ArticlePDF Available

The Motivating-Uncertainty Effect: Uncertainty Increases Resource Investment in the Process of Reward Pursuit

Authors:

Abstract and Figures

Can a reward of an uncertain magnitude be more motivating than a reward of a certain magnitude? This research documents the motivating-uncertainty effect and specifies when this effect occurs. People invest more effort, time, and money to qualify for an uncertain reward (e.g., a 50% chance at $2 and a 50% chance at $1) than a certain reward of a higher expected value (e.g., a 100% chance at $2). This effect arises only when people focus on the process of pursuing a reward, not when they focus on the outcome (the reward itself). When the focus is on the process of reward pursuit, uncertainty generates positive experience such as excitement and hence increases motivation. Four studies involving real rewards lend support to the motivating-uncertainty effect. This research carries theoretical implications for research on risk preference and motivation and practical implications for how to devise cost-efficient consumer incentive systems.
Content may be subject to copyright.
Journal of Consumer Research, Inc.
The Motivating-Uncertainty Effect: Uncertainty Increases Resource Investment in the
Process of Reward Pursuit
Author(s): Luxi Shen, Ayelet Fishbach, and Christopher K. Hsee
Source:
Journal of Consumer Research,
(-Not available-), p. 000
Published by: The University of Chicago Press
Stable URL: http://www.jstor.org/stable/10.1086/679418 .
Accessed: 21/11/2014 15:04
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.
.
The University of Chicago Press and Journal of Consumer Research, Inc. are collaborating with JSTOR to
digitize, preserve and extend access to Journal of Consumer Research.
http://www.jstor.org
This content downloaded from 128.135.212.25 on Fri, 21 Nov 2014 15:04:57 PM
All use subject to JSTOR Terms and Conditions
000
2014 by JOURNAL OF CONSUMER RESEARCH, Inc. Vol. 41 February 2015
All rights reserved. 0093-5301/2015/4105-0010$10.00. DOI: 10.1086/679418
Please use DOI when citing. Page numbers are not final.
The Motivating-Uncertainty Effect:
Uncertainty Increases Resource Investment
in the Process of Reward Pursuit
LUXI SHEN
AYELET FISHBACH
CHRISTOPHER K. HSEE
Can a reward of an uncertain magnitude be more motivating than a reward of a
certain magnitude? This research documents the motivating-uncertaintyeffect and
specifies when this effect occurs. People invest more effort, time, and money to
qualify for an uncertain reward (e.g., a 50% chance at $2 and a 50% chance at
$1) than a certain reward of a higher expected value (e.g., a 100% chance at $2).
This effect arises only when people focus on the process of pursuing a reward,
not when they focus on the outcome (the reward itself). When the focus is on the
process of reward pursuit, uncertainty generates positive experience such as ex-
citement and hence increases motivation. Four studies involving real rewards lend
support to the motivating-uncertainty effect. This research carries theoretical im-
plications for research on risk preference and motivation and practical implications
for how to devise cost-efficient consumer incentive systems.
Consumers often invest resources such as effort, time, or
money toward a reward of an uncertain magnitude. We
seek to answer a foundational question in consumer psy-
chology: how reward uncertainty affects resource investment.
Does uncertainty increase or decrease motivation?
Previous research offers contradictory answers. On the
one hand, both normative theories (e.g., Expected-Utility
Theory; Bernoulli 1738; von Neumann and Morgenstern
Luxi Shen (luxi.shen@cuhk.edu.hk) is assistant professor of marketing,
Chinese University of Hong Kong, Hong Kong. Ayelet Fishbach (ayelet
.fishbach@chicagobooth.edu) is the Jeffrey Breakenridge Keller Professor of
Behavioral Science and Marketing, University of Chicago. Christopher K. Hsee
(chris.hsee@chicagobooth.edu) is the Theodore O. Yntema Professor of Be-
havioral Science and Marketing, University of Chicago, Chicago, IL 60637,
USA. Address correspondence to Luxi Shen. This research is based on the first
author’s doctoral dissertation under the supervision of the second and third
authors. The authors thank Dan Bartels, Bill Goldstein, Scott Hawkins, David
Levari, Tony Lian, Devin Pope, Dilip Soman, Abby Sussman, Richard Thaler,
Oleg Urminksy, and George Wu for their thoughtful comments and advice; the
editors, Mary Frances Luce and Eileen Fischer, the associate editor, and the
reviewers for their many helpful suggestions; and the John Templeton Foun-
dation for research support.
Mary Frances Luce and Eileen Fischer served as editors and Stijn Van
Osselaer served as associate editor for this article.
Electronically published November 19, 2014
1944) and descriptive theories (e.g., Prospect Theory; Kahn-
eman and Tversky 1979) predict that when facing gain op-
tions, people are risk averse and prefer a reward of a certain
magnitude over a reward of an uncertain magnitude. This
preference is robust and universal in evaluations of out-
comes, and thus it is possible that people would express
higher motivation toward a reward of a certain magnitude.
On the other hand, research on affective experience shows
that people sometimes enjoy uncertainty (Marschak 1950;
Pascal 1670; Vosgerau, Wertenbroch, and Carmon 2006;
Wagenaar 1989; Wilson et al. 2005), and thus it is also
possible that people would express higher motivation toward
a reward of an uncertain magnitude.
In this research, we propose that a reward of an uncertain
magnitude can be more motivating than a reward of a certain
magnitude, even if the uncertain reward is objectively worse.
We focus on the uncertainty about the magnitude of the
reward; that is, we compare rewards of a fixed and known
magnitude with rewards of a probabilistic (i.e., uncertain)
magnitude. We define motivation as the investment of re-
sources, including time, money, and effort, that people ex-
pend to acquire a reward (Toure´-Tillery and Fishbach 2014).
For example, we compare how high bidders will bid for an
item of an uncertain magnitude (e.g., a bag that contains
either 10 or 5 chocolate truffles with even chances) with
how high they will bid for an item of a certain magnitude
This content downloaded from 128.135.212.25 on Fri, 21 Nov 2014 15:04:57 PM
All use subject to JSTOR Terms and Conditions
000 JOURNAL OF CONSUMER RESEARCH
Please use DOI when citing. Page numbers are not final.
(e.g., a bag that contains 10 chocolates for sure). Meanwhile,
we hold constant the degree of certainty with which they
will successfully complete the actions and thereby qualify
for the reward.
In the following sections, we first introduce our “domi-
nated-uncertainty paradigm,” in which we compare the ef-
fectiveness of certain rewards with uncertain rewards of a
lower expected value. We explore why and when uncertainty
increases motivation in this paradigm. We then report four
studies that demonstrate a boost in motivation when people
focus on the process of pursuing a reward of an uncertain
magnitude, but not when they focus on the outcome of
reward pursuit. We further identify that the positive expe-
rience during the reward pursuit underlies the boost in mo-
tivation.
REWARDS OF AN UNCERTAIN
MAGNITUDE
People invest effort, time, and money in pursuit of re-
wards (Amir and Ariely 2008; Hsee, Yu, et al. 2003; Kivetz
2005; Kivetz, Urminsky, and Zheng 2006; Koo and Fish-
bach 2010; Nunes and Dre`ze 2006; Soman 1998; Zhang
and Huang 2010; Zhang et al. 2011). We distinguish between
two basic types of rewards that people pursue: certain and
uncertain. A certain reward has a fixed and known mag-
nitude, V; an uncertain reward has several (i.e., at least two)
potential magnitudes with known or unknown probabilities.
We compare motivation toward a certain reward, for ex-
ample, a 100% chance of getting V, to motivation toward
an uncertain reward, for example, a 50% chance of getting
Vand a 50% chance of getting c. In our paradigm, we only
compare positive values (i.e., Vis always positive), and the
largest reward magnitude is held constant across the certain-
and uncertain-reward conditions (i.e., cis always a positive
value smaller than V).
This paradigm provides a strong test for the motivational
impact of uncertainty because the best-case scenario—re-
ceiving V—is the same across both the certain and uncertain
rewards and the certain reward always has a higher expected
value. Alternatively, if we had kept the expected value con-
stant by introducing a probable option of a higher magnitude
(e.g., certain reward: a 100% chance at $10; uncertain re-
ward: a 50% chance at $15 and a 50% chance at $5), a
stronger motivation toward the uncertain reward could have
been the result of the desire for the best possible outcome
($15 in the example above; see research on consumer op-
timism: Dhar, Gonzales-Vallejo, and Soman 1995, 1999;
Goldsmith and Amir 2010).
In our dominated-uncertainty paradigm, a person needs to
meet a performance standard in order to qualify for a reward.
The performance standard can be known, unknown, or re-
vealed in the course of pursuing the reward. For example,
people may need to meet a given standard, do better than
others without information on others’ performance, or exceed
others’ bids in a multiple-round unsealed auction. Regardless
of performance standards, a greater investment of resources
increases the chance of qualifying for a reward and indicates
stronger motivation. Although the investment of resources is
under personal control, the size of the reward is not. For those
qualifying for an uncertain reward, the exact reward mag-
nitude is determined by chance. It is completely independent
of their motivation, and all people are aware that the mag-
nitude of the reward is out of their personal control. In fact,
our paradigm reflects many everyday situations in which the
final reward depends on a combination of motivation (effort)
and uncontrollable factors (luck).
Existing motivation research has explored other types of
uncertainty. Specifically, research on achievement motiva-
tion has varied the probability of successful goal attainment
and found that people express strong motivation for mod-
erately difficult tasks (Atkinson 1957; Brehm and Self 1989;
Locke and Latham 2006). A moderate probability of success
increases motivation because the task poses a challenge,
which increases physiological arousal (Brehm et al. 1983).
Our paradigm is different because we do not manipulate the
likelihood of successful performance; however, we share the
intuition that uncertainty, albeit from a different source, can
trigger positive feelings and physiological arousal.
FOCUS ON THE PROCESS VERSUS THE
OUTCOME OF REWARD PURSUIT
We suggest that uncertainty is exciting and predict that
uncertainty increases motivation by generating positive ex-
perience during the pursuit of a reward of an uncertain mag-
nitude. Research has shown that uncertainty about positive
outcomes stimulates positive feelings and arousal (i.e., ex-
citement and enjoyment; Bar-Anan, Wilson, and Gilbert
2009; Berns et al. 2001; Lee and Qiu 2009; Moon and
Nelson 2014; Schultz, Dayan, and Montague 1997; Wilson
et al. 2005; Zillmann 1983). Thus, uncertainty can be a
source of positive experience. Other research, in turn, has
suggested that positive feelings increase motivation (Erez
and Isen 2002; Fishbach, Shah, and Kruglanski 2004; Klein
and Fishbach 2014; Kuhl and Kaze´n 1999). That is, people
prefer to pursue actions that are associated with positive
feelings and work harder in those actions than in affect-
neutral actions (Custers and Aarts 2005; Czikszentmihalyi
1990; Ferguson 2008; Fishbach and Choi 2012).
Importantly, positive experience should matter when peo-
ple focus on pursuing the reward and not when they focus
on evaluating reward outcomes. People naturally attend to
the process during the reward pursuit (e.g., in goal striving),
and during this phase, consummatory and affect-rich aspects
increase motivation (Andrade and Iyer 2009; Choi and Fish-
bach 2011; Deci and Ryan 1985; Jeffrey 2009; Kivetz 2003;
Le Menestrel 2001; Millar and Tesser 1986; Sansone et al.
