, J. A., Otnes, C. C., & Brunel, F. F. (1999). Gift receipt and the reformulation of
interpersonal relationships. Journal of Consumer Research, 25(4), 385–402.
https://doi.org/10.1086/209546
Sandstrom, G. M., & Dunn, E. W. (2014a). Is efficiency overrated?: Minimal social interactions
lead to belonging and positive affect. Social Psychological and Personality Science, 5(4),
437–442. https://doi.org/10.1177/1948550613502990
Sandstrom, G. M., & Dunn, E. W. (2014b). Social interactions and well-being: The surprising
power of weak ties. Personality and Social Psychology Bulletin, 40(7), 910–922.
https://doi.org/10.1177/0146167214529799
Savitsky, K., Epley, N., & Gilovich, T. (2001). Do others judge us as harshly as we think?
Overestimating the impact of our failures, shortcomings, and mishaps. Journal of
Personality and Social Psychology, 81(1), 44–56. https://doi.org/10.1037/0022-
3514.81.1.44
Savitsky, K., & Gilovich, T. (2003). The illusion of transparency and the alleviation of speech
anxiety. Journal of Experimental Social Psychology, 39(6), 618–625.
https://doi.org/10.1016/S0022-1031(03)00056-8
Teigen, K. H., & Keren, G. (2003). Surprises: Low probabilities or high contrasts? Cognition,
87(2), 55–71. https://doi.org/10.1016/s0010-0277(02)00201-9
Valenzuela, A., Mellers, B., & Strebel, J. (2010). Pleasurable surprises: A cross-cultural study of
consumer responses to unexpected incentives. Journal of Consumer Research, 36(5),
792–805. https://doi.org/10.1086/605592
THE SURPRISE OF REACHING OUT
60
Williams, K. D., Cheung, C. K. T., & Wilma, C. (2000). Cyberostracism: Effects of being
ignored over the Internet. Journal of Personality and Social Psychology, 5(79), 748–762.
https://doi.org/10.1037//0022-3514.79.5.748
Yang, A. X., & Urminsky, O. (2018). The smile-seeking hypothesis: How immediate affective
reactions motivate and reward gift giving. Psychological Science, 29(8), 1221–1233.
https://doi.org/10.1177/0956797618761373
Yzerbyt, V., Muller, D., Batailler, C., & Judd, C. M. (2018). New recommendations for testing
indirect effects in mediational models: The need to report and test component
paths. Journal of Personality and Social Psychology, 115(6), 929–
943. https://doi.org/10.1037/pspa0000132
Zhang, Y., & Epley, N. (2009). Self-centered social exchange: Differential use of costs versus
benefits in prosocial reciprocity. Journal of Personality and Social Psychology, 97(5),
796–810. https://doi.org/10.1037/a0016233
Zhang, Y., & Epley, N. (2012). Exaggerated, mispredicted, and misplaced: When “it’s the
thought that counts” in gift exchanges. Journal of Experimental Psychology: General,
141(4), 667–681. https://doi.org/10.1037/a0029223
Zhao, X., & Epley, N. (2021a). Insufficiently complimentary?: Underestimating the positive
impact of compliments creates a barrier to expressing them. Journal of Personality and
Social Psychology, 121(2), 239-256. https://doi.org/10.1037/pspa0000277
Zhao, X., & Epley, N. (2021b). Kind words do not become tired words: Undervaluing the
positive impact of frequent compliments. Self and Identity, 20(1), 25–46.
https://doi.org/10.1080/15298868.2020.1761438
THE SURPRISE OF REACHING OUT
1
WEB APPENDIX
The Surprise of Reaching Out: Appreciated More than We Think
TABLE OF CONTENTS
Web Appendix
Description
Page
Numbers
A
Supplemental Study S1: The Underestimation of Reach-Out Appreciation for
Brief Message Reach-Outs and Small Gift Reach-Outs
2-6
B
Supplemental Study S2: Mediation by a Differential Focus on the Responder’s
Surprise
7-10
C
Supplemental Study S3: Correlating Focus on the Responder’s Positive
Surprise, Appreciation Predictions, and Reach-out Interest
11-14
D
Supplemental Study S4: Increasing Interest in Reaching Out?
15-20
E
Exploratory Analyses on Attrition in Dyadic Experiments 2, 3, and 5a
21-24
F
Pre-registered Linear Regression Analysis in Experiment 6
25
All pre-registrations, survey materials, data, and syntax are available at ResearchBox
(https://researchbox.org/289).
THE SURPRISE OF REACHING OUT
2
WEB APPENDIX A
Supplemental Study S1: The Underestimation of Reach-Out Appreciation for Brief
Message Reach-Outs and Small Gift Reach-Outs
Supplemental study S1 is a replication of the finding that the underestimation of reach-
out appreciation occurs both for message reach-outs and for gift reach-outs. A noteworthy
observation from experiments 4a-4b is that the effect size of the underestimation of reach-out
appreciation was roughly twice as large in experiment 4b (d = .65) as in experiment 4a (d = .31).
