You are on page 1of 9

Journal of Economic Psychology 33 (2012) 686694

Contents lists available at SciVerse ScienceDirect

Journal of Economic Psychology


journal homepage: www.elsevier.com/locate/joep

Review

Trust as a social and emotional act: Noneconomic considerations


in trust behavior q
David Dunning a,, Detlef Fetchenhauer b, Thomas M. Schlsser b
a
Department of Psychology, Uris Hall, Cornell University, Ithaca, NY 14853-7601, United States
b
Institute for Economic and Social Psychology, Patrizia Tower, Venloer Str 151-153 (8th oor), 50672 Cologne, University of Cologne, Germany

a r t i c l e i n f o a b s t r a c t

Article history: We review research suggesting that decisions to trust strangers may not depend on eco-
Available online 14 September 2011 nomic dynamics as much as emotional and social ones. Classic treatments of trust empha-
size its instrumental or consequential nature, proposing that people trust based on
JEL classication: expectations that their trust will be honored and the size of reward if it is. Data from
C72 our labs, however, focusing on the trust or investment game, suggest that people trust even
C91 when their expectations of reward fall below their general tolerance for risk. Further data
D81
from our lab suggests that people trust not out of a concern for the consequences of their
PsycINFO classication: actions as much as for the actions themselves. The emotions people report feeling about
3020 trusting versus withholding trust predicts their decisions much more strongly than the
3660 emotions they attach to the potential outcomes. Social dynamics, such as whether partic-
ipants have been assigned to a specic counterpart in the game, inuence whether they
Keywords: trust, even though their economic expectations and payoffs remain unchanged. The
Trust
dynamics surrounding decisions to trust are complex, and involve social and emotional
Economic games
considerations beyond economic ones.
Risk behavior
Exchange relationships 2011 Elsevier B.V. All rights reserved.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 687
2. Trust as an economic act . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 687
2.1. Expectations and decisions to trust . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 688
2.2. Payoffs and decisions to trust. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 689
3. Trust as an emotional act . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 689
4. Trust as a social act . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 690
4.1. Norms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 691
4.2. Signals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 691
4.3. Relationship status . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 691
5. Concluding remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 692
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 693

q
This manuscript is based on an invited keynote address given by David Dunning at the joint conference of the International Association for Research in
Economic Psychology, the Society for the Advancement of Behavioral Economics, and the International Confederation for the Advancement of Behavioral
Economics and Economic Psychology, Cologne, Germany, September 2010.
Corresponding author.
E-mail address: dad6@cornell.edu (D. Dunning).

0167-4870/$ - see front matter 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.joep.2011.09.005
D. Dunning et al. / Journal of Economic Psychology 33 (2012) 686694 687

1. Introduction
One must be fond of people and trust them if one is not to make a mess of life.
Foster (English novelist and essayist, 18791970).
Fosters assertion makes good sense. Trust is an essential underpinning of any type of social relationship, whether it is
between two people in marriage (Deutsch, 1958; Fehr, 1988), between boss and employee in a business rm (Kramer,
1998; Kreps, 1990), or between a government and its citizens in a democracy (Fukuyama, 1995). Trust among strangers
is also essential within any given collective of individuals. With trust, institutions can be built that benet all. Ebay, for
example, would not be possible if buyers and sellers had frequent fears about whether they could trust one another. Without
trust, a restaurant patron would have to follow the server to the back after handing over the credit card just to make sure
that the server is not ripping the credit card number for his or her own use.
As another example, in the town where one of us (Dunning) lives, Ithaca, New York, a new non-prot automobile sharing
program, called Ithaca CarShare, has emerged that runs on trust as much as it does on gasoline. Participants in this program
can forego owning their own car. Instead, they can pay an annual fee allowing them to borrow fuel-efcient cars that have been
strategically parked around the city. Participants can specify a time that they want to use a car, pay an additional fee for every
hour and mile they use that car, and then make sure to return the car to its proper place when they are done. If any car runs low
on gas, participants are asked to ll the cars gas tank using one of the programs credit cards.
Programs like this allow people to gain access to a car when they need it, without the chronic costs of automobile loan pay-
ments, insurance, gasoline, parking, washing, or maintenance. It allows them to choose the specic car for the task they need to
complete: a pickup truck to move heavy loads, a small compact for a trip to the supermarket, or a convertible for a scenic week-
end trip (Boutin, 2006). It is also arguably environmentally friendly, in that it keeps surplus cars off the roadwith claims that up
to ten car are kept off the road for each automobile in the shared eet (Dribben, 2007). However, in order for this program to
survive, people do have to rely on the benevolence and cooperation of strangers. They have to trust that others will return
the cars when they claim they willand will do so without damage. They have to trust that others use the gasoline credit cards
to ll only the cars in the program instead of some other car owned by a family member. To the extent that people abuse the
program, and the trust of its participants, it runs the risk of decay and failure.
Carshare programs like this one reveal that trust lies on the heart of social capital. To the extent that people can trust
others, they can work together to create benets that each individually cannot generate alone. Thus, it should come as no
surprise that a World Value Survey discovered that levels of trust within a country are correlated with that countrys rate
of economic growth (Fetchenhauer & Vegt, 2001; Knack & Keefer, 1997).
Not all treatments of trust, however, are so positive. There is a strain of thought in Western philosophy (see Hobbes, 1660/
1997 and Machievelli, 1515/2003) and neoclassical economics that trust in others, particularly strangers, is an action that
should be avoided at all costs. Consider a traditional neoclassical analysis of trust for the following economic game, which
can be presented to participants in the laboratory. Suppose the experimenter gave each participant US$5 and a choice: the
participant can keep the money or hand it over to a stranger, also taking part in the experiment, who will remain mutually
anonymous to the participant. If the participant decides to hand over the money, the experimenter will inate the amount to
US$20 and give the other participant his or her own choice to make: to keep the entire amount or give US$10 back to the
participant. This situation, called the trust or the investment game (Berg, Dickhaut, & McCabe, 1995; Camerer, 2003; Snijders
& Keren, 2001), presents a choice containing the central feature of what it means to trust a stranger, which in behavioral
economics is dened as making ones self vulnerable to exploitation by that stranger with the expectation of some benet
in return (Rousseau, Sitkin, Burt, & Camerer, 1998).
According to a strict neoclassical analysis, no participant should ever trust (i.e., give up the US$5) in this situation. This is
because the interaction between participant and the other individual is anonymous and one-off. As such, the other person is
under no pressure to act against his or her material self-interest, and thus will with certainty keep the entire US$20. Because
of this, the participant should never make himself or herself vulnerable to this certain exploitation (see Berg et al., 1995).
However, people do trust in this economic game. Moreover, people in the position of the other person frequently honor that
trust by returning money backproviding a dual-vexation for the traditional neoclassical analysis (Berg et al., 1995; Camer-
er, 2003; Dunning & Fetchenhauer, 2010a, 2010b; Snijders & Keren, 2001).
It is the focus of this manuscript to discuss the possible dynamics of this situation that may produce decisions to trust
another person. Although our research has not pointed yet to a specic answer about what motivates trust, it has given
us many clues about what types of dynamics to focus on. To our surprise, and the astonishment, we presume, of many the-
orists, variables that can be considered to be economic in nature play a much more minor role in decisions to trust than one
would expect. Instead, trust seems to be responsive to social or emotional factors much more than one would forecast. In the
material that follows, we discuss the case against trust as an economic decision and the case for it being a more social and
emotional one. We hasten to add that we do not wish to dismiss economic concerns entirely. They matter. They just do not
matter as much as most theorists typically think.