1992). By contrast, people spontaneously attend to the out-
come of their actions when prospectively deciding whether
and how much to invest in pursuit of a reward (e.g., in goal
setting; Bagozzi and Dholakia 1999; Toure´-Tillery and Fish-
bach 2012). The focus on the outcome elicits a deliberative
mind-set in which instrumental and affect-poor aspects are
This content downloaded from 128.135.212.25 on Fri, 21 Nov 2014 15:04:57 PM
All use subject to JSTOR Terms and Conditions
SHEN, FISHBACH, AND HSEE 000
Please use DOI when citing. Page numbers are not final.
central to the evaluation of actions (Gollwitzer 2012; Heck-
hausen 1987; Higgins, Kruglanski, and Pierro 2003; Hsee
and Rottenstreich 2004). Thus, uncertainty decreases mo-
tivation when people focus on the outcome of the pursuit.
Preliminary evidence for our theory comes from work by
Goldsmith and Amir (2010). These authors documented that
promotions offering uncertain rewards often have a similar
appeal to promotions offering certain rewards because people
are innately optimistic. For example, participants in their stud-
ies were similarly interested in an uncertain promotion that
offered either a desirable can of soda or a less desirable bag
of microwave popcorn as they were in a certain promotion
that offered a can of soda, because those faced with the un-
certain promotion optimistically expected to receive the de-
sirable outcome. The authors also mentioned excitement as
an additional mechanism other than optimism but left it un-
tested. By contrast, we predict greater resource investment in
pursuit of an uncertain reward than a certain reward. That is,
we explore situations in which people respond even more
positively to the dominated uncertain reward than to the cer-
tain reward. Moreover, we empirically show that uncertainty
brings excitement and boosts motivation.
Notably, an increase in motivation when people pursue a
reward of uncertain magnitude might also reflect an illusion
of control (Langer 1975), or an illusory belief that good
things happen to those who work hard (Callan, Ellard, and
Nicol 2006). Although such illusory beliefs can also result
in an increase in motivation when uncertainty is involved,
we argue that based on research on the unsealed-fate su-
perstition (Strickland, Lewicki, and Katz 1966), these illu-
sory beliefs would predict that uncertainty increases moti-
vation only if the size of the reward is yet to be determined.
By contrast, we predict that uncertainty increases motivation
regardless of whether the size of the uncertain reward is
determined before or after the qualifying action is taken.
We further note that our hypothesis may seem incongruent
with theory and research on risk aversion, which has doc-
umented a reliable preference for certainty in choice and
evaluation of gains (Arrow 1965; Bernoulli 1738; Gneezy,
List, and Wu 2006; Holt and Laury 2002; Hsee and Weber
1997; Kivetz 2003; Rabin 2000; Rabin and Thaler 2001;
Simonsohn 2009). However, research on risk aversion as-
sesses evaluations of outcomes; we, on the other hand, study
motivation in the process of reward pursuit.
To recapitulate, we use a dominated-uncertainty paradigm
in which we measure individuals’ motivation in an activity
that, if they succeed, will qualify them for a reward of either
a certain or an uncertain magnitude. The uncertain reward
always has a lower expected value than the certain reward,
and we assess motivation by the amount of effort, time, and
money individuals invest. We propose three specific hy-
potheses:
H1:The motivating-uncertainty effect: a reward of an
uncertain magnitude can be more motivating than
a reward of a certain magnitude, even if the un-
certain reward has a lower expected value.
H2: The motivating-uncertainty effect occurs when the
focus is on the process of reward pursuit rather
than on the outcome (the reward itself).
H3: The positive experience in the process of reward
pursuit mediates the effect of reward uncertainty
on motivation.
We next report four studies that tested these hypotheses
across a variety of tasks in which successful performance,
if achieved, was rewarded with a real reward of either a
certain or an uncertain magnitude. Study 1 provides an initial
demonstration that uncertainty can increase motivation (hy-
pothesis 1). Study 2 generalizes the motivating-uncertainty
effect to a range of reward probabilities (from 1% to 99%).
Study 3 demonstrates that uncertainty increases motivation
during rather than before pursuing a reward (hypothesis 2).
Finally, study 4 examines the mechanism underlying the
motivating-uncertainty effect: uncertainty generates positive
experience in the process, thereby boosting motivation (hy-
potheses 2 and 3).
STUDY 1: UNCERTAINTY BOOSTS
MOTIVATION
To test whether uncertain rewards increase motivation
(hypothesis 1), study 1 compared the percentages of partic-
ipants who completed a water-drinking task across two con-
ditions: in the certain-reward condition, the reward was $2;
in the uncertain-reward condition, the reward was either $2
or $1, determined by a coin toss. The task difficulty was
held constant across the two conditions, but the probability
of getting the more desirable reward ($2) varied and was
neither contingent on the participants’ efforts nor within
their control. We predicted that a higher percentage of par-
ticipants would complete the task for the uncertain reward
than for the certain reward, even though the uncertain reward
had a lower expected value.
Method
Eighty-seven college students (47 women, 40 men) from
the University of Chicago took part in this study. Their task
was to drink 1.4 liters (about 1.48 quarts) of water in 2
minutes for different cash rewards. Drinking this amount of
water is challenging but possible for most people and does
not pose health risks, as would be the case for a significantly
larger amount of water.
The study adopted a 2 (reward certainty: certain vs. un-
certain) between-participants design. Participants completed
the study in individual sessions. They were sitting in front
of a table on which we presented a $1 bill, a stack of two
$1 bills, a straw, and a large (3.8 liter) pitcher of water. The
experimenter explained the participants’ task: they must
drink some of the water using the straw over a period of 2
minutes in return for a cash reward. In the certain-reward
condition, if the participants drank down to an indicated line
on the pitcher (corresponding to 1.4 liters of water, although
This content downloaded from 128.135.212.25 on Fri, 21 Nov 2014 15:04:57 PM
All use subject to JSTOR Terms and Conditions
000 JOURNAL OF CONSUMER RESEARCH
Please use DOI when citing. Page numbers are not final.
participants did not know the exact amount), they would
receive $2. While giving these instructions, the experimenter
pointed to the stack of two $1 bills. In the uncertain-reward
condition, if the participants drank down to the indicated
amount, they would receive a reward. The experimenter
would then flip a coin to determine whether the reward
would be $2 or $1. While giving these instructions, the
experimenter pointed to both the $1 bill and the stack of
two $1 bills. Because a $2 reward could appear larger when
one compares it to $1 (Ariely, Loewenstein, and Prelec 2003;
Hsee and Zhang 2010; Hsee et al. 2009, 2013; Zhang 2014),
we let participants in both conditions see the $2 reward with
the $1 reward as a reference. In a pretest (Np76) that
used the same presentation formats, the subjective evalua-
tion of the $2 reward did not vary across two reward con-
ditions (on a 7-point scale of reward attractiveness: M
uncertain
p5.27, SD
uncertain
p1.35; M
certain
p5.30, SD
certain
p1.24;
t!.01, p1.90). As our dependent variable, we measured
whether the participant successfully completed the task by
drinking up to or more than the indicated amount.
Results and Discussion
In support of hypothesis 1, 70% of the participants in the
uncertain-reward condition successfully completed the drink-
ing task, whereas only 43% of the participants in the certain-
reward condition successfully completed the task (x
2
(1, Np
87) p6.25, pp.012). This finding suggests that uncertainty
increases motivation even though the uncertain reward had a
lower expected value and was therefore dominated by the
certain reward. Merely adding reward uncertainty motivated
an additional 27% of participants to finish drinking the in-
dicated amount of water.
This study employed a performance standard and mea-
sured whether participants met that standard (a discrete var-
iable). The motivating-uncertainty effect should also apply
to “do your best” goals, such as drinking as much water as
possible, which further allows us to measure motivation as
a continuous variable. In a follow-up study (Np50), we
used a similar procedure, except that it contained an un-
known performance standard to make participants’ likeli-
hood of qualifying for a reward increase continuously as
they worked harder. Specifically, participants learned that if
they drank more than the average of the participants who
already completed the study, they would receive a reward.
To implement the unknown standard, we informed partici-
pants about the average amount only after they had finished
the task. The findings in this follow-up study were consistent
with the main study. Participants who were incentivized by
an uncertain reward drank more water than those who were
incentivized by a certain reward (M
uncertain
p1.20 liter,
SD
uncertain
p.33 liter; M
certain
p.95 liter, SD
certain
p.43 liter;
t(48) p2.21, pp.032); that is, the objectively worse
uncertain reward increased consumption by 25%.
Taking the findings in the main and follow-up studies
together, we found that the prospect of acquiring an uncer-
tain reward increased effort investment. Because the un-
certain options had lower expected values, and because re-
search participants (and people in general) aspire for higher
expected values, we can further conclude that uncertainty
increased motivation even more than lower expected value
decreased motivation.
Study 1 adopted 50% as the probability for the uncertain
reward. A question remains of whether the motivating-un-
certainty effect is specific to a particular reward probability,
such as 50%, or is generalizable to different reward proba-
bilities. Because the motivating-uncertainty effect is based on
the positive feelings stimulated by uncertainty, we predict that
any specific probabilities, as well as an unspecified proba-
bility, should lead to a boost in motivation.
STUDY 2: THE MOTIVATING-
UNCERTAINTY EFFECT APPLIES TO
DIFFERENT REWARD PROBABILITIES
Study 2 explored different probabilities of receiving abetter
reward outcome and compared the motivational effect of a
certain reward with the motivational effect of each of these
uncertain rewards. Specifically, participants evaluated a series
of print advertisements in return for either a reward of acertain
magnitude (two certain-reward conditions: $0.20 and $0.50)
or a reward of an uncertain magnitude (six uncertain-reward
conditions: a chance at $0.50 instead of $0.20 ranging from
1% to 99% or an unspecified probability). Based on hypoth-
esis 1, we predicted that the participants would invest more
effort and time (i.e., evaluate more advertisements) in the
pursuit of an uncertain reward, regardless of reward proba-
bility, than in pursuit of a certain reward.
Method
Five hundred and thirty native English speakers (284
women, 246 men) in the United States completed the study
on Amazon Mechanical Turk. They were asked toparticipate
in a marketing survey in which they evaluated a number of
print advertisements in return for $1, and they had the op-
portunity to receive a bonus of either $0.20 or $0.50 by
completing extra work. This study adopted an eight-con-
dition (chance of receiving the larger instead of smaller
bonus: 0%, 1%, 40%, 50%, 60%, 99%, 100%, and unspe-
cified) between-participants design. We used the same task
and possible incentives across all conditions and only varied
the probability of getting the higher bonus.
The participants’ task was to view and evaluate print ad-
vertisements. Each print advertisement was displayed on the
screen for 12 seconds before participants could provide their
evaluation on a 10-point scale (1 p“it sucks”; 10 p“it’s
awesome”). Participants received a flat payment of $1 for
the first 60 print advertisements evaluated. Afterward, they
could continue evaluating advertisements or stop at any
point. We measured participants’ motivation as a continuous
variable, using a procedure similar to the one in the follow-
up to study 1. Participants read that to receive a bonus on
top of their flat payment, they would need to do better than
the average; that is, they would need to evaluate more ad-
This content downloaded from 128.135.212.25 on Fri, 21 Nov 2014 15:04:57 PM
All use subject to JSTOR Terms and Conditions
SHEN, FISHBACH, AND HSEE 000
Please use DOI when citing. Page numbers are not final.
