64 Scientific American, September 2019
Corruption, big or small, impedes the socioeco-
nomic development of nations. It affects economic ac-
tivities, weakens institutions, interferes with democ-
racy and erodes the public’s trust in government offi-
cials, politicians and their neighbors. Understanding
the underlying psychology of bribery could be crucial
to tackling the problem. Troublingly, our studies sug-
gest that mere exposure to corruption is corrupting.
Unless preventive measures are taken, dishonesty can
spread stealthily and uninvited from person to person
like a disease, eroding social norms and ethics—and
once a culture of cheating and lying becomes en-
trenched, it can be difficult to dislodge.CONTAGION
SUPPOSE YOU REFUSED the City Hall employee’s request
for a bribe. How would the experience influence your
response to a subsequent ethical dilemma? In lab-
oratory studies we conducted with behavioral re-
search ers Vladimir Chituc, Aaron Nichols, Heather
Mann, Troy Campbell and Panagiotis Mitkidis, which
are currently under review at an academic journal, we
sought an answer to that question.
We invited individuals to the behavioral lab in the
university to play a game that involved throwing a vir-
tual die for a reward. Everyone was told that they
would be compensated based on the outcome of mul-
tiple rolls. In practice, however, they could misreport
their rolls to earn more money. So all participants
faced a conflict between playing the game by the rules
and behaving dishonestly to earn more. We created
this setup to assess how individuals balance external
and internal—or psychological—rewards when mak-
ing ethical decisions. Research that Nina Mazar, On
Amir and one of us (Ariely) published in 2008 indi-
cates that most people act unethically to the extent
that they can benefit while also preserving their mor-
al self-image—an observation they described as the
theory of self-concept maintenance.
Our game involved rolling a virtual die 30 times on
iPads. Many behavioral economists have used similar
paradigms involving physical dice and coins to assess
dishonesty in so-called decontextualized games—that
is, games that are not affected by social or cultural
norms. Prior to each roll, participants were instructed
to choose a side of the die in their mind—top or bot-
tom—and report their choice after seeing the out come
of the roll. They would earn a fixed amount of money
per dot on the side they reported each time. So every-
one had a financial incentive to cheat by reporting
the high-paying side. For example, if the outcome
of the roll was two on the top of the die and five on
the bottom of the die, people might be tempted to
say they had chosen “bottom” before the roll even if
they had not.
This paradigm does not allow us to know whether
someone cheated in a specific roll. Nevertheless, when
results are aggregated across all rolls and participants
in a group, the proportion of favorable rolls chosencan be compared against chance (50 percent) to as-
sess the magnitude of dishonesty.
After participants received instructions about the
game and how they would make money in the session,
which they would get to take home, they were ran-
domly assigned to a low- or a high-payment version.
Those in the high-payment game would do exactly the
same thing as those in the low-payment game but
earn 10 times more. Everyone was told about the exis-
tence of the other game. Then, half the participants in
the low-payment condition were offered the possibili-
ty of paying a bribe to be switched to the high-pay-
ment game.
The research assistant administering the session
framed that opportunity as illegal to engender a mor-
al dilemma similar to one that might arise in real life.
The person mentioned that the boss was not around
and that the participant could easily be switched to the
high-paying game without anyone finding out. Thus,
we ended up with three groups of people: low-pay-
ment no bribe, high-payment no bribe, and bribe ex-
posed; the last group could be further split into bribe
payers and bribe refusers. This arrangement allowed
us to assess how ethically those exposed to the bribe
would behave after having encountered the offer.
We administered three versions of the test to a to-
tal of 349 individuals in our behavioral lab. In the first
two studies, some participants were offered the possi-
bility of paying a $2 bribe to be placed in the high-pay-
ment version of the game, and 85 percent of them paid.
Crucially, we observed that in the games they went on
to play, bribe-exposed participants cheated more than
participants who did not receive such a request. In the
second study, for example, bribe-exposed participants
cheated 9 percent more than those who played the
high-payment version of the game and 14 percent
more than participants who played the low-payment
version of the game but had not been asked for a bribe.
In a third study, we tested whether people act more
immorally when they pay a bribe or when they are
merely exposed to one. We made the bribe costlier at
$12, and 82 percent turned down the request, giving
us a large sample size of bribe refusers. Disturbingly,
even when we limited our analysis to this group of
apparently ethical individuals, we found that bribe-
exposed individuals cheated more than those who did
not receive the illegal request. Taken together, results
from these three experiments suggest that receiving a
bribe request erodes individuals’ moral character,
prompting them to behave more dishonestly in sub-
sequent ethical decisions.ERODING NORMS
OUR WORK SUGGESTS that bribery is like a contagious
disease: it spreads quickly among individuals, often
by mere exposure, and as time passes it becomes hard-
er and harder to control. This is because social norms—
the patterns of behavior that are accepted as normal—
impact how people will behave in many situations, in-IN BRIEF
Corruption damag-
es economies, insti-
tutions and demo-
cratic structures.
Exposure to brib-
ery can, in and of
itself, be corrupt-
ing—suggesting
a mechanism by
which unethical
behavior may spread
through society.
Social norms ́ù-
ence ethical behav-
ior. Surprisingly,
however, the innate
tendency to cheat
(or not) is the same
across countries,
myåÈïyÿDåïmyà-
ences in actual cor-
ruption levels.
Further research
is required into what
drives bribery and
corruption, how it
spreads and how
it can be controlled.