2. Trust as an economic act

Traditional treatments of trust in both behavioral economics and psychology are implicitly economic in nature, in that
they suggest that people trust when they harbor sufcient expectations that their trust will be reciprocated. A hard-line neo-
688 D. Dunning et al. / Journal of Economic Psychology 33 (2012) 686694

classical analysis of the trust game described above suggests that other participants would never give money back, but peo-
ple know that their peers do, at times, act in cooperative and pro-social wayseven to complete strangers. People know that
other people are not always driven by pure material selshness. For example, on January 2010 a 7.2 earthquake hit the Carib-
bean island of Haiti, killing over 200,000 people. By July on that year, US$1.3 billion had been raised and donated in the Uni-
ted States alone for earthquake relief (Smith, 2010). Thus, people have ample experience telling them that others, at times,
do not necessarily act in completely selsh ways. Thus, there is a non-trivial chance that people will act in a more cooper-
ative way when trust is placed in them. This chance may make it rational for people to extend their trust.
Key to this economic analysis is the notion that people trust because they are concerned, rst and foremost, with the out-
comes their trust might bring about. They are concerned about the likelihood that their trust will be reciprocated, as well as
the size of the benet they can potentially gain. In effect, their decisions to trust are consequentialist and instrumental. People
do not decide to trust as an end in itself, but rather as a means toward some goal (e.g., more money) they would like to attain.
This consequentialist and instrumental treatment of trust is transparent in economic thought. Recall the denition of
trust provided by Rousseau et al. (1998) above that stated that people make themselves vulnerable to exploitation by others
if they have reasonable expectations that their trust will be honored. The psychological literature on trust also focuses rst
and foremost on the issue of expectation (Barber, 1983; Deutsch, 1973; Kramer & Carnevale, 2001; Pruitt & Rubin, 1986). For
example, in a recent review on the psychological literature, Simpson (2007) stated that:
Intergroup researchers . . . have dened trust as a specic set of socially learned expectations that people hold about var-
ious social systems, ranging from other people to social organizations to the larger moral social order. . .Trust has also
been dened as a constellation of beliefs regarding the extent to which others are or will be concerned about ones per-
sonal welfare and best interests (p. 589).
Indeed, the very measurement of individual differences in trust within psychology focuses primarily on expectation. The
classic Rotter (1967, 1971) scale of interpersonal trust mostly assesses what expectations people have of their contemporar-
ies. Respondents, for example, indicate whether they agree with the proposition that Other people are primarily interested
in their own welfare.
Our work, however, suggests that decisions to trust strangers in the laboratory are not necessarily driven by such eco-
nomic factors as expectation and the size of the potential benet. To be sure, these factors do inuence decisions to trust,
but the inuence is rather meager in strength, and peoples decisions often fail to follow the logic of an economic, instrumen-
tal analysis.