Although multiple aspects differed between the two experiments, we conjectured based on our
responder surprise account that the most influential factor accounting for the different effect
sizes might be the nature of the particular reach-outs: whether it was via the brief message or the
small gift described in the scenarios. Specifically, the brief message in experiment 4a was
described as, “Hi there! I was just thinking about you, and wanted to see how you’re doing. I
hope all is well with you and your family!”, and the small gift in experiment 4b was described as
“consisting of a box of cookies, a box of tea bags, a bag of coffee, or a mini flower bouquet.”
Whereas this brief message may be somewhat surprising, we conjectured that a small
gift of cookies, tea, coffee, or flowers would likely be even more surprising between
acquaintances. In supplemental study S1, we first report a pre-test confirming that these small
gifts are indeed more surprising for responders than these brief messages. Then, in our main
experiment, we randomly assigned participants both to a role (initiator vs. responder) and to a
reach-out mode (brief message vs. small gift). We tested whether the underestimation of reach-
out appreciation is magnified when the reach-out is done via a small gift versus a brief message,
supporting our framework based on differential focus on responder surprise.
THE SURPRISE OF REACHING OUT
3
Pretest
A pretest was conducted to establish whether the small gift reach-outs as described in
experiment 4b are indeed more surprising to responders than the brief message reach-outs as
described in experiment 4a. Fifty participants recruited from Prolific Academic (42% female;
Mage = 33.76) indicated the initials of a person within their social circle who lives nearby, whom
they would consider to be an acquaintance or friend with whom they have typically had pleasant
encounters but with whom it has been a while since they have spent any time together.
Participants then indicated which gesture from this person would surprise them more: a brief
message (“Hi there! I was just thinking about you, and wanted to see how you’re doing. I hope
all is well with you and your family!”) or a small gift consisting of a box of cookies, a box of tea
bags, a bag of coffee, or a mini flower bouquet. Participants responded on a 7-point scale (-3 =
Definitely the brief message, 3 = Definitely a small gift). The presentation order of the two
gestures, as well as the scale labels, were counterbalanced and accordingly re-coded. A one-
sample t-test showed that the small gift was indeed perceived as more surprising than the brief
message (M = 1.96, SD = 1.71), t(49) = 8.09, p < .001, d = 1.14. In the main experiment, we thus
examined both kinds of reach-outs within the same experiment, enabling us to directly compare
the magnitude of the underestimation of reach-out appreciation for each kind of reach-out.
Methods
Participants and design. This experiment was pre-registered at
https://aspredicted.org/2tb33.pdf. U.S.-based adult participants (N = 801; 51.8% female; Mage =
35.54) recruited from Prolific Academic were randomly assigned to one of four conditions in a 2
(role: initiator vs. responder) ´ 2 (reach-out mode: brief message vs. small gift) between-subjects
design.
THE SURPRISE OF REACHING OUT
4
Procedure. All participants were first asked to think about a friend or an acquaintance
who lives nearby and to indicate the initials of the person they thought about in response to the
prompt. Next, initiators were asked to imagine that they happened to be thinking about this
person and that they are considering reaching out with a brief text message (brief message
condition) or a small gift (small gift condition). By contrast, responders were asked to imagine
that this person happened to be thinking about them and is considering reaching out with a brief
text message (brief message condition) or a small gift (small gift condition). The content of the
text message (“Hi there! I was just thinking about you, and wanted to see how you’re doing. I
hope all is well with you and your family!”) and the nature of the small gift (i.e., a bag of
cookies, a box of tea bags, a bag of coffee, or a mini flower bouquet) were held constant across
initiators and responders. Finally, both initiators and responders rated the same four items (
α
=
.95) composing the appreciation index as in experiments 4a-4b on a 7-point Likert scale (1 = not
at all, 7 = to a great extent).
Results and Discussion
Results. We conducted a 2 (role: initiator vs. responder) ´ 2 (reach-out mode: brief
message vs. small gift) Analysis of Variance (ANOVA) on the appreciation index, which
revealed a significant interaction (F(1, 797) = 5.27, p = .022,
η!
"
= .007). Given that the
interaction was significant, we conducted simple-effects tests to examine the magnitude of the
underestimation effect in each reach-out context. Initiators significantly underestimated
responders’ appreciation of being reached out to, both in the case of brief messages (Minitiator =
5.69, SD = 1.19 vs. Mresponder = 5.99, SD = 1.12; Mdifference = -.30 (95% CI [-.52, -.09]), F(1, 797)
= 7.48, p = .006,
η!
"= $ .009
), and as evidenced by the significant interaction, to an even greater
THE SURPRISE OF REACHING OUT
5
extent in the case of small gifts (Minitiator = 5.82, SD = 1.25 vs. Mresponder = 6.48, SD = .82;
Mdifference = -.66 (95% CI [-.88, -.44]), F(1, 797) = 35.71, p < .001,
η!
"= $ .043
). See figure S1.
Figure S1. Underestimation of Reach-Out Appreciation by Reach-Out Mode
Note. The figure depicts means and standard errors of the mean.
Discussion. Supplemental study S1 shows that the underestimation of reach-out
appreciation occurs both with brief messages and small gifts, replicating the findings in
experiments 4a-4b. We also found that the underestimation effect was larger for small gift reach-
outs than for brief message reach-outs. We suggest that these moderation findings are in line
with our surprise account: a small gift is more surprising to a responder in this situation than a
brief message and thus may lead to even greater misalignment between initiator and responder
due to their differential focus on responder surprise.