TABLE 1
PERFORMANCE RESULTS FOR STUDY 2
Separate analyses
Combined analysis:
Number of extra ads evaluated
Number of extra
ads evaluated
among those
evaluating extra ads
Bonus Mean SD NPercentage of those
evaluating extra ads Mean SD
0% certain ($0.20) 1.58 2.84 61 41 8.23 3.69
1% uncertain (1% chance at
$0.50 and 99% chance at $0.20) 6.43** 3.66 62 69*16.29 2.80
40% uncertain (40% chance at
$0.50 and 60% chance at $0.20) 5.27*4.34 75 63 16.92*3.36
50% uncertain (50% chance at
$0.50 and 50% chance at $0.20) 7.27*** 3.40 68 75*** 15.01 2.93
60% uncertain (60% chance at
$0.50 and 40% chance at $0.20) 5.53** 3.42 62 73** 11.31 3.55
99% uncertain (99% chance at
$0.50 and 1% chance at $0.20) 6.61** 4.45 69 67*18.89*3.42
p% uncertain (either $0.50 or
$0.20, but the chance of receiv-
ing either is unknown) 5.61** 2.70 62 77*** 9.77 2.98
100% certain ($0.50) 2.24 3.31 71 48 9.63 3.84
N
OTE
.—Means and standard deviations are presented after reversed transformation. For each reported variable, numbers with asterisks are
significantly different at the given level from the 100%-certain-reward condition (i.e., our comparison standard).
*p!.05.
**p!.01.
***p!.001.
vertisements than the average number evaluated by others
who completed the survey. Participants did not know what
that average was, and thus, by putting more work into eval-
uating advertisements, they increased their chances of re-
ceiving a bonus.
We manipulated the bonus; table 1 presents the bonus op-
tions. Of the eight conditions, two were certain-reward con-
ditions (i.e., eligible participants received $0.20 for sure and
$0.50 for sure) and six were uncertain-reward conditions. Of
the six uncertain-reward conditions, five had a specified re-
ward probability (e.g., eligible participants had a 50% chance
of receiving $0.50 and a 50% chance of receiving $0.20), and
one had an unspecified reward probability (“you will receive
either $0.50 or $0.20, but the chance of receiving one or the
other is unknown”). The dependent variable was the number
of advertisements evaluated beyond the required 60 adver-
tisements.
Results and Discussion
Because the distribution of the number of extra advertise-
ments evaluated was highly skewed (Skewness p3.24, SD p
.11, p!.001), we log-transformed the responses prior to anal-
yses and report means and standard deviations after reversed
transformation. Notably, we obtained similar results using un-
transformed responses here and in subsequent studies. The
number of extra advertisements varied across conditions (F(7,
530) p5.24, p!.001; table 1). In support of hypothesis 1, a
contrast analysis revealed that participants in the uncertain-
reward conditions (1%, 40%, 50%, 60%, 99%, and unspecified
probability) evaluated more extra advertisements than those in
the certain-reward conditions (0% and 100%; M
uncertain
p6.08,
SD
uncertain
p3.65; M
certain
p1.92, SD
certain
p3.09; t(528) p
5.86, p!.001). In addition, participants in the uncertain-reward
conditions evaluated more extra advertisements than those in
the 100%-certain-reward condition (M
uncertain
p6.08, SD
uncertain
p3.65; M
100%-certain
p2.24, SD
100%-certain
p3.31; t(467) p3.97,
p!.001), suggesting that an uncertain $0.30 increment had a
larger effect on task motivation than a certain $0.30 increment.
Interestingly, although increasing the bonus by $0.30 was not
sufficient to motivate participants to evaluate significantly more
advertisements (for the difference between the two certain re-
wards: t!1), adding uncertainty, regardless of probabilities,
to the bonus size increased motivation: all paired comparisons
between each uncertain-reward condition and the 100%-cer-
tain-reward condition were significant (t’s 12.5, p’s !.05),
and all paired-comparisons between any two uncertain con-
ditions were not significant (t’s !1.5, p’s 1.3). The similar
results across all uncertain conditions could be due to proba-
bility neglect (Rottenstreich and Kivetz 2006; Sunstein 2003),
in which case, although participants were sensitive to the prob-
able nature of an outcome, they were insensitive to different
probabilities.
We conducted further analyses to separate the effect on
(a) the percentage of participants who continued the task
beyond the required first 60 advertisements and (b) the av-
erage number of extra advertisements evaluated by those
who continued. Results of these analyses, also presented in
This content downloaded from 128.135.212.25 on Fri, 21 Nov 2014 15:04:57 PM
All use subject to JSTOR Terms and Conditions
000 JOURNAL OF CONSUMER RESEARCH
Please use DOI when citing. Page numbers are not final.
table 1, suggest that, compared to the economically better
and strictly dominating certain reward ($0.50 for sure), all
of the uncertain rewards increased either the percentage of
participants who continued the task, the number of extra
advertisements evaluated by those who continued the task,
or both.
An uncertain reward magnitude can mean either that the
probability of the large reward is known between 0 and 1
or that the probability of the large reward is unknown
(Knight 1921; Tversky and Fox 1995). In this study, we
examined both. We also explored different specific known
probabilities. Regardless of whether the reward probability
is specified, and regardless of what the specified probability
is, adding uncertainty to rewards boosts motivation.
Studies 1 and 2 assessed motivation in the process of
pursuing a reward. In the next study, we further assess mo-
tivation in planning an activity when the focus is on the
outcome. Comparing the effect of uncertainty when the fo-
cus is on the process versus the outcome of reward pursuit
allows us to explore the cause of the motivating-uncertainty
effect—that uncertainty generates positive experience in the
process. If uncertainty feels good, it should increase resource
investment when people decide whether or not to continue
investing resources but not when people evaluate the po-
tential rewards prior to the pursuit of these rewards. Fur-
thermore, it might be the case that putting participants in a
situation in which goal achievement was uncertain made
them more responsive to an uncertain reward because of a
better perceived fit. In the next study, we rule out this al-
ternative by introducing a moderator—the focus on outcome
versus process—and predicting that uncertainty would only
boost motivation when the focus is on the process.
STUDY 3: UNCERTAINTY INCREASES
MOTIVATION IN THE PROCESS OF
PURSUING A REWARD
Study 3 tested whether the focus on the process versus
the outcome of working toward a reward moderates the
motivating-uncertainty effect (hypothesis 2). We hypothe-
sized that a reward of an uncertain magnitude is more mo-
tivating than a reward of a certain magnitude in the process
of pursuing the reward but that the opposite pattern holds
beforehand when the focus is on the outcome. To test this
prediction, we set up an auction in which participants bid
against each other to win an auction item. The auction item
was either a bag of a certain number of Godiva chocolate
truffles or a bag of an uncertain number of Godiva chocolate
truffles. We manipulated how willingness to pay (WTP) was
elicited. In one condition, participants generated their WTP
through actual bidding (i.e., focus on the process) and in
the other condition, they generated their WTP in advance,
stating their reservation price before bidding (i.e., focus on
the outcome). Presumably, actual bidding draws people’s
attention to the process (the experience of the auction),
whereas an advance statement draws people’s attention to
the outcome (the value of the item; Cheema, Chakravarti,
and Sinha 2012; Hsee, Zhang et al. 2003). We predicted
higher WTP for the uncertain (vs. certain) auction item when
prices were solicited in the actual bidding rather than in the
advance statement.
Our paradigm in study 3 contains several notable features.
First, participants paid cash out-of-pocket for the auction
item so the bids were consequential; either participants won
the auction and paid for the auction item or they lost it.
Second, we determined the outcome of the uncertain reward
before the auction. The uncertain chocolate bag already had
either a small or a large number of truffles, but the exper-
imenter did not reveal the actual number to the participants
until the end of the auction. This latter feature allowed us
to rule out as a possible explanation the superstitious belief
that good things happen to those who work hard (Callan et
al. 2006; Converse, Risen, and Carter 2012; Langer 1975),
that is, that by paying more, one can increase his/her chances
of getting the larger number of truffles. Such magical belief
should not occur if the reward is determined before the
person invests resources (Strickland et al. 1966).
Method
One hundred and thirty-eight college students (79 women,
59 men) from the University of Chicago took part in this
study for a nominal payment of $2. This study adopted a 2
(reward certainty: certain vs. uncertain) #2 (elicitation
method: actual bidding vs. advance statement) between-par-
ticipants design. Participants completed the study in groups
of three. Their task was to bid on a bag of Godiva chocolate
truffles.
The experimenter assumed the role of the auctioneer and
ran each group separately. All participants saw a bag on the
table. In the certain-reward condition, the auctioneer un-
wrapped the bag to reveal four truffles. In the uncertain-
reward condition, the auctioneer did not reveal the content
of the bag; instead, he announced that there was an equal
chance that the bag contained two or four truffles. Partici-
pants in this condition saw a sample of a Godiva chocolate
truffle and learned that they would find out the exact number
of truffles in the bag by the end of the auction.
Participants then read the auction instructions: they would
bid for eight rounds, and in each round, they could either
repeat the bid from their last round or give a higher bid.
They would write down their bid on the paper and show
the bid to the auctioneer and other bidders after everyone
had finished writing. No other communication was allowed
during the auction. The winner of the auction would be the
person who submitted the highest bid in the final round in
the group. Then he/she would pay the second-highest bid
in the final round in return for the truffles (this procedure
incentivized bidders to bid on their true value; Vickrey
1961).
After reading the instructions, participants in the advance-
statement condition stated their WTP in the upcoming auc-
tion (“the highest bid I would give to the auction item,” i.e.,
their reservation price). Any bid that each participant sub-
sequently gave should have been lower than or equal to this
This content downloaded from 128.135.212.25 on Fri, 21 Nov 2014 15:04:57 PM
All use subject to JSTOR Terms and Conditions
SHEN, FISHBACH, AND HSEE 000
Please use DOI when citing. Page numbers are not final.
FIGURE 1
BIDDING RESULTS IN STUDY 3
N
OTE
.—Bidders set a lower advance WTP for the uncertain auction item but paid a higher price after an eight-round auction.
WTP. Participants listed their WTP on a piece of paper and
kept the paper with them until the end of the study. Partic-
ipants in the actual-bidding condition did not list their WTP
in advance. All participants then proceeded to the auction.
For our dependent variables, we measured the second-
highest WTP of each group in all conditions. We used only
one price per group because the process of bidding created
a dependency between the values solicited by participants
within each three-person group. Although advance state-
ments were independent of each other, we needed to use the
same measurement approach to compare them with final
bids. Finally, although participants in the advance-statement
condition continued to complete the auction, we used their
actual bidding data only in an exploratory analysis because
the procedure of advance statement meant that both outcome
focus and process focus influenced these bids.
Results and Discussion
Because the distribution of the WTP responses was highly
skewed (Skewness p2.79, SD p.35, p!.001), we log-
transformed the responses prior to analyses and report means
and standard deviations after reversed transformation. In
support of hypothesis 2, an ANOVA of WTP revealed an
interaction between reward certainty and elicitation method
(F(1, 46) p15.24, p!.001; see fig. 1). The ANOVA
revealed no main effects for reward certainty (F!1), al-
though a main effect for elicitation method did arise (F(1,
46) p12.97, pp.001), indicating that advance statements
were higher than actual bids (consistent with findings on
inflation of stated WTP; Voelckner 2006; Wertenbroch and
Skiera 2002).
A contrast analysis revealed that the WTP through actual
bidding was higher for the uncertain bag (two or four truf-
fles; Mp$1.49, SD p$1.55) than for the certain bag
(four truffles; Mp$0.66, SD p$1.85; t(21) p3.62, p!
.01). By contrast, the WTP elicited through advance state-
ments was higher for the certain bag (four truffles; Mp
$2.37, SD p$1.67) than for the uncertain bag (two or four
truffles; Mp$1.41, SD p$2.00; t(21) p2.01, pp.06).
These results suggest that advance statements appear to be
sensitive to expected values and potentially reflect a pref-
erence for certainty. But, as a person went through the pro-
cess of bidding, uncertainty led to higher bids. We note that
in this design, a comparison of the certain-reward conditions
across the advance-statement and actual-bidding conditions
is not meaningful because we used a different method for
eliciting WTP, which in turn had a main effect on elicited
prices.
