Content uploaded by Taya R. Cohen
Author content
All content in this area was uploaded by Taya R. Cohen on Jul 18, 2019
Content may be subject to copyright.
YOU CAN HANDLE THE TRUTH 1
YOU CAN HANDLE THE TRUTH:
MISPREDICTING THE CONSEQUENCES OF HONEST COMMUNICATION
Emma Levinea & Taya R. Cohenb
aThe University of Chicago
bCarnegie Mellon University
***Forthcoming at the Journal of Experiment Psychology: General***
Levine, E. E., & Cohen, T. R. (in press, 2018). You can handle the truth: Mispredicting the
consequences of honest communication. Journal of Experimental Psychology: General.
This paper is not the copy of record and may not exactly replicate the authoritative document
published in the APA journal.
This research is based on a chapter of the first author’s dissertation. We are grateful for
feedback and advice from Maurice Schweitzer, Katherine Milkman, Andrew Carton, Adam
Grant, and Li Jiang, as well as research assistance from Soaham Bharti, Lucy Huang, Adam
Picker, Barbara Toizer, and Emory Richardson. Studies were conducted at the University of
Pennsylvania (the first author’s former institution), the University of Chicago, and Carnegie
Mellon University. We are grateful for financial support and resources provided by the William
S. Fishman Faculty Research Fund at the University of Chicago Booth School of Business, The
Wharton Behavioral Lab, the Center for Behavioral and Decision Research at Carnegie Mellon
University, and a Russell Sage Foundation Behavioral Economics small grant. This research was
presented at the 2017 Academy of Management Annual Conference and the 2017 International
Association of Conflict Management Annual Conference.
Address correspondence to Emma Levine, The University of Chicago Booth School of
Business, 5807 S Woodlawn Ave, Chicago, IL 60637. Email: Emma.Levine@chicagobooth.edu,
Phone: 773-834-2861.
Total word count: 21,866
YOU CAN HANDLE THE TRUTH 2
YOU CAN HANDLE THE TRUTH:
MISPREDICTING THE CONSEQUENCES OF HONEST COMMUNICATION
ABSTRACT
People highly value the moral principle of honesty, and yet, they frequently avoid being
honest with others. In the present research, we explore the actual and predicted consequences of
honesty in everyday life. We utilize field and laboratory experiments that feature two types of
honesty interventions: 1) instructing individuals to focus on complete honesty across their
interactions for a period of time, and 2) instructing individuals to engage in specific honest
conversations that they frequently avoid in everyday life. In Studies 1a and 1b, we randomly
assigned individuals to either be (or imagine being) honest, kind, or conscious of their
communication in every conversation with every person in their life for three days. We find that
people significantly mispredict the consequences of honesty: focusing on honesty (but not
kindness or communication-consciousness) is more pleasurable, meaningful, socially connecting,
and does less relational harm than individuals expect. We extend our investigation by examining
the consequences of specific well-controlled honest conversations for both communicators and
their relational partners in two preregistered laboratory experiments. In Study 2 we examine the
predicted and actual consequences of honestly disclosing personal information, and in Study 3
we examine the predicted and actual consequences of honestly sharing negative feedback. Our
results suggest that individuals broadly misunderstand the consequences of increased honesty
because they overestimate how negatively others will react to their honesty. Overall, this
research contributes to our understanding of affective forecasting processes and uncovers
fundamental insights on how communication and moral values shape well-being.
Abstract word count: 247
Key words: honesty, affective forecasting, ethics, communication, well-being, social connection
YOU CAN HANDLE THE TRUTH 3
YOU CAN HANDLE THE TRUTH:
MISPREDICTING THE CONSEQUENCES OF HONEST COMMUNICATION
Honesty is one of the most fundamental moral values in human life. Honesty is among
the most important traits for interpersonal judgment (Anderson, 1968; Goodwin, Piazza, &
Rozin, 2014) and dominates philosophical and religious teachings across time and cultures.
Given the importance of honesty to individuals’ sense of self and morality, why do individuals so
frequently avoid being honest with others? One reason is that being honest often feels unkind or
uncomfortable. Even seemingly mundane questions, such as “How are you?” can trigger a
difficult conversation if a communicator believes that answering the question honestly might be
uncomfortable for them or another party.
People routinely face difficult conversations – conversations in which honesty could be
aversive (Stone, Patton, & Heen, 2010). They face these conversations in their personal lives
when sharing their feelings with friends and family members, and in their professional lives
when deciding how to deliver negative news and critical feedback. Though these difficult
conversations are part of everyday life, navigating them can elicit distress and anxiety (e.g.,
Margolis & Molinsky, 2008; Molinsky & Margolis, 2005). As such, individuals often choose to
be dishonest during difficult conversations, or avoid engaging in such conversations altogether,
and instead focus on being pleasant and creating smooth social interactions (Lee, 1993; Rosen &
Tesser, 1970; Tesser, Rosen, & Tesser, 1971).
Is forgoing honesty a good decision? On the one hand, it is possible that avoiding honesty
– particularly during difficult conversations – promotes social connection and leads to more
enjoyable interactions. On the other hand, it is possible that individuals expect honesty to be far
YOU CAN HANDLE THE TRUTH 4
more unpleasant and socially damaging than it truly is. And, perhaps being completely honest
also yields unexpected benefits, such as a heightened sense of meaning.
In one field experiment (as well as a pilot field experiment reported in the online
supplemental materials) and two preregistered laboratory experiments, we examine the predicted
and actual effects of honesty on psychological well-being and social connection. We examine the
consequences of two types of honesty interventions: 1) instructing individuals to focus on
complete honesty across their interactions for a period of time, and 2) instructing individuals to
engage in specific honest conversations that they frequently avoid in everyday life. We compare
honest communication to kind communication, as well as a neutral control condition, and we
examine how honesty influences three distinct facets of well-being: hedonic well-being,
eudaimonic well-being, and social connection. Hedonic well-being is characterized by pleasure,
enjoyment, and happiness. In the hedonic view, well-being consists of the presence of pleasure
and the absence of pain (Ryan & Deci, 2001). Eudaimonic well-being is characterized by
meaning, fulfillment, and individual autonomy. In the eudaimonic view, well-being consists of
the actualization of human potentials, rather than pleasure (Waterman, 1990, 1993). Social
connection, in addition to one’s sense of pleasure and meaning, is an important component of
overall physical and subjective well-being (Helliwell & Putnam, 2004; Myers & Diener, 1995).
This research has important theoretical and practical contributions. First, we document
the psychological forces that push people away from communicating honestly. As will be
revealed through our studies, individuals find honest communication to be more enjoyable,
meaningful and socially connecting than they expect. Our results suggest that individuals
misunderstand the personal consequences of increased honesty because they misunderstand the
YOU CAN HANDLE THE TRUTH 5
social consequences of honesty. Specifically, communicators assume that others will react more
negatively to honest conversations than they actually do.
Our studies also reveal that individuals’ aversion to honesty may be misguided. The
choice to be less than fully honest with others is driven by individuals’ inaccurate expectations of
how unenjoyable honest conversations will be. However, individuals’ post hoc appreciation for
honest conversations is driven by how meaningful those conversations are. These results suggest
that affective forecasting failures can lead individuals to miss out on meaningful experiences.
Additionally, by doing a deep dive into the intrapersonal and interpersonal consequences
of honesty, this work helps to bridge the gap between our understanding of normative and
behavioral ethics. For centuries, ethicists have touted the moral significance of different virtuous
behaviors, but psychologists have only recently begun to examine the experience and
consequences of enacting or violating these virtues (Aknin et al., 2013; Aknin, Dunn, & Norton,
2012; Dunn, Aknin, & Norton, 2008; Emmons & McCullough, 2003; Gino, Kouchaki, &
Galinsky, 2015; Hofmann, Wisneski, Brandt, & Skitka, 2014; Lyubomirsky, Shelden, &
Schkade, 2005; Rudd, Norton, & Aaker, 2014). To our knowledge, our work is the first to
examine how specific moral principles and styles of communication influence all three
fundamental forms of well-being (hedonic, eudaimonic, and social connection).
Dening Honesty
A key reason that being honest during difficult conversations is a source of stress for so
many people is that there is uncertainty about how one’s conversational partner will react to the
information that is divulged. Rather than engage in such conversations, many people choose to
cloak their personal thoughts and feelings in social niceties or avoid having difficult
conversations altogether (Stone et al., 2010). The result is a glaring lack of honesty.
YOU CAN HANDLE THE TRUTH 6
We define honesty as speaking in accordance with one’s own beliefs, thoughts and
feelings. Our definition of honesty builds on a recent definition of authenticity by Gino,
Kouchaki, and Galinsky (2015), who define authenticity as acting “in accordance with one’s own
sense of self, emotions, and values.” According to Gino, Kouchaki, and Galinsky, inauthenticity
entails being untrue to oneself, whereas being dishonest entails being untrue to others. In other
words, authenticity and inauthenticity are not limited to communication, and do not require a
target. Conversely, we define honesty and dishonesty as forms of communication, which require
a target (i.e. listener).
We note three important facets about our definition of honesty. First, a communicator’s
beliefs, thoughts and feelings are independent of reality. That is, individuals can be honest by
communicating something that they believe to be true, even if that belief is, in fact, false. Thus,
individuals can be honest about an objective fact, but they can also be honest about their inner
experience and opinions (DePaulo & Bell, 1996). We are primarily interested in honesty about
individuals’ inner experiences (i.e., their private opinions and histories) in the present
manuscript. It is important to understand the consequences of honestly communicating one’s
inner experience because sharing this information is relatively discretionary (compared to
sharing objective negative news, such as layoffs), but is nonetheless the basis for important
social and organizational processes (e.g., self-disclosure and interpersonal feedback).
Second, we define honesty based on the content that a communicator shares, rather than
the tone with which they share that content. In other words, individuals can, but need not, be
honest in a blunt or harsh manner. Third, we conceptualize honesty as a communication style that
YOU CAN HANDLE THE TRUTH 7
involves openness and truth-telling, rather than a set of actions that involves the avoidance of
cheating and duplicity.1
Honesty, Pleasure, and Social Connection
One reason people may avoid being honest is that honesty often appears to conflict with
kindness: truthfully sharing one’s opinions and feelings can hurt others and create social tension
during difficult conversations. Human beings are hardwired to detect pain and suffering in others
(Craig, 2009), and consequently, they often prioritize the prevention of interpersonal harm over
other principles, such as honesty (Gray, Young, & Waytz, 2012; Haidt & Graham, 2007; Schein
& Gray, 2017). The primacy of harm avoidance influences a host of moral and social decisions,
including everyday communication. Most people recognize that one goal of social
communication is to avoid embarrassing others and to help others “save face” (Goffman 1967),
which can lead people to be overly polite and avoid sharing negative opinions, rather than being
completely honest (DePaulo & Bell, 1996).
Individuals also prioritize harm avoidance when they judge others’ communication. For
example, individuals often judge those who tell prosocial lies (i.e., statements that prioritize
benevolence over honesty) as more ethical and trustworthy than those who are honest but harm
others (Levine & Schweitzer, 2014, 2015). Despite the strength of individuals’ intuitions about
the social harm associated with difficult truths, we know very little about when these intuitions
are correct or incorrect. Although recent research demonstrates that targets do indeed resent
communicators who speak truths that cause objective harm (Levine & Schweitzer, 2014, 2015),
much less is known about the perceived and actual harm associated with the truths that
individuals withhold in everyday life.
1 We note that we are not focused on reactions to the word “honesty” per se, but rather we
investigate the consequences of engaging in more honest communication than people typically
do (see Study S1 in our online supplement for more on this distinction).
YOU CAN HANDLE THE TRUTH 8
The belief that honesty during difficult conversations will cause interpersonal harm leads
communicators to worry about how enacting this harm will affect their own emotional state and
their relationships. When individuals are faced with the prospect of harming others, they often
anticipate feeling negative emotions such as guilt and shame (Cohen, Wolf, Panter, & Insko,
2011; Tangney, Stuewig, & Mashek, 2007). This also occurs in the domain of communication.
For example, individuals who have to deliver bad news to others often avoid doing so because
they feel guilty about not sharing the person’s fate (Heider, 1958; Tesser & Rosen, 1972). In their
classic research on the “MUM effect” (keeping mum about undesirable information), Tesser and
Rosen (1972) found that individuals were reluctant to tell another (fictitious) participant that they
would receive painful electric shocks during an experiment, particularly when the
communicators did not share the same bad fate. Work by the same author team revealed that the
reluctance to deliver bad news persists even in situations in which communicators and targets
have no preexisting relationships and anticipate no future interactions, and when the information
that needs to be delivered is objective and urgent (Rosen & Tesser, 1970). Tesser and Rosen
interpret these results as suggesting that the MUM effect occurs because communicators
anticipate internal discomfort rather than relational costs (Tesser & Rosen, 1972).
Related research, however, argues that relational costs do play a role. For example, Bond
and Anderson (1987) demonstrate that individuals only exhibit the MUM effect when they are
visible to the target, suggesting that the reluctance to share negative information is not driven
solely by one’s internal emotions, but rather a concern about how the other person will react and
judge the communicator. Similarly, within organizations, Fisher (1979) finds that individuals
exaggerate the positivity of feedback to poor performers, which is mediated by the belief that
poor performers will like them less. Research within close relationships also demonstrates that
YOU CAN HANDLE THE TRUTH 9
friends avoid telling each other that they do not like their friend’s romantic partner because they
fear harming their relationship with their friend (Mayer, 1957) and that they avoid honestly
disclosing personal information (i.e., self-disclosure), in part, because they worry about others’
judgments and social distance (Rosenfeld, 1979). Across contexts, individuals also tend to avoid
honestly sharing their opinions in the presence of others who do not share these views (e.g.,
Tetlock, 1985; Tetlock, Skitka, & Boettger, 1989), presumably because they fear interpersonal
judgment.
Despite the large literature on the anticipated costs of honesty during difficult
conversations, little is known about whether these anticipated costs are accurate. That is, existing
research demonstrates that people avoid being honest because they associate honesty with
interpersonal harm and discomfort, and as a result, expect to incur affective (i.e., hedonic) and
social costs when they are completely honest. Thus far, however it is unclear whether individuals
are well-calibrated in their estimation of such costs. Does honesty cause as much interpersonal
pain and tension as people expect? Is it unpleasant and socially isolating to be honest with
others? It is possible that communicators overestimate the costs of honesty, and thus, could
afford to be more honest than they are without suffering negative affective or social
consequences. On the other hand, it is possible that individuals understand the costs of honesty
with reasonable accuracy and their reluctance to engage in difficult conversation reflects a
rational cost-benefit calculation. In the present research, we examine these possibilities.
We hypothesize that communicators overestimate both the hedonic and social costs of
honesty. Research on emotion-regulation, secrecy, and the performance of necessary evils
suggest that this may be the case. For example, individuals who honestly express their emotions
experience lower stress and blood pressure, and develop higher levels of intimacy than
YOU CAN HANDLE THE TRUTH 10
individuals who hide their emotions (Butler et al., 2003; Srivastava, Tamir, McGonigal, John, &
Gross 2009). On the other hand, individuals who harbor secrets have poorer health than
individuals who do not (Slepian, Chun, & Mason, 2017). Research on the experience of
performing “necessary evils,” such as delivering terminal prognoses or critical performance
feedback (e.g., Margolis & Molinsky, 2008; Molinsky & Margolis, 2005), also suggests that
communicators’ concerns about the costs of honesty might be overstated. This body of research
finds that despite prior assumptions that the prospect of harm-doing triggers psychological
disengagement (Bandura, 1999), many professionals who are taxed with communicating
unpleasant truths to others (e.g., doctors, managers, corrections officers) are able to
psychologically engage and connect with the targets of their communication. Furthermore,
communicating conflict directly rather than indirectly can in many circumstances help de-
escalate negative conflict spirals and improve interpersonal relationships, (Weingart, Behfar,
Bendersky, Todorova, & Jehn, 2015).
Theoretically, our prediction of an affective forecasting error about honesty is informed
by a growing body of research documenting other ways in which individuals mispredict how
their conversations will affect them. In particular, individuals frequently underestimate the
pleasure they derive from conversing with outgroup members (Mallet, Wilson, & Gilbert, 2008)
and strangers (Dunn, Biesanz, Human, & Finn, 2007; Epley & Shroeder, 2014). Mallet et al.
(2008) found that individuals mispredict the consequences of interacting with outgroup members
because they expect outgroup members to be more different from themselves than they actually
are. Epley and Schroeder (2014) found that communicators mistakenly predict that talking to a
stranger will be less pleasant than sitting in isolation, when in fact, connecting with a stranger is
more pleasant than sitting in isolation because people expect others to be less interested in
YOU CAN HANDLE THE TRUTH 11
connecting than they actually are. These papers document two reasons people systematically
mispredict the consequences of social interactions: 1) they misunderstand the nature of who they
will interact with, and 2) they misunderstand whether other individuals are interested in social
interaction in the first place.