However, we do not claim that all gifts will lead to a greater misalignment than all
messages. For instance, within some relationships, a reach-out consisting of a very lengthy,
heartfelt message might be even more surprising than a small gift, which could mean that there
would be greater misalignment between initiators and responders for the message. Rather, what
1
2
3
4
5
6
7
Small Gift Reach-out Brief Message Reach-out
Responder Appreciation
Initiator Responder
THE SURPRISE OF REACHING OUT
6
we do suggest (consistent with our pre-test) is that for these common instantiations of reach-
outs—a brief message as described herein (a common way of reaching out to acquaintances) and
a small gift as described herein (consisting of items that are relatively common to gift to
others)—the results are compatible with and supportive of our differential focus on responder
surprise account. However, a possible future direction would involve systematically examining
the role of the magnitude and sincerity of message and gift reach-outs.
THE SURPRISE OF REACHING OUT
7
WEB APPENDIX B
Supplemental Study S2: Mediation by a Differential Focus on the Responder’s Surprise
Supplemental study S2 was conducted prior to experiment 5b and has a smaller sample
size than experiment 5b, but otherwise has the same design and procedure as experiment 5b.
Supplemental study S2 tested for mediation evidence of the underlying role of a differential
focus on the responder’s surprise between initiators and responders. Specifically, we tested
whether the responder (vs. initiator) role is more focused on the responder’s surprise when
predicting how much the responder would appreciate being reached out to, and whether a greater
focus on the responder’s surprise is linked to predictions of higher appreciation. We also
examined a potential alternative explanation involving a focus on the effort undertaken by the
initiator. Specifically, we examined whether the responder (vs. initiator) is more focused on the
effort and work that the initiator invested to reach out, thereby bolstering appreciation. To
preview the findings, we observed a trending but nonsignificant mediation effect in supplemental
study S2, which led us to test for this effect in a subsequent study with a larger sample size (i.e.,
experiment 5b).
Methods
Participants and Design. This experiment was pre-registered at
https://aspredicted.org/9we6x.pdf. U.S.-based adult participants (N = 200; 70.5% female; Mage =
36.01) recruited from Prolific Academic were randomly assigned to one of two roles: initiator
versus responder.
Procedure. All participants were asked to think about someone in their social circle with
whom they typically have pleasant encounters but with whom they haven’t interacted in a while,
THE SURPRISE OF REACHING OUT
8
who they know to some extent and who knows them to some extent, and with whom they are
friendly. Participants then indicated the initials of the person they thought about in response to
the prompt. Initiators were then asked to consider a situation in which they happened to be
thinking about this person and that they are considering reaching out to this person, whereas
responders were asked to think of a situation in which this person happened to be thinking about
them and was considering reaching out to them. In both conditions, reaching out involved
sending a small gift of cookies, a box of tea bags, a bag of coffee, or a mini plant. Both initiators
and responders then rated four appreciation items, similar to those used in experiments 4a-4b and
5b (1 = not at all, 7 = to a great extent), which were averaged to form an appreciation index (
α
=
.94).
Next, participants completed measures assessing two potential mediators: a focus on the
responder’s surprise and a focus on the initiator’s effort. Each potential mediator was assessed
with two items. The two items for the focus on surprise mediator were the same as in experiment
5b: “How much were you thinking about how surprised you ([initials]) would feel by this gift?”
and “How much were you thinking about how unexpected this gift feels (would feel for
[initials])?” The two items for the focus on effort mediator were: How much were you thinking
about the effort involved behind this gift” and “How much were you thinking about the work
involved behind this gift?” All four mediator items were on a 7-point scale (1 = not at all, 7 =
very much). We averaged the two surprise items to create a focus on surprise index (r = .73, p <
.001), and we averaged the two effort items to create a focus on effort index (r = .83, p < .001).
Results and Discussion
Underestimation of Appreciation. We conducted an independent-samples t-test with
THE SURPRISE OF REACHING OUT
9
role as the independent variable and the appreciation index as the dependent variable.
15
Initiators
(M = 6.16, SD = .94) significantly underestimated responder appreciation (M = 6.59, SD = .80)
of being reached out to, Mdifference = -.43 (95% CI [-.67, -.18]), t(198) = -3.43, p < .001, d = .49.
Parallel Mediation. We then examined whether the underestimation of reach-out
appreciation was mediated by a focus on the responder’s surprise, a focus on the initiator’s
effort, or both. A parallel mediation analysis (5,000 bootstrapped samples; Model 4; Hayes,
2013) showed that the mediation pathway through a focus on the responder’s surprise was
trending though nonsignificant (B = .07, SE = .04, 90% CI [.01, .14]) and that the mediation
pathway through a focus on the initiator’s effort was nonsignificant (B = .01, SE = .06, 90% CI
[-.08, .11], 95% CI [-.10, .14]). See figure S2.
Figure S2. Mediation Model
Note. Regression weights are unstandardized; SE in parentheses; 90% confidence intervals of the
regression coefficients in brackets.