2.1. Expectations and decisions to trust

Consider expectations. We have found that people often trust strangers even when their expectations suggest they should
do anything but. Across several studies, we have presented participants with the trust game described above, in which they
have to consider, for example, whether to give US$5 to a complete stranger for an opportunity, no way assured, of receiving
US$10 back (Fetchenhauer & Dunning, 2009). This game appears to be an appropriate vehicle to study trust, in that it ts the
denition provided by Rousseau et al. (1998) above (i.e., to gain the US$10, people must rst make themselves vulnerable to
the exploitation of others).
It also ts peoples intuitive ideas about trust. In one informal experiment, we asked participants to play the trust game,
and then presented them with 13 words that could be used to describe the game, based on previous debrieng interviews.1
Each participant could endorse up to six words as relevant, but tended to endorse far fewer (M = 4.5, or 35%). Only four words
were endorsed by a majority of participants. Two were risk (72.5%) and gambling (52%). The nal two were trust (61%) and faith
in others (63%), in which we should note that faith is often dened as belief or trust in the truth or trustworthiness of a person,
idea, or thing (Wikipedia, 2010). A full 78% of participants endorsed at least either trust or faith in their characterization of the
trust game, and 47% endorsed both, although we described the trust game in as neutral terms as possible and never came close
to alluding to these terms.
In studies in our lab, we have asked participants to assess the likelihood that their trust would be reciprocated by an
anonymous stranger in the trust game described above. We also asked a question aimed at assessing participants tolerance
for risk. Specically, we asked the minimum odds of winning they would need to induce them to bet US$5 in a lottery in
which they can win US$10. From this information, we could see if participants made decisions to trust that were rational,
in that they trusted only when their expectations of reciprocation were as high or higher than the minimum odds they would
accept to gamble in the lottery. For example, if participants thought there was a 60% chance that their trust would be recip-
rocated, and they would accept any lottery gamble with at least a 51% chance of winning, it would be appropriate to trust.
However, if they thought their trust would be honored only 30% of the time, and would pursue a lottery gamble only if they
had an 80% chance of winning, they should not trust.
Our studies revealed that participants had rather cynical expectations about their peers. Although nearly 80% of their
peers indicated they would reciprocate any trust placed in them by giving US$10 back, the average participant thought there

1
The full list of thirteen words, presented in random order to participants, was: intelligence, risk, trust, gambling, faith in others, investment, morality,
respecting others, altruism, hope, ethics, curiosity, and emotion.
D. Dunning et al. / Journal of Economic Psychology 33 (2012) 686694 689

was roughly only a 45% chance that their trust would be reciprocated (Fetchenhauer & Dunning, 2009). In subsequent stud-
ies, we have discerned at least one reason for this mistaken cynicism. People believe other people are more likely to be
untrustworthy than trustworthy because of asymmetries in the feedback they receive about other people. When people trust
others, sometimes their trust is reciprocated and sometimes that trust is violated. The other person abuses their trust,
leading them to become more cynical about their peers. However, when they decide to withhold trust, they receive no
comparable corrective feedback when that decision is in error. The other person never has a chance to show that he or
she is, indeed, trustworthy. Thus, people fail to correct overly cynical perceptions in a more optimistic direction. However,
if given this feedback, people are quick to erase much of their undue cynicism about their peers (Fetchenhauer & Dunning,
2010).
However, although cynical, participants in our studies prove to be quite trustingfar more trusting than their expecta-
tions should allow. In these studies, we calculated the proportion of participants who should trust, in that their expectations
about that their trust would be honored met or exceeded their tolerance for risk. Roughly 30% should have trusted given
their expectations and risk tolerance, but across studies nearly 65% didmeaning roughly 35% of people trusted others in
this economic game when they arguably should have not. They accepted levels of risk in their dealing with another person
that they claimed they would not accept if betting on nature (Fetchenhauer & Dunning, 2009).
Other work shows that expectations fail to play as large a role in trust decisions as they do in other decisions under risk. In
another study, participants were given a chance to bet US$5 in a lottery in which they had either a 46% versus an 80% chance
of winning US$10. Needless to say, the odds of winning inuenced decisions to gamblewhich those willing to take the gam-
ble rising from 29% to 78% as the odds of winning rosea 49% difference. However, this change in odds did not have a similar
effect in a trust game in which participants could risk US$5 with a chance that another person gave them US$10 back. Par-
ticipants were paired up with peers who had already made their decisions about whether they would reciprocate trust. They
were either told that they would be paired up with a person from a group in which 46% reciprocated, or from one in which
80% reciprocated. These probabilities were exact ones; having already collected the decisions of the partners; we knew with
precision the chance that a participants trust would be honored. The rate of trust across the two conditions rose only from
54% to 70%a 16% difference (Fetchenhauer & Dunning, in press). Note that the rate of trust at low oddsthat is, at 46%was
signicantly higher than the percentage willing to bet in a lottery. Again, people accepted risks on other people that they did
not take when betting on nature.