We posit a third reason that people mispredict the pleasure of particular conversations:
they misunderstand how others will react to the content of the conversation. Specifically, we
hypothesize that individuals overestimate how negatively others will react when they share
honest information that is typically left unsaid (e.g., intimate personal information or
interpersonal critiques). This may happen for at least two reasons. First, it is possible that people
believe that the norm of not being completely honest with others reflects a preference. That is,
just as individuals’ infer that others are disinterested in talking to strangers from the norm of
silence among strangers (Epley & Schroeder, 2014), people may infer that others are uninterested
in – or worse yet, would be offended by – hearing their intimate secrets or critical opinions,
based on norms of politeness and conflict avoidance. Furthermore, individuals may suffer from a
focusing illusion (e.g., Gilbert & Wilson, 2000; Loewenstein & Schkade, 1999; Wilson,
Wheatley, Meyers, Gilbert, & Axsom, 2000) and narrowly focus on the exact information shared,
rather than recognizing the broader context in which that information is shared. For example,
individuals may overestimate the costs of providing negative feedback because they assume the
listener will focus on the specific criticism provided, rather than the speaker’s intent or the
relationship in which the feedback is shared. Either, or both, of these processes could give rise to
systematic mispredictions of others’ reactions to honesty.
The central purpose of the current research is to examine whether individuals’
assumptions regarding the hedonic and social costs of increased honesty in everyday life are
YOU CAN HANDLE THE TRUTH 12
misguided. We hypothesize that they are, and we explore whether mispredictions of others’
reactions to honesty underlie this error.
Honesty and Meaning
A secondary goal of this research is to broadly examine the consequences of honest
communication. Thus, in addition to examining the relationship between honesty and hedonic
well-being and social connection, we consider how honesty influences one’s sense of meaning,
or eudaimonic well-being. To communicate honestly, individuals must look inwards and consult
their personal feelings and opinions. This process may increase self-actualization and produce
feelings of personal control and autonomy, key components of eudaimonia (Deci & Ryan, 2008).
Research on the experience of inauthenticity is consistent with this proposition. Behaving
inauthentically by misrepresenting one’s emotions or by conforming to social norms that are
inconsistent with one’s personal beliefs lowers moral self-regard and one’s sense of moral purity
(Gino et al., 2015). Moral identity is closely linked to sense of self (Aquino & Reed, 2002), thus
decrements in moral identity may undermine one’s sense of meaning and purpose as well (i.e.,
eudaimonic well-being). Consequently, we expect honesty to increase eudaimonic well-being.
We did not have any a priori predictions about whether individuals would accurately
predict the relationship between honesty and eudaimonic well-being, thus this aspect of the
current investigation is more exploratory and inductive. Nonetheless, examining eudaimonic
well-being forecasts provides insight into individuals’ relative ability to predict hedonic versus
eudaimonic outcomes, thus shedding new light on the types of outcomes individuals accurately
and inaccurately forecast.
Overview of Studies
YOU CAN HANDLE THE TRUTH 13
We rely on field experiments and laboratory experiments to examine the predicted and
actual consequences of increased honesty in interpersonal communication. First, we conducted a
pilot field experiment in which we randomly assigned 117 participants to be completely honest,
kind, or conscious of their communication (the control condition) in every social interaction for
three days. Participants in this study forecasted the experience and then judged the experience
every day for three days. Because Study 1a replicates the main results of the Pilot Study, we
report this Pilot Study in our online supplemental materials only.
Our Pilot Study informed our first field experiment (Study 1a). Study 1a featured the
same experimental manipulations that we used in our Pilot Study (i.e., we randomly assigned
participants to be completely honest, kind, or conscious of their communication for three days)
but we refined our protocol and our measures. We complemented the field experiment with a
forecasting experiment (Study 1b) in which a separate sample of participants made predictions
about the experience of being honest, kind, or conscious of their communication in every social
interaction for three days.
In Study 1, we compare the consequences of focusing on honesty, focusing on kindness,
and being conscious of one’s communication (our neutral control condition) for several reasons.
First, we wanted to examine whether individuals mispredict the consequences of honesty in
particular, or if they mispredict the consequences of focusing on any moral principle when
communicating. We chose kindness as a comparison because it is the natural counterpart to
honesty during difficult conversations, the context in which we expect individuals to avoid
honesty most frequently. Second, the kindness and neutral control conditions in our field
experiments allow us to draw conclusions about the absolute consequences of honesty, consistent
with our secondary research goal. By including a neutral control condition in the research design,
YOU CAN HANDLE THE TRUTH 14
we can examine how honesty influences well-being above and beyond participants’ reactions to
participating in the study itself. Note, however, that the control condition does not perfectly
reflect individuals’ routine communication (i.e., in the absence of an intervention).
Two weeks after the field experiment ended, participants responded to open-ended
questions about the nature of their interactions during the experiment and the consequences
thereof. To better understand the mechanisms underlying our effects and the nature of
participants’ experiences, we analyzed participants’ open-ended responses (Study 1c). These
analyses suggest that focusing on honesty increased the frequency with which individuals
engaged in difficult conversations and genuine self-expression in their everyday lives, and that
individuals were surprised that others reacted more positively to their honesty than they
anticipated. To extend these initial findings, in Studies 2 and 3, we examine the predicted and
actual consequences of engaging in specific, well-controlled, honest conversations with pairs of
close relational partners. In Study 2, individuals had an honest conversation about a series of
difficult topics with a close relational partner, and in Study 3, individuals honestly provided
negative interpersonal feedback to a close relational partner. In Study 3, we also test whether
communicators’ (mis)predictions about the personal consequences of honesty are mediated by
their (mis)predictions about others’ reactions to honesty. Across our studies, we decided our
sample sizes in advance, and we report all manipulations and measures we collected. The
institutional review board (IRB) of the University of Pennsylvania approved all aspects of the
Pilot Study and Study 1, and the IRB of the University of Chicago approved all aspects of
Studies 2 and 3. Data and materials from our studies can be found on the Open Science
Framework: https://tinyurl.com/y8kaj356.2
2 Due to privacy concerns, we stripped participants’ open-ended responses (about their difficult
conversations) from our posted data sets. These data are available from the authors upon request.
YOU CAN HANDLE THE TRUTH 15
Study 1
Study 1 contained two separate samples: Experiencers (Study 1a) and Forecasters (Study
1b). Experiencers were randomly assigned to be completely honest, kind, or conscious of their
communication in every interaction for three days. Forecasters made predictions about all three
experiences, but did not actually engage in the experience. Because our hypotheses involve the
contrast between forecasts and experiences and we only collected forecasts in Study 1b, we first
describe the methods of both studies, and then present our analyses comparing Study 1a to 1b.
To test the feasibility of the design and explore initial predictions, we ran a very similar
Pilot Study prior to running Study 1. The main difference between the Pilot Study and Study 1
was that the Pilot Study included additional exploratory measures and featured a within-subjects
design, such that participants made forecasts about the three-day communication study before
engaging in it. The Pilot Study and Study 1 yielded nearly identical results. We report full
information about the Pilot Study in our online supplemental materials.
Study 1a: Experiencers
Sample. Our goal was to recruit 50 participants per cell in Study 1 based on our Pilot
Study and recommended research practices (Simmons, Nelson, & Simonsohn, 2013). One-
hundred fifty-seven adults initially enrolled in our study, 154 of whom completed at least one
daily survey (98.1%; 70.8% female3, mean age = 22), and 144 of whom completed the two-week
follow-up (91.7%).
3 We conducted exploratory analyses in every study to examine potential gender effects.
Although gender occasionally interacts with our manipulations, we do not find consistent
patterns of gender effects across studies. Because we did not develop hypotheses about gender,
and because the findings were not consistent across the studies, we do not discuss gender further.
Further information about gender in our studies is available on the Open Science Framework:
https://tinyurl.com/y8kaj356.
YOU CAN HANDLE THE TRUTH 16
Method. We ran Study 1a on three consecutive Thursdays during March and April of
2016. We conducted the study over consecutive weeks so that all participants would participate
in the study for two weekdays (Thursday, Friday) and one weekend day (Saturday), in order to
maximize the number of different types of relational partners participants interacted with.
We advertised Study 1 to a panel of adults and students in Philadelphia, Pennsylvania. All
individuals who had ever completed a study at this laboratory received an email informing them
that a new study was available and that to learn more, they should complete a short introductory
survey. The introductory survey contained information about a study called the “Challenging
Exercise” study, which was an optional 3-day experiment that would challenge the way they
communicate with others. In exchange for their participation, participants would earn $20 and
the chance to win an iPad mini. Participants were informed of the time commitment of the study
and the potential distress that could be caused by participating. However, they were not provided
any information about the experimental conditions at this time.
Participants had to provide their participant ID to indicate their interest and that they
understood the nature of the study. Participants who provided their participant ID were eligible to
sign up for the study online through the laboratory’s online portal.
Assignment to condition. Participants who signed up for the study were required to come
to the laboratory to receive specific instructions about the study. When participants arrived at the
lab, they were assigned to one of three experimental conditions: honesty, kindness, or a control
condition. We randomized condition at the session-level. That is, each session of participants
(i.e., the group of participants that arrived at the lab during the same time) was assigned to the
same condition. Participants learned about the experimental condition verbally, and had the
opportunity to ask questions. Because of the verbal instructions, it was necessary to have each
YOU CAN HANDLE THE TRUTH 17
session of participants assigned to a single condition. There were 30 sessions in all; 10 honesty
sessions, 10 kindness sessions, and 10 control sessions.
A trained research assistant, who was blind to our hypotheses, first provided some basic
information about the study and prompted individuals to consider the tradeoff between honesty
and kindness during difficult conversations. Then s/he instructed participants how to behave for
the next three days according to their experimental condition and invited questions from
participants to ensure their understanding of the protocol.
Specifically, the research assistant explained:
[Honesty condition]
Throughout the next three days – that means today, tomorrow, and the following day - please
strive to be absolutely honest in every conversation you have with every person you talk to.
Really try to be completely candid and open when you are sharing your thoughts, feelings, and
opinions with others. You should be honest in every conversation you have, in every interaction,
with every person in your life. Even though this may be difficult, you should do your absolute
best to be honest.
[Kindness condition]
Throughout the next three days – that means today, tomorrow, and the following day - please
strive to be kind in every conversation you have with every person you talk to. Really try to be
caring and considerate when you are sharing your thoughts, feelings, and opinions. You should
be kind in every conversation you have, in every interaction, with every person in your life. Even
though this may be difficult, you should do your absolute best to be kind.
[Communication-consciousness– Control condition]
Throughout the next three days – that means today, tomorrow, and the following day - please be
conscious of the way you communicate with others. Please act as you normally would
throughout the length of this study. You should not change your behavior, but you should be
conscious of it.
Participants were instructed not to tell anyone about the experiment, including their
relational and conversational partners. We include the full script for each condition in Appendix
A.
YOU CAN HANDLE THE TRUTH 18
Participants were then directed to a computerized survey. Within the survey, participants
first provided demographic information. Then, they re-read the instructions associated with their
condition and responded to a one-item comprehension check, asking them what their goal in the
study was (response-options: “To be honest in all of my communication”, “To be kind in all my
communication,” or “To communicate as I normally do, but be conscious of my
communication”). Participants had to answer the comprehension check correctly to proceed with
the study.
Next, participants provided their email address to indicate their continued consent, and to
allow us to contact them with their nightly surveys. Finally, participants were asked to confirm
their commitment to the study by typing the following statement into the survey, “For the next
three days, I will [communicate honestly, communicate kindly, be conscious of my
communication].”
Before leaving the laboratory, participants were reminded of their study condition and
were instructed to begin the study immediately. Participants were told that they would receive
their first nightly survey that evening at 6pm. Participants had to say aloud, “I agree to
participate.” Upon leaving the laboratory, participants received a small card that read
“Communicate Honestly,” “Communicate Kindly,” or “Be Conscious of Your Communication”
and the dates of the study, and directed participants to email the experimenter if they had any
questions.
Nightly surveys. We tracked participants’ behavior over three consecutive days. We
emailed participants a nightly survey on Thursday and Friday at 6pm, and on Saturday at 12pm.
We emailed participants earlier in the day on Saturday to ensure they would not miss the survey.
YOU CAN HANDLE THE TRUTH 19
When completing the nightly surveys, participants first responded to several open-ended
questions. We list all questions that we asked and discuss the coding and results associated with
these questions in the online supplemental materials.
Then, participants responded to our focal measures. Using five-point bipolar rating
scales, participants rated the degree to which their experience in the study that day was difficult-
easy, unpleasant-pleasant, unenjoyable-enjoyable, stressful-relaxing, meaningless-meaningful,
constraining-liberating, unfulfilling-fulfilling, futile-enriching, socially isolating-socially
connecting, and alienating-uniting. We combined the first four items into a single measure of
enjoyment (daily αs > .82; this is our measure of hedonic well-being). We combined the middle
four items (meaningful, liberating, fulfilling, enriching) into a single measure of meaning (daily
αs > .77; this is our measure of eudaimonic well-being). We combined the final two items into a
single measure of social connection (daily rs > .60). Our measures of hedonic and eudaimonic
well-being are consistent with existing research (e.g., Huta & Ryan, 2010; 2015) that
conceptualizes the hedonic motives as seeking enjoyment, pleasure, and relaxation, and
eudaimonic motives seeking to learn something or do what you believe in.
Next, participants rated their agreement (1 = strongly disagree, 7 = strongly agree) with
two manipulation check items: “I was completely honest and candid in every conversation I had
today” and “I was kind and compassionate during every conversation I had today.” Finally, we
asked participants to reflect on their experience that day and to explain how they either did or did
not comply with the experiment. We also asked them to write about any challenges they faced
and how it felt to focus on [honesty, kindness, their communication]. Participants were given the
lead experimenter’s email address and were invited to reach out to her with questions or concerns
at any time.
YOU CAN HANDLE THE TRUTH 20
Reflection survey. Two weeks after participants completed the third and final day of the
experiment, they were emailed a final reflection survey. Participants first responded to several
open-ended questions, asking them what they learned, how their behavior and communication
had changed, what difficulties they had, any surprises they faced, and how their relationships
changed. We provide more details about these open-ended questions in Study 1c and Appendix
B.
In the two-week follow-up survey, participants also rated their agreement with 23
statements intended to assess the degree to which the experiment had lasting impact on their
honesty and kindness, improved their hedonic and eudaimonic well-being, as well as improved
and harmed their relationships. Importantly, participants also assessed the degree to which they
appreciated the experience. We provide all items and the scale reliabilities in Appendix C (all rs
> .71, αs > .91).
Before exiting the survey, participants indicated whether they would prefer their $20
payment via PayPal, Venmo or by receiving an Amazon.com gift card. We randomly selected one
participant to win the iPad mini and we compensated all participants within one week.
--- Table 1 about here ---
Study 1b: Forecasters
Sample. We ran Study 1b shortly after we ran Study 1a, recruiting from the same
participation pool, with the stipulation that participants who completed Study 1a were not
eligible to complete Study 1b. Two-hundred twenty-three participants (65.0% female, mean age
= 22) participated in Study 1b and were included in our analyses. An additional four participants
completed the survey, but indicated they had already participated in Study 1a (despite passing
YOU CAN HANDLE THE TRUTH 21
our initial screener). These participants were deleted from the dataset before any analyses were
conducted.
Method. Following the procedure of Epley and Schroeder (2014), we included the same
conditions in the forecasting study (Study 1b) as we included in the experience study (Study 1a),
but we manipulated the conditions within-, rather than between-, subjects. Specifically,
participants learned about an experiment that was taking place, called “the Challenging
Exercise” study. We described the protocol of the Challenging Exercise study (Study 1a) very
closely. Participants learned that individuals who enrolled in the Challenging Exercise study
would have to make modifications to their communication for three days and complete nightly
surveys, and that the experience might cause discomfort. Then, all participants learned about all
three conditions of the study—honesty, kindness, and communication-consciousness—and read
the exact instructions that participants in the Challenging Exercise study (Study 1a) actually
received. After reading about each of the conditions of the study, participants made a
hypothetical choice and selected the one condition they would want to participate in. Note that
participants did not actually engage in the Challenging Exercise study (i.e., the three-day
communication interventions).
Then, participants were asked to imagine participating in each condition of the study and
to imagine being honest [being kind, being conscious of their communication] for three days.
Participants forecasted their level of enjoyment, social connection, and meaning in each of the
experimental conditions using the same items we used in Study 1a (all rs > .65, αs > .80).
Next, we asked participants to imagine that they had completed the study and were
evaluating the experience of being honest [being kind, being conscious of their communication]
two weeks later. Participants then responded to the same 23 items we collected during the
YOU CAN HANDLE THE TRUTH 22
follow-up reflection survey of Study 1a (see Appendix B; all rs > .45, αs > .73). Then, we asked
participants to imagine they actually had to participate in one condition in the study. Participants
selected the one condition they would want to participate in. Participants provided demographic
information before they were dismissed.
Results
Analytical approach. Within each construct (e.g., enjoyment, social connection,
meaning), we present Forecasters’ predictions (Study 1b), followed by Experiencers’ actual
ratings (Study 1a). To analyze Experiencers’ actual ratings in Study 1a we created daily average
variables by taking the three-day average of the respective dependent variables (e.g., daily
enjoyment, social connection, and meaning) that we collected during the experiment. We conduct
our manipulation check and experienced enjoyment, social connection, and meaning analyses
using all participants who responded to at least one daily survey and we conduct our subjective
long-term impact analyses using all participants who completed at least one daily survey as well
as completed the two-week follow-up. Table 1 depicts the number and percentage of participants
who began and completed each stage of the experiment across our conditions. To check the
robustness of our results, we also replicated our findings using missing data analyses with five
sets of multiple imputation data, and none of our results are substantively different when
analyzing the raw data compared to the imputed data.
We present the results of planned t-tests between predictions (Study 1b) and actual ratings
(Study 1a) of each construct, within each condition. Finally, we examine Forecasters’ choices.
We present all descriptive statistics, as well as all non-focal results (daily manipulation checks,
Long-term honesty and kindness) in Table 2.