+ p < .10. **p < .01. ***p < .001.
15
We also conducted an independent-samples Welch t-test to address heterogeneity of variances between initiators
and responders. The results were very similar, t(193.148) = -3.43, p < .001.
THE SURPRISE OF REACHING OUT
10
Discussion. Supplemental study S2 finds trending but non-significant mediation evidence
for our proposed account involving a differential responder (vs. initiator) focus on the surprise
anticipated from the reach-out. We did not find mediation evidence compatible with an
alternative account wherein the underestimation of reach-out appreciation is driven by a
differential focus on the effort undertaken by the initiator to reach out. Given the trending but
non-significant mediation by a differential responder (vs. initiator) focus on the surprise
anticipated from the reach-out, we subsequently conducted a new study with a larger sample size
(i.e., experiment 5b). In experiment 5b, we found significant mediation evidence: Responders
were more focused than initiators on the surprise of being reached out to when predicting how
much they would appreciate the reach out, and this differential focus on surprise was one reason
why initiators underestimated the extent to which their gesture of reaching out was appreciated.
THE SURPRISE OF REACHING OUT
11
WEB APPENDIX C
Supplemental Study S3: Correlating Focus on the Responder’s Positive Surprise,
Appreciation Predictions, and Reach-out Interest
In experiments 1-7, we have shown that initiators underestimate how much responders
appreciate being reached out to, as they are less focused on responders’ feelings of positive
surprise. We suggest that predictions about how much responders will appreciate being reached
out to are important to examine because they are one driver of people’s interest in reaching out:
believing that someone will appreciate a reach-out more should increase interest in reaching out
to them. In supplemental study S3, we thus asked participants to consider reaching out to
someone in their lives with whom they haven’t talked in a while and to indicate how much they
think that person would appreciate their reaching out, how much they were focused on that
person's positive surprise as they considered how much they thought that person would
appreciate being reached out to, and their interest in reaching out. We predicted that participants'
focus on the recipient's pleasant surprise would be positively related to their predictions about
the recipient's appreciation, which would be positively related to their interest in reaching out.
Support for this prediction would help establish the relevance of examining the underestimation
of appreciation of reach-outs, as we have suggested that this underestimation is one reason why
people do not reach-out to others as much as may be ideal. This study’s approach was adapted
from supplemental experiment S2 of Zhao and Epley (2021).
Methods
Participants and Design. This study was pre-registered at:
https://aspredicted.org/5dr5e.pdf. U.S.-based adult participants (N = 199; 54.8% female; Mage =
THE SURPRISE OF REACHING OUT
12
39.44) recruited from Amazon Mechanical Turk through the CloudResearch Platform completed
this study.
Procedure. All participants were asked to think about someone in their social circle with
whom they typically have pleasant encounters but haven’t interacted in a while, who they know
to some extent and who knows them to some extent, with whom they are friendly, but haven’t
seen or talked to in a while. Participants indicated this person’s first initial. They were then asked
to consider a situation in which they happened to be thinking about this person and that they are
considering reaching out to this person, by sending a brief note to say hello and check in on
them. They were asked to imagine actually reaching out to this person and to predict how this
person would feel after being reached out to. Specifically, they responded to the question, “To
what extent would [person’s first initial] appreciate it?” (1 = not at all, 5 = extremely).
Next, participants indicated what they were focusing on as they made this prediction
about how much this person would appreciate being reached out to: “As you were indicating
how much you thought [person’s first initial] would appreciate it, to what extent were you
thinking about how pleasantly surprised [person’s first initial] would be by your reach-out?” (1 =
not at all, 5 = extremely).
Finally, participants indicated their interest in actually reaching out: “How interested are
you in actually reaching out to [person’s first initial]? (could be in an email, a text message, a
handwritten note, or a virtual or in-person conversation)” (1 = not at all interested, 5 = extremely
interested).
Results and Discussion
Results. First, to examine how participants’ appreciation predictions are related to their
reach-out intentions, we conducted a linear regression with appreciation predictions predicting
THE SURPRISE OF REACHING OUT
13
interest in reaching out. We found that the more participants thought the person would appreciate
their reach-out, the more interested they were in reaching out (B = .82, SE = .08, 95% CI [.67,
.97], t = 10.67, p < .001).
Second, to examine how participants' focus on the responder's pleasant surprise is related
to their appreciation predictions, we conducted a linear regression with focus on the other
person’s pleasant surprise predicting appreciation predictions. We found that the more
participants were focused on the other person’s pleasant surprise, the more they predicted that
the other person would appreciate their reach-out (B = .43, SE = .05, 95% CI [.33, .53], t = 8.30,
p < .001).
Finally, to examine participants’ focus on the other person’s pleasant surprise,
appreciation predictions, and reach-out intentions together, we conducted a mediation analysis
(PROCESS model 4; Hayes, 2013) with interest in reaching out as the “outcome variable,” focus
on the responder's pleasant surprise as the “predictor,” and appreciation predictions as the
“mediator.” This analysis showed that there was a significant “indirect effect” wherein a greater
focus on the responder’s pleasant surprise increased appreciation predictions, which increased
interest in reaching out (B = .34, SE = .05, 95% CI [.24, .45]). See figure S3. Of note, this model
is not causal given the nature of the data (Fiedler et al., 2011, 2018), as Zhao and Epley (2021,
supplemental expeirment S2) acknowledges.