2.2. Payoffs and decisions to trust

Now, consider size of payoff. If trust is not inuenced that much by expectation, perhaps it is inuenced by the out-
comes or benets that people are considering. To be sure, the last study suggested that people treated a decision to trust
differently than a decision to gamble even when the trust decision and gamble presented the same odds and same pay-
offs to the self. The trust game, however, differs from the lottery gamble in one important respect. It typically contains a
possible benet not only for the participant but for the person the participant has been paired with. According to a strict
neoclassical analysis, people should not be concerned about that strangers outcomes at all, but perhaps they do. People
show social preferences, being concerned about the payoffs for all rather than just for themselves (e.g., Charness & Rabin,
2002), and this added benet to others may have been enough to push people toward an afrmative decision to trust.
People may be willing to take a risk if they can enlarge the pie for all, even if they potentially receive none of that
pie themselves.
Our work, however, suggests that a simple desire to enlarge the pie is not behind decisions to trust. For example, consider
a chance to risk 5 to receive a 50% chance of winning 10. People are much more willing to take that risk when it means
trusting another person with than when it means ipping a coin. But what if they consider a coin ip that offers benets to
another person as well as the self? Suppose if they won 10 that some other random person would receive 10, but that if
they lost, some random person would receive 20. Note that this set of payoffs exactly mirrors the payoff structure for self
and other contained in the trust game. Does this expanded set of payoffs induce people to be more likely to ip the coin? The
answer is a clear No. A chance to expand the pie does nothing to prompt people to take on a person risk (Mensching,
Schlsser, Dunning, & Fetchenhauer, 2011).
To be sure, we want to remind the reader that expectations do matter. In our work, decisions to trust are correlated with
expectations that trust will be reciprocated. The correlations, however, are modest in nature, ranging from .16 to .41 across
studies (Fetchenhauer & Dunning, 2009). And size of potential reward also does matter (e.g., Evans & Krueger, 2011). If peo-
ple can double their money by trusting another person, they are much more likely to trust than if they can, at best, recoup
their money (Schlsser, Mensching, Dunning, & Fetchenhauer, 2011). Our work, however, suggests that decisions to trust are
inuenced by a host of other, more noneconomic, considerations.

3. Trust as an emotional act

Trust, like any gamble, need not be only about the money in play. Part of the rewards or costs people might encounter are
the emotional reactions that wrap around the decision to trust. For example, it is increasingly recognized, in both economics
and psychology, that people may derive utility for their actions, in part, by the emotions they feel once the consequences of
690 D. Dunning et al. / Journal of Economic Psychology 33 (2012) 686694

those actions are known. People may feel disappointment if they fail to win, and they are reasonable in incorporating that
potential cost in the calculation their preferences (Bell, 1985; Loomes & Sugden, 1986). Similarly, people may feel regret if
the alternative action they could have chosen would have led to a much more favorable outcome (Bell, 1982; Loomes &
Sugden, 1982). They may feel surprise if they encounter a low probability outcome, whether that outcome is positive or neg-
ative, which will ratchet up the degree of emotion of satisfaction or dissatisfaction they feel (Mellers, Schwartz, Ho, & Ritov,
1997; Mellers, Schwartz, & Ritov, 1999).
Emotions, however, may play into decisions to trust in two very different ways. First, people may base their decisions, at
least in part, on how they think they will feel once they know how their partner has responded and the outcome of their
decision is known. For example, they may anticipate that they will feel elated and relieved if the person honors their trust,
but embarrassed and angry if their trust is violatedand use these anticipations as an input into whether they should make
themselves vulnerable to that other person. Emotions at this level can be termed anticipated emotions, and refer to the fore-
casts people make about their emotional reactions to the outcomes of their actions. These emotions, however, are different
from immediate emotions, after Loewenstein, Weber, Hsee, and Welch (2001), which we dene as the emotions people feel
about the decision options in front of them at the moment they consider whether or not to trust. For example, people may
feel nervous or excited about giving their money to a stranger as they contemplate it, or may feel calm or guilty as they mull
over the option of keeping the money.
In two studies, we have examined the emotions participants experience as they weigh whether to trust another person,
looking to see how well anticipated and immediate emotions predict who will extend versus withhold trust. For anticipated
emotions, we asked participants how positive, aroused, and in control they will feel if they trust another person and that
trust is rewarded. They answered similar inquiries about all other possible outcome scenarios as well (trusting and having
that trust violated; withholding trust and nding out the person would have rewarded that trust; withholding that trust and
discovering the other person would have taken all the money). For immediate emotions, we asked participants directly how
they feltalong the same emotional dimensionsabout giving the money and about keeping the money before they
committed to their actual decision.
Across both studies, we found that immediate emotions, particularly about keeping the money, strongly predicted who
would trust versus who would not (Schlsser et al., 2011). In one study, of those who felt better about keeping the money
and more negatively about giving it, only 35% decided to trust. Of those who felt the reverse, a full 91% decided to trust. (In
the other study, the comparable gures were 12% versus 80%). Immediate emotions were a strong predictor of the course of
action participants took, explaining 21% and 33% of variance in Studies 1 and 2, respectivelygures that are much higher
than the variance explained by expectation of return found in previous studies, ranging from 2.5% to 16% (Fetchenhauer &
Dunning, 2009). It is interesting that it was immediate emotions surrounding keeping the money that better predicted trust
decisions than emotions attached to giving the money to the other person. There is no obvious explanation, at least to us,
about why this would be so. Clearly, this suggests that different considerations or potential goals come to mind as partici-
pants consider keeping the money rather than giving it away. But, at the moment, we do not have a clear account of what
those different considerations might be.
In addition, and to our surprise, across both studies, anticipated emotions, once immediate emotions were controlled for,
failed to predict who would trust versus who would not, even though participants were remarkably accurate in predicting
the emotions they would feel once those trust games were played and their outcomes revealed. These results suggest that
emotions do predict who will trust, but in a way, once again, different from what an economic analysis would suggest.
According to an economic analysis that allows emotion to matter, emotions should factor into trust decisions by determin-
ing, in part, how people evaluate the potential outcomes of their actions. Our data suggest, however, that this does not mat-
ter as much as how people feel directly about the actions they have to choose between, not the outcomes those actions may
produce. That is, if peoples emotions feed into the trust decisions they make, they feed into it by determining the utility one
achieves by performing one action over the other, absent any utility derived from outcomes.
In addition, although only a correlational study, one observation indirectly suggests that participants immediate emo-
tions were a causal input into their decisions rather than just some by-product rationalization of a decision already achieved.
If reports of immediate emotions were mere justications of trust decisions reach via other means, one would expect imme-
diate emotions attached to giving the money to be correlated with those emotions associated with keeping it. Emotional re-
ports for either action, keeping versus giving, were uncorrelated in both studies, suggesting they were potentially
independent inputs into the decisions participants would make. Immediate emotions were also generally uncorrelated with
the anticipatory emotions that participants expressed (Schlsser et al., 2011).