---Tables 1 and 2 about here---
YOU CAN HANDLE THE TRUTH 23
Daily experience measures.
Enjoyment.
Forecasted Enjoyment. A repeated measures ANOVA revealed a significant effect of
Condition, F(2,442) = 96.23, p < .001,
ηp
2
= .30. Forecasters (i.e., participants in Study 1b)
expected the Honesty condition to be less enjoyable than the Control and Kindness conditions (ts
> 9.48, ps < .001, ds > .85).
Actual Enjoyment. Actual enjoyment from Experiencers (i.e., participants in Study 1a)
did not mirror the forecasts. A one-way ANOVA revealed a marginal effect of Condition,
F(2,151) = 2.76, p = .066,
ηp
2
= .035. Unlike the Forecasters, Experiencers rated the Honesty
condition as no more or less enjoyable than the Control or Kindness conditions (ps > .17). In
fact, the marginal trend was such that the Control condition was slightly less enjoyable than the
Honesty and Kindness conditions.
Consistent with our hypothesis, participants in Study 1b expected Honesty to be
significantly less enjoyable than participants in Study 1a actually experienced it to be (t(270) =
6.86, p < .001, d = 1.15). Importantly, Forecasters did not mispredict the enjoyment associated
with the Kindness or Control conditions (ps > .68). We depict these results in Figure 1.
---Figure 1 about here---
Social connection.
Forecasted social connection. A repeated measures ANOVA revealed a significant effect
of Condition, F(2,442) = 141.12, p < .001,
ηp
2
= .39. Forecasters (i.e., participants in Study
1b) expected the Honesty condition to be less socially connecting than the Control condition and
the Kindness condition (ts > 7.8, ps < .001, ds > .64).
YOU CAN HANDLE THE TRUTH 24
Actual social connection. Actual social connection did not mirror these forecasts. A one-
way ANOVA revealed a significant effect of Condition, F(2,151) = 4.89, p = .009,
ηp
2
= .061,
but unlike the forecasts, participants in the Honesty condition of Study 1a actually felt marginally
more socially connected than participants in the Control condition (t(94) =1.79, p = .076, d = .
36), and no more or less connected than participants in the Kindness condition (p = .20). We
depict these results in Figure 2.
Consistent with our predictions, participants in Study 1b expected Honesty to be
significantly less socially connecting than participants in Study 1a actually experienced it to be
(t(270) = 5.46, p < .001, d = 0.97). Interestingly, participants overestimated the social connection
associated with Kindness (t(278) = 2.87, p = .004, d = 0.45), but they did not mispredict the
social connection associated with the Control condition (p = .84). These results are shown in
Figure 2.
---Figure 2 about here---
Meaning.
Forecasted meaning. A repeated measures ANOVA revealed a marginal effect of
Condition, F(2,442) = 2.47, p = .086,
ηp
2
= .01. Forecasters (i.e., participants in Study 1b)
expected the Honesty condition to be less meaningful than the Kindness condition (t(221) = 2.12,
p = .035, d = .17) and no different than the Control condition (p = .67).
Actual meaning. A one-way ANOVA revealed a significant effect of Condition on
meaning, F(2,151) = 7.05, p = .001,
ηp
2
= .085, but unlike the forecasts, participants in the
Honesty condition experienced similar levels of meaning compared to participants in the
Kindness condition (p = .17), and greater meaning than participants in the Control condition
(t(94) = 3.70, p < .001, d = .79).
YOU CAN HANDLE THE TRUTH 25
Interestingly, participants in Study 1b expected Honesty to be somewhat less meaningful
than participants in Study 1a actually experienced it to be (t(270) = 1.76, p = .088, d = 0.30).
Participants did not mispredict the meaning associated with the Kindness condition (p = .44), and
they actually overestimated the meaning associated with the Control condition (t(268) = 2.09, p
= .037, d = 0.37). We depict these results in Figure 3.
Although Forecasters marginally underestimated the meaning associated with honesty
than the Experiencers reported, the magnitude of this misprediction was substantially smaller (d
= .30) than mispredictions of enjoyment (d = 1.12) and social connection (d = .97).
---Figure 3 about here---
Subjective long-term impact.
Long-term hedonic well-being.
Predicted long-term hedonic well-being. A one-way ANOVA revealed a significant effect
of Condition on predicted long-term enjoyment, F(2,422) = 52.54, p < .001,
ηp
2
= .19.
Forecasters (i.e., participants in Study 1b) expected to derive less long-term enjoyment in the
Honesty condition than in the Control and Kindness conditions (ts > 4.21, ps < .001, ds > .36).
Actual long-term hedonic well-being. Actual long-term hedonic well-being did not mirror
these forecasts. A one-way ANOVA revealed a significant effect of Condition on long-term
enjoyment, F(2,141) = 3.64, p = .029,
ηp
2
= .049, such that participants in the Honesty
condition in Study 1a derived marginally greater enjoyment than participants in the Control
condition (t(84) = 1.94, p = .054, d = .43), and similar levels of enjoyment as participants in the
Kindness condition (p = .59).
Further supporting our predictions, Forecasters (i.e., participants in Study 1b)
underestimated the long-term enjoyment associated with Honesty (t(264) = 2.26, p = .025, d =
YOU CAN HANDLE THE TRUTH 26
0.39). However, they actually overestimated the long-term enjoyment associated with the
Control condition (t(264) = 2.45, p = .015, d = 0.41) and the Kindness condition (t(279) = 2.47,
p = .014, d = 0.37).
Long-term relational improvement.
Predicted long-term relational improvement. A one-way ANOVA revealed a significant
effect of Condition on predicted long-term relational improvement, F(2,422) = 26.82, p < .001,
ηp
2
= .11. Forecasters (i.e., participants in Study 1b) expected to improve their relationships
less in the Honesty condition than in the Control and Kindness conditions (ts > 5.98, ps < .001,
ds > .51).
Actual long-term relational improvement. Actual relational improvement did not mirror
these forecasts. A one-way ANOVA revealed a marginal effect of Condition, F(2,141) = 2.37, p =
.097,
ηp
2
= .03, but unlike the forecasts, participants in Study 1a reported greater relationship
improvement in the Honesty condition than in the Control condition (t(84) = 2.15, p = .03, d = .
49), and there was no difference between the Honesty condition and Kindness conditions (p = .
16).
Forecasters (i.e., participants in Study 1b) did not mispredict the relational improvements
associated with Honesty (p = .21). However, they overestimated the relational improvement
associated with the Control condition (t(264) = 5.63, p < .001, d = 0.85) and the Kindness
condition (t(279) = 4.18, p < .001, d = 0.59).
Long-term relational harm.
Predicted long-term relational harm. A one-way ANOVA revealed a significant effect of
Condition on predicted long-term relational harm, F(2,422) = 138.39, p < .001,
ηp
2
= .39.
YOU CAN HANDLE THE TRUTH 27
Forecasters (i.e., participants in Study 1b) expected to harm their relationships more in the
Honesty condition than in the Kindness and Control conditions (ts > 13.38, ps < .001, ds > 1.23).
Actual long-term relational harm. Actual relational improvement did not mirror these
forecasts. A one-way ANOVA revealed no main effect of Condition on actual relational harm,
F(2,141) = 1.57, p = .21,
ηp
2
= .02.
Forecasters (i.e., participants in Study 1b) expected that Honesty would cause greater
relational harm than participants in Study 1a actually experienced (t(264) = 6.49, p < .001, d =
1.01). Participants also overestimated relational harm in the Control condition (t(264) = 2.08, p =
.039, d = .32), but they did not mispredict relational harm in the Kindness condition (p = .68).
Long-term eudaimonic well-being.
Predicted long-term eudaimonic well-being. A one-way ANOVA revealed a significant
effect of Condition on predicted long-term meaning, F(2,422) = 15.54, p < .001,
ηp
2
= .07.
Participants in the Honesty condition of Study 1b expected to derive less long-term meaning than
in the Kindness and Control conditions (ts > 2.24, ps < .025, ds > .15).
Actual long-term eudaimonic well-being. A one-way ANOVA revealed no main effect of
Condition, F(2,141) = 1.45, p = .24,
ηp
2
= .02.
Participants did not mispredict the long-term meaning they would derive from Honesty (p
= .90) or Kindness (p = .22). Participants overestimated the long-term meaning they would
derive from the Control condition (t(264) = 5.05, p < .001, d = .78).
Appreciation of the experience.
Predicted appreciation of the experience. A one-way ANOVA revealed a significant
effect of Condition, F(2,422) = 33.84, p < .001,
ηp
2
= .13. Participants in Study 1b expected to
YOU CAN HANDLE THE TRUTH 28
appreciate their experience less in the Honesty condition than in the Control and the Kindness
conditions (ts > 6.43, ps < .001, ds > 0.52).
Actual appreciation for the experience. Actual appreciation did not mirror these forecasts.
A one-way ANOVA revealed no main effect of Condition, F(2,141) = 1.71, p = .19,
ηp
2
= .02.
Importantly, Forecasters (i.e., participants in Study 1b) underestimated how much
Experiencers (i.e., participants in Study 1a) actually appreciated the experience of Honesty
(t(264) = 3.38, p < .001, d = 0.62). Participants in Study 1b did not mispredict their appreciation
of Kindness (p = .66) and they marginally overestimated their appreciation of the Control
condition (t(264) = 1.84, p = .067, d = 0.30).
Choice. Using the Study 1b data, we assessed participants’ choices for which condition to
participate in by conducting a chi-square goodness of fit test against the null hypothesis that
there were no significant differences in preferences across the three conditions (i.e., expected
proportion of 33.3% for each of the three conditions). Participants were not equally likely to
choose to participate in all three conditions; χ2(2) = 15.75, p < .001. Specifically, they were less
likely to choose the Honesty condition (21.1%) compared to the Kindness (37.2%) and Control
conditions (41.7%).
The drivers of choice versus appreciation. We conducted regression analyses using the
data from Studies 1a and 1b to examine what predictions and experiences underlie participants’ a
priori decisions to avoid honesty and their post-hoc appreciation for honesty. Specifically, a
logistic regression on the choice not to engage in honesty in Study 1b (1 = chose kindness or
control, 0 = chose honesty), using forecasted enjoyment, meaning, and social connection (within
the honesty condition) as independent variables revealed that the choice to avoid honesty was
associated with predicted enjoyment (b = -.53, SE = .21, p = .014, OR = .59), but not predicted
YOU CAN HANDLE THE TRUTH 29
meaning (b = -.45, SE = .28, p = .11, OR = .64) and social connection (b =- .19, SE = .22, p = .
39, OR = .83). These results suggest that mispredictions of the hedonic consequences of honesty
may be the strongest deterrent for being honest.
On the other hand, a regression on appreciation of the experience (in the honesty
condition) in Study 1a, using judgments of enjoyment, meaning, and social connection as
independent variables revealed that reported meaning significantly predicted participants’
appreciation of honesty (b = 1.13, SE = .44, p = .014), but enjoyment (b = .13, SE = .30, p = .
65) and social connection (b =- .49, SE = .47, p = .30) did not. In other words, individuals’
appreciation of an experience is driven primarily by the meaning associated with it, but their
choice of which experience to engage in is driven by their predictions of enjoyment.
Study 1a and 1b Discussion
Mispredicting the consequences of honesty. Study 1 provides evidence consistent with
our hypotheses. Importantly, we find that honesty is significantly more enjoyable (during the
experiment and two weeks later) than people expect. Furthermore, honesty yielded significantly
greater social connection during the experiment, and caused less relational harm two weeks later
than participants expected. We find marginal evidence that people underestimate the immediate
eudaimonic consequences of honesty, and no evidence that they mispredict the long-term
eudaimonic consequences of honesty.
Interestingly, we also find that individuals mispredict their appreciation of honesty.
Forecasters (i.e., participants in Study 1b) predicted they would be much less likely to
recommend or want to repeat the experience of honesty than Experiencers (i.e., participants in
Study 1a) indicated two weeks after completing the study. Participants’ level of experienced
meaning drove this appreciation for honesty. However, similar to our Pilot Study, participants’
YOU CAN HANDLE THE TRUTH 30
level of (anticipated) enjoyment, not meaning, predicted their choice not to participate in the
honesty condition (Study 1b). Taken together, these results suggest that individuals may be
making a mistake. Individuals’ miscalibrated hedonic expectations lead them to avoid a behavior
that they appreciate post hoc.
The consequences of honesty. A secondary, and more exploratory, goal of this study was
to examine the consequences of honesty, broadly. We find that focusing on honesty often has
some positive consequences relative to simply being conscious of your communication (our
control condition), but, in many ways, does not differ from focusing on kindness. For example,
participants in our honesty condition experienced greater meaning during the three-day
experiment and greater relational improvements two-weeks later than participants in our control
condition. However, participants in our honesty condition did not experience significantly
different consequences than participants in our kindness condition on any measure. These results
replicate the results of our Pilot Study, in which we also find evidence that honesty yields some
benefits relative to our control condition, but is no different than kindness. These results suggest
that being intentional in one’s communication, and focusing on any moral value – whether it is
honesty or kindness – may yield psychological and relational benefits, relative to one’s usual
communication. Furthermore, these results provide further evidence that individuals may be
making a mistake when they choose to communicate as usual, rather than communicating
honestly.
Study 1c: Text analysis
We analyzed the text from the two-week reflection surveys that the participants in Study
1a provided. These data provide us with insight into participants’ actual thought processes and
conversations, and thus are useful for understanding the mechanisms underlying our effects. We
note, however, that our analyses of these data are limited by the retrospective self-report nature
YOU CAN HANDLE THE TRUTH 31
of participants’ responses. Thus, we interpret them as primarily illustrative and inductive, rather
than as providing us with conclusive evidence of the nature of our participants’ experiences.
Survey questions. Two weeks after the experiment ended, participants completed a
survey that included several open-ended questions about participants’ experiences, whether the
three-day experiment caused participants to implement changes in their life, whether their
experience surprised them, and what they learned during the experiment (see Appendix C for
exact questions). We coded participants’ responses from both Study 1 and the Pilot Study
according to three meta-categories: What happened during the experiment, How the experience
differed from expectations, and Long-term impact of the experiment. We coded each
participant’s responses to our focal questions as a single transcript, and coded the transcript for
the presence or absence of 14 categories (7 categories about What happened during the
experiment, 4 categories about How the experience different from expectations, and 3 Long-term
impact categories). Descriptions of our coding procedures, all coding categories, examples of
participant responses, and the frequency with which the categories occurred in each experimental
condition are provided in Table 3, along with inter-rater reliability statistics.
---Table 3 about here---
What happened during the experiment. Participants’ two-week reflections indicate that
they took the experiment seriously and followed the instructions of each condition. The
qualitative coding revealed significant differences in the extent to which participants expressed
themselves honestly, overall χ2(2) = 45.26, p < .001, which we consider to be a manipulation
check. Specifically, participants were significantly more likely to express themselves honestly in
the Honesty condition (48.8%) than in the Kindness (3.4%, χ2(1) = 18.93, p < .001) and Control
conditions (2.4%, χ2(1) = 23.90, p < .001).
YOU CAN HANDLE THE TRUTH 32
Interestingly, we also found that participants were more likely to discuss difficult topics
in the Honesty condition (11.6%) than in the Control condition (0.0%; χ2(1) = 5.19, p = .023),
but not compared to the Kindness condition (6.9%, χ2(1) = .68, p = .41), overall χ2(2) = 4.93, p =
.085. Participants were also more likely to generate conflict in the Honesty condition (11.6%)
than in the Kindness (0.0%; χ2(1) = 7.10, p = .008) and Control conditions (0.0%; χ2(1) = 5.07, p
= .024), overall χ2(2) = 11.93, p = .003.
Importantly, we did not find that the experiment caused people to avoid social
interactions. This point is important because one possible explanation of our results could be that
individuals expect to engage in difficult conversations (particularly in the Honesty condition), but
avoid those conversations during the actual experience. We do not find support for this idea.
Relatively few people reported avoiding social interactions or limiting their conversations, and
this did not differ significantly across conditions (Honesty = 2.3% vs. Kindness = 3.4% vs.
Control = 4.8%; overall χ2(2) = 0.37, p = .83).
To illustrate the experiences participants had in Study 1, consider the quotes below. These
reflections from participants in the Honesty condition highlight the nature of their experiences
and how it affected their lives.
Participant 94 (Honesty condition): I remember one of the three days I was
asked to be honest, I ended up lashing out at my coworker. This caused a
little bit of tension between us that I feel is still there. Nonetheless, I feel it
was worth it. What's the point of pretending when someone asks how you
are feeling. Additionally, being able to ask whatever question I desired
and to answer in a truthful manner allowed me and the other person I was
speaking to to be more open and comfortable. I feel that generally, being
honest allows for better relationships and more trust. Since that
experience, I have been trying to be more honest in my daily life.
Participant 128 (Honesty condition): I felt less fake when I didn't have to
lie about what I thought about some things or say I felt fine when I didn't.
Tension was building up with my roommate because couldn't bring myself
to tell her the things she was doing that were annoying and while doing
the study I told her all those things I had been avoiding to tell her for a
YOU CAN HANDLE THE TRUTH 33
while and it felt kind of liberating. Some of the things were kind of
awkward but others felt good and it helped ease some of the tension. I
learned that it feels better to say those kind of things instead of keeping
them inside until it explodes.
How the experience differed from expectations. We explored four possible ways in
which participants’ experiences may have differed from their expectations (and thus could have
caused them to mispredict the consequences of honesty): the experience was easier than they
expected, the experience led to better social interactions than they expected, the experience was
harder than they expected, and the experience led to worse social interactions than they expected.