Figure S3. Mediation Model
THE SURPRISE OF REACHING OUT
14
Note. Regression weights are unstandardized; SE in parentheses; 95% confidence intervals of the
regression coefficients in brackets.
NS p > .10. ***p < .001.
Discussion. Supplemental study S3 examined the relationship between participants’
thoughts, their predictions, and their interest in behavioral reach-out action, showing that the
more participants were focused on the responder’s pleasant surprise, the more they thought the
responder would appreciate their reach-out, and the more they were interested in reaching out.
These findings are thus compatible with our account that the underestimation of appreciation of
reach-outs is important to examine as this may be one reason why some people may not reach-
out to others as much as may be ideal.
THE SURPRISE OF REACHING OUT
15
WEB APPENDIX D
Supplemental Study S4: Increasing Interest in Reaching Out?
Building on the correlational findings in supplemental study S3, supplemental study S4
tested whether an intervention on potential initiators informing them about our main findings
(i.e., the tendency for people to underfocus on others’ pleasant surprise on being reached out to
and thus to underestimate how much others appreciate being reached out to) increases their
interest in reaching out, via increasing their predictions about appreciation. We also explored
whether this intervention increased actual reach-out during a two-minute window during the
study (using a procedure adapted from Zhao and Epley (2021)’s experiment 4).
Methods
Participants and Design. This experiment was pre-registered:
https://aspredicted.org/7wy73.pdf. Undergraduate students (N = 118; 52.5% female; Mage =
19.78) recruited from a U.S. university’s behavioral lab participated in this experiment for course
credit or extra credit. All participants were randomly assigned to one of two roles: initiator-
control versus initiator-intervention.
Procedure. This study was called the “Reaching Out Study,” and aspects of the
procedure were adapted from Zhao and Epley (2021)’s experiment 4. All participants were first
asked to think about someone in their circle with whom they typically have pleasant encounters
but haven’t interacted in a while. They were told that they would be asked to write a brief
message reaching out to them to say hi and check in and that this brief message should be
something they could give someone without notice, as an act of checking in. They were provided
with advance notice that after writing this message, we would give them an opportunity to decide
THE SURPRISE OF REACHING OUT
16
whether they would actually want to deliver their reach-out message right away. All participants
then wrote down the first name of the person they were considering reaching out to right now to
say hi and check in on them. They also indicated their relationship with this person (friend,
acquaintance, family member, classmate, or other) and how long they have known each other (in
days). Participants then wrote a brief message that they could send this person, saying hi and
checking in on them.
All participants were then informed that now that they had written a reach-out message,
they had an opportunity to actually deliver it right now or they could wait to reach-out on another
occasion. All participants further read that when deciding whether to reach out, it is important to
consider how much they think the other person will appreciate their reach-out. In the initiator-
intervention condition, participants were presented with a brief paragraph summarizing our
research findings (adapting an intervention tested by Zhao and Epley (2021)). Specifically, they
were asked to read about the latest social science research on this topic and subsequently were
presented with a paragraph on social science research showing that people often do not fully
realize how much the people in their lives appreciate being reached out to, with an explanation
focusing on how those who are reached out to are extremely positively surprised, which
increases their feelings of appreciation well beyond what people expect. To ensure that
participants engaged with this research summary, participants were asked to apply these findings
to answer a question about a scenario. Participants in the initiator-control condition did not read
this brief paragraph and did not apply these findings to answer a question about a scenario.
All participants then completed the two focal dependent variables, as pre-registered. First
participants provided appreciation predictions: “How much do you think [first name of person
they listed] would appreciate being reached out to by you with the message you wrote earlier?”
THE SURPRISE OF REACHING OUT
17
(1 = not at all, 2 = slightly, 3 = moderately, 4 = very, 5 = extremely). Second, participants
indicated their current reach-out interest: “How interested are you in sending this reach-out
message you wrote earlier to [first name of person they listed]?” (1 = not at all interested, 2 =
slightly interested, 3 = moderately interested, 4 = very interested, 5 = extremely interested).
All participants were then asked to take out their cellphones in the lab for the next two
minutes and given the opportunity in the midst of the study to actually reach out to the person
whose name was specified. They were given a template if they wanted to use it, wherein they
could tell the other person that they were doing a research study at [name of university] and were
asked to reach-out to someone to say hi and check in and had decided to reach-out to them. They
were also shown the message they had typed earlier. Participants were informed that their
decision to deliver their reach-out message would not influence their credit or extra credit for this
lab session. After two minutes, participants were able to proceed and responded to the third
(exploratory) dependent variable, which gauged whether they had actually sent their reach-out
message to this person right then (yes, no). Participants were then asked to put away their
cellphones for the remainder of the lab session. Participants who indicated that they had reached
out indicated how they sent their reach-out message (text message, email, social media, other).
Participants who indicated that they had not reached out filled out a free-response question
asking, “Why did you choose not to send your reach-out message today?” Finally, all
participants provided their demographic information.