As an exploratory analysis, we looked at the final actual
bid data of those who listed their WTP before the auction
(in the advance-statement condition). An ANOVA of WTP
on elicitation method (this time, within-subjects) #reward
certainty revealed a main effect for elicitation method (F(1,
21) p13.56, pp.001), no main effect for reward certainty
(F!1), and no interaction between elicitation method and
reward certainty (F(1, 21) p2.80, pp.11). The WTP
elicited through advance statements for the certain bag was
higher than that for the uncertain bag (as noted above), but
the WTP through actual bidding was similar for certain and
uncertain bags (M
certain
p$0.87, SD
certain
p$2.93; M
uncertain
p$0.97, SD
uncertain
p$2.27; t!1). The latter null effect
could be a combined effect of uncertainty undermining
motivation in advance statements (i.e., outcome focused)
and increasing motivation in actual bidding (i.e., process
focused).
The results of study 3 demonstrate that uncertainty in-
creases motivation when people invest resources in the ac-
This content downloaded from 128.135.212.25 on Fri, 21 Nov 2014 15:04:57 PM
All use subject to JSTOR Terms and Conditions
000 JOURNAL OF CONSUMER RESEARCH
Please use DOI when citing. Page numbers are not final.
tual pursuit of the reward rather than when they focus on
the outcome of their actions and make decisions in prospect.
Thus, the focus on the process versus outcome of pursuing
a reward moderates the motivating-uncertainty effect.
By including the advance-statement condition, we were
able to replicate the previously widely documented uncer-
tainty aversion (Gneezy et al. 2006; Kahneman and Tversky
1979), but only when the focus was on the outcome of the
action. In fact, almost all situations in which previous re-
searchers found uncertainty aversion were in outcome-fo-
cused choices and judgments (e.g., Gneezy et al. 2006) and
before pursuing a reward (e.g., Kivetz 2003). In addition,
moderation by how the decision is elicited (advance judg-
ment vs. actual action) hints at the mechanism for the mo-
tivating-uncertainty effect: the positive experience of pur-
suing a reward. In the next, and final, study, we seek more
direct evidence that the positive experience from uncertainty
underlies the motivating-uncertainty effect. In study 4, all
participants are working toward the reward, but we direct
their attention to either the process or the outcome of their
actions. We expect uncertainty to generate positive experi-
ence, such as excitement and interest, and to boost moti-
vation only when participants focus on the process.
STUDY 4: UNCERTAINTY GENERATES
POSITIVE EXPERIENCE
In study 4, we examined whether uncertainty in a reward
indeed makes the pursuit of the reward more exciting and
interesting and hence more motivating (hypothesis 3). We
used a mediation approach to test this hypothesis.
This study used an auction similar to the one used by
shopping websites such as eBay. We elicited participants’
WTP gradually until it reached the maximum, via actual bid-
ding. Participants successfully purchased the auction item if
their WTP was higher than the predetermined asking price
(a procedure similar to the Becker-DeGroot-Marschak
method; see Becker, DeGroot, and Marschak 1964). In this
paradigm, a higher WTP reflected a greater motivation to get
the auction item. We manipulated the focus on the process,
the outcome, or neither (control condition) during the auction,
and we expected this variable to moderate the motivating-
uncertainty effect. We expected control (no treatment) par-
ticipants to focus on the process as those in the previous
studies did. Based on hypothesis 2, we predicted that,whereas
participants who focused on the process spontaneously (con-
trol condition) or experimentally would be willing to pay a
higher final price to purchase an item of an uncertain mag-
nitude than an item of a certain magnitude, participants who
focused on the outcome would be willing to pay a lower price
for the uncertain item than the certain item.
To test whether positive experience underlies the moti-
vating-uncertainty effect (hypothesis 3), we measured both
the experience of the bidding process and the attractiveness
of the auction item. According to our theorizing, process
experience, rather than reward attractiveness, would predict
bidding prices in the control and process conditions. That
is, process experience would mediate the effect of uncer-
tainty on bidding prices in these conditions. By contrast,
reward attractiveness, rather than process experience, would
predict bidding prices in the outcome condition. That is,
reward attractiveness would mediate the effect of uncertainty
on bidding prices in this condition.
Method
One hundred and eighty-five Chicago residents (102
women, 83 men) visited a downtown lab of the University
of Chicago and participated in the study for a nominal pay-
ment of $2. The study adopted a 2 (reward certainty: certain
vs. uncertain) #3 (focus: control vs. process vs. outcome)
between-participants design. Participants bid on chocolate
coins and purchased the chocolate coins if their WTP was
equal to or higher than a preset (but unknown) seller’s asking
price.
As a buyer, each participant completed the study with
only the experimenter, who assumed the role of the seller,
in the room. Each participant (i.e., buyer) sat in front of a
wide table, with the experimenter (i.e., seller) sitting on the
side. The participant completed three auctions. In each auc-
tion, the experimenter placed an opaque cup covering some
chocolate coins, a closed envelope containing the seller’s
price, and a bid card on the table. The participant placed a
bid for the chocolate coins. All auctions used the same price-
eliciting procedure, in which the experimenter announced a
sequence of prices starting at $0.10 and increasing in in-
crements of $0.05 (e.g., $0.10, $0.15, $0.20, etc.). The par-
ticipant answered “yes” if he/she found the price acceptable
and “no” if he/she would not pay the price. The experimenter
stopped asking after the first price that the participant in-
dicated he/she would not be willing to pay. Then the ex-
perimenter asked the participant to specify his/her price for
the auction item, which was between the last and the second-
to-last prices on the bid card (e.g., $1.08 between $1.10 and
$1.05). The experimenter then moved on to the next auction.
The experimenter informed the participants that when set-
ting the asking price, the seller had the same information
as they did regarding the number of chocolate coins under
the cup. The participants would not know the seller’s asking
price for each cup of chocolate coins until the end of all
three auctions. At the end of all of the auctions, the partic-
ipants opened the envelopes to find out the seller’s asking
price for each cup (cup 1: $3.63, cup 2: $3.45, and cup 3:
$3.66). If any of the buyer’s elicited prices were as high as
or higher than the seller’s asking price, the participants paid
the seller’s asking price for the chocolate coins and took
the chocolate coins home. If any of the buyer’s elicited
prices were lower than the seller’s price, no transaction oc-
curred. All transactions were real and binding.
We manipulated reward certainty. In the certain-reward
conditions, each cup contained five chocolate coins, and in
the uncertain-reward conditions, each cup contained either
three or five chocolate coins, with equal probability. In the
certain-reward conditions, the experimenter removed the cup
before each auction and let the participant see the five choc-
This content downloaded from 128.135.212.25 on Fri, 21 Nov 2014 15:04:57 PM
All use subject to JSTOR Terms and Conditions
SHEN, FISHBACH, AND HSEE 000
Please use DOI when citing. Page numbers are not final.
FIGURE 2
WILLINGNESS-TO-PAY RESULTS IN STUDY 4
N
OTE
.—Bidding prices were a function of focus and uncertainty.
olate coins, whereas in the uncertain-reward conditions, the
experimenter did not remove the cups until the end of all
three auctions. The participant read about the entire pro-
cedure before starting the auctions, and the experimenter
further read the instructions aloud to make sure everything
was clear to the participant.
We manipulated the focus on process, outcome, or control.
In the control conditions, the participants received the afore-
mentioned auction instructions. In the process conditions,
the participant further read, “Enjoy the auction” just before
the auctions began. In the outcome conditions, the partici-
pant further read, “The auctions are a way to get the coins
at a good price.”
Our main dependent variable was total WTP for all three
auctions. A higher WTP for an auction reflects a higher
investment (money) incentivized by the reward. Because
participants in both the certain and uncertain conditions did
not find out the seller’s price until the end of all auctions,
and because the participants in the uncertain conditions did
not find out the number of chocolate coins in each cup until
the end, the three auctions combined were similar to the
auction in study 3. Therefore, we used the sum of the WTP
across all three auctions as our primary measurement of
motivation.
To assess the experience of the bidding process, partici-
pants rated the extent to which they found the auction (a)
interesting (1pNot interesting at all, 9 pVery interesting)
and (b) exciting (1 pNot exciting at all, 9 pVery exciting).
To assess the attractiveness of the reward, we asked partic-
ipants to rate the attractiveness of the auction item (i.e., the
reward; 1 pNot attractive at all, 9 pVery attractive).
Participants completed these questions after the three auc-
tions were completed and before we revealed the auction
items. (We also asked participants to rate the extent to which
they found the auction item valuable, but this rating was
generally low across all conditions (Mp3.91, SD p2.22)
and thus was dropped from further analyses.)
A secondary measure of motivation was whether partic-
ipants wanted to continue the bidding activity. Immediately
after the three auctions, the experimenter provided an op-
tional fourth auction for all participants. The fourth round
followed the same bidding rules and (un)certainty settings
as the previous rounds, and all participants remained in their
original condition. We coded whether participants chose to
continue to a fourth round as a binary variable.
Results and Discussion
Willingness to Pay. We log-transformed WTP due to its
positively skewed distribution (Skewness p3.47, SD p.18,
p!.001), and we report means and standard deviations after
reversed transformation. As predicted, an ANOVA of WTP
yielded an interaction between reward certainty and focus
(F(2, 185) p9.44, p!.001). The ANOVA further yielded
a main effect for reward certainty (F(1, 185) p3.92, pp
.049), indicating a higher WTP in the uncertain- versus cer-
tain-reward conditions and a marginal main effect for focus
(F(2, 185) p2.56, pp.08), indicating different WTPs in
the control, process, and outcome conditions. Figure 2 dis-
plays the results.
Contrast analyses within each focus condition indicate
that the motivating-uncertainty effect occurred in both the
control and the process conditions but not in the outcome
condition. Specifically, participants in the control condition
set higher bids for the uncertain chocolate coins (Mp$1.89,
SD p$1.87) than for the certain chocolate coins (Mp
$1.25, SD p$1.81; t(53) p2.51, pp.015), as did the
participants in the process condition (M
uncertain
p$1.56,
SD
uncertain
p$2.00; M
certain
p$0.83, SD
certain
p$2.61; t(62)
p3.03, p!.01). However, participants in the outcome
This content downloaded from 128.135.212.25 on Fri, 21 Nov 2014 15:04:57 PM
All use subject to JSTOR Terms and Conditions
000 JOURNAL OF CONSUMER RESEARCH
Please use DOI when citing. Page numbers are not final.
TABLE 2
RATING RESULTS FOR STUDY 4
Process
experience Reward
attractiveness
Mean SD Mean SD
Control condition:
Certain reward 4.87*2.04 5.04 2.49
Uncertain reward 5.98 1.81 5.21 2.64
Process condition:
Certain reward 5.03*2.06 5.37 2.30
Uncertain reward 6.16 1.79 4.53 2.58
Outcome condition:
Certain reward 5.50 2.11 5.78*2.38
Uncertain reward 5.72 1.96 4.62 2.41
*For each comparison between certain and uncertain rewards the
asterisk indicates a significant difference at p!.05.
condition set lower bids for the uncertain chocolate coins
(Mp$1.11, SD p$1.67) than for the certain chocolate
coins (Mp$1.68, SD p$2.19; t(64) p2.46, pp.017).
These findings confirm that uncertainty boosts motivation
for those focusing on the process rather than the outcome
of bidding.