Our coding suggests that one mechanism underlying participants’ mispredictions could be that
honesty caused them to have better social interactions than they expected, overall χ2(2) = 8.93, p
= .01. Participants were more likely to indicate that they had better interactions with others than
they expected in the Honesty condition (18.6%) than in the Control condition (0.0%; χ2(1) =
8.63, p = .003), but not the Kindness condition (8.6%, χ2(1) = .2.20, p = .14). In other words,
participants expected their conversational and relational partners to react more poorly to their
honesty than they actually did. The following quotes from participants in the Honesty condition
illustrate this sentiment:
Participant 34 (Honesty condition): I was particularly surprised when
being honest got me further in my position in an organization because
voicing my honest opinion made others think about the situation more and
come to the conclusion that I was thinking as well.
Participant 49 (Honesty condition): People reacted differently than what I
thought. They liked and appreciated the honesty and honestly I did not
believe that would happen. It was refreshing meaning that I was happy to
talk what was on my mind and not worry about what was said.
We do not find evidence that participants were more likely to believe that the experience
was easier than they expected, harder than they expected, or lead to worse social interactions
than they expected in the honesty condition relative to our other conditions (ps > .19).
YOU CAN HANDLE THE TRUTH 34
Long-term impact of the experiment. Many (but not all) participants across the
conditions believed they learned about the importance of communication (Honesty = 32.6% vs.
Kindness = 31.0% vs. Control = 35.7%), that the experiment positively impacted their lives
(Honesty = 11.6% vs. Kindness = 5.5% vs. Control = 4.8%), and that the experiment prompted
them to change their life in the future (Honesty = 39.5% vs. Kindness = 41.4% vs. Control =
45.2%). None of these categories significantly differed across experimental conditions.
Study 1c Discussion. Our analysis of participants’ two-week reflections reveals several
key insights. First, as illustrated by the quotes, and supported by the qualitative coding of the
open-ended responses, participants appeared fully committed to the study and followed the
instructions associated with each condition. Accordingly, participants in the Honesty condition
shared information that they would have otherwise left unsaid, which lead to more difficult
conversations, and increased conflict. Importantly, participants in the Honesty condition did not
simply avoid social interactions or limit their conversations. Instead, they challenged themselves
to pursue the goals of the study, and ultimately found their experience rewarding despite
whatever difficulties they may have faced. Finally, our results highlight an important potential
mechanism underlying individuals’ mispredictions of honesty: participants’ relational partners
reacted more positively to their honest communication than participants expected, which may
have influenced both enjoyment and social connection during the study.
Study 1 Post-test
Although we carefully crafted our manipulation in the field experiment to parallel our
theoretical definition of honesty (speaking in accordance with one’s own beliefs, thoughts and
feelings), we recognize that we also explicitly mentioned the trade-off between honesty and
kindness in the study instructions. Thus, upon reading our instructions, participants may have
YOU CAN HANDLE THE TRUTH 35
associated honesty with harshness, but upon engaging in honesty, realized that they did not
actually need to speak harshly. In other words, participants’ subjective meaning of honesty may
have shifted during different phases of the experiment. We ran a post-test to examine this
possibility.
Specifically, we presented the instructions from our honesty condition in Study 1a to a
new set of participants from the same laboratory pool used in Study 1 (N = 144). We asked them
to identify the primary goal of the study (by selecting one option from a selection of six possible
choices: “To encourage participants to speak bluntly”; “To encourage participants to speak
harshly”; “To encourage participants to share their private opinions”; “To encourage participants
to be truthful”; “To encourage participants to engage in self disclosure”; “To encourage
participants to avoid lying”). We found that the majority of participants (68.1%) believed that the
instructions were designed to encourage participants to be truthful. Respectively, 18.8%, 7.6%,
3.5%, and 2.1% of participants believed that the primary goal of the study was to encourage
participants to engage in self-disclosure, to avoid lying, to speak bluntly, and to share their
private opinions. No participants believed that the primary goal of the study was to encourage
participants to speak harshly.
We also had participants rate their agreement with each of the following statements (1=
strongly disagree, 7 = strongly agree): “These instructions are asking participants to…speak
bluntly, to speak harshly, to share private opinions, to be truthful, to engage in self-disclosure, to
avoid lying.” Participants generally agreed that the instructions asked participants to be truthful
(M = 6.45, SD =.75), to avoid lying (M = 6.06, SD =1.12), and to engage in self-disclosure (M =
6.45, SD =.75), consistent with our theoretical conceptualization of honesty (all means were
significantly above the scale midpoint, ps < .001). To some extent, participants also believed that
YOU CAN HANDLE THE TRUTH 36
the instructions were asking them to share their private opinions (M = 4.53, SD =1.55), and to
speak bluntly (M = 4.48, SD = 1.57; these means were also significantly above the scale
midpoint, but significantly below the means of the prior three items, ps < .001). Participants
generally disagreed that the instructions were asking them to speak harshly (M = 2.72, SD =1.23;
significantly below the scale midpoint, ps < .001). These results suggest that participants seem to
have interpreted our instructions as intended: they associated honesty with truthfulness and self-
disclosure.
Study 2: A deeper look into the consequences of honesty
Study 1 examined the consequences of focusing on honesty in one’s everyday life across
one’s social interactions. Our qualitative coding and our post-test of Study 1 suggest that our
three-day honesty intervention caused individuals to engage in greater self-disclosure and initiate
more difficult conversations than they would have otherwise. In Studies 2 and 3, we examine the
consequences of these honest conversations (involving self-disclosure and difficult topics) in
greater detail, using well-controlled laboratory paradigms. In doing so, we gain greater insight
into the mechanisms underlying our findings and rule out several potential confounds present in
the field experiments.
First, we address the possibility that forecasters mispredicted the types of conversations
that would actually occur during the three-day field study, rather than the experience of engaging
in those conversations. Just as individuals imagine that they would have the courage to confront
racism but rarely do (Kawakami, Dunn, Karmali, & Dovidio, 2014), it is possible that when
considering the consequences of honesty, participants imagined that they would have the courage
to state their most ardent criticisms or most embarrassing secrets, but perhaps did not do so
during the actual three-day experiment. Although our qualitative analyses in Study 1c suggest
YOU CAN HANDLE THE TRUTH 37
that focusing on honesty did prompt more difficult conversations, and our Study 1 Post-test
suggests that participants did not necessarily associate honesty with harshness, we do not yet
fully understand whether individuals expected the content and delivery of their conversations
(rather than the experience of these conversations) to be much different than they were.
Furthermore, given that participants were not directly monitored during the field experiment, we
cannot be positive that their self-reports (e.g., of how honest they were) were not biased by
experimental demand.
In Study 2, we had forecasters and experiencers consider the exact same conversational
topics (a series of questions that promote interpersonal evaluation and self-disclosure), thereby
ruling out differences in the topics of predicted conversations as the key driver of our results.
Furthermore, we monitor participants’ behavior in Study 2 to address the possibility that
participants in Study 1 did not fully commit to the experiment, and that their self-reports simply
reflect experimental demand. In Study 2, research assistants watched participants engage in
honest conversations and reminded them to follow the experimental instructions if they were not
fully engaged.
In Study 2, we also expand our investigation by examining how honest communication
influences both communicators and targets. Although Study 1 suggests that communicating
honestly can create meaning for communicators (relative to our control condition), and does not
harm enjoyment or social connection, it is not clear that honesty will influence targets (i.e., those
on the receiving end of the honesty) in the same way. In fact, recent research demonstrates that
relational partners often resent painful honesty (Levine & Schweitzer, 2015). Investigating
targets’ reactions allows us to provide further insight into the mechanism we identified in the
qualitative analyses: that individuals expect targets to react more poorly to honesty than they
YOU CAN HANDLE THE TRUTH 38
actually do. In Study 2, we also randomly assign participants to Forecaster and Experiencer
roles, rather than comparing them across studies.
Method
Sample. We ran Study 2 in the fall of 2016 in Chicago, Illinois. We aimed to recruit 50
pairs of Experiencers and 50 pairs of Forecasters and ended up with a final sample of 50
Experiencer-pairs and 51 Forecaster-pairs (42.1% female, mean age = 29). We preregistered all
aspects of this study at AsPredicted.org (http://aspredicted.org/blind.php?x=jqfjrp). A total of
31.2% of dyads were female-female pairs, 32.1% were mixed-sex pairs, and 36.7% were male-
male pairs.
A total of 67.0% of participants participated in the study with someone they considered to
be a friend, 15.3% with a roommate, 20.2% with a non-married romantic partner, 8.4% with a
colleague or teammate, 3.9% with a spouse, 4.9% with a relative, and 6.4% with a neighbor.
Note that these categories were not mutually exclusive, so participants could have participated
with someone they considered to be both a friend and a roommate.
Recruitment. We advertised Study 2 to a panel of adults in downtown Chicago (using a
laboratory listserv and Craigslist) and students at a Chicago-area university (using a laboratory
listserv and flyers). The recruitment materials informed potential participants that they would be
required to come to the laboratory with a close relational partner (such as a roommate, romantic
partner, close friend, or colleague) and that they would either be asked to have (or imagine
having) a conversation in which they would discuss personal information and sensitive topics
with this person. Participants knew they would earn $8-$10 for participation and that they would
have the chance to win a $100 bonus. Participants who were assigned to the Experiencer
YOU CAN HANDLE THE TRUTH 39
condition ended up earning $10 for 40 minutes of participation, whereas participants who were
assigned to the Forecaster condition ended up earning $8 for 20 minutes of participation.
Assignment to condition. Four research assistants, who were blind to our hypotheses,
were responsible for running the study. Participants were randomly assigned to be either
Forecasters or Experiencers when they arrived to the lab, based on a predetermined
randomization schedule. Each pair of participants was seated in a private room at computers that
were separated by a divider.
Experiencers. Experiencers began the study by completing a consent form on the
computer. In their consent form, Experiencers could opt in to having their conversation filmed
for research purposes (79% of experiencer-pairs were video-taped).
After participants completed their consent form, they read the study instructions on a new
page of the survey, which the research assistant also read aloud. Specifically, participants learned
that they would be required to have a 20-minute conversation, in which they would discuss 20
pre-specified questions with the partner they signed up with (the other person in the room).
Furthermore, participants learned that one person would be randomly assigned to the role of
Asker and one person would be randomly assigned to the role of Responder. The Asker would
ask questions, and the Responder would answer them.
Then, participants saw the list of 20 questions that they would ask or respond to. The first
10 questions were inspired by the topics of conversation that arose in Study 1 (e.g., “How is
school/work? Are you having any issues?”, “How long do you expect our relationship to last?”,
“Do you have any (positive or negative) opinions about me that you have been hesitant to share?
What are they?”). The second group of 10 questions were adapted from Aron, Melinat, Aron,
Vallone, and Bator’s (1997) Interpersonal Closeness Task (e.g., “When did you last cry in front
YOU CAN HANDLE THE TRUTH 40
of another person?”). Taken together, the questions were selected to increase both self-disclosure
(sharing thoughts, feelings, and opinions about the self) and interpersonal feedback (sharing
thoughts, feelings and opinions about one’s relational partner). We expected that both sets of
questions could elicit difficult conversations, similar to those that were prompted by the honesty
intervention in Study 1. We provide the full question list in Appendix D. Participants were
instructed to discuss every question on the list.
Next, the research assistant randomly assigned participants to the role of Asker or
Responder. Participants then indicated their role in the computerized survey, and were directed to
role-specific instructions.
The Asker read:
You will ask the RESPONDER questions. Your goal is to listen to the
RESPONDER’s responses and respond as if you are having a natural
conversation. You can respond to follow-up questions the RESPONDER may ask
you. So, ask a question, listen to and process the RESPONDER’s response
carefully, and have a discussion about the response as needed. Then, move on to
the next question.
The Responder read:
The ASKER will ask you the questions on the list you received. Your goal in
answering all questions is to be as honest as possible. Think carefully about your
answers and focus on sharing your completely honest opinions, feelings, and
reactions. This means that when the ASKER asks you a basic question like, “How
are you?” you should openly share your feelings and speak authentically. TRY TO
BE AS HONEST AS POSSIBLE. Please focus on being entirely honest, but do not
do anything you are uncomfortable with.
These role assignments allowed us to examine the consequences of honesty for both
communicators (Responders, who were explicitly asked to be honest) and targets (Askers, who
simply listened to the Responder’s honesty).
YOU CAN HANDLE THE TRUTH 41
When both participants were finished reading their instructions, the research assistant
directed them to sit at a table, facing each other, and begin their 20-minute conversation. The
research assistant then set up the video-camera (if both individuals had consented to it) and left
the room. The research assistant checked on the participants and if they did not seem engaged in
the study reminded them to follow the instructions and take the task seriously. The research
assistant returned after 20 minutes to instruct them to end their conversation and complete their
exit survey.
Forecasters. Forecasters followed a similar protocol (however, their consent form did not
include a section on being video-taped). Forecasters’ instructions specified that participants
would imagine having a 20-minute conversation, and imagine being the Asker or Responder.
Forecasters did not actually engage in a 20-minute conversation. Rather, they imagined the
experience for at least two minutes, and then completed the exit survey.
Exit survey. After completing [imagining] their conversation, participants rated the
conversation on enjoyment (Asker: α = .90, Responder: α = .85), social connection (Asker: r = .
63, Responder: r = .60), and meaning (Asker: α = .86, Responder: α = .82) using the same scales
we used in Study 1. Then we asked participants to briefly explain their answers (free response).
Next, participants answered three questions about how the experiment influenced their
relationship: “This experience deepened [would deepen] my relationship with my conversational
partner” (reverse-scored), “This experience strained [would strain] my relationship with my
conversational partner,” and “This experience caused [would cause] relational conflict with my
conversational partner.” We combined these three items into a single measure of relational harm
(Asker: α = .81, Responder: α = .73). Participants were then asked to recall [imagine] what they
[their partner] had [would have] answered in response to each of the 20 questions.
YOU CAN HANDLE THE TRUTH 42
Experiencers also answered three additional exploratory questions about how the
experience differed from their expectations: “Was your conversation more or less enjoyable than
you expected?” (1 = Much less enjoyable than I expected, 4 = As enjoyable as I expected, 7 =
Much more enjoyable than I expected), “Was your conversation more or less meaningful than
you expected?” (1 = Much less meaningful than I expected, 4 = As meaningful as I expected, 7 =
Much more meaningful than I expected), and “Did your conversation have a better or worse
effect on your relationship than you expected?” (1 = Much worse for my relationship than I
expected, 4 = As good for my relationship as I expected, 7 = Much better for my relationship
than I expected).
At the end of the survey, all participants answered a series of open-ended questions about
their partner and their experience: “Who was your conversational partner (e.g., friend, roommate,
significant other)?”, “How long have you known each other?”, “What was [do you think would
be] the most surprising thing revealed during your conversation?” and “Do you have any other
reactions to, or thoughts about your conversation [the conversation you imagined]?”
Participants then provided demographic information and were informed that they could
earn a $100 bonus for completing a follow-up survey. Participants provided their email address if
they were willing to complete the follow-up survey.
Experiencer follow-up survey. One week after Experiencers completed the study, they
were emailed a follow-up survey. A total of 55% of the initial Experiencers (n = 55) completed
the follow-up survey. The follow-up survey contained questions measuring Long-term hedonic
well-being (α = .82), Long-term eudaimonic well-being (α = .96), Long-term relational harm (α
= .81), and Appreciation for the experience (α = .88) using similar items to those we used in
Study 1a. Before exiting the survey, participants indicated whether they would prefer to receive
YOU CAN HANDLE THE TRUTH 43
their $100 payment (if they were the randomly selected winner) via PayPal, Venmo or by
receiving an Amazon.com gift card.
After the study closed, we also sent Forecasters a follow-up survey, so that they could
also be entered into the $100 raffle. The forecaster follow-up survey simply asked them to
confirm their email address and to indicate whether they would prefer to receive their $100
payment (if they were the randomly selected winner) via PayPal, Venmo or by receiving an
Amazon.com gift card. We randomly selected one participant in the study to win the $100 bonus,
and we compensated the participant within one week of the survey closing.
Results
We conducted mixed within-between subjects ANOVAs at the dyad level on our main
dependent variables (enjoyment, social connection, relational harm, and meaning) using
Perspective (Experiencer vs. Forecaster) as a between-subjects factor and Role (Asker vs.
Responder) as a within-subjects factor.4 We provide descriptive statistics associated with our
main dependent variables in Table 4.
In addition to these analyses, for Experiencers, we descriptively examined judgments of
how the honest conversation differed from their expectations, and we examined the long-term
consequences of the honest conversation. Specifically, we conducted a repeated-measures
ANOVA at the dyad level using Role (i.e., Asker vs Responder) as a within-subjects factor
(among Experiencers only).
---Table 4 about here---
Forecasters versus Experiencers.
4 We preregistered between-subjects analyses to compare Communicators to Targets. However,
given that Communicators and Targets are nested within dyads, and observations within dyad are
not independent, we decided that within-subjects analyses (with dyad representing the subject)
would be more appropriate. Our results are unchanged if we use between-subjects ANOVAs per
our preregistration. We follow the same procedure for reporting results in Study 3.
YOU CAN HANDLE THE TRUTH 44
Enjoyment. There was a significant effect of Perspective, F(1, 99) = 26.91, p < .001,
ηp
2
= .21; Forecasters expected their honest conversation to be less enjoyable than
Experiencers actually experienced it to be. There was also an effect of Role, F(1,99) = 4.66, p = .