Results and Discussion
Intervention on Appreciation Predictions. An independent-samples t-test with
intervention (vs. control) as the independent variable and appreciation predictions as the
dependent variable showed that the intervention (M = 4.15, SD = .76) significantly increased
THE SURPRISE OF REACHING OUT
18
predictions of responder appreciation (M = 3.81, SD = .86) of being reached out to, Mdifference =
.34 (95% CI [.04, .63]), t(116) = 2.24, p = .027, d = .42.
Intervention on Reach-out Interest. An independent-samples t-test with intervention
(vs. control) as the independent variable and interest in reaching out as the dependent variable
showed that the intervention (M = 3.05, SD = 1.28) did not significantly increase interest in
reaching out (M = 2.62, SD = 1.21), though the effect trended in the predicted direction, Mdifference
= .44 (95% CI [-.02, .89]), t(116) = 1.90, p = .060, d = .35.
Actual Reach-out Behavior. An exploratory binary logistic regression with intervention
(vs. control) predicting actual reach-out (yes vs. no) was non-significant (B = .16, SE = .38,
Wald χ2 = .17, p = .680, OR = 1.17, 95% CI of OR [.59, 2.44]). Given that this intervention was
non-significant (and given the significant effect on appreciation predictions and the non-
significant but trending effect on interest in reaching out, respectively), we conducted the
following follow-up tests for exploratory purposes
16
.
First, we conducted a mediation analysis (5,000 bootstrapped samples; Model 4; Hayes,
2013) to explore whether the effect of the intervention on reach-out interest was mediated by
appreciation predictions, though with the caveat that we were testing for an indirect effect in the
absence of a significant direct effect (see also, Yzerbyt et al., 2018). This analysis showed that
the mediation pathway through appreciation predictions was significant (B = .28, SE = .13, 95%
CI [.03, .56]). Specifically, besides the significant effect of the intervention (vs. control) on
appreciation predictions (p = .027), a regression on interest in reaching out with both condition
16
In our pre-registration, we stated that we would conduct these follow-up tests if there was a significant effect on
both appreciation predictions and on interest in reaching out. We had not specified whether we would conduct these
follow-up tests if there was a significant effect on appreciation predictions and a non-significant but trending effect
on interest in reaching out (i.e., p = .060). We thus conducted these follow-up tests for exploratory purposes and
completeness of reporting.
THE SURPRISE OF REACHING OUT
19
and appreciation predictions as predictors showed that appreciation predictions significantly
predicted interest in reaching out (B = .84, SE = .12, 95% CI [.60, 1.07], t = 6.98, p < .001) and
that condition was non-significant (B = .15, SE = .20, 95% CI [-.24, .55], t = .78, p = .435)
suggesting full mediation through appreciation predictions (though this mediation should be
interpreted with caution; see Yzerbyt et al., 2018).
Second, we also conducted a binary logistic regression with reach-out interest predicting
actual reach-out, to explore whether reach-out interest predicts actual reach-out and if so, what
percentage of variance in people's actual behavior is accounted for by reach-out interest. This
regression was significant (B = .85, SE = .19, Wald χ2 = 20.15, p < .001, OR = 2.34, 95% CI of
OR [1.61, 3.38]), though interest in reaching out accounted for only 19.3% of the variance in
actual reach-out behavior during the two-minute window of opportunity provided during the
experiment (Cox and Snell R2 = .19), indicating that other factors mainly drove the decision to
actually reach-out during this window.
Discussion. Supplemental study S4 examines an intervention informing potential
initiators about our findings regarding the tendency to underestimate reach-out appreciation.
Although this intervention significantly increased their predictions about how much others would
appreciate their reach-out attempts, it did not significantly increase their interest in reaching out
despite trending in that direction. There was a significant indirect effect from the intervention to
appreciation predictions to interest in reaching out but see Yzerbyt et al. (2018) on interpreting
this indirect effect with caution. Finally, the intervention did not significantly increase actual
reach-out during a two-minute window provided during the experiment, with the interest in
reaching out only explaining approximately 20% of the variance in actual reach-out behavior.
Interestingly, this explanatory variance is quite similar to the degree of variance explained by
THE SURPRISE OF REACHING OUT
20
interest in complimenting for actual complimenting during a similar window of opportunity in
Zhao and Epley (2021, experiment 4). Indeed, Zhao and Epley (2021, experiment 4) also found
no significant effect of their intervention on actual complimenting behavior and found that
interest in complimenting explained approximately 20% of the variance in actual complimenting.
In light of their findings regarding the difficulty of changing actual behavior (particularly within
a prescribed brief window of opportunity in an experiment), we had purposely pre-registered
analyses on the actual reach-out behavior outcome as exploratory in supplemental study S4, with
the primary focus on appreciation predictions and reach-out intentions. We also explored the
free-responses from participants who opted not to reach-out during the window of opportunity in
supplemental study S4. Participants expressed various reasons, including that they would prefer
to reach-out to the other person later outside of the study, that they felt it was awkward to reach-
out in the context of doing a study, that they would rather call than text, or that the other person
was likely currently busy with class or work. Altogether, it is thus apparent that reach-out
behavior is driven by many inputs, of which interest in reaching out is one input. Overall, we
found correlational evidence that the underestimation of reach-out appreciation is one barrier to
interest in reaching out (supplemental study S3), but there was not much evidence in favor of the
intervention that we tested using a brief reach-out window paradigm in this study (i.e.,
supplemental study S4). Future research might examine other interventions and paradigms for
increasing reaching out.