Another ANOVA of WTP on the control and process
conditions yielded no interaction between reward certainty
and focus (F!1), which suggests similar effects across
these conditions. This result is consistent with our assump-
tion that participants in the control condition may have nat-
urally focused on process. Therefore, it allowed us to col-
lapse the control and process conditions in subsequent
mediation analyses.
Mediations by Process Experience and Reward Attrac-
tiveness. We averaged the ratings of excitement and inter-
est into an index of process experience (ap.90). Table 2
displays process experience ratings as well as reward at-
tractiveness ratings by condition. To test whether process
experience mediated the effects of reward certainty on WTP,
and whether focus moderated such mediation, we tested the
moderated-mediation model by following the procedure out-
lined in Preacher, Rucker, and Hayes (2007). Theconditional
indirect effect of reward certainty on WTP through process
experience was significant in the combined control and pro-
cess condition (B
conditional indirect
p.07, SE p.03; 95% C.I.
p[.02, .13]; based on 10,000 bootstrap samples) but not
in the outcome condition (B
conditional indirect
p.01, SE p.02;
95% C.I. p[.03, .07]). Specifically, in the combined con-
trol and process condition, an uncertain (vs. certain) reward
directly increased WTP (Bp.23, SE p.06, p!.001). In
addition, an uncertain (vs. certain) reward increased positive
process experience (Bp1.19, SE p.35, pp.001), which
in turn increased WTP (Bp.07, SE p.01, p!.001). This
result indicates that increased process experience mediated
the effect of reward certainty on WTP but did so only when
participants focused on the process of the auctions.
We further tested for attractiveness of the reward in a
similar moderated-mediation model. We found that the con-
ditional indirect effect of reward certainty on WTP through
reward attractiveness was not significant in the combined
control and process condition (B
conditional indirect
p.02, SE p
.02; 95% C.I. p[.06, .02]; based on 10,000 bootstrap
samples) but that it was significant in the outcome condition
(B
conditional indirect
p.04, SE p.03; 95% C.I. p[.12,
.00]). Specifically, in the outcome condition, a certain (vs.
uncertain) reward directly increased WTP (Bp.18, SE
p.07, pp.017). In addition, a certain (vs. uncertain)
reward directionally (though not significantly) increased re-
ward attractiveness (Bp.87, SE p.54, pp.11), which
in turn increased WTP (Bp.04, SE p.34, p!.01). The
result indicates that increased reward attractiveness mediated
the effect of reward certainty on WTP only when partici-
pants focused on the outcome of the auctions.
Willingness to Continue. The results for willingness to
continue, although less strong, were also consistent with
hypothesis 2 (fig. 3). We found a marginally significant main
effect for reward certainty on willingness to continue (cer-
tain vs. uncertain: Wald’s x
2
(1, Np185) p3.03, pp
.082), along with no main effect for focus (control vs. pro-
cess vs. outcome: Wald’s x
2
(1, Np185) p2.47, pp
.116) and no interaction (Wald’s x
2
(1, Np185) p2.45,
pp.117). Contrast analyses revealed that more participants
chose to have another uncertain (vs. certain) auction in the
process condition (uncertain vs. certain: 62% vs. 33%, x
2
(1,
Np64) p5.16, pp.023) but not in the outcome condition
(uncertain vs. certain: 45% vs. 62%, x
2
(1, Np66) p1.97,
pp.16). When the control and the process conditions were
collapsed, the motivational effect of uncertainty again oc-
curred for willingness to continue (uncertain vs. certain:
60% vs. 40%, x
2
(1, Np119) p4.44, pp.035), although
the control condition alone only directionally replicated the
effect (uncertain vs. certain: 57% vs. 48%, x
2
!1). These
results provide additional support for the motivating-uncer-
tainty effect (hypothesis 1) and its moderation effect (hy-
pothesis 2).
In summary, study 4 not only demonstrated the motivat-
ing-uncertainty effect but also identified when and why this
effect happens. Drawing people’s attention to the process
(vs. outcome) of reward pursuits makes them evaluate this
experience more positively (i.e., increases the value of the
process attribute) and care more about this experience (i.e.,
increases the decision weight of the process attribute). As
a result, people invest more resources in pursuing uncertain
rewards than certain rewards of a higher expected value.
Although we found similar results in the control and the
process conditions, we want to be cautious about general-
izing the similarities between the two. We do not want to
speak for a null-result finding; the fact that the difference
is minimal might be due to the subtle manipulation in the
process conditions. We also do notclaim that everyone spon-
taneously finds the experience of every activity positive.
Some activities (e.g., auctions and gambles) have greater
excitement potential than others, and some people have a
stronger disposition to seek positive experience (e.g., sen-
sation-seeking individuals; Zuckerman 2007).
This content downloaded from 128.135.212.25 on Fri, 21 Nov 2014 15:04:57 PM
All use subject to JSTOR Terms and Conditions
SHEN, FISHBACH, AND HSEE 000
Please use DOI when citing. Page numbers are not final.
FIGURE 3
WILLINGNESS-TO-CONTINUE RESULTS IN STUDY 4
N
OTE
.—Choices of an additional round were a function of focus and uncertainty.
GENERAL DISCUSSION
Life is full of uncertainties. Previous literature documents
a preference for certain over uncertain rewards in evaluation
of outcomes, specifically in the domain of gains (Arrow
1965; Bernoulli 1738; Kahneman and Tversky 1979). Mov-
ing beyond which outcomes people prefer, we ask what type
of reward increases their motivation in the process of reward
pursuit, and we find that a reward of an uncertain magnitude
is more motivating than a reward of a certain magnitude,
even when the uncertain reward has a lower expected value.
We demonstrated this motivating-uncertainty effect in a
series of studies involving activities with real consequences.
We found that more participants completed a water-drinking
task in the pursuit of a reward of an uncertain magnitude
than in the pursuit of a reward of a certain and larger mag-
nitude (study 1) and that participants rated more advertise-
ments in return for a reward of an uncertain magnitude,
regardless of the specific reward probability (study 2). We
further found in study 3 that the motivational boost from
uncertainty occurred only for those focusing on the process
of pursuing a reward, rather than those focusing on the
outcome. Finally, in study 4, we demonstrated that in the
process of bidding, uncertainty increased investment as long
as participants’ focus was on the process and not on the
outcome of their actions. When participants focused on the
outcome, uncertainty (coupled with lower expected value)
decreased their investment. Importantly, we also found that
uncertainty induced positive experience such as excitement,
which in turn increased motivation to invest in the pursuit
of uncertain rewards. In sum, uncertainty in rewards induces
positive experience and increases investment of effort, time,
and money in pursuing rewards.
In addition to the moderating conditions tested in the
studies, other background conditions may exist for the mo-
tivating-uncertainty effect, and they await further research.
For example, the effect may be more likely to arise for
tedious activities than for interesting activities because bor-
ing tasks may have more room to benefit from uncertainty-
generated excitement (though, notably, we observed the ef-
fect both for tedious activities such as evaluating ads in
study 2 and for more interesting activities such as bidding
in study 3). Another possible background condition for the
effect is that the procedure of the task does not focus
people’s attention on the outcome. Indeed, in study 4, the
motivating-uncertainty effect was attenuated when partici-
pants were instructed to attend to the outcome. A third pos-
sible background condition for the motivating-uncertainty
effect is reward magnitude. We suspect that as the magnitude
of the reward scales up, the effect might become weaker
because large outcomes may shift people’s attention from
the process to the outcome.
Relationship with Prior Research
Decision research has largely viewed uncertainty as a
negative influence in decision making (e.g., Gneezy et al.
2006; Rabin 2000; von Neumann and Morgenstern 1944).
Although some have shown that uncertainty can have a
positive impact on affective experience (Bar-Anan et al.
2009; Moon and Neilson 2014; Wilson et al. 2005; Whit-
church, Wilson, and Gilbert 2011), this impactdoes not seem
to extend to choice or other consequential behaviors (Lee
and Qiu 2009).
Recent research by Gneezy and colleagues (2006) showed
that in the gain domain, people value an uncertain prospect
even less than its worst possible outcome. Studies have rep-
licated this effect across various choice and evaluation tasks
(Markle et al. 2014; Moon and Nelson 2014; Simonsohn
This content downloaded from 128.135.212.25 on Fri, 21 Nov 2014 15:04:57 PM
All use subject to JSTOR Terms and Conditions
000 JOURNAL OF CONSUMER RESEARCH
Please use DOI when citing. Page numbers are not final.
2009; Sonsino 2008; Wang, Feng, and Keller 2014; Yang,
Vosgerau, and Loewenstein 2013). These studies have ex-
plained it in terms of distaste for uncertainty such that un-
certainty has a negative utility (Simonsohn 2009) and in
terms of differential outcome framing (e.g., “lottery tickets”
for uncertain gains and “gift certificate” for certain gains;
Yang et al. 2013). The finding that an uncertain prospect is
valued less than its worst possible outcome may appear to
contradict our finding that an uncertain reward is more mo-
tivating than its best possible outcome. To explain this dis-
crepancy, we distinguish between decisions with a focus on
outcome and decisions with a focus on process (see also
Higgins et al. 2003; Toure´-Tillery and Fishbach 2011). We
speculate that uncertainty aversion emerges in the former
type of decisions, whereas motivating uncertainty emerges
in the latter type. Focusing on process rather than outcome
can change people’s feelings toward uncertainty from neg-
ative (aversion) to positive (excitement, interest). Indeed, in
study 3, bidders valued the uncertain item less than the
certain item before the auction, but they found the uncertain
item more motivating than the certain item during the auc-
tion, although the comparison in our study is somewhat
different than that in Gneezy et al.’s (2006).
Our excitement account has its root in existing literature.
Cognitive research on achievement motivation finds that
tasks with a moderate level of uncertainty energize people
and stimulate their motivation (Atkinson 1957; Brehm and
Self 1989; Brehm et al. 1983; Locke and Latham 2006).
These researchers find that uncertainty about the likelihood
of qualifying for a reward can be challenging and hence
motivating. We add to this literature by suggesting that for
those who qualified, uncertainty about reward magnitude
also boosts motivation. In addition, personality research
finds that gamblers seek excitement through uncertain games
(Wagenaar 1989; Zuckerman 2007), and these findings on
individual differences complement our (and other) research
on situations in which uncertainty boosts motivation.
Using a similar paradigm of comparing a certain reward
to a dominated uncertain reward, Goldsmith and Amir (2010)
found no difference between these reward conditions, whereas
we found a positive effect of uncertain rewards. Importantly,
however, we only found this effect when participants focused
on the process and not when they focused on the outcome.
If we think of decisions as varying on a continuum, from
those that are completely process-focused to those that are
completely outcome-focused, the decisions in Goldsmith and
Amir (2010) may have fallen somewhere in the middle of
the continuum. We speculate that this would explain why
those authors found no difference between the certain and
uncertain reward conditions. Consistent with this speculation,
those authors noted that when participants focused on the
outcome, such as deliberately considering outcome probabil-
ities in their study 2, they favored the certain gain over the
uncertain gain.
Furthermore, most decision models have assumed inde-
pendence between the value function and the probability
function (e.g., Expected Utility Theory: Bernoulli 1738; von
Neumann and Morgenstern 1944; Prospect Theory: Tversky
and Kahneman 1979, 1992; also see Coombs and Huang
1970; Markowitz 1987; Sharpe 1970). However, our re-
search suggests that these two functions may be interde-
pendent such that uncertainty brings additional value to the
pursuit of the outcome (see Jia and Dyer [1996] and Slovic
et al. [2007] for related arguments.) We look forward to
better models to capture how people behave under uncer-
tainty.
Marketing Implications
Our research documented the motivating-uncertainty ef-
fect, and in this section, we discuss its implications for real-
world consumer behaviors. We argue that uncertain rewards
can be beneficial for marketers and policy makers, and we
analyze the conditions under which including uncertain re-
wards would likely be most beneficial.