033,
ηp
2
= .045; Askers found the conversation to be somewhat more enjoyable than
Responders did. Importantly, there was no significant Perspective x Role interaction, F(1,99) = .
38, p = .54,
ηp
2
= .004. We depict these results in Figure 4, Panel A.
Social connection. There was a significant effect of Perspective, F(1, 99) = 13.29, p < .
001,
ηp
2
= .11; Forecasters expected their honest conversation to be less socially connecting
than Experiencers actually experienced it to be. There was also a significant main effect of Role,
F(1,99) = 5.39, p = .022,
ηp
2
= .05; Askers found the conversation to be more socially
connecting than Responders. There was no significant Perspective x Role interaction, F(1,99) = .
61, p = .44,
ηp
2
= .006. We depict these results in Figure 4, Panel B.
Relational harm. There was also a significant effect of Perspective on judgments of
relational harm, F(1, 99) = 11.22, p = .001,
ηp
2
= .102; Forecasters expected their honest
conversation to cause more relational harm than Experiencers actually experienced. There was
no main effect of Role, F(1,99) = .94, p = .34,
ηp
2
= .009, nor was there a significant
Perspective x Role interaction, F(1,99) = .57, p = .45,
ηp
2
= .006.
Meaning. There was a significant effect of Perspective, F(1, 99) = 5.40, p = .022,
ηp
2
= .052; Forecasters expected their honest conversation to be less meaningful than Experiencers
actually experienced it to be. There was no main effect of Role, F(1,99) = 1.15, p = .29,
ηp
2
= .
YOU CAN HANDLE THE TRUTH 45
011, nor was there a significant Perspective x Role interaction, F(1,99) = .31, p = .58,
ηp
2
= .
003. We depict these results in Figure 4, Panel C.
Although participants significantly underestimated the meaning associated with honesty
in this study, participants underestimated enjoyment to a greater degree (consistent with Study
1). Indeed, a mixed within-between subjects ANOVA using Perspective as a between-subjects
factor and Measure (enjoyment vs. social connection vs. relational harm vs. meaning) and Role
as within-subjects factors, revealed a significant Measure x Perspective interaction, F(3, 297) =
15.81, p < .001,
ηp
2
= .14, such that the meaning misprediction was significantly smaller in
magnitude relative to the enjoyment and social connection mispredictions, and significantly
different in both magnitude and direction relative to the relational harm misprediction. We find
no evidence of a significant Measure x Perspective x Role interaction, suggesting that Askers and
Responders did not misforecast meaning (relative to other measures) to different degrees.
---Figure 4 about here---
Experiencers only: Experience versus expectations and subjective long-term impact.
Experience versus expectations. A repeated-measures ANOVA with Role as the within-
subjects factor revealed no differences between Askers and Responders (Fs < .74, ps > .39).
Across roles, Experiencers indicated that their honest conversations were significantly more
enjoyable (M = 5.03, SD = 1.47), meaningful (M = 5.14, SD = 1.54), and better for their
relationship (M = 5.24, SD = 1.33) than they expected (all means are significantly above 4, the
midpoint of the scale, ps < .001).
Subjective long-term impact. A repeated-measures ANOVA with Role as the within-
subjects factor revealed no differences between Askers and Responders (Fs < 3.14, ps > .09). In
their one-week follow-up survey, Experiencers (collapsed across Role) indicated that their honest
YOU CAN HANDLE THE TRUTH 46
conversations had significant positive effects for their long-term eudaimonic well-being (M
= 4.67, SD = 1.58; significantly above 4, the midpoint of the scale, ps < .001), that their
conversations did not cause relational harm (M = 2.44, SD = 1.26; significantly below 4, the
midpoint of the scale, p < .001), and that they had high appreciation for the experience (M
= 5.54, SD = 1.57; significantly above 4, the midpoint of the scale, p < .001). Conversations did
not have a significant effect on long-term hedonic well-being (M = 3.79, SD = 1.09; p = .15
compared to 4, the midpoint of the scale).
Discussion
Even when we guided the content of participants’ honest conversations via discussion
prompts and monitored their conversations, we found that individuals mispredicted the
consequences of honesty. Thus, we can be reasonably confident that individuals’ expectations
about the topics that they would discuss when being completely honest are not the sole driver of
the forecasting error that we document in this research. Participants’ open-ended responses
suggest that individuals’ mispredictions are driven, at least in part, by their failure to understand
others’ reactions to honesty during difficult conversations. Take for example, the following
quotes provided by Forecasters in Study 2 when explaining their predictions:
Participant 108: I feel that, the questions pertaining to my darker past and
my weaknesses may be harder to talk about with my significant other
because I would not want her to alienate me.
Participant 124: These questions ask me to reveal much about myself that
I've held back. I hold back some of these things for good reason: They will
harm other people, they will result in me being judged and socially
isolated.
Several participants shared similar sentiments, worrying that their honesty would burden
their partner, hurt their feelings, and cause them to be aggressive. In reality, however, most
Experiencers felt that the conversation, though difficult, helped them achieve mutual
YOU CAN HANDLE THE TRUTH 47
understanding with their relational partners and brought them closer together. The following
quotes from Experiencers illustrate this idea:
Participant 50: the asker got to know more about me; more about how I
think/feel, and how I feel about him. So while this exercise was difficult
and unpleasant for me at times,I think perhaps we strenghten our
friendship, at least I hope we did. / It was also kinda liberating to be able
to say exactly what I think about him and our friendship. I think he
learned more about me, probably more than he wanted, but I also learned
about him as well, from the responses to my answers. So while this
experience was a stressful, difficult and somewhat unpleasant it was also
a positive one.
Participant 25: It felt good to tell her what I had been holding in, and I think it really
helped to strengthen our bond, for me to open up to her about something in my past that I
had always kept separate from my day to day life. There was a little emotional distress,
but I focused mostly on not allowing myself to cry, because I know she would cry too.
Finally, one striking finding from Study 2 is how long-lasting the benefits of a single
honest conversation can be. We found that a single honest conversation led to significant gains in
eudaimonic well-being one week later, and did not significantly harm people’s relationships.
Moreover, participants indicated that they were grateful for and would want to repeat the
experience. Although our interpretation of the results from the follow-up survey is limited by
attrition and potential self-selection issues—only 55% of participants completed the follow-up
survey—when the results of Study 2 are combined with Study 1 (as well as our Pilot Study), they
provide converging evidence that a small dose of honesty injected into one’s relationships may
have long-lasting and unexpected positive effects.
Study 3
Study 3 extends our investigation in four ways. First, we directly measure our proposed
mechanism. Specifically, we measure whether individuals mispredict how others will react to
their honesty, and we examine whether that is associated with mispredictions of enjoyment and
social connection. Second, we disentangle mispredictions about the content of one’s honest
YOU CAN HANDLE THE TRUTH 48
conversation from mispredictions about others’ reactions to one’s conversation. In Study 2, we
control for the topics of conversation, but it is still possible that forecasters expected to say
different responses to the questions than experiencers actually did. In Study 3, we had all
participants write down exactly what they would say, make a forecast before reading their
statement aloud to their partner, and finally, make judgments after reading their statement aloud.
We also conducted Study 3 within-subjects, rather than between-subjects, which gave us greater
control over the content participants shared (imagined sharing) across experiences and forecasts.
Method
Sample. We ran Study 3 in the spring of 2018 in Chicago, Illinois. We intended to recruit
110 dyads. To determine our sample size, we conducted a power analysis, based on our results
from Study 2, in which we obtained an effect size (d) of .35 for our smallest focal effect: the
difference between Experiencers’ and Forecasters’ judgments of meaning. We used G*Power to
calculate the required sample size for a within-subjects design and 80% power, which yielded a
recommendation of N = 109. We preregistered all aspects of this study at AsPredicted.org
(http://aspredicted.org/blind.php?x=6hg6pj).
We ultimately ended up with 106 dyads5 who completed our entire experiment (44.3%
female, mean age = 26.0). A total of 25.2% of dyads were female-female pairs, 39.8% were
mixed-sex pairs, and 34.95% were male-male pairs. A total of 67.5% of participants participated
in the study with someone they considered to be a friend, 17.0% with a roommate, 21.7% with a
non-married romantic partner, 4.7% with a colleague or teammate, 3.3% with a spouse, 4.7%
with a relative, and 1.9% with a neighbor. Note that these categories were not mutually
5 When we ended data collection, we believed that 110 dyads had completed the study.
However, when we began to analyze the data, we realized that there had been an error in
assigning dyad numbers to each pair of participants; four numbers were skipped. Thus, we only
had 106 dyads complete the study.
YOU CAN HANDLE THE TRUTH 49
exclusive, so participants could have participated with someone they considered to be both a
friend and a roommate.
Recruitment. We advertised Study 3 to a panel of adults in downtown Chicago (using a
laboratory listserv and Craigslist) and students at a Chicago-area university (using a laboratory
listserv and flyers). The recruitment materials informed potential participants that this study
examined interpersonal communication and that they would be required to come to the
laboratory with a close relational partner (such as a roommate, romantic partner, close friend, or
colleague). Participants knew that they would be asked to discuss a sensitive topic with this
person. Participants earned $15 for participation in exchange for a 30-minute study.
Sixty-one dyads were recruited in this manner. However, recruitment began to slow after
roughly one month. We recruited the remaining dyads using a slightly altered recruitment
protocol. Our laboratory hosted a dinner-and-a-study event, for which we recruited pairs of close
relational partners to come to the lab for 40 minutes one evening, in exchange for $6 and free
dinner. We find no differences in results based on recruitment type; therefore, we do not discuss
this feature of our design in our results.
Assignment to condition. Research assistants, who were blind to our hypotheses, were
responsible for running the study. For regular recruitment, each pair of participants who arrived
to the lab was seated in a private room at computers that were separated by a divider.
For event recruitment (those recruited during the “dinner-and-a-study” event), all pairs of
participants arrived to the event at the same time and were seated next to their partner in a large
classroom. Participants completed the survey on their phones or tablets. Participants completed
the initial portion of the survey in the classroom, and then, each pair found a private space in the
YOU CAN HANDLE THE TRUTH 50
building in which to have their honest conversation. Following their conversation, each pair
returned to the classroom to complete their surveys.
All participants began by completing a consent form. Then, we randomly assigned one
participant in each pair to the roles of Communicator and Target. Communicators learned that
they would have to deliver negative feedback to the Target. Specifically, we instructed them to,
“Provide one piece of critical feedback to the person you came to the lab with today (the
“Target” in this study). Specifically, you will share your honest opinions, feelings, and reactions
about one thing you think this person should do differently, change about themselves, or improve
upon.” Communicators knew that they would have to read their message aloud to the Target and
engage in a conversation with the Target about it.
Targets learned that the Communicator would deliver a personal message to them, but
they did not know the nature of that message when they began their survey. We instructed targets
to listen to the Communicator’s message and respond as if they were having a natural
conversation.
Forecast judgments. Before Communicators read their message aloud to the Target, we
asked Communicators to forecast the experience. Specifically, we asked Communicators to
predict their enjoyment (α = .84), social connection (r = .39), relational harm (α = .65), and
Meaning (α = .55), using nearly identical scales to those we used in Study 2.6 The only change
was that we replaced “conversational partner” with “target” in the Relational Harm items. As in
Study 2, we consider relational harm as another dependent variable in addition to enjoyment,
social connection, and meaning.
6 We note the relatively low internal consistency reliability coefficients for some of our
measures in this study. This was unexpected given that we used the same items as we did in the
previous studies. We discuss this limitation of our study in the General Discussion.
YOU CAN HANDLE THE TRUTH 51
Next, Communicators answered a series of questions about how they expected the Target
to react to the conversation. Specifically, participants rated their agreement (1 = strongly
disagree, 7 = strongly agree) with the following eight items, which we combined into a single
measure of “target’s negative reaction” (α = .75): “This conversation will offend the target,”
“This conversation will make the target defensive,” “The target will be receptive to this
conversation (reverse-scored),” “This conversation will cause the target to be judgmental,” “The
target will appreciate this conversation (reverse-scored),” “This conversation will make the target
relieved (reverse-scored),” “This conversation will be awkward,” “What I say during this
conversation will surprise the target.” Whereas relational harm captures participants’ beliefs
about how a conversation will impact the relatively long-term status of their relationship, target’s
negative reaction captures participants’ beliefs about their partners’ moment-to-moment response
to the conversation itself.
Experience judgments. Communicators read their message aloud to the Target. The
Communicator’s message was piped back to them on their computer (phone, tablet) screen and
Communicators were instructed to allow the Targets to see the message. This procedure helped
us to ensure that Communicators read exactly what they had anticipated saying to the Targets.
We informed participants that the Target could respond to the Communicator’s message and the
two parties could engage in a conversation as they saw fit. Communicators then judged the
experience on all of the same measures they forecasted (αs for enjoyment, meaning, and
relational harm > .56, r for social connection = .69).
YOU CAN HANDLE THE TRUTH 52
Target judgments. After the conversation, targets also rated their experience using the
same enjoyment (α = .82), social connection (r = .55), relational harm (α = .62), meaning (α = .
81), and target’s negative reaction (α = .70) scales that Communicators used.7
At the end of the survey, Communicators and Targets answered a series of open-ended
questions about their partner and their experience: “Who was your conversational partner (e.g.,
friend, roommate, significant other)?”, “How long have you known each other?”, “What was the
most surprising thing revealed during your conversation?” and “Do you have any other reactions
to, or thoughts about your conversation?” All participants provided demographic information at
the end of the study, and then received payment.
Results
Consistent with our preregistration, we conducted two sets of analyses. First, we used
repeated-measure ANOVAs to examine how prediction and experience judgments differed
among Communicators. Specifically, we used Perspective (Forecast vs. Experience) as our
within-subjects factor and each of our measures as dependent variables. We followed this
analysis with within-subjects mediation analyses to test whether Communicators’ mispredictions
of enjoyment, social connection, relational harm, and meaning were driven by their
mispredictions of the targets’ negative reactions to their conversations.
Second, we conducted two within-subjects ANOVAs to examine 1) how Communicators’
forecasts of targets’ negative reactions differed from Targets’ ratings of their own negative
reactions, and 2) how Communicators’ experiences differed from Targets’ experiences. In the
first analysis we used Communicators’ forecasts vs. Targets’ experiences as with the within-
7 We note that in our preregistration, we only indicated that we would measure “Target’s
Negative Reaction” from the target perspective. However, after filing our preregistration, but
before collecting data, we decided to add measures of enjoyment, social connection, meaning,
and relational harm from the target perspective.
YOU CAN HANDLE THE TRUTH 53
subjects variable, and in the second analysis, we used Communicators’ experiences vs. Targets’
experiences as the within-subjects variable. We depict all means and standard deviations in Table
5.
---Table 5 about here ---
Communicators’ predictions versus experiences.
Enjoyment. There was a significant effect of Perspective, F(1, 105) = 87.46, p < .001,
ηp
2
= .45; such that Communicators expected their conversations to be less enjoyable than
they actually were. We depict these results in Figure 5, Panel A.
Social connection. There was a significant effect of Perspective, F(1, 105) = 40.02, p < .
001,
ηp
2
= .28; such that Communicators expected their conversations to be less socially
connecting than they actually were. We depict these results in Figure 5, Panel B.
Relational harm. There was a significant effect of Perspective, F(1, 105) = 21.71, p < .
001,
ηp
2
= .17; such that Communicators expected their conversations to harm their
relationship more than it actually did.
Meaning. There was a significant effect of Perspective, F(1, 105) = 13.49, p < .001,
ηp
2
= .11; such that Communicators expected their conversations to be less meaningful than
they actually were. We depict these results in Figure 5, Panel C.
Although we found that Communicators did mispredict the meaning associated with
honesty, consistent with our prior studies, we found that the magnitude of this misprediction was
smaller than the misprediction of either enjoyment or social connection. Specifically, a within-
subjects ANOVA using perspective (Forecast vs. Experience) and measure (enjoyment vs. social
connection vs. relational harm vs. meaning) as the factors revealed a significant Perspective x
YOU CAN HANDLE THE TRUTH 54
Measure interaction, F(3, 315) = 44.31, p < .001,
ηp
2
= .29, such that the meaning
misprediction was significantly smaller in magnitude relative to the enjoyment and social
connection mispredictions, and significantly different in both magnitude and direction relative to
the relational harm misprediction.
--- Figure 5 about here---
Targets’ Negative Reactions. There was a significant effect of Perspective, F(1, 105) =
58.50, p < .001,
ηp
2
= .36; such that Communicators expected the Target to react more poorly
to the conversation than they actually did.
Mediation analyses. To explore whether Communicators’ mispredictions of their own
experiences were driven by their mispredictions of Targets’ reactions to their conversations, we
ran several mediation analyses. We used the MEMORE macro for within-subjects mediation
using SPSS (Montoya & Hayes, 2017), in which the independent variable is represented by the
repeated measurements of the mediator variable and the dependent variable (in this case,
Forecaster vs. Experiencer judgments). In each model, we examined whether the
Forecaster/Experiencer difference in judgments of targets’ negative reactions mediates the
Forecaster/Experiencer difference in enjoyment (model 1), social connection (model 2),
relational harm (model 3), and meaning (model 4). As indicated in our preregistration, we
predicted significant mediation in Models 1-3, but not in Model 4. However, we found
significant evidence of mediation in all four models (see Table 6). The confidence interval
around the indirect effect of target’s negative reaction excluded zero in every model, suggesting
that Communicators’ mispredictions of their relational partners’ reactions to honest feedback at
least partially accounts for mispredictions of the enjoyment, social connection, and meaning they
derive from the conversation, as well as the relational harm caused by the conversation.