4. Trust as a social act

These data suggest that decisions to trust hinge importantly on issues surrounding the act of trust itself rather than on its
potential outcomes. To be sure, outcomes matter, but there is something about the action itself that triggers a decision to-
ward or away from trust. In a word, that means that the decision to trust is importantly expressive as well as instrumental. By
expressive, we mean that people decide to trust because of the immediate payoff they receive once they perform the ac-
tion. What might those payoffs be?
D. Dunning et al. / Journal of Economic Psychology 33 (2012) 686694 691

4.1. Norms

Here, we regret to say that we have many candidates but no nal and denitive answer as of yet. Whatever might be in
play, we suspect that the considerations people have are social in nature. People might simply be following the dictates of
some social norm (Henrich & et al., 2010). For example, research on dictator games suggests that people strive to avoid
knowingly acting in a selsh way. People often prefer a division of money in a dictator game that equally divides the money
between themselves and another person over one in which they receive more money than the other person. However, if the
experimenters hide information about potential payoffs for the other person by merely placing a PostIt note over that infor-
mation, people chose the option that benets them the most, unconcerned about whether that produces a benecial or equal
payoff to the other person. They are told they can remove the PostIt note and look at their partners payoffs if they wish, but
they largely ignore the opportunity (Dana, Weber, & Kuang, 2007).
In a similar vein, people give large amounts of money to other people in dictator games, avoiding the selsh action of
keeping all the money for themselves. However, if the experimenter adds an even more selsh option for participants to
considerfor example, the dictator can take money away from the other personpeoples preferences to give money to
the other person largely evaporates. Indeed, a signicant plurality of participants choose to take money from the other
person (Bardsley, 2008; List, 2007). In short, the inclusion of the selsh option appears to change the norms surrounding
the choice presented to participantsor at least removes the norm to share money and to share it equally with the other
person.

4.2. Signals

Beyond possibly implicating social norms, a decision to extend or withhold trust also provides an immediate host of
signals, to both self and other, that about the personality and character of the person making the decision. People like
to think of themselves as altruistic and moral beings (Bodner & Prelec, 2002; Dunning, 2007; Monin & Miller, 2001),
and so might choose to trust to send a signal to themselves that they are, indeed, moral and altruistic by following the
moral dictates or norms suggested by the situation (Lotz, Schlsser, Cain, & Fetchenhauer, 2011). People also might be
motivated by what signals they send to other people, and thus choose to trust in order to achieve a reputation among oth-
ers as a reliable and cooperative person (Andreoni & Bernheim, 2009; Fehr, Brown, & Fehnder, 2009; Milinski, Semmann, &
Krambeck, 2002).

4.3. Relationship status

Whatever considerations may be in play, we are becoming aware that their inuence is not universal but begin only when
people are placed in a social relationship with another individual. They do not arise when the decision to trust is merely
hypothetical and involves no other person. For example, in one study in which people were asked to play the trust game,
a full 56% decided to give up their money when there was a real and non-trivial chance that they would have to act on that
decision, but only 38% decided to trust when the decision was completely hypothetical (Fetchenhauer & Dunning, 2009,
Study 2; see Holm & Nystedt, 2008, for similar ndings). This nding completely contradicts a long lineage of past research
showing the reverse: that people in hypothetical situations portray themselves as kind and generous individuals who just
never seem to show up when faced with a real opportunity to be seless (Balcetis & Dunning, 2008; Epley & Dunning,
2000; Epley & Dunning, 2006).
Why the reverse result? We can provide two speculative reasons. The rst is that people fail to feel strong immediate
emotions when the decision is hypothetical, a conjecture consistent with past research (Loewenstein, ODonoghue, &
Rabin, 2003; Van Boven, Loewenstein, & Dunning, 2005; Van Boven, Loewenstein, Welch, & Dunning, in press). Without
such strong emotions, people may fall back to a more conservative and cautious stance about what to do with their
money.
Second, the result may reverse because a hypothetical situation presents no social relationship with another whereas a
potentially real choice does. Real social relationships bring with them norms and practices that are more prosocial in avor.
For example, consider the following scenario from Kahneman, Knetsch, and Thaler (1986, p. S296):
A small photocopying shop has one employee who has worked in the shop for 6 months and earns $9 per hour. Business
continues to be satisfactory, but a factory in the area has closed and unemployment has increased. Other small shops have
now hired reliable workers at $7 an hour to perform jobs similar to those done by the photocopy shop employee. The
owner of the photocopying shop reduces the employees wage to $7 an hour.
Of participants reading this scenario, 83% thought it unfair for the shop owner to reduce the employees wages. However,
if the employee left and the shop owner offered the new employee only $7, a full 73% thought this perfectly acceptable.
Apparently, the presence of an ongoing relationship between owner and employee was enough to bring a different set of
norms and practices with it about how the owner and employee should treat each other.
We have found something similar in our studies of the trust game. The presence of a social relationship between two peo-
ple must be in place for high rates of trust to be observedeven if that relationship is minimal, eeting, and involves people
who will never learn the identity of one another. In one such study, students in one group were asked if they would trust
692 D. Dunning et al. / Journal of Economic Psychology 33 (2012) 686694