THE SURPRISE OF REACHING OUT
21
WEB APPENDIX E
Exploratory Analyses on Attrition in Experiments 2, 3, and 5a
Experiments 2, 3, and 5a were dyadic studies in which we recruited initiators to reach-out
to someone in their lives and then contacted those people on their behalf, serving as responders.
Such dyadic reach-out studies thus involve attrition, based on the response rates of those who
were reached out to (experiment 2: 54% response rate from 54 responders to 100 valid initiators;
experiment 3: 74% response rate from 67 responders to 91 valid initiators; experiment 5a: 78%
response rate from 85 responders to 109 valid initiators; see footnote 5 in the main paper). Given
that attrition can lead to biased sampling, we thus conducted follow-up exploratory analyses to
examine if there were any systematic differences across the three experiments between valid
initiators in the retained dyads (i.e., the initiator’s chosen responder completed our study) and
valid initiators in the dyads that dropped out (i.e., the initiator’s chosen responder was sent a
recruitment email but did not complete our study). Our exploratory analyses below compare the
valid initiators whose responders did versus did not complete our study (i.e., experiment 2: 54
initiators with complete dyads vs. 46 incomplete initiator-only data; experiment 3: 67 initiators
with complete dyads vs. 24 incomplete initiator-only data; experiment 5a: 85 initiators with
complete dyads vs. 24 incomplete initiator-only data) in terms of the word counts of initiators’
reach-out messages, their appreciation predictions, how much they thought about how pleasantly
surprised the responder would feel by their reach-out (experiment 5a only), and their
demographic characteristics (age, gender).
Experiment 2 Exploratory Analyses of Attrition
THE SURPRISE OF REACHING OUT
22
First, we examined whether attrition status predicts the word count of the initiator’s note.
We first log-transformed the word count of the initiator’s note because the raw distribution was
positively skewed, and we then conducted a one-way ANOVA with attrition status as the
independent variable and the log transformed word count as the dependent variable, which
revealed no significant difference between the notes written by initiators in retained dyads (Mraw
= 26.19, SD = 21.43; Mtransformed = 1.27, SD = .44, n = 54) and initiators in dyads that dropped out
(Mraw = 30.93, SD = 21.22; Mtransformed = 1.36, SD = .41, n = 46), F(1, 98) = 1.08, p = .302.
Second, we conducted a one-way ANOVA with attrition status as the independent variable and
predictions of responder appreciation as the dependent variable, which revealed no significant
difference in predictions of responder appreciation between initiators in dyads that remained in
the study (M = 5.57, SD = 1.35) and initiators in dyads that dropped out (M = 5.11, SD = 1.49),
F(1, 98) = 2.67, p = .106. Third, we conducted a one-way ANOVA with attrition status as the
independent variable and age as the dependent variable, which revealed no significant difference
in age between initiators in dyads that remained in the study (M = 19.61, SD = 1.17) and
initiators in dyads that dropped out (M = 19.89, SD = 1.37), F(1, 98) = 1.22, p = .273. Finally, a
chi-square test of independence revealed no significant relationship between attrition and
gender
17
(Retained: 44.2% male, 55.8% female; Dropped off: 39.1% male, 60.9% female),
𝒳"
(1,
N = 98) = .26, p = .609.
Experiment 3 Exploratory Analyses of Attrition
First, we examined whether the word count of the initiator’s note differed across attrition
status. We log-transformed the word count of the initiator’s note because the raw distribution
was positively skewed, and then we conducted a one-way ANOVA with attrition status as the
17
For this exploratory analysis, we excluded the two participants who reported gender as “other” due to small cell
sizes.
THE SURPRISE OF REACHING OUT
23
independent variable and the log transformed word count as the dependent variable, which
revealed that there was no significant difference in the word counts of notes written by initiators
in retained dyads (Mraw = 25.51, SD = 17.85; Mtransformed = 1.28, SD = .38, n = 67) and initiators in
dyads that dropped out (Mraw = 20.38, SD = 13.13; Mtransformed = 1.20, SD = .35, n = 24), F(1, 89)
= .895, p = .347. Second, a one-way ANOVA with attrition status as the independent variable
and predictions of responder appreciation as the dependent variable revealed that there was no
significant difference in responder appreciation predictions between initiators in dyads that
remained in the study (M = 4.25, SD = .86) and initiators in dyads that dropped out (M = 3.92,
SD = 1.05), F(1, 89) = 2.40, p = .125. Third, a one-way ANOVA also revealed no significant
difference in age between initiators in dyads that remained in the study (M = 20.49, SD = 1.04)
and initiators in dyads that dropped out (M = 20.71, SD = 1.12), F(1, 89) = .735, p = .394.