Uncertain Rewards in Marketing. The benefits of un-
certain rewards are threefold. First, as we have shown, from
the motivation perspective, uncertain rewards can increase
resource investment. Second, from an economic perspective,
uncertain rewards can be less expensive because, as dem-
onstrated in our studies, an uncertain reward can be more
motivating than a certain reward of a higher expected value.
Third, from a hedonic perspective, uncertainty can be a
source of positive experience and hence can increase con-
sumer enjoyment and satisfaction. For example, compared
to a promotion that offers a certain reward for qualified
shoppers (e.g., “$50 off if you spend over $200”), a pro-
motion that offers an uncertain reward (e.g., “$30 or $50
off if you spend over $200”) feels like a game, making
shopping fun and engaging. Thus, adding uncertainty into
rewards, like other nudging ideas (Huang and Soman 2013;
Thaler and Sunstein 2008), can be an effective marketing
strategy.
How to Design Uncertain-Reward Programs. Although
uncertainty might become less motivating as the reward
magnitude increases, this does not undermine the importance
or limit the applications of our research. In fact, uncertain
rewards could still be effective at the aggregate level. For
example, when adding uncertainty to small-value coupons,
marketers may slightly increase each consumer’s spending,
and such a slight increase in individual spending could have
a large impact on the economy, such as a positive signal of
recovery from a recession when Black Friday sales increase.
In applying our findings to reward programs, one should
keep in mind that the motivating-uncertainty effect requires
consumers to focus on the process of reward pursuit. Our
findings suggest that uncertainty motivates people only
while and not before engaging in the activity. This finding
implies that before “working” (e.g., shopping in a store,
eating in a restaurant, donating to a charity), consumers
would be less likely to opt in to a reward program with
uncertain rewards than to one with certain rewards, although
they would work harder in the program with uncertain re-
This content downloaded from 128.135.212.25 on Fri, 21 Nov 2014 15:04:57 PM
All use subject to JSTOR Terms and Conditions
SHEN, FISHBACH, AND HSEE 000
Please use DOI when citing. Page numbers are not final.
wards. Accordingly, marketers should target different con-
sumers with different types of incentives: certain rewards
to potential new customers who are not yet part of the reward
program and uncertain rewards to existing customers who
are already in the “process” (the reward program). In fact,
to maximize their customer base and to fully utilize both
rewards, marketers can even combine both certain and un-
certain rewards in a reward program, with the certain re-
wards highlighted when recruiting new members and the
uncertain rewards highlighted when motivating existing cus-
tomers.
Closing Remarks
For centuries, research on uncertainty has mainly inves-
tigated the negative impact on judgment (Gneezy et al. 2006)
and choice (Bernoulli 1738; Kahneman and Tversky 1979).
The present research shows that uncertainty can have a pos-
itive impact on behavior: it brings excitement and boosts
motivation. We hope the current research will stimulate fu-
ture researchers to further explore the topic and imbue them
with the uncertainty, hence the excitement, of new discov-
eries.
DATA COLLECTION INFORMATION
The first author supervised the collection of data by re-
search assistants at the University of Chicago’s Decision Re-
search Lab (Campus) between the winter and spring of 2013
and the spring of 2014 for study 1 and between the winter
and spring of 2011 for study 3. The first author also managed
the collection of data for study 2 on Amazon Mechanical
Turk in the spring of 2013 and supervised the collection of
data for study 4 by research assistants at the University of
Chicago’s Chicago Research Lab (Downtown) between the
spring and summer of 2012. These data were analyzed by the
first author with the support of a statistician at the University
of Chicago.
REFERENCES
Amir, On, and Dan Ariely (2008), “Resting on Laurels: The Effects
of Discrete Progress Markers as Subgoals on Task Perfor-
mance and Preferences,” Journal of Experimental Psychol-
ogy: Learning, Memory, and Cognition, 34 (5), 1158–71.
Andrade, Eduardo B., and Ganesh Iyer (2009), “Planned versus
Actual Betting in Sequential Gambles,” Journal of Marketing
Research, 46 (3), 372–83.
Ariely, Dan, George Loewenstein, and Drazen Prelec (2003),“‘Co-
herent Arbitrariness’: Stable Demand Curves without Stable
Preferences,” Quarterly Journal of Economics, 118 (1), 73–
106.
Arrow, Kenneth J. (1965), Aspects of the Theory of Risk Bearing,
Helsinki: Yrjo¨ Jahnssonin Sa¨a¨tio¨.
Atkinson, John W. (1957), “Motivational Determinants of Risk-
Taking Behavior,” Psychological Review, 64 (6), 359–72.
Bagozzi, Richard P., and Utpal Dholakia (1999), “Goal Setting
and Goal Striving in Consumer Behavior,” Journal of Mar-
keting, 63 (Special Issue), 19–32.
Bar-Anan, Yoav, Timothy D. Wilson, and Daniel T. Gilbert (2009),
“The Feeling of Uncertainty Intensifies Affective Reactions,”
Emotion, 9 (1), 123–27.
Becker, Gordon M., Morris H. DeGroot, and Jocabo Marschak
(1964), “Measuring Utility by a Single-Response Sequential
Method,” Behavioral Science, 9 (3), 226–32.
Bernoulli, Daniel (1738), “Specimen Theoriae Novae de Mensura
Sortis,” Comentarii Academiae Scientiarum Imperialis Pe-
tropolitanae, 5, 175–92 (translated by Louise Sommer [1954]
in Econometrica, 22 [1], 23–36).
Berns, Gregory S., Samuel M. McClure, Giuseppe Pagnoni, and
P. Read Montague (2001), “Predictability Modulates Human
Brain Response to Reward,” Journal of Neuroscience,21(8),
2793–98.
Brehm, Jack W., and Elizabeth A. Self (1989), “The Intensity of
Motivation,” Annual Review of Psychology, 40, 109–31.
Brehm, Jack W., Rex A. Wright, Sheldon Solomon, Linda Siika,
and Jeff Greenberg (1983), “Perceived Difficulty, Energiza-
tion, and the Magnitude of Goal Valence,” Journal of Ex-
perimental Social Psychology,19(1),2148.
Callan, Mitchell J., John H. Ellard, and Jennifer E. Nicol (2006),
“The Belief in a Just World and Immanent Justice Reasoning
in Adults,” Personality and Social Psychology Bulletin,32
(12), 1646–58.
Cheema, Amar, Dipankar Chakravarti, and Atanu R. Sinha (2012),
“Bidding Behavior in Descending and Ascending Auctions,”
Marketing Science, 31 (5), 779–800.
Choi, Jinhee, and Ayelet Fishbach (2011), “Choice Is an End versus
a Means,” Journal of Marketing Research, 48 (3), 544–54.
Converse, Benjamin A., Jane L. Risen, and Travis J. Carter (2012),
“Investing in Karma: When Wanting Promotes Helping,” Psy-
chological Science, 23 (8), 923–30.
Coombs, Clyde H., and Lily Huang (1970), “Tests of a Portfolio
Theory of Risk Preference,” Journal of Experimental Psy-
chology,85(1),2329.
Custers, Ruudm, and Henk Aarts (2005), “Positive Affect as Im-
plicit Motivator: On the Nonconscious Operation of Behav-
ioral Goals,” Journal of Personality of Social Psychology,89
(2), 129–42.
Czikszentmihalyi, Mihalyi (1990), Beyond Boredom and Anxiety:
The Experience of Flow in Work and Games, San Francisco,
CA: Jossey-Bass.
Deci, Edward L., and Richard M. Ryan (1985), Intrinsic Motivation
and Self-Determination in Human Behavior, New York: Ple-
num.
Dhar, Sanjay K., Claudia Gonzalez-Vallejo, and Dilip Soman
(1995), “Brand Promotions as a Lottery,” Marketing Letters,
6 (3), 221–33.
—— (1999), “Modeling the Effects of Advertised Price Claims:
Tensile versus Precise Claims?” Marketing Science,18(2),
154–77.
Erez, Amir, and Alice M. Isen (2002), “The Influence of Positive
Affect on the Components of Expectancy Motivation,” Jour-
nal of Applied Psychology, 87 (6), 1055–67.
Ferguson, Melissa J. (2008), “On Becoming Ready to Pursue a
Goal You Don’t Know You Have: Effects of Nonconscious
Goals on Evaluative Readiness,” Journal of Personality and
Social Psychology, 95 (6), 1268–94.
Fishbach, Ayelet, and Jinhee Choi (2012), “When Thinking about
Goals Undermines Goal Pursuit,” Organizational Behavior
and Human Decision Processes, 118 (2), 99–107.
Fishbach, Ayelet, James Y. Shah, and Arie W. Kruglanski (2004),
“Emotional Transfer in Goal Systems,” Journal of Experi-
mental Social Psychology, 40 (6), 723–38.
This content downloaded from 128.135.212.25 on Fri, 21 Nov 2014 15:04:57 PM
All use subject to JSTOR Terms and Conditions
000 JOURNAL OF CONSUMER RESEARCH
Please use DOI when citing. Page numbers are not final.
Gneezy, Uri, John A. List, and George Wu (2006), “The Uncer-
tainty Effect: When a Risky Prospect Is Valued Less than Its
Worst Possible Outcome,” Quarterly Journal of Economics,
121 (4), 1283–1309.
Goldsmith, Kelly, and On Amir (2010), “Can Uncertainty Improve
Promotions?” Journal of Marketing Research, 47 (6), 1070–77.
Gollwitzer, Peter M. (2012), “Mindset Theory of Action Phases,”
in Handbook of Theories of Social Psychology, Vol. 1, ed.
Paul A. M. Van Lange, Arie W. Kruglanski, and E. Tory
Higgins, London: Sage, 526–45.
Heckhausen, Heinz (1987), “Wishing-Weighing-Willing,” in Jen-
seits des Rubikon: Der Wille in den Humanwissenschaften,
ed. Heinz Heckhausen, Peter M. Gollwitzer, and Franz E.
Weinert, Heidelberg: Springer-Verlag, 3–9.
Higgins, E. Tory, Arie W. Kruglanski, and Antonio Pierro (2003),
“Regulatory Mode: Locomotion and Assessment as Distinct
Orientations,” in Advances in Experimental Social Psychol-
ogy, Vol. 35, ed. James M. Olson and Mark P. Zanna, New
York: Academic Press, 293–344.
Holt, Charles A., and Susan K. Laury (2002), “Risk Aversion and
Incentive Effects,” American Economic Review, 92 (5), 1644–
55.
Hsee, Christopher K., and Yuval Rottenstreich (2004), “Music,
Pandas, and Muggers: On the Affective Psychology of Value,”
Journal of Experimental Psychology: General, 133 (1), 23–
30.
Hsee, Christopher K., Fang Yu, Jiao Zhang, and Yan Zhang (2003),
“Medium Maximization,” Journal of Consumer Research,30
(1), 1–14.
Hsee, Christopher K., and Elke U. Weber (1997), “A Fundamental
Prediction Error: Self-Other Discrepancies in Risk Prefer-
ence,” Journal of Experimental Psychology: General, 126 (1),
45–53.
Hsee, Christopher K., Yang Yang, Naihe Li, and Luxi Shen (2009),
“Wealth, Warmth, and Well-Being: WhetherHappiness Is Rel-
ative or Absolute Depends on Whether It Is about Money,
Acquisition, or Consumption,” Journal of Marketing Re-
search, 46 (3), 396–409.
Hsee, Christopher K., and Jiao Zhang (2010), “General Evalua-
bility Theory,” Perspectives on Psychological Science,5(4),
343–55.