another person in the regular trust game described above; a full 55% said they would. In a second group, we described the
trust game, but instructed participants that their counterpart in the game had not been assigned yet and would not be as-
signed if they decided to keep their money. Thus, participants were not yet in a social relationship. In this situation, only 35%
of participants opted to trust. However, this decision was also a completely private one, and so we had to run a third con-
dition in which a person had been assignedand thus a social relationship had been establishedbut that this other person
would not learn about the trust game unless the participant gave up the money. Here, with a social relationship established,
62% of participants opted to give up their moneyeven though they could have kept the US$5 in complete privacy with the
other person never knowing (Dunning & Fetchenhauer, 2011).
A second study echoed these results. Participants were given a choice between playing a trust game or taking part in a
lottery. Given that simple choice, most participants opted for the lottery and only 28% choose to play the trust game and
to give their US$5 to the other person. However, in a second condition, participants were presented the choice is a slightly
more complicated way. To opt for the lottery, participants rst had to play a trust game with an assigned other person and to
indicate that they would keep their US$5. If they kept the money, they were then free to play the lottery if they wished. This
procedure still presented participants a choice between playing the lottery and the trust game, but it also compelled partic-
ipants to record a decision in a trust game involving another person. In this circumstance, 59% of participants chose to trust
the other person (with the rest, save one participant, opting to play the lottery). That is, placing the trust decision within a
relationship prompted participants to trust. Making it a decision that people, instead, had to volunteer for caused far fewer
people to trust another (Dunning & Fetchenhauer, 2011).
Other results echo these ndings. People will not give as much money in a dictator game if they rst have to make a
choice about entering into that game (Lazear, Malmendier, & Weber, 2009). People will donate to a victim of a loss in a lab-
oratory game if they have been assigned to a specic victim versus not having been assigned. People will donate more money
to a charity if there is a specic individual they have been assigned to help (Small & Loewenstein, 2003), particularly if that
individual falls inside their social group (Kogut & Ritov, 2005). In prisoner dilemma games, participants cooperate less if they
think through their decision before they are matched with a partner rather than after they have been matched (Yamagishi,
Terai, Kiyonari, Mifune, & Kanazawa, 2007).
There are many reasons to suspect that people would develop norms that prompt them to treat others in social rela-
tionships in cooperative or at least benign ways. For the individual, classic work has shown that cooperating rst in a pris-
oners dilemma gameand then treating others in a tit-for-tat manneris a strategy that survives across a wide range of
circumstances (Axelrod, 1984). Thus, individuals may have been shaped by human evolution to be nice as a rst move in
an exchange relationships, even when they harbor internal skepticism and would not make the move if considered
hypothetically.
Entire cultures, as well, may be shaped by evolution to demand that its members follow pro-social norms in their daily
exchanges, even among strangers. Henrich et al. (2010) has argued that strangers within a culture must follow pro-social dic-
tates, even if it is at times personally costly, to maintain a stable social structure that allows its members to thrive. Supporting
their argument, within cultures that depend on strangers for such daily necessities like food, people give to strangers in dic-
tator games and punish others who transgress norms for sharing even if the punisher is left worse off in the process. Cultures
in which members depend, instead, almost exclusively on kind or friends for daily necessities fail to exhibit such behavior.

5. Concluding remarks

In sum, our research suggests that any narrow analysis of trust, as measured in laboratory economic games as well as
in real life, would be misplaced. Trust appears to hinge on any number of dynamics, and our work is only beginning to
specify what those dynamics might be. Traditional treatments of trust assume that people are focused, rst and foremost,
on the potential outcomes of their actionsand we concede that peoples expectations about whether their trust will be
reciprocated, and well as the potential size of the gain if their trust is reciprocated, does hold some sway over their
decisions.
However, we increasingly assert that concerns about outcomes are only part of the story about why people trust. Trust is
not merely an instrumental or economic decision. It is also expressive, in that concerns about the act itself appear to matter a
great deal. Peoples decisions to trust are best predicted by the emotions they attach to the action itself rather than by emo-
tions they attach to possible outcomes. People are more likely to trust if there is an ongoing relationshipno matter how
eeting, minimal, or anonymouswith a counterpart than when no relationship exists, even though the odds and potential
payoffs of trusting are the same under both circumstances.
We regret we have yet to specify completely the emotional and social dynamics governing trust, but we are excited to be
in the huntand are pleased that although there are many candidate dynamics to consider, the number, at least for now, is
nite. But, feeling that we end the story in the middle, without a moral with which to end the tale, let us end our discussion
with an old German proverb that we offer as a tentative piece of advice about trust:
Vertrauen Sie, aber nicht zu viel.
Which translates in English to:
Trust, but not too much.
D. Dunning et al. / Journal of Economic Psychology 33 (2012) 686694 693