YOU CAN HANDLE THE TRUTH 55
--- Table 6 about here---
Communicators vs. Targets.
Communicators’ predictions of Targets’ negative reactions vs. Target’s negative
reactions. There was a significant effect of Role, F(1, 102) = 106.09, p < .001,
ηp
2
= .51; such
that Communicators expected Targets to react more negatively to their honest conversations than
Targets reported.
Communicators’ experiences of Targets’ negative reactions vs. Target’s negative
reactions. After the conversation took place, Communicators continued to overestimate how
negatively Targets had reacted compared to what the Targets reported, F(1, 102) = 9.27, p = .003,
ηp
2
= .08, albeit this misprediction was smaller than before the conversation took place.
Communicators’ experiences vs. Targets’ experiences. There was a marginal effect of
Role, F(1, 102) = 3.46, p = .066,
ηp
2
= .51, on enjoyment such that Communicators enjoyed
the honest conversation somewhat less than Targets did. There were no effects of Role on
judgments of social connection, F(1, 102) = 1.00, p = .32,
ηp
2
= .01, relational harm, F(1, 102)
= .04, p = .85,
ηp
2
< .01, or meaning, F(1, 102) = .43, p = .51,
ηp
2
= .01.
Discussion
Study 3 reveals several key insights. First, Study 3 demonstrates that communicators
mispredict the consequences of honest, difficult conversations even when they know the exact
information they will communicate ahead of time. This result is compelling because it suggests
that communicators misunderstand the experience of being honest with others, rather than (or
perhaps in addition to) the information that they will share when they focus on being honest.
Study 3 also provides further evidence that communicators mispredict the affective and social
YOU CAN HANDLE THE TRUTH 56
consequences associated with honesty during difficult conversations more than they mispredict
its meaning.
Third, Study 3 provides further insight into targets’ reactions to honesty. Honest
conversations – in this case, conversations involving interpersonal criticism – did not yield
significantly different levels of enjoyment, social connection, or meaning for targets relative to
communicators. However, communicators and targets do have different impressions of how
negatively targets react to these conversations. Communicators are particularly likely to
overestimate their relational partners’ reactions to honesty before a conversation has occurred. In
Study 3, we find that communicators expected their relational partners to react more negatively
to their interpersonal critique before having the conversation than participants indicated after
having the conversation, and that this misprediction at least partially accounts for mispredictions
of enjoyment, social connection, meaning, and relational harm.
Interestingly, communicators in Study 3 overestimated the negativity of their relational
partner’s reactions, relative to the reports made by their relational partners, both before the
conversation and after the conversation. This result suggests that communicators may not receive
accurate feedback from their relational partners, or they may not believe the feedback they
receive, after engaging in difficult conversations. For example, it is possible that targets do not
directly express their appreciation for constructive criticism, and therefore, communicators fail to
fully learn targets’ true reactions. On the other hand, it is possible that targets do express their
appreciation, but communicators assume that targets are just trying to be polite. Communicators’
failure to fully understand targets’ experiences likely perpetuates the avoidance of honesty.
General Discussion
YOU CAN HANDLE THE TRUTH 57
In this research, we break new ground by exploring how honesty, one of the most basic
moral principles and facets of human communication, influences – and is expected to influence –
psychological well-being. We accomplished this by conducting an intensive three-day field
experiment in which individuals had to be honest (versus kind or conscious of their
communication) in all of their social interactions, one laboratory experiment in which individuals
had to be honest with a close relational partner while answering personal and potentially difficult
discussion questions, and a final laboratory experiment in which individuals had to provide
honest, critical feedback to a close relational partner.
Our findings make several important contributions to our understanding of morality,
affective forecasting, and human communication. First, we provide insight into why people avoid
being honest with others. Our results suggest that individuals’ aversion to honesty is driven by a
forecasting failure: Individuals expect honesty to be less pleasant and less socially connecting
than it is. Furthermore, our studies suggest this is driven by individuals’ misguided fear of social
rejection. Whereas prior work on mispredictions of social interactions has primarily examined
how individuals misunderstand others or their preferences for interaction, the present research
examines how individuals misunderstand others’ reactions to honest disclosure of thoughts and
feelings, and how this shapes social communication.
Second, this research documents the broader consequences of being honest. Individuals’
predictions that honest communication would be less enjoyable and socially connecting than
kind communication or one’s baseline communication were generally wrong. In the field
experiment (Study 1a), participants in the honesty condition either felt similar or higher levels of
social connection relative to participants in the kindness and control conditions. Participants in
the honesty condition also derived greater long-term hedonic well-being and greater relational
YOU CAN HANDLE THE TRUTH 58
improvements relative to participants in the control condition. Furthermore, participants in Study
2 reported increased meaning in their life one week after engaging in their brief, but intense,
honest conversation. Scholars have long claimed that morality promotes well-being, but to our
knowledge, this is the first research to document how enacting specific moral principles promote
different types of well-being.
Taken together, these findings suggest that individuals’ avoidance of honesty may be a
mistake. By avoiding honesty, individuals miss out on opportunities that they appreciate in the
long-run, and that they would want to repeat. Individuals’ choices about how to behave – in this
case, whether or not to communicate honestly – seem to be driven primarily by expectations of
enjoyment, but appreciation for these behaviors is driven by the experience of meaning. We
encourage future research to further examine how affective forecasting failures may prevent
individuals from finding meaning in their lives.
Practically, our research has implications for the delivery of performance appraisals in
organizations, and the establishment of honest work cultures. Prior work has documented that
managers are often overly positive when attempting to deliver critical feedback, leading
subordinates to hold misperceptions about how their work is actually evaluated by their
supervisors (Schaerer, Kern, Berger, Medvec, & Swaab, 2018). A suggestion from our work is
that one factor underlying this phenomenon might be miscalibrated expectations about the
difficulty of giving honest feedback. Like the participants in our studies, we suspect that many
managers have erroneous expectations about the pain associated with providing honest feedback
to direct reports. We hope our findings shine a light on this potential inaccuracy, and provide an
impetus for those tasked with providing performance appraisals to do so with greater honesty and
clarity.
YOU CAN HANDLE THE TRUTH 59
Our findings also suggest that prioritizing honesty may indeed benefit companies and
their employees, though future research is needed to fully understand the nature and boundaries
of these benefits. Recently, several organizations – such as Netflix (McCord, 2017) and
Bridgewater Associates (Dalio, 2017) – and public figures have celebrated the practice of
“radical candor” (Scott, 2017) and “radical honesty” (Blanton, 2005), but there have been no
systematic empirical investigations of the consequences of such practices for personal or
organizational relationships.8 Many lay people believe that radical honesty in organizations is a
terrible idea—something that is likely to be painful to experience and a recipe for destructive
conflict among employees. Our research suggests that such beliefs may be misguided.
Nonetheless, we caution against organizations endorsing radical honesty as a general practice
given that other research suggests that, at least in some circumstances, people prefer those who
speak “prosocial lies” to those who offer painful truths (Levine & Schweitzer, 2014, 2015).
There is important work to be done to examine the relational and organizational consequences of
complete honesty, particularly over time and within work contexts.
Finally, the current research has implications for understanding barriers to successful
conflict management. In the presence of disagreement, honest communication affects whether
conflict will escalate into unpleasant, negative spirals, or whether it will de-escalate into more
reflective and productive conversations (Weingart et al., 2015). To the extent that the parties are
reluctant to communicate honestly because of the faulty assumptions highlighted in the current
research, they are unlikely to communicate their disagreements effectively, which, of course,
puts them at risk for destructive conflict escalation.
Limitations and Future Directions
8 However, there has been at least one entertaining report from a single person who
experimented with radical honesty in his own life (Jacobs, 2007).
YOU CAN HANDLE THE TRUTH 60
Many questions remain about how individuals and their conversational partners react to
honesty. Although Studies 2 and 3 demonstrate that individuals underestimate the benefits of
honesty even when the topics and content of conversation are held constant, we still do not know
precisely which types of honest conversations yield the greatest benefits. For example, does self-
disclosing personal information or does sharing long-hidden critiques yield greater enjoyment,
social connection, and meaning for the two parties? And, how do these outcomes compare with
people’s expectations of these conversations? We also do not yet know what types of
conversational strategies yield the greatest benefits. For example, does avoiding explicit lies or
does unburdening oneself of secrets (e.g., Slepian, Chun, & Mason, 2017) yield greater personal
benefits? We encourage future researchers to more deeply explore the conversational topics and
tactics that are most beneficial for communicators, and most misunderstood. We also encourage
future researchers to examine our proposed mechanism in greater detail. We find that individuals
misunderstand others’ reactions to the information they share during honest conversations, but
other potential mechanisms may also contribute to mispredictions of honesty. Furthermore,
mispredictions of others’ reactions to specific information may also contribute to forecasting
errors in many other conversational domains, such as story-telling and expressing gratitude (see
Cooney, Gilbert, & Wilson, 2017; Kumar & Epley, 2018).
It will also be important for future work to examine the types of relationships that most
benefit from (or are harmed by) honesty. Although we sampled from many different types of
relationships across our studies, we did not have sufficient power to explore how honesty
influenced different relationships differently. One possibility is that honesty harmed some
relationships and benefited others. For example, perhaps honesty within insecure or
dysfunctional relationships leads to the dissolution of the relationship, whereas honesty within
YOU CAN HANDLE THE TRUTH 61
secure and functional relationships makes the relationship even stronger. And perhaps,
individuals only focus on the former possibility when making forecasts, when in reality, both
processes occur. We conducted exploratory analyses to examine this idea. In particular, with the
data from the field experiments (the Pilot Study and Study 1), we examined whether there was
greater variance in social connection and relational improvement and harm in the honesty
condition than in the kindness or control conditions. If this were the case, it would suggest that
honesty leads to more extreme relational outcomes than other communication tactics. However,
we find no evidence of greater variance in the honesty condition (all ps > .167 for Levene’s test
for equality of variance). Despite these null results, we encourage future researchers to test this
idea more thoroughly and systematically.
Our research also suggests that there is interesting work to be done on the nature of
individuals’ mispredictions, and in particular, the specific outcomes that individuals do and do
not have insight into. Existing research has focused almost entirely on affective forecasting,
concluding that individuals often lack insight into the affective consequences of future
experiences (e.g., Dunn et al., 2007; Gilbert, Driver-Linn, & Wilson, 2002; Van Boven,
Loewenstein, Dunning & Nordgren, 2013; Wilson & Gilbert, 2005). Our findings are consistent
with this body of research. Interestingly, we also find that individuals are often more accurate
when predicting the eudaimonic consequences of future experiences—individuals were relatively
more accurate when predicting the meaning (compared to the enjoyment, social connection, and
relational harm) associated with the communication interventions. These results suggest that
individuals may generally be more accurate when making more cognitive predictions, like
predictions of personal meaning, than affective or social predictions. The forecasting literature
has not explored this possibility, so our results suggest an interesting area for future research to
YOU CAN HANDLE THE TRUTH 62
explore. Perhaps individuals who experience human suffering – through breakups, death, and
defeat (Gilbert et al., 1998) – do recognize that with hardship comes meaning. We encourage
future scholars to pursue this idea and to employ more reliable scales to capture hedonic,
eudaimonic and social well-being. We note that one limitation of the current work was the
relatively low internal consistency for some of the measures in some of the studies. We used the
same or highly similar measures across studies, so it is unclear why internal consistency
coefficients differed across studies, with strong evidence of reliability in some and more modest
reliability in others.
It will also be important to more deeply explore targets’ reactions to honesty. In Study 2,
we find that targets, like communicators, underestimate the benefits of honesty. Research on
targets’ positive reactions towards prosocial lies (Levine & Schweitzer, 2014, 2015), however,
suggests that targets and observers often have negative reactions to hurtful truths. In our studies,
unlike Levine and Schweitzer’s economic games, the costs and benefits of honesty were
ambiguous. In the context of difficult conversations, people expect honesty to do more harm
(e.g., cause relational strain, offend another person) than it actually does, and in fact, this
contributes to communicators’ misguided forecasts of enjoyment and social connection. We
expect that in circumstances in which honesty does indeed cause harm, targets would judge truth-
tellers negatively (as in Levine and Schweitzer, 2014, 2015). Thus, the present research extends
existing research by demonstrating how forecasts of harm might diverge from reality and by
highlighting the importance of perspective (communicator versus target) when studying
judgments of honesty and deception. Future research is needed to fully understand targets’
reactions towards honesty, kindness, and prosocial lies, particularly in naturalistic settings.
YOU CAN HANDLE THE TRUTH 63
Finally, it would be worthwhile for future research to more carefully disentangle the
consequences of focusing on honesty, relative to the consequences of engaging in difficult
conversations. An additional experiment reported in the online supplement suggests that honesty
per se does not drive our effects. Specifically, in Study S1, individuals misunderstood the
consequences of delivering negative feedback regardless of whether individuals were specifically
prompted to be completely honest when delivering it. This suggests that individuals
misunderstand the consequences of honesty because they misunderstand the consequences of
difficult conversations, not because of the associations they have with the word honesty. More
research, however, is needed to understand the full range of difficult conversations that
individuals misunderstand. In the present research, we directly examined self-disclosure and
negative feedback. Would other difficult conversations, such as delivering bad news, confronting
another person, or proposing a bold new idea be similarly misunderstood? Do people
overestimate the risks and interpersonal costs of any conversation that could lead to negative
interpersonal judgment? We encourage future scholars to answer these questions.
Conclusion
Individuals often shy away from communicating honestly during difficult conversations,
fearing the hedonic and social costs of honesty. Our findings suggest these fears are misguided.
Honesty is not as unpleasant or isolating as it seems, and in fact, may promote meaning and
long-term growth. People can handle the truth (more than you think).
YOU CAN HANDLE THE TRUTH 64
Context
This paper is based on a chapter of Emma Levine’s dissertation. Levine’s research
focuses primarily on how individuals navigate the tension between honesty and kindness. In
previous studies, she has found that targets often prefer kindness to honesty and thus reward
prosocial lies. In the current paper, she sought to extend her prior work by exploring how
communicators react to honesty and kindness, and whether they make accurate predictions about
these experiences.
Taya Cohen served as a member of Levine’s dissertation committee and actively
collaborated with her on the current project. Cohen’s research focuses primarily on
understanding the antecedents of moral behavior. In previous studies, she has found that some
people are reliably and predictably more honest than others across different situations they
encounter in their lives, and that such people can be identified via standardized personality
assessments and behavior-based interview questions. In the current paper, she sought to extend
her prior work by exploring the consequences of an honesty intervention on people’s daily lives,
and to test whether making people act in a more honest way is beneficial for their well-being and
social relationships.
YOU CAN HANDLE THE TRUTH 65
References
Anderson, N. H. (1968). Likableness ratings of 555 personality-trait words. Journal of
Personality and Social Psychology, 9(3), 272.
Aknin, L.B., Barrington-Leigh, C.P., Dunn, E.W., Helliwell, J.F., Burns, J., Biswas-Diener, R.,
Kemeza, I., Nyende, P., Ashton-James, C., & Norton, M.I. (2013). Prosocial spending and
well-being: Cross-cultural evidence for a psychological universal. Journal of Personality
and Social Psychology, 104(4), 635-652.
Aknin, L.B., Dunn, E.W., & Norton, M.I. (2012). Happiness runs in a circular motion: Evidence
for a positive feedback loop between prosocial spending and happiness. Journal of
Happiness Studies, 13(2), 347-355.
Aquino, K., & Reed II, A. (2002). The self-importance of moral identity. Journal of Personality
and Social Psychology, 83(6), 1423.
Aron, A., Melinat, E., Aron, E. N., Vallone, R. D., & Bator, R. J. (1997). The experimental
generation of interpersonal closeness: A procedure and some preliminary
findings. Personality and Social Psychology Bulletin, 23(4), 363-377.
Ashton, M. C., & Lee, K. (2009). The HEXACO–60: A short measure of the major dimensions
of personality. Journal of Personality Assessment, 91(4), 340-345.
Bandura, A. (1999). Moral disengagement in the perpetration of inhumanities. Personality and
Social Psychology Review, 3(3), 193-209.
Bentham, J. (1843/1948). An Introduction to the Principles of Morals and Legislation. Oxford,
United Kingdom: Blackwell. (Original work published 1843)
Blanton, B. (2005). Radical Honesty: How to Transform Your Life By Telling the Truth. Virginia:
Sparrowhawk Publications.
YOU CAN HANDLE THE TRUTH 66
Bond, C. F., & Anderson, E. L. (1987). The reluctance to transmit bad news: Private discomfort
or public display?. Journal of Experimental Social Psychology, 23(2), 176-187.
Butler, E. A., Egloff, B., Wlhelm, F. H., Smith, N. C., Erickson, E. A., & Gross, J. J. (2003). The
social consequences of expressive suppression. Emotion, 3(1), 48.
Cohen, T. R., Kim, Y., Jordan, K. P., & Panter, A. T. (2016). Guilt-proneness is a marker of integrity and
employment suitability. Personality and Individual Differences, 92, 109-112.
Cohen, T. R., Wolf, S. T., Panter, A. T., & Insko, C. A. (2011). Introducing the GASP scale: A
new measure of guilt and shame proneness. Journal of Personality and Social
Psychology, 100, 947-966.
Cooney, G., Gilbert, D. T., & Wilson, T. D. (2017). The Novelty Penalty: Why Do People Like
Talking About New Experiences but Hearing About Old Ones?. Psychological
science, 28(3), 380-394.