Hsee, Christopher K., Jiao Zhang, Fang Yu, and Yiheng Xi (2003),
“Lay Rationalism and Inconsistency between Predicted Ex-
perience and Decision,” Journal of Behavioral Decision Mak-
ing, 16 (4), 257–72.
Hsee, Christopher K., Jiao Zhang, Liangyan Wang, and Shirley
Zhang (2013), “Magnitude, Time, and Risk Differ Similarly
between Joint and Single Evaluations,” Journal of Consumer
Research, 40 (1), 172–84.
Huang, Wendy Hsin-Yuan, and Dilip Soman (2013), “A Practi-
tioner’s Guide to Gamification of Education,” Behavioural
Economics in Action Report Series, Rotman School of Man-
agement, University of Toronto.
Jeffrey, Scott A. (2009), “Justifiability and the Motivational Power
of Tangible Noncash Incentives,” Human Performance,22
(2), 143–55.
Jia, Jianmin, and James S. Dyer (1996), “A Standard Measure of
Risk and Risk-Value Models,” Management Science,42(12),
1691–1705.
Kahneman, Daniel, and Amos Tversky (1979), “Prospect Theory:
An Analysis of Decision under Risk,” Econometrica,47(2),
263–92.
Kivetz, Ran (2003), “The Effects of Effort and Intrinsic Motivation
on Risky Choice,” Marketing Science, 22 (4), 477–502.
—— (2005), “Promotion Reactance: The Role of Effort-Reward
Congruity,” Journal of Consumer Research, 31 (4), 725–36.
Kivetz, Ran, Oleg Urminsky, and Yuhuang Zheng (2006), “The
Goal-Gradient Hypothesis Resurrected: Purchase Accelera-
tion, Illusionary Goal Progress, and Customer Retention,”
Journal of Marketing Research,43(1),3958.
Klein, Nadav, and Ayelet Fishbach (2014), “Feeling Good at the
Right Time: Why People Value Predictability in Goal At-
tainment,” Journal of Experimental Social Psychology,55
(2014), 21–30.
Knight, Frank H. (1921), Risk, Uncertainty, and Profit, Boston:
Hart, Schaffner & Marx, Houghton Mifflin.
Koo, Minjung, and Ayelet Fishbach (2010), “A Silver Lining of
Standing in Line: Queuing Increases Valueof Products,” Jour-
nal of Marketing Research, 47 (August), 713–24.
Kuhl, Julius, and Miguel Kaze´n (1999), “Volitional Facilitation of
Difficult Intentions: Joint Activation of Intention Memory and
Positive Affect Removes Stroop Interference,” Journal of Ex-
perimental Psychology: General, 128 (3), 382–99.
Langer, Ellen J. (1975), “The Illusion of Control,” Journal of Per-
sonality and Social Psychology, 32 (2), 311–28.
Lee, Yih Hwai, and Cheng Qiu (2009), “When Uncertainty Brings
Pleasure: The Role of Prospect Imageability and Mental Im-
agery,” Journal of Consumer Research, 36 (4), 624–33.
Le Menestrel, Marc (2001), “A Process Approach to the Utility
for Gambling,” Theory and Decision, 50 (3), 249–61.
Locke, Edwin A., and Gary P. Latham (2006), “New Directions
in Goal-Setting Theory,” Current Directions in Psychological
Science, 15 (5), 265–68.
Markle, Alex, Yuval Rottenstreich, Jeff Galak, and Leif Nelson
(2014), “Upper and Lower Violations of Internality in Risky
Decisions,” Working Paper, Hong Kong University of Science
and Technology.
Markowitz, Harry M. (1987), Mean-Variance Analysis in Portfolio
Choice and Capital Markets, Oxford: Basil Blackwell.
Marschak, Jacob (1950), “Rational Behavior, Uncertain Prospects,
and Measurable Utility,” Econometrica, 18 (2), 111–41.
Millar, Murray G., and Abraham Tesser (1986), “Effects of Af-
fective and Cognitive Focus on the Attitude–Behavior Re-
lation,” Journal of Personality and Social Psychology,51(2),
270–76.
Moon, Alice, and Leif Nelson (2014), “Dissecting the Uncertainty
Effect: Do People Really Dislike Uncertainty?” Working Pa-
per, University of California, Berkeley.
Nunes, Joseph C., and Xavier Dre`ze (2006), “The Endowed Pro-
gress Effect: How Artificial Advancement Increases Effort,”
Journal of Consumer Research, 32 (4), 504–12.
Pascal, Blaise (1670), Pense´es, Paris: Port-Royal.
Preacher, Kristopher J., Derek D. Rucker, and Andrew F. Hayes
(2007), “Addressing Moderated Mediation Hypotheses: The-
ory, Methods, and Prescriptions,” Multivariate Behavioral Re-
search, 42 (1), 185–227.
Rabin, Matthew (2000), “Risk Aversion and Expected-Utility The-
ory: A Calibration Theorem,” Econometrica, 68 (5), 1281–
92.
Rabin, Matthew, and Richard H. Thaler (2001), “Anomalies: Risk
Aversion,” Journal of Economic Perspectives,15(1),219
32.
Rottenstreich, Yuval, and Ran Kivetz (2006), “On Decision Mak-
ing without Likelihood Judgment,” Organizational Behavior
and Human Decision Processes, 101 (1), 74–88.
This content downloaded from 128.135.212.25 on Fri, 21 Nov 2014 15:04:57 PM
All use subject to JSTOR Terms and Conditions
SHEN, FISHBACH, AND HSEE 000
Please use DOI when citing. Page numbers are not final.
Sansone, Carol, Charlene Weir, Lora Harpster, andCarolyn Morgan
(1992), “Once a Boring Task Always a Boring Task? Interest
as a Self-Regulatory Mechanism,” Journal of Personality and
Social Psychological, 63 (3), 379–90.
Schultz, Wolfram, Peter Dayan, and P. Read Montague (1997), “A
Neural Substrate of Prediction and Reward,” Science, 275
(March 14), 1593–99.
Sharpe, William F. (1970), Portfolio Theory and Capital Markets,
New York: McGraw-Hill.
Simonsohn, Uri (2009), “Direct Risk Aversion: Evidence from
Risky Prospects Valued below Their Worst Outcome,” Psy-
chological Science, 20 (6), 686–92.
Slovic, Paul, Melissa L. Finucane, Ellen Peters, and Donald G.
MacGregor (2007), “The Affect Heuristic,” European Journal
of Operational Research, 177 (3), 1333–52.
Soman, Dilip (1998), “The Illusion of Delayed Incentives: Eval-
uating Future Effort-Money Transactions,” Journal of Mar-
keting Research, 35 (4), 427–37.
Sonsino, Doron (2008), “Disappointment Aversion in Internet Bid-
ding-Decisions,” Theory and Decision, 64 (2), 363–93.
Strickland, Lloyd, Roy J. Lewicki, and Arnold M. Katz (1966),
“Temporal Orientation and PerceivedControl as Determinants
of Risk-Taking,” Journal of Experimental Social Psychology,
2 (2), 143–51.
Sunstein, Cass R. (2003), “Terrorism and Probability Neglect,”
Journal of Risk and Uncertainty, 26 (2–3), 121–36.
Thaler, Richard H., and Cass R. Sunstein (2008), Nudge: Improving
Decisions about Health, Wealth, and Happiness, New Haven,
CT: Yale University Press.
Toure´-Tillery, Maferima, and Ayelet Fishbach (2011), “The Course
of Motivation,” Journal of Consumer Psychology,21(4),
414–23.
——— (2012), “The End Justifies the Means, but Only in the
Middle,” Journal of Experimental Psychology: General, 141
(3), 570–83.
——— (2014), “How to Measure Motivation: A Guide for the
Experimental Social Psychologist,” Social and Personality
Psychology Compass, 8 (7), 328–41.
Tversky, Amos, and Craig R. Fox (1995), “Weighing Risk and Un-
certainty,” Psychological Review, 102 (2), 269–83.
Tversky, Amos, and Daniel Kahneman (1992), “Advances in Pros-
pect Theory: Cumulative Representation of Uncertainty,”
Journal of Risk and Uncertainty, 5 (4), 297–323.
Vickrey, William (1961), “Counterspeculation, Auctions, and Com-
petitive Sealed Tenders,” Journal of Finance,16(1),837.
Voelckner, Franziska (2006), “An Empirical Comparison of Meth-
ods for Measuring Consumers’ Willingness to Pay,” Market-
ing Letter, 17 (2), 137–49.
Von Neumann, John, and Oskar Morgenstern (1944), Theory of
Games and Economic Behavior, Princeton, NJ: Princeton Uni-
versity Press.
Vosgerau, Joachim, Klaus Wertenbroch, and Ziv Carmon (2006),
“Indeterminacy and Live Television,” Journal of Consumer
Research, 32 (4), 487–95.
Wagenaar, Willem Albert (1989), Paradoxes of Gambling Behav-
ior, New York: Psychology Press.
Wang, Yitong, Tianjun Feng, and L. Robin Keller (2014), “De-
biasing the Uncertainty Effect: An Empirical Investigation,”
Working Paper, Tsinghua University.
Wertenbroch, Klaus, and Bernd Skiera (2002), “Measuring Con-
sumers’ Willingness to Pay at the Point of Purchase,” Journal
of Marketing Research, 39 (2), 228–41.
Whitchurch, Erin R., Timothy D. Wilson, and Daniel T. Gilbert
(2011), “‘He Loves Me, He Loves Me Not’: Uncertainty Can
Increase Romantic Attraction,” Psychological Science 22 (2),
172–75.
Wilson, Timothy D., David B. Centerbar, Deborah A. Kermer, and
Daniel T. Gilbert (2005), “The Pleasures of Uncertainty: Pro-
longing Positive Moods in Ways People Do Not Anticipate,”
Journal of Personality and Social Psychology, 88 (1), 5–21.
Yang, Yang, Joachim Vosgerau, and George Loewenstein (2013),
“Framing Influences Willingness to Pay but Not Willingness
to Accept,” Journal of Marketing Research, 50 (6), 725–38.
Zhang, Jiao (2014), “Joint versus Separate Modes of Evaluation:
Theory and Practice,” in Blackwell Handbook of Judgment
and Decision Making, 2nd ed., ed. Gideon Keren and George
Wu, Malden, MA: Blackwell.
Zhang, Ying, and Szu-Chi Huang (2010), “How Endowed versus
Earned Progress Impacts Consumer Goal Commitment and
Motivation,” Journal of Consumer Research, 37 (4), 641–54.
Zhang, Ying, Jing Xu, Zixi Jiang, and Szu-chi Huang (2011),
“Been There, Done That: The Impact of Effort Investment on
Goal Value and Consumer Motivation,” Journal of Consumer
Research,38(1),7893.
Zillmann, Dolf (1983), “Transfer of Excitation in Emotional Be-
havior,” in Social Psychophysiology: A Sourcebook, ed. John
T. Cacioppo and Richard E. Petty, New York: Guilford,
215–40.
Zuckerman, Marvin (2007), Sensation Seeking and Risky Behavior,
Washington, DC: American Psychology Association.
This content downloaded from 128.135.212.25 on Fri, 21 Nov 2014 15:04:57 PM
All use subject to JSTOR Terms and Conditions
... Goldsmith and Amir (2010) believe that taking uncertainty into account in marketing is a method to reduce marketing costs and maintain consumers' interest, which has a good marketing effect. Shen et al. (2015) argued that people would invest more time, money, and energy to obtain uncertain rewards. Based on the positive emotion theory, Lee and Qiu (2009) proposed that when consumers face an uncertainty related to a positive event, they will experience greater and more lasting positive emotions. ...