References

Andreoni, J., & Bernheim, B. D. (2009). Social image and the 5050 norm: A theoretical and experimental analysis of audience effects. Eonometric, 77,
16071636.
Axelrod, R. (1984). The evolution of cooperation. New York: Basic Books.
Balcetis, E., & Dunning, D. (2008). A mile in moccasins: How situational experience reduces dispositionism in social judgment. Personality and Social
Psychology Bulletin, 34, 102114.
Barber, B. (1983). The logic and limits of trust. New Brunswick, NJ: Rutgers University Press.
Bardsley, N. (2008). Dictator game giving: Altruism or artefact? Experimental Economics, 11, 122133.
Bell, D. E. (1982). Regret in decision making under uncertainty. Operations Research, 30, 961981.
Bell, D. E. (1985). Disappointment in decision making under uncertainty. Operations Research, 33, 127.
Berg, J., Dickhaut, J., & McCabe, K. (1995). Trust, reciprocity, and social history. Games and Economic Behavior, 10, 122142.
Bodner, R., & Prelec, D. (2002). Self-signaling and diagnostic utility in everyday decision making. In I. Brocas & J. Carillo (Eds.), Collected essays in psychology
and economics (pp. 122). Oxford University Press.
Boutin, P. (2006). Sell your car: Car sharing is the ultimate pedal-to-the-metal lifestyle option for upwardly mobile, status-conscious yuppies like me.
National Post (September 19), A21.
Camerer, C. F. (2003). Behavioral game theory: Experiments on strategic interaction. Princeton: Princeton University Press.
Charness, G., & Rabin, M. (2002). Understanding social preferences with simple tests. Quarterly Journal of Economics, 117, 817869.
Dana, J., Weber, R. A., & Kuang, J. X. (2007). Exploiting moral wiggle room: Behavior inconsistent with a preference for fair outcomes. Economic Theory, 33,
6780.
Deutsch, M. (1958). Trust and suspicion. Conict Resolution, 2, 265279.
Deutsch, M. (1973). The resolution of conict: Constructive and destructive processes. New Haven, CT: Yale University Press.
Dribben, M. (2007). Philly CarShare becomes well-traveled; With 30,000 members in 5 years, its a success story in its eld. Philadelphia Inquirer (October
29), A01.
Dunning, D. (2007). Self-image motives and consumer behavior: How sacrosanct self-beliefs sway preferences in the marketplace. Journal of Consumer
Psychology, 17, 237249.
Dunning, D., & Fetchenhauer, D. (2010a). Trust as an expressive rather than an instrumental act. In S. Thye & E. Lawler (Eds.). Advances in group processes
(Vol. 27, pp. 97127). New York: Emerald.
Dunning, D., & Fetchenhauer, D. (2010b). Understanding the psychology of trust. In D. Dunning (Ed.), Social motivation (pp. 147170). New York: Psychology
Press.
Dunning, D., & Fetchenhauer, D. (2011). The minimal relationship effect in decisions to trust. Unpublished manuscript, Cornell University University of
Cologne.
Epley, N., & Dunning, D. (2000). Feeling holier than thou: Are self-serving assessments produced by errors in self or social prediction? Journal of Personality
and Social Psychology, 79, 861875.
Epley, N., & Dunning, D. (2006). The mixed blessings of self-knowledge in behavioral prediction: Enhanced discrimination but exacerbated bias. Personality
and Social Psychology Bulletin, 32, 641655.
Evans, A. M., & Krueger, J. I. (2011). Elements of trust: Risk and perspective-taking. Journal of Experimental Social Psychology, 47, 171177.
Fehr, B. (1988). Prototype analysis of the concepts of love and commitment. Journal of Personality and Social Psychology, 4, 557579.
Fehr, E., Brown, M., & Fehnder, C. (2009). On reputation: A microfoundation of contract enforcement and price rigidity. The Economic Journal, 119, 333353.
Fetchenhauer, D., & Dunning, D. (2009). Do people trust too much or too little? Journal of Economic Psychology, 30, 263276.
Fetchenhauer, D., & Dunning, D. (2010). Why so cynical? Asymmetric feedback underlies misguided skepticism in the trustworthiness of others.
Psychological Science, 21, 189193.
Fetchenhauer, D., & Dunning, D. (in press). Betrayal aversion versus principled trustfulness: How to explain risk avoidance and risky choices in trust games.
Journal of Economic and Behavior & Organization.
Fetchenhauer, D., & Vegt, G. v. d. (2001). Honesty, trust and economic growth. A cross-cultural comparison of Western industrialized countries. Zeitschrift fr
Sozialpsy-chologie, 32, 189200.
Fukuyama, F. (1995). Trust: The social virtues and the creation of prosperity. New York: Free Press.
Henrich, J. et al. (2010). Markets, religion, community size, and the evolution and punishment. Science, 327, 14801484.
Hobbes, T. (1660/1997). The Leviathan. New York: Touchstone Press.
Holm, H., & Nystedt, P. (2008). Trust in surveys and gamesA methodological contribution on the inuence of money and location. Journal of Economic
Psychology, 29, 522542.
Kahneman, D., Knetsch, J. L., & Thaler, R. H. (1986). Fairness as a constraint on prot seeking: Entitlements in the market. American Economic Review, 76,
728741.
Knack, S., & Keefer, P. (1997). Does social capital have an economic payoff? A cross-country investigation. The Quarterly Journal of Economics, 112,
12511288.
Kogut, T., & Ritov, I. (2005). The identied victim effect: An identied group, or just a single individual? Journal of Behavioral Decision Making, 18, 157167.
Kramer, R. M. (1998). Paranoid cognition in social systems. Thinking and acting in the shadow of doubt. Personality and Social Psychology Review, 2, 251275.
Kramer, R. M., & Carnevale, P. J. (2001). Trust and intergroup negotiation. In R. Brown & S. Gaertner (Eds.), Blackwell handbook of social psychology,
Intergroup relations (Vol. 4, pp. 431-450). Oxford, UK: Blackwell Publishers.
Kreps, D. M. (1990). Corporate culture and economic theory. In J. Alt & K. Shepsle (Eds.), Perspectives on positive political economy (pp. 90143). Cambridge
University Press.
Lazear, E. P., Malmendier, U., & Weber, R. A. (2009). Sorting and social preference. Unpublished manuscript, Stanford University.
List, J. A. (2007). On the interpretation of giving in dictator games. Journal of Political Economy, 115, 482493.
Loewenstein, G., ODonoghue, T., & Rabin, M. (2003). Projection bias in predicting future utility. Quarterly Journal of Economics, 118(4), 12091248.
Loewenstein, G. F., Weber, E. U., Hsee, C. K., & Welch, N. (2001). Risk as feelings. Psychological Bulletin, 127, 267286.
Loomes, G., & Sugden, R. (1982). Regret theory: An alternative theory of rational choice under uncertainty. Economic Journal, 92, 805824.
Loomes, G., & Sugden, R. (1986). Disappointment and dynamic consistency in choice under uncertainty. Review of Economic Studies, 53, 271282.
Lotz, S., Schlsser, T., Cain, D., Fetchenhauer, D. (2011). What you dont know may hurt me: The effects of variations in justice sensitivity on altruism.
Unpublished manuscript, University of Cologne, Germany.
Machievelli, N. (1515/2003). In R. Goodwin (Ed.), The Prince. Wellesley, MA: Dante University Press. 2003.
Mellers, B. A., Schwartz, A., Ho, K., & Ritov, I. (1997). Decision affect theory: How we feel about risky options. Psychological Science, 8, 423429.
Mellers, B. A., Schwartz, A., & Ritov, I. (1999). Emotion-based choice. Journal of Experimental Psychology: General, 128, 114.
Mensching, O., Schlsser, T., Dunning, D., & Fetchenhauer, D. (2011). Why do people trust? A multi-perspective approach. Unpublished manuscript,
University of Cologne.
Milinski, M., Semmann, D., & Krambeck, H-J. (2002). Reputation helps solve the tragedy of the commons. Nature, 415, 424426.
Monin, B., & Miller, D. T. (2001). Moral credentials and the expression of prejudice. Journal of Personality and Social Psychology, 81, 3343.
Pruitt, D. G, & Rubin, J. Z (1986). Social conict: Escalation, stalemate, and settlement. New York: McGraw-Hill.
Rotter, J. B. (1967). A new scale for the measurement of interpersonal trust. Journal of Personality, 35, 651665.
Rotter, J. B. (1971). Generalized expectancies of interpersonal trust. American Psychologist, 26, 443452.
694 D. Dunning et al. / Journal of Economic Psychology 33 (2012) 686694