Craig, K. D. (2009). The social communication model of pain. Canadian
Psychology/Psychologie canadienne, 50(1), 22.
Dalio, R. (2017). Principles: life and work. New York, NY.
Deci, E. L., & Ryan, R. M. (2008). Hedonia, eudaimonia, and well-being: An
introduction. Journal of Happiness Studies, 9(1), 1-11.
DePaulo, B. M., & Bell, K. L. (1996). Truth and investment: Lies are told to those who
care. Journal of Personality and Social Psychology, 71(4), 703.
Diener, E. D., Emmons, R. A., Larsen, R. J., & Griffin, S. (1985). The satisfaction with life
scale. Journal of Personality Assessment, 49(1), 71-75.
Dunn, E. W., Aknin, L. B., & Norton, M. I. (2008). Spending money on others promotes
happiness. Science, 319(5870), 1687-1688.
YOU CAN HANDLE THE TRUTH 67
Dunn, E. W., Biesanz, J. C., Human, L. J., & Finn, S. (2007). Misunderstanding the affective
consequences of everyday social interactions: The hidden benefits of putting one’s best
face forward. Journal of Personality and Social Psychology, 92(6), 990 –1005.
Van Boven, L., Loewenstein, G., Dunning, D., & Nordgren, L. F. (2013). Changing places: A
dual judgment model of empathy gaps in emotional perspective taking. Advances in
Experimental Social Psychology, 47, 117-171.
Emmons, R. A., & McCullough, M. E. (2003). Counting blessings versus burdens: an
experimental investigation of gratitude and subjective well-being in daily life. Journal of
Personality and Social Psychology, 84(2), 377.
Epley, N., & Schroeder, J. (2014). Mistakenly seeking solitude. Journal of Experimental
Psychology: General, 143(5), 1980.
Fisher, C. D. (1979). Transmission of positive and negative feedback to subordinates: A
laboratory investigation. Journal of Applied Psychology, 64(5), 533.
Gilbert, D. T., Pinel, E. C., Wilson, T. D., Blumberg, S. J., & Wheatley, T. P. (1998). Immune
neglect: a source of durability bias in affective forecasting. Journal of Personality and
Social Psychology, 75(3), 617.
Gilbert, D. T., Driver-Linn, E., & Wilson, T. D. (2002). The trouble with Vronsky: Impact bias in
the forecasting of future affective states. In L. Feldman Barrett & P. Salovey (Eds.), The
Wisdom in Feeling: Psychological Processes in Emotional Intelligence (pp. 114 –143).
New York, NY: Guilford Press.
Gilbert, D. T., & Wilson, T. D. (2000). Miswanting: Some problems in the forecasting of future
affective states. In J. P. Forgas (Ed.), Studies in emotion and social interaction, second
series. Feeling and thinking: The role of affect in social cognition (pp. 178-197). New
York, NY, US: Cambridge University Press.
YOU CAN HANDLE THE TRUTH 68
Gino, F., Kouchaki, M., & Galinsky, A. D. (2015). The Moral Virtue of Authenticity How
Inauthenticity Produces Feelings of Immorality and Impurity. Psychological
Science, 26(7), 983-996.
Goffman, E. (1967). On face-work. Interaction ritual, 5-45.
Goodwin, G. P., Piazza, J., & Rozin, P. (2014). Moral character predominates in person
perception and evaluation. Journal of Personality and Social Psychology, 106(1), 148.
Gray, K., Young, L., & Waytz, A. (2012). Mind perception is the essence of
morality. Psychological Inquiry, 23(2), 101-124.
Haidt, J., & Graham, J. (2007). When morality opposes justice: Conservatives have moral
intuitions that liberals may not recognize. Social Justice Research, 20(1), 98-116.
Helliwell, J. F., & Putnam, R. D. (2004). The social context of well-being. Philosophical
transactions-royal society of London series B biological sciences, 1435-1446.
Hofmann, W., Wisneski, D. C., Brandt, M. J., & Skitka, L. J. (2014). Morality in everyday life.
Science, 345(6202), 1340 – 1343
Heider, F. (1958). The psychology of interpersonal relations. New York: Wiley.
Huta, V. (2015). The complementary roles of eudaimonia and hedonia and how they can be
pursued in practice. Positive psychology in practice: Promoting human flourishing in
work, health, education, and everyday life, 159-182.
Huta, V., & Ryan, R. M. (2010). Pursuing pleasure or virtue: The differential and overlapping
well-being benefits of hedonic and eudaimonic motives. Journal of Happiness
Studies, 11(6), 735-762.
Jacobs, A. (2017). I Think You're Fat. Retrieved from https://www.esquire.com/news-
politics/a26792/honesty0707/
YOU CAN HANDLE THE TRUTH 69
Kawakami, K., Dunn, E., Karmali, F., & Dovidio, J. F. (2014). Mispredicting affective and
behavioral responses to racism. Science. 323, 276-278.
Kifer, Y., Heller, D., Perunovic, W. Q. E., & Galinsky, A. D. (2013). The good life of the
powerful the experience of power and authenticity enhances subjective well-
being. Psychological Science, 24(3), 280-288.
Kumar, A. & Epley, N. (in press). Undervaluing gratitude: Expressors
misunderstand the consequences of showing
appreciation. Psychological Science.
Landis, J. R., & Koch, G. G. (1977). An application of hierarchical kappa-type statistics in the
assessment of majority agreement among multiple observers. Biometrics, 363-374.
Lee, F. (1993). Being Polite and Keeping MUM: How Bad News is Communicated in
Organizational Hierarchies1. Journal of Applied Social Psychology, 23(14), 1124-1149.
Levine, E. E., & Schweitzer, M. E. (2014). Are liars ethical? On the tension between
benevolence and honesty. Journal of Experimental Social Psychology, 53, 107-117.
Levine, E. E., & Schweitzer, M. E. (2015). Prosocial lies: When deception breeds
trust. Organizational Behavior and Human Decision Processes, 126, 88-106.
Loewenstein, G. & Schkade, D. (1999). Wouldn't it be nice?: Predicting future feelings. In D.
Kahneman, E. Diener, & N. Schwartz (Eds.), Well-being: The foundations of hedonic
psychology (pp. 85-105). New York: Russell Sage Foundation.
Lyubomirsky, S., Sheldon, K. M., & Schkade, D. (2005). Pursuing happiness: the architecture of
sustainable change. Review of General Psychology, 9(2), 111.
YOU CAN HANDLE THE TRUTH 70
Mallett, R. K., Wilson, T. D., & Gilbert, D. T. (2008). Expect the unexpected: failure to anticipate
similarities leads to an intergroup forecasting error. Journal of Personality and Social
Psychology, 94(2), 265.
Margolis, J. D., & Molinsky, A. (2008). Navigating the bind of necessary evils: Psychological
engagement and the production of interpersonally sensitive behavior. Academy of
Management Journal, 51(5), 847-872.
Mayer, J. E. (1957). The self-restraint of friends: A mechanism in family transition. Social
Forces, 230-238.
McCord, P. (2017). Powerful: Building a culture of freedom and responsibility. Silicon Guild.
Molinsky, A., & Margolis, J. (2005). Necessary evils and interpersonal sensitivity in
organizations. Academy of Management Review, 30(2), 245-268.
Montoya, A. K., & Hayes, A. F. (2015, May). Estimating and testing indirect effects in
withinsubject mediation analysis: A path-analytic framework. In Annual Convention of
Myers, D. G., & Diener, E. (1995). Who is happy?. Psychological Science, 6(1), 10-19.
Rosen, S., & Tesser, A. (1970). On reluctance to communicate undesirable information: The
MUM effect. Sociometry, 253-263.
Rosenfeld, L. B. (1979). Self‐disclosure avoidance: Why I am afraid to tell you who I
am. Communications Monographs, 46(1), 63-74.
Rudd, M., Norton, M., & Aaker, J. (2014) Getting the Most Out of Giving: Pursuing Concretely-
Framed Prosocial Goals Maximizes Happiness. Journal of Experimental Social
Psychology, (54), September, 11-24.
Ryan, R. M., & Deci, E. L. (2001). On happiness and human potentials: A review of research on
hedonic and eudaimonic well-being. Annual Review of Psychology, 52(1), 141-166.
YOU CAN HANDLE THE TRUTH 71
Schaerer, M., Kern, M., Berger, G., Medvec, V., & Swaab, R. I. (2018). The illusion of
transparency in performance appraisals: When and why accuracy motivation explains
unintentional feedback inflation. Organizational Behavior and Human Decision
Processes, 144, 171-186.
Schein, C., & Gray, K. (2017). The theory of dyadic morality: Reinventing moral judgment by
redefining harm. Personality and Social Psychology Review, 1088868317698288.
Scott, K. (2017). Radical Candor: Be a Kick-ass Boss Without Losing Your Humanity. Pan
Macmillan.
Simmons, J., Nelson, L., Simonsohn, U. (2013) Life after P-hacking. Meeting Soc. Personality
Soc. Psych., New Orleans. http://ssrn .com/abstract=2205186
Slepian, M. L., Chun, J. S., & Mason, M. F. (2017). The experience of secrecy. Journal of
Personality and Social Psychology, 113(1), 1.
Srivastava, S., Tamir, M., McGonigal, K. M., John, O. P., & Gross, J. J. (2009). The social costs
of emotional suppression: a prospective study of the transition to college. Journal of
Personality and Social Psychology, 96(4), 883.
Steger, M. F., Kashdan, T. B., & Oishi, S. (2008). Being good by doing good: Daily eudaimonic
activity and well-being. Journal of Research in Personality, 42(1), 22-42.
Stone, D., Heen, S., & Patton, B. (2010). Difficult conversations: How to discuss what matters
most. Penguin.
Strauss, A., & Corbin, J. M. (1990). Basics of qualitative research: Grounded theory procedures and techniques.
Sage Publications, Inc.
Tangney, J. P., Stuewig, J., & Mashek, D. J. (2007). Moral emotions and moral behavior. Annual
Review of Psychology, 58, 345-372.
YOU CAN HANDLE THE TRUTH 72
Tesser, A., & Rosen, S. (1972). Similarity of objective fate as a determinant of the reluctance to
transmit unpleasant information: The MUM effect. Journal of Personality and Social
Psychology, 23(1), 46.
Tesser, A., Rosen, S., & Tesser, M. (1971). On the reluctance to communicate undesirable
messages (the MUM effect): A field study. Psychological Reports, 29(2), 651-654.
Tetlock, P. E. (1985). Accountability: The neglected social context of judgment and
choice. Research in organizational behavior, 7(1), 297-332.
Tetlock, Philip E., Linda Stitka, and Richard Boettger (1989), "Social and Cognitive Strategies
for Coping with Accountability: Conformity, Complexity, and Bolstering," Journal of
Personality and Social Psychology, 57 (4), 632-640.
Urry, H. L., Nitschke, J. B., Dolski, I., Jackson, D. C., Dalton, K. M., Mueller, C. J., ... &
Davidson, R. J. (2004). Making a life worth living neural correlates of well-
being. Psychological Science, 15(6), 367-372.
Van Boven, L., Loewenstein, G., Dunning, D. & Nordgren, L.L. (2013). Changing places:
Empathy gaps in emotional perspective taking. In J. Olson & M. P. Zanna
(Eds.), Advances in Experimental Social Psychology, (pp. 117-171).
Waterman, A. S. (1990). The relevance of Aristotle’s conception of eudaimonia for the
psychological study of happiness. Theoretical and Philosophical Psychology, 10, 39–44.
Waterman, A. S. (1993). Two conceptions of happiness: Contrasts of personal expressiveness
(eudaimonia) and hedonic enjoyment. Journal of Personality and Social Psychology,
64(4), 678–691.
YOU CAN HANDLE THE TRUTH 73
Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of brief measures
of positive and negative affect: the PANAS scales. Journal of Personality and Social
Psychology, 54(6), 1063.
Weingart, L. R., Behfar, K. J., Bendersky, C., Todorova, G., & Jehn, K. A. (2015). The
Directness and Oppositional Intensity of Conflict Expression. Academy of Management
Review, 40(2), 235-262.
Wilson, T. D., & Gilbert, D. T. (2005). Affective forecasting knowing what to want. Current
Directions in Psychological Science, 14(3), 131-134.
Wilson, T. D., Wheatley, T., Meyers, J. M., Gilbert, D. T., & Axsom, D. (2000). Focalism: a
source of durability bias in affective forecasting. Journal of Personality and Social
Psychology, 78(5), 821.
Wood, A. M., Linley, P. A., Maltby, J., Baliousis, M., & Joseph, S. (2008). The authentic
personality: A theoretical and empirical conceptualization and the development of the
Authenticity Scale. Journal of Counseling Psychology, 55(3), 385.
YOU CAN HANDLE THE TRUTH 74
Tables
Table 1. Enrollment and attrition across conditions (Study 1a)
Assignme
nt to
condition
Day 1 Day 2 Day 3
Complete
d any
daily
surveys
Completed
any daily
survey and
the two-
week
follow-up
Honesty 50 49 45 45 49 43
98.00% 90.00% 90.00% 98.00% 86.00%
Kindnes
s59 57 55 55 58 58
96.61% 93.22% 93.22% 98.31% 98.31%
Control 48 47 44 44 47 43
97.92% 91.67% 91.67% 97.92% 89.58%
Total 157 153 144 144 154 144
97.45% 91.72% 91.72% 98.09% 91.72%
Note. Percentages reflect the proportion of individuals assigned to condition that completed
surveys at each subsequent time-point.
YOU CAN HANDLE THE TRUTH 75
Table 2. The anticipated and experienced effects of honesty and kindness on communicators (Study 1)
Experience
Manipulation
checks Subjective Impact two-weeks later
Enjoyment Social
Connection Meaning
Daily
Honest
y
Daily
Kindness
Long-
term
honest
y
Long-
term
kindnes
s
Long-
term
hedonic
well-
being
Long-term
relational
Improvement
Long-
term
relational
harm
Long-term
eudaimonic
well-being
Appreciation
for the
experience
Forecasts (Study 1b)
Honesty M2.41 2.83 3.57 5.59 3.66 3.62 4.52 4.82 4.73 4.39
SD 1.01 1.09 0.88 N/A N/A 1.26 1.51 1.41 1.37 1.30 1.08 1.52
n222 222 222 222 222 222 222 222 222 222
Kindnes
sM3.57 4.22 3.72 3.62 6.15 4.75 5.19 2.98 4.91 5.15
SD 1.00 0.84 0.90 N/A N/A 1.65 1.01 1.35 1.22 1.57 1.21 1.23
n222 222 222 222 222 222 222 222 222 222
Control M3.21 3.45 3.60 4.79 4.85 4.09 5.14 3.27 5.19 5.09
SD 0.86 0.81 0.81 N/A N/A 1.26 1.26 1.16 0.99 1.21 0.95 1.17
n222 222 222 222 222 222 222 222 222 222
Experiencers (Study 1a)
Honesty M3.46 3.71 3.79 5.92 5.02 5.44 3.70 4.13 4.80 3.34 4.74 5.20
SD 0.82 0.69 0.59 1.13 1.24 1.44 1.50 1.24 1.34 1.58 1.18 1.11
n49 49 49 49 49 43 43 43 43 43 43 43
Kindnes
sM3.59 3.88 3.63 4.55 5.80 3.45 5.51 4.26 4.41 2.90 4.69 5.07
SD 0.71 0.66 0.64 1.47 0.86 1.40 1.34 1.29 1.46 1.34 1.08 1.36
n58 58 58 58 58 58 58 58 58 58 58 58
Control M3.27 3.47 3.34 5.05 4.81 4.37 4.47 3.62 4.16 2.85 4.36 4.73
SD 0.56 0.64 0.54 1.16 1.03 1.29 1.49 1.10 1.29 1.41 1.16 1.20
n47 47 47 47 47 43 43 43 43 43 43 43
Note. Experience ratings were made on a five-point rating scale. Subjective impact ratings were made on seven-point rating scale.
YOU CAN HANDLE THE TRUTH 76
Table 3. Two-week Reflections: Qualitative coding of responses from Study 1a/c
Coding
category Definition Example Kappa Total Honesty Kindness Control p
What happened during the experiment
Avoided people or
spoke less
Participant explicitly discusses avoiding
social interaction (either by talking less or
avoiding people) to cope with experiment.
-In order to deal with the
demands of this experiment,
I basically did not
communicate nearly as
much as I normally do.
0.75 3.5% 2.3% 3.4% 4.8% .83
Expressed myself
more honestly
(manipulation
check)
Participant discusses how they opened up
to others more, filtered themselves less,
expressed their true feelings or opinions.
This is mutually exclusive with "Avoided
people or spoke less."
-I lowered how much I
filtered in the way I speak to
others
0.73 16.8% 48.8% 3.4% 2.4% <.01
Expressed myself
less honestly
Participant discusses how they opened up
to other less, filtered themselves more,
expressed their true feelings or opinions
less, bit their tongue, etc. This is mutually
exclusive with "Avoided people or spoke
less."
-I felt like I was being
restrained in communicating
how I truly felt 0.91 17.5% 4.7% 27.6% 16.7% .01
Discussed
difficult topics
Participant mentions that they engaged in
challenging or difficult conversations, or
mention topics of conversations that seem
difficult or challenging.
-It effected my relationship
with my boyfriend. I told him
the truth about how i felt
sometimes, which lead to
our break up
0.88 6.3% 11.6% 6.9% 0.0% .09
Engaged in self-
reflection
Participant discusses consulting their own
feelings, looking inwards, thinking about
their own feelings, desires and
relationships, or learning about
themselves during the experiment.