Finally, a chi-square test of independence revealed a non-significant (though marginal)
relationship between attrition and gender (Retained: 25.4% male, 74.6% female; Dropped off:
45.8% male, 54.2% female),
𝒳"
(1, N = 91) = 3.47, p = .062.
Experiment 5a Exploratory Analyses of Attrition
First, we examined whether the word count of the initiator’s note differed by attrition
status. We log-transformed the word count of the initiator’s note because the raw distribution
was positively skewed, and then we conducted a one-way ANOVA with attrition status as the
independent variable and the log transformed word count as the dependent variable, which
revealed that there was no significant difference in the word counts of notes written by initiators
in retained dyads (Mraw = 29.80, SD = 33.78; Mtransformed = 1.28, SD = .45, n = 85) and initiators in
dyads that dropped out (Mraw = 22.58, SD = 19.03; Mtransformed = 1.18, SD = .43, n = 24), F(1, 107)
= .89, p = .348. Second, a one-way ANOVA with attrition status as the independent variable and
THE SURPRISE OF REACHING OUT
24
predictions of responder appreciation as the dependent variable revealed that there was no
significant difference in responder appreciation predictions between initiators in dyads that
remained in the study (M = 3.91, SD = .95) and initiators in dyads that dropped out (M = 4.08,
SD = 1.02), F(1, 107) = .64, p = .427. Third, a one-way ANOVA with attrition status as the
independent variable and the extent to which initiators thought about how pleasantly surprised
the responder would feel by their reach-out as the dependent variable revealed that there was no
significant difference in focus on the responder’s pleasant surprise between initiators in dyads
that remained in the study (M = 3.62, SD = 1.08) and initiators in dyads that dropped out (M =
3.46, SD = 1.29), F(1, 107) = .40, p = .527. Fourth, a one-way ANOVA also revealed no
significant difference in age between initiators in dyads that remained in the study (M = 20.04,
SD = 1.17) and initiators in dyads that dropped out (M = 20.38, SD = 1.64), F(1, 107) = 1.31, p =
.255. Finally, a chi-square test of independence revealed no significant relationship between
attrition status and gender
18
(Retained: 33.3% male, 66.7% female; Dropped off: 50.0% male,
50.0% female),
𝒳"
(1, N = 108) = 2.22, p = .136.
Summary of Exploratory Analyses
Overall, one exploratory comparison (i.e., gender) was marginally significantly different
in experiment 3 but not in experiment 2 or in experiment 5a. These exploratory analyses thus
suggest that there are not consistent biased sampling effects that account for the underestimation
of appreciation effect.
18
For this exploratory analysis, we excluded one participant who reported gender as “other” due to small cell sizes.
THE SURPRISE OF REACHING OUT
25
WEB APPENDIX F
Pre-registered Linear Regression Analysis in Experiment 6
As pre-registered, we conducted a linear regression with role (initiator vs. responder),
reach-out context (surprising vs. unsurprising), and their interaction on the appreciation index,
which revealed a significant interaction (B = .40, SE = .08, 95% CI [.23, .56]; t = 4.83, p < .001).
Given that the interaction was significant, we conducted simple-effects tests to examine the
pattern of the interaction. In the surprising reach-out context, we replicated the underestimation
of appreciation effect: Initiators significantly underestimated responders’ appreciation of being
reached out to (Minitiator = 4.14, SD = .81 vs. Mresponder = 4.46, SD = .67; B = -.32, SE = .06, 95%
CI [-.44, -.21], t = -5.55, p < .001, d = .43). By contrast, in the unsurprising reach-out context, the
underestimation of appreciation effect was eliminated (Minitiator = 3.93, SD = .85 vs. Mresponder =
3.86, SD = .90; B = .07, SE = .06, 95% CI [-.04, .19], t = 1.28, p = .202, d = .08).
THE SURPRISE OF REACHING OUT
26
References
Fiedler, K., Harris, C., & Schott, M. (2018). Unwarranted inferences from statistical mediation
tests–An analysis of articles published in 2015. Journal of Experimental Social
Psychology, 75, 95-102. https://doi.org/10.1016/j.jesp.2017.11.008
Fiedler, K., Schott, M., & Meiser, T. (2011). What mediation analysis can (not) do. Journal of
Experimental Social Psychology, 47(6), 1231-1236.
https://doi.org/10.1016/j.jesp.2011.05.007
Hayes, A. F. (2013). Methodology in the social sciences. Introduction to mediation, moderation,
and conditional process analysis: A regression-based approach. Guilford Press.
Yzerbyt, V., Muller, D., Batailler, C., & Judd, C. M. (2018). New recommendations for testing
indirect effects in mediational models: The need to report and test component
paths. Journal of Personality and Social Psychology, 115(6), 929–
943. https://doi.org/10.1037/pspa0000132
Zhao, X., & Epley, N. (2021). Insufficiently complimentary?: Underestimating the positive
impact of compliments creates a barrier to expressing them. Journal of Personality and
Social Psychology, 121(2), 239–256. https://doi.org/10.1037/pspa0000277
A preview of this full-text is provided by American Psychological Association.
Content available from Journal of Personality and Social Psychology
This content is subject to copyright. Terms and conditions apply.