Rousseau, D. M., Sitkin, S. B., Burt, R., & Camerer, C. (1998). Not so different after all: A cross-discipline view of trust. Academy of Management Review, 23,
393404.
Schlsser, T., Mensching, O., Dunning, D., & Fetchenhauer, D. (2011). Is trust driven by efciency: Effects of mutual gain on the decision to trust. Unpublished
manuscript, University of Cologne.
Simpson, J. A. (2007). Foundations of interpersonal trust. In A. W. Kruglanski & E. T. Higgins (Eds.), Social psychology: Handbook of basic principles (2nd ed.,
pp. 587607). New York: Guilford.
Small, D. A., & Loewenstein, G. (2003). Helping A victim or helping THE victim: Altruism and identiability. Journal of Risk and Uncertainty, 26(1), 516.
Smith, A. (2010). Haiti donations: $1.3 billion. CNNMoney.com. <http://money.cnn.com/2010/07/09/news/international/haiti_donation/index.htm>
Accessed 17.10.10 (July 9).
Snijders, C., & Keren, G. (2001). Do you trust? Whom do you trust? When do you trust? Advances in Group Processes, 18, 129160.
Van Boven, L., Loewenstein, G., & Dunning, D. (2005). The illusion of courage in social predictions: Underestimating the impact of fear of embarrassment on
other people. Organizational Behavior and Human Decision Processes, 96, 130141.
Van Boven, L., Loewenstein, G., Welch, E., & Dunning, D. (in press). The illusion of courage in self-predictions: Predicting ones own behavior in embarrassing
situations. Journal of Behavioral Decision Making.
Wikipedia (2010). Faith. <http://en.wikipedia.org/wiki/Faith#cite_note-dict-0> Accessed 20.10.10.
Yamagishi, T., Terai, S., Kiyonari, T., Mifune, N., & Kanazawa, S. (2007). The social exchange heuristic: Managing errors in social exchange. Rationality and
Society, 19, 259291.