-…enabled me to recognize
things about myself and my
interactions that I hadn't
before
0.9 27.3% 20.9% 20.7% 42.8% .03
Generated
conflict
Participant mentions disagreements,
arguments, or upsetting/offending others
as a result of their behavior in the
experiment.
-…she felt like she was
being attacked when it was
anything but
0.8 3.5% 11.6% 0.0% 0.0% <..01
Table continued…
YOU CAN HANDLE THE TRUTH 77
Coding
category Definition Example Kappa Total Honesty Kindness Control p
How the experience compared to expectations
Easier for self
Participant mentions that they were
already honest [kind], that they did not
find themselves in situations or
discussing topics that required them to
alter their behavior compared to their
normal, or that this experience was in
general, not as difficult as they
expected.
-It wasn't hard because I did not
come across any uncomfortable
questions that would have made
me uneasy if I had to be honest
about
0.75 5.6% 7.0% 5.2% 4.8% .89
Better
interaction with
others
Participant mentions that others reacted
more positively to their honesty
[kindness] than they expected,
discusses feeling surprised when
someone reacted positively, or that the
experiment caused less conflict than
expected.
-Was surprised how little anyone
cared about my honest feedback;
was well received
0.83 9.1% 18.6% 8.6% 0.0% .01
Harder for self
Participant mentions that they were
surprisingly dishonest [unkind] before
the experiment, that they found
themselves in situations or discussing
topics that required them to alter their
behavior more than they expected, or
that this experience was in general,
more difficult than they expected.
- I thought it would have been
easier than it was. 0.76 18.2% 14.0% 20.7% 19.0% .68
Worse
interaction with
others
Participant mentions that others reacted
more negatively to their honesty
[kindness] than they expected, or
discusses feeling surprised when
someone reacted negatively, or that the
experiment caused more conflict than
expected.
-I was surprised to find that one of
my friends got detached from the
conversation
1 4.2% 9.3% 3.4% 0.0% .19
Long-term Impact
Learned about
importance of
communication
Participant says they have become
more conscious of their communication
as a result of the experiment
- help me learn to be more aware
of how I interact with people and
the tone I set by interacting a
certain way and the choices I
make
0.62 32.9% 32.6% 31.0% 35.7% .89
Experiment had
positive life
impact
Participant says that the experiment
caused a positive change in their life
-I have smiled more and felt much
happier than before. Also, I have
had more conversations with
people that I didn't regularly talk
to.
0.82 11.2% 11.6% 5.5% 4.8% .24
Experiment
caused them to
seek out life
changes in
future
Participant says that they learned new
things about how they want to
behave/change in the future
-I have told myself to be more
open and honest with people
because it's the best way to do
things with the most tangible
results from what Ive seen.
0.80 42.0% 39.5% 41.4% 45.2% .86
YOU CAN HANDLE THE TRUTH 78
Note. In the two-week reflections, participants answered 5 open-ended questions about how they behaved during the study, what
surprised them, and how the study impacted them. We used an iterative coding procedure (Strauss & Corbin, 1990) to code
participants’ responses. For coding, we combined the data from Study 1 with the data from our pilot study (see online supplemental
materials). We first developed an initial coding scheme, based on our hypotheses and our knowledge of participants’ responses.
Then, we had two research assistants, blind to the study hypotheses, independently code 15 participants’ responses. Then we met to
discuss the codes and made edits to the guide to clarify the categories and resolve disagreement. We repeated this procedure three
times; in round one, the research assistants coded 15 responses, in round 2, they coded 30, and in round three, they coded 50. In
each round, we revised the guide and the research assistants’ previous codes were deleted. After round three, we developed the
final coding guide. The research assistants used the final guide to code 100 responses independently, which we used to test for
agreement. The research assistants then coded the remaining participant responses independently and resolved disagreement
through discussion. In the table, kappa reflects the level of agreement between the two coders’ independent coding of the 100
participant transcripts for each category. Kappa values above .81 reflect excellent agreement; Kappa values above .61 reflect
substantial agreement (Landis & Koch, 1977). Frequencies in columns labeled Total, Honesty, Kindness, and Control reflect the
percentage of participants who mentioned each coding category within each condition (in each study). “P” reflects the p-value
associated with a chi-square test of proportions of the null hypothesis that each code appeared in equal frequency across the three
experimental conditions.
YOU CAN HANDLE THE TRUTH 79
Table 4. The anticipated and experienced effects of honesty on communicators and targets (Study 2)
Askers Responders
Enjoyment
Social
Connectio
n
Relational
Harm Meaning Enjoyment
Social
Connectio
n
Relational
Harm Meaning
Forecasts
M3.3 4.1 2.77 3.95 3.12 3.75 3.05 3.89
S
D1.01 0.9 1.41 0.78 1 0.93 1.22 0.81
n51 51 51 51 51 51 51 51
Experiencers
M4.17 4.48 2.23 4.28 3.85 4.31 2.26 4.11
S
D0.9 0.71 1.21 0.78 0.96 0.85 1.27 0.78
n50 50 50 50 50 50 50 50
YOU CAN HANDLE THE TRUTH 80
Table 5. The anticipated and experienced effects of honesty on communicators and targets (Study 3)
Enjoyme
nt
Social
Connecti
on
Relation
al Harm
Meani
ng
Target'
s
Reacti
on
Communicator
s' Forecast
M2.91 3.63 3.08 3.67 3.73
S
D1 0.88 1.15 0.65 0.95
n106 106 106 106 106
Communicator
s' Experience
M3.87 4.17 2.47 3.97 2.91
S
D1.02 0.9 1.21 0.93 1.06
n106 106 106 106 106
Targets'
Experience
M4.07 4.29 2.46 3.92 2.63
S
D0.9 0.81 1.21 0.83 0.95
n103 103 103 103 103
YOU CAN HANDLE THE TRUTH 81
Table 6. Mediation results: Indirect effect of Target's reaction, mediating the effect of Perspective (Forecast vs. Experience)
on DV (Study 3)
Model 1 Model 2 Model 3 Model 4
DV:
Enjoymen
t
Social
Connectio
n
Relational
Harm Meaning
Lower CI 0.2136 0.1426 -.7259 0.0844
Upper CI 0.6769 0.5112 -.3373 0.3727
Estimate 0.4173 0.3100 -.5314 0.2178
Note. Results depict bootstrapped confidence intervals and point estimates of the indirect effect of Target’s reaction as a mediator for
the within-subjects difference between each DV at each time-point (forecast before the conversation and experience after the
conversation). We used the MEMORE macro for within-subjects mediation using SPSS (Montoya & Hayes, 2017), in which the
independent variable is represented by the repeated measurements of the mediator variable and the dependent variable.
HANDLE THE TRUTH 82
Figures
Figure 1. The anticipated and experienced effects of honesty and kindness on hedonic
well-being during the 3-day experience (Study 1).
1
2
3
4
5
HedonicWell-being
(Enjoyment)
Note. Ratings were made on five-point rating scales. Box plots depict the lower quartile, median,
and upper quartile of each set of ratings. Whiskers end at minimum and maximum ratings
(Spear style). Forecasters (Study 1b) significantly underestimated Experiencers’ (Study 1a)
enjoyment within the Honesty condition (p < .001), but not the Kindness or Control conditions.
HANDLE THE TRUTH 83
Figure 2. The anticipated and experienced effects of honesty and kindness on social
connection during the 3-day experience (Study 1).
1
1.5
2
2.5
3
3.5
4
4.5
5
Social Connection
Note. Ratings were made on five-point rating scales. Box plots depict the lower quartile, median,
and upper quartile of each set of ratings. Whiskers end at minimum and maximum ratings
(Spear style). Forecasters (Study 1b) significantly underestimated Experiencers’ (Study 1a)
social connection within the Honesty condition (p < .001), but not the Kindness or the Control
conditions.
HANDLE THE TRUTH 84
Figure 3. The anticipated and experienced effects of honesty and kindness on
eudaimonic well-being during the 3-day experience (Study 1).
1
1.5
2
2.5
3
3.5
4
4.5
5
Eudaimonic Well-being
(Meaning)
Note. Ratings were made on five-point rating scales. Box plots depict the lower quartile, median,
and upper quartile of each set of ratings. Whiskers end at minimum and maximum ratings
(Spear style). Forecasters (Study 1b) marginally underestimated Experiencers’ (Study 1a)
meaning associated with Honesty condition (p= .088), but not the Kindness or the Control
conditions.
HANDLE THE TRUTH 85
Figure 4. The anticipated and experienced effects of honest conversations among
communicators (Responders) and targets (Askers) (Study 2).
Panel A. Enjoyment
1
2
3
4
5
Hedonic Well-being
(Enjoyment)
Panel B. Social Connection
1
2
3
4
5
Social Connection
Panel C. Meaning
1
2
3
4
5
Eudaimonic Well-being
(Meaning)
HANDLE THE TRUTH 86
Figure 5. The anticipated and experienced effects of honest conversations among
communicators (Study 3).
Panel A. Enjoyment
Forecast Experience
1
2
3
4
5
Hedonic Well-being
(Enjoyment)
Panel B. Social Connection
Forecast Experience
1
2
3
4
5
Social Connection
Panel C. Meaning
Forecast Experience
1
2
3
4
5
Meaning
HANDLE THE TRUTH 87
Note. Ratings were made on five-point rating scales. Box plots depict the lower quartile,
median, and upper quartile of each set of ratings. Whiskers end at minimum and maximum
ratings (Spear style). For every measure, the difference between Experiencers and Forecasters
is significant, ps < .001. That is, Communicators underestimated the enjoyment, social
connection, and meaning associated with honesty.
HANDLE THE TRUTH 88
Appendix A. Verbal Instructions for Recruitment and Three-Day Communication
Intervention in Study 1a
Please listen carefully.
The study is about communication in everyday life. In this study, you will be asked to be very
conscious of your interpersonal communication. We expect that as a result of participating in this
study, you will learn about the way they communicate with and relate to others. However, you
may be asked to communicate in ways that could cause discomfort. You should only participate
if you are truly willing to be thoughtful about your communication and are open to
communicating in different ways.
To participate, you must show up to the Wharton Behavioral Lab and take an introductory
survey, which will take roughly 5 minutes. Then, you will learn more about the study. In order to
participate in the full study, you will have to follow detailed for three days, which will specify
how you should communicate with others in your daily interactions. You will also have to
complete 3 surveys (on Thurs night, Friday night, and Saturday afternoon), which will each take
about five minutes. You will take one survey about your experience two weeks later. You will
receive each survey via email.
In exchange for participating in this study, you will earn a $5 show-up fee during the lab session,
and $20 and the chance to win an iPad mini after you have completed the entire study (all nightly
surveys and 2-week follow up). Your $20 payment will be paid either directly to you by the
experimenter, through PayPal, Venmo or you can choose to receive a $20 amazon e-gift card
instead.
In addition to the payment of $20, we will run a lottery for an iPad mini. Thus, you will also
have a chance to win an iPad mini in exchange for your participation.
If you do not want to join this study, you can check out of the lab at this time.
Please take a moment to think about your decision. You are in no way obligated to participate in
this research and you can choose to leave the study at any time. You can head to check out if you
do not want to participate.
[wait before proceeding to study instructions]
MISPREDICTING HONESTY 89
All conditions:
In this study, you will be asked to reflect upon your social communication. Often, speaking with
others requires balancing honesty and kindness. Being completely open and honest about our
thoughts, feelings, and opinions, can sometimes upset others and be unkind. Alternatively, being
kind, considerate, and helpful towards others sometimes means not being 100% honest.
Control:
Throughout the next three days – that means today, tomorrow, and the following day - please be
conscious of the way you communicate with others. Please act as you normally would
throughout the length of this study. You should not change your behavior, but you should be
conscious of it.
You should act as you normally would with your closest relational partners. However, you
should NOT tell them, or anyone else, any specific information about this study. They can only
know that you were asked to pay special attention to your interpersonal communication. After
the study has ended, you can share any information you’d like about this study.
Please think about what it means to be conscious of your communication. Feel free to raise your
hand if you have questions. [field questions, wait for a moment] Is everyone ready to continue?
If so, you can complete the next link on your computer.
Honesty:
Throughout the next three days – that means today, tomorrow, and the following day - be honest
in every conversation you have with every person you talk to. Really try to be completely candid
and open when you are sharing your thoughts, feelings, and opinions with others. You should be
honest in every conversation you have, in every interaction, with every person in your life. Even
though this may be difficult, try your best to be honest.
Being authentic, honest, and true to oneself are important virtues. Embrace these virtues every
day for the next three days. When someone asks you how you feel, tell them the truth. That
means saying you feel happy only when you feel happy and saying you feel sad when you feel
sad. When you are giving your opinion, be completely honest. You should provide positive
opinions only when you truly feel positive, and you should provide negative opinions when you
feel negative.
You should be particularly honest with your closest relational partners. However, you should
NOT tell them, or anyone else, any specific information about these instructions. They can only
know that you were asked to pay special attention to your interpersonal communication. After
the study has ended, you can share any information you’d like about this study.
Please think about what it means to be completely honest. Feel free to raise your hand if you
have questions. [field questions, wait for a moment] Is everyone ready to continue? If so, you
can complete the next link on your computer.
MISPREDICTING HONESTY 90
Kindness:
Throughout the next three days – that means today, tomorrow, and the following day - please
strive to be kind in every conversation you have with every person you talk to. Really try to be
caring and considerate when you are sharing your thoughts, feelings, and opinions. You should
be kind in every conversation you have, in every interaction, with every person in your life. Even
though this may be difficult, you should do your absolute best to be kind.
Being kind and helpful, and avoiding harming others are important virtues. Embrace these
virtues every day for the next three days. When someone asks you how you feel, give a kind
answer. That means taking their feelings and state of mind into consideration. When you are
giving your opinion, be kind. You should provide opinions kindly and focus on the needs and
feelings of those around you.
You should be particularly honest with your closest relational partners. However, you should
NOT tell them, or anyone else, any specific information about these instructions. They can only
know that you were asked to pay special attention to your interpersonal communication. After
the study has ended, you can share any information you’d like about this study.
Please think about what it means to be kind. Feel free to raise your hand if you have questions.
[field questions, wait for a moment] Is everyone ready to continue? If so, you can complete the
next link on your computer.
MISPREDICTING HONESTY 91
Appendix B. Open-ended reflection Questions in the Survey Administered Two Weeks after
the Three-Day Communication Intervention (Study 1a/c)
Please think about your experience in this study. How did focusing on communicating
kindly [communicating honestly, being conscious of your communication] change the
way you thought, felt, behaved, and communicated with others? What were the key
difficulties you experienced during the study? What have you learned?
Please describe your experience and insights in as much detail as possible, using the
space below.
Did your experience in the study cause you to implement any changes in your life? What
have you done since the study to enact kindness in your daily life? What, if any, long-
term changes have you made?
Please describe how your communication has or has not changed as a result of this study,
using the space below.
Did your experience in the study surprise you in any way? Think back to your
expectations of this experience. Was there anything particularly interesting or unexpected
that occurred during your participation?
What did you learn in this study - either about yourself, others, or the way you
communicate?
Is there anything else you would like to share with the research team, about your
experience in the study?
MISPREDICTING HONESTY 92
Appendix C. Measures in the Reflection Survey Administered Two Weeks after the Three-
Day Communication Intervention in Study 1a
Participants rated their agreement with the following statements (1 = Strongly disagree, 7 =
Strongly agree):
Long-term honesty
This experience made me communicate more honestly.
This experience made me communicate more directly.
Long-term kindness
This experience made me communicate more kindly.
This experience made me communicate more nicely.
Long-term hedonic well-being
This experience made me less stressed.
This experience made me happier.
This experience made my life more pleasurable.
Long-term eudaimonic well-being
This experience gave me greater meaning.
This experience made me more thoughtful.
This experience brought me self-awareness.
This experience led to personal growth.
This experience brought me self-improvement.
This experience created purpose in my life.
This experience made me a better person.
This experience made me a better communicator.
Long-term relational improvement
This experience improved how I interact with others.
This experience deepened my relationships.
Long-term relational harm
This experience strained my relationships.
This experience caused relational conflict.
Appreciation for the experience
I am grateful for the experience
I would want to repeat the experience
I would recommend the experience to others
I am glad I participated in the study
MISPREDICTING HONESTY 93
Appendix D. Questions asked in Study 2
1. How are you?
2. What’s new in your life?
3. How is school/work? Are you having any issues?
4. Are you having any issues in any of your relationships?
5. Is there anything you need my advice on? What is it?
6. What is your least favorite thing about me? / What behavior of mine bothers you most?
7. Are you happy in our relationship? Is there anything you would like to change?
8. How long do you expect our relationship to last?
9. Have you ever been frustrated with something I did but not told me about it? What is it?
10. Do you have any (positive or negative) opinions about me that you have been hesitant to
share? What are they?
11. Is there anything you have never told me about yourself or anything you have been
hiding from me? What is it?
12. What is your most terrible memory?
13. If you knew that in one year you would die suddenly, would
you change anything about the way you are now living? Why?
14. What does friendship mean to you?
15. What roles do love and affection play in your life?
16. How close and warm is your family? Do you feel your childhood
was happier than most other people’s?
17. How do you feel about your relationship with your mother?
18. What is your most embarrassing moment?
19. When did you last cry in front of another person? By yourself?
20. If you were to die this evening with no opportunity to
communicate with anyone, what would you most regret not having told
someone? Why haven’t you told them yet?