... In this paper, consumption objectives are divided into pursuit process and pursuit result (Shen et al., 2015). When people pursue a process, positive experience is crucial, and thus, it is possible that completion and emotional enrichment increase motivation (Fishbach and Choi, 2012). ...
... Purchase intention refers to the viewpoint of Dodds et al. (1991), with a total of four items. Consumption purpose is divided into process-oriented and results-oriented (Shen et al., 2015). ...
Article
Full-text available
As the younger generation, who like to pursue novelty and excitement, becomes the main consumer and the traditional consumption culture changes in China, the blind box has become a popular product among young people with its uncertain characteristics. Previous studies have mainly explored the role of uncertainty in promotion, while this paper focuses on the role of uncertainty in daily sales of blind box products. Based on the stimulus–organism–response (SOR) theory, this paper conducted an online questionnaire survey and an empirical analysis in China, which examined the mechanism of the positive impact of uncertainty and the moderating effect of consumption purpose. The results show that uncertainty affects consumers’ purchase intention mainly through affecting their emotional value, which is one dimension of perceived value; consumer purpose also moderates the effect of uncertainty on perceived value, and the effect of perceived value on purchase intention. The results of this study are not only of great significance for understanding the uncertain marketing and blind box products, but also have management implications for enterprises to make use of the uncertain marketing.
... Recent research (see Web Appendix A for a summary) suggests that offerings involving positive uncertainty may be appealing to consumers by virtue of its ability to induce positive affect and stimulate cognitive processing. First, the element of surprise in the domain of gains creates positive emotions (Mellers and McGraw 2000 ;Wilson et al. 2005 ), thereby enhancing the appeal of surprise outcomes ( Loewenstein 1987 ;Shen, Fishbach, Hsee 2015 ), particularly for affect-driven purchase decisions ( Laran and Tsiros 2013 ). More specifically, research has shown that optimism and curiosity drive preferences for surprise rewards ( Goldsmith and Amir 2010 ;Hill et al. 2015 ;Shen et al. 2015 ;van Dijk and Zeelenberg 2007 ). ...
... First, the element of surprise in the domain of gains creates positive emotions (Mellers and McGraw 2000 ;Wilson et al. 2005 ), thereby enhancing the appeal of surprise outcomes ( Loewenstein 1987 ;Shen, Fishbach, Hsee 2015 ), particularly for affect-driven purchase decisions ( Laran and Tsiros 2013 ). More specifically, research has shown that optimism and curiosity drive preferences for surprise rewards ( Goldsmith and Amir 2010 ;Hill et al. 2015 ;Shen et al. 2015 ;van Dijk and Zeelenberg 2007 ). Second, positive uncertainty enhances cognitive engagement and information processing ( Golman, Loewenstein, and Gurney 2015 ;Tormala, Jia, and Norton 2012 ) and often leads to mental simulation and savoring of potential positive outcomes ( Lee and Qiu 2009 ;Loewenstein 1987 ). ...
Article
Full-text available
In the last few years, retailers have introduced numerous products that intentionally conceal some information from the consumer at the time of decision making. While prior research has identified contexts in which customers are attracted to such offerings in the aggregate, heterogeneity in customer proclivities is not well-understood. In the present paper, we examine the effect of gender on choice of surprise (vs. certain) offerings at the point of purchase. We propose and find that, on average, men are less likely to opt for a surprise offering compared to women. We examine multiple mechanisms that could explain this effect – emotionality, desire for exploration, and desire for control – and find the strongest support for the latter, demonstrating that it is men's stronger desire for control over the purchase outcome that drives their preference for certain (vs. surprise) offerings. Consequently, contexts or product categories that make it acceptable for men to let go of control attenuate the observed gender difference. We present data from a travel services firm, an online product catalog, and both field and lab studies, providing robust support for this theory across multiple product categories and participant populations. This work concludes with a discussion of the potential boundary effects of the observed gender difference, a managerial roadmap that delineates the ways in which marketers can offer surprise offerings more fruitfully to both men and women, and recommendations for future research.
... Laran and Tsiros (2013) explored the promotional effect of uncertainty on freebies under different consumption decisions, and the results showed that if consumers made decisions based on emotion, uncertain rewards would bring them surprise and pleasure, which will promote the improvement of marketing performance. Shen et al. (2015) investigated the possible positive effects and applicable conditions of uncertain rewards. It was found that when people pay more attention to the process of pursuing uncertain rewards rather than the results, they will get excited. ...
Article
Full-text available
Since 2019, China has gradually seen a “blind box” boom, and young people have quickly become the main buying force of blind boxes, promoting the continuous development of the blind box industry. Previous studies have shown that uncertainty in events with positive prospects can play a more positive role than certainty. However, how does uncertainty in the blind box affect consumers’ emotions and cognition and trigger subsequent consumption decisions? To clarify the internal mechanism of this process, this paper takes the blind box as the research object and constructs the mechanism model of perceived uncertainty on consumers’ impulsive purchase intention, based on Stimulus-Organism-Response (SOR) theory. In addition, the curiosity variable and perceived luck variable are introduced according to the information gap theory and optimism theory. On this basis, we conduct an empirical analysis by means of a questionnaire survey. The results show that perceived uncertainty has a positive impact on consumers’ impulsive purchase intentions, in which curiosity plays a mediating role. Besides, perceived luck positively moderates the impact of perceived uncertainty on impulsive purchase intention. This study clarifies the internal impact of perceived uncertainty on impulsive purchase intention of the blind box and enriches the basic theory of uncertainty reward and purchase intention. At the same time, we also offer related recommendations for future enterprises to learn from the marketing model of uncertain rewards.
... Storytellers have long intuited that surprising, unexpected events capture attention and keep it, which is why the most engaging stories contain an element of surprise (Heath & Heath, 2007). Given the psychological pull to resolve curiosity and answer open questions (Kruglanski & Webster, 1996;Shen et al., 2015), surprising events draw engagement. Modern TV advertisers use this fact to their advantage, for example, by creating TV ads that only reveal the name of the product at the end of the commercial to hold the viewer's attention (Loewenstein, 1994). ...
Article
Full-text available
We present a new consequence of stereotypes: they affect the length of communications. People say more about events that violate common stereotypes than those that confirm them, a phenomenon we dub surprised elaboration. Across two public data sets, government officials wrote longer reports when negative events befell White people (stereotype-inconsistent) than when the same events befell Black or Hispanic people (stereotype-consistent). Officers authored longer missing child reports of White (vs. Black or Hispanic) children (Study 1a), and medical examiners wrote longer reports of unidentified White (vs. Black or Hispanic) bodies (Study 1b). In follow-up experiments, communicators found stereotype-inconsistent events more surprising and this prompted them to elaborate (Study 2). Surprised elaboration occurred for negative events (i.e., crimes, misdemeanors) and also positive ones (i.e., weddings; Study 3). We found that surprised elaboration has policy implications. Observers preferred to funnel government and media resources toward White victims, since their case reports were longer, even when longer reports were not more informative (Studies 4-6). Together, these studies introduce surprised elaboration, a new theoretical phenomenon with implications for public policy. (PsycInfo Database Record (c) 2022 APA, all rights reserved).
... However, it should be noted that an excessive amount of variability is undesirable because it gives rise to unpredictability and uncertainty, which can induce physiological reactions and incur health problems (e.g., Romero et al., 2009). Consistent with the theory of optimal stimulation and arousal (Berlyne, 1960), people generally do not like excessive uncertainty and are therefore inherently motivated to resolve it even if more resource investment is needed and when negative consequences are possible (Hsee & Ruan, 2016;Shen et al., 2015). ...
Article
Full-text available
Much of everyday life consists of obligatory long-term behaviors, from work itself to doing dishes. Although some activities (e.g., exercise) are harder than others, even the seemingly easy activities can turn into hard ones when repeated monotonously day after day. It is proposed that this paradox has its roots in two fundamental human tendencies: (a) following the path of least resistance and (b) avoiding monotony, boredom, and even stimulus deprivation. How does the human mind operate in the presence of these tendencies to get everyday tasks and activities completed? The first tendency is manifested in gradual reduction of conscious processing and incremental increase in nonconscious processing. Constant repeats of demanding behaviors reduce cognitive strain and energy consumption, enhancing automaticity through the strengthened cue-behavior link. Automaticity in turn makes task performance easier and more efficient, resulting in a greater likelihood of getting everyday behaviors done. However, the sheer repeating of behavior is not enough; rather, a key is to repeat-with-variety, which posits that conscious interjection of variability into routine patterns facilitates the completion of both demanding and nondemanding behaviors in the long run. It is important, however, that such conscious intervention does not activate a sense of freedom about engagement because if it did, the elevated sense of freedom would undermine attempts to repeat behaviors and complete tasks. Understanding task completion also requires a consideration of the object of consciousness: Being unconscious of the mental processes motivating an action but conscious of the experience of doing the action.
... Typically, the closer consumers are to receiving a reward, the more motivated they are (Kivetz et al. 2006). Research has uncovered a variety of factors that influence the effectiveness of rewards (DellaVigna and Pope 2018; Shen, Fishbach, and Hsee 2015), such as how rewards interact with social or intrinsic motivation (Gneezy, Meier, and Rey-Biel 2011;Woolley and Sharif 2021), as well as who the recipient of the reward is (i.e., the self vs. a friend; Gershon et al. 2020). ...
Article
Full-text available
Eight studies (N = 5,025) demonstrate that consumers persist more when they must complete a target number of goal-related actions before receiving continuous rewards (i.e., what we term work-to-unlock rewards) than when they receive continuous rewards for their effort right away (i.e., what we term work-to-receive rewards). The authors suggest that the motivating power of work-to-unlock rewards arises because these rewards (1) naturally encourage consumers to set an attainable goal to start earning rewards, motivating consumers initially through goal setting and (2) keep consumers engaged after reaching this goal due to low perceived progress in earning rewards. A work-to-unlock reward structure increases persistence relative to standard continuous rewards across a variety of consumer-relevant domains (e.g., exercising, flossing, evaluating products), and even when work-to-unlock rewards offer rewards of a lower magnitude. Further, a work-to-unlock reward structure outperforms other reward structures that encourage goal setting. Lastly, the authors identify a theoretically consistent boundary condition of this effect: the length of the unlocking period.
Article
Drawing on people’s motivation to whet their curiosity, we tested a previously unexplored solution to reconciling want/should conflicts. Past work has shown that people are motivated to satisfy their curiosity and find enjoyment in doing so. Our work shows that piquing people’s curiosity can be leveraged to influence their choices, by steering them away from tempting “want” options (e.g., choosing unhealthy foods, watching lowbrow films, taking the elevator), and toward less-than-tempting, though normatively desirable “should” options. In two lab and two field studies, we created curiosity lures—incentives that pique people’s curiosity and deliver its closure on the condition people choose the “should” option over the “want” option. In all, our nudges were successful and highlight the external validity of our research. Notably, we observed a 9.8% increase in stairwell-use, and a 10% increase in fruit-and-vegetable purchases when we tested curiosity lures in large-scale field experiments totaling over 100,000 observations.
Article
Mysterious consumption items represent products that are chosen or purchased without knowing the exact nature of the product(s). In contrast to what we know about uncertainty aversion, the present research shows that consumers prefer uncertainty over certainty in the context of mysterious consumption. Across a variety of products (stress-balls, ice cream, songs, teas, snacks, hotel rooms, masks, rental cars), participants preferred mysterious consumption items over non-mysterious consumption items of equal expected value. The value of mysterious consumption lies at least in part in the uncertainty about the nature of the outcome among objectively similar outcomes. Specifically, the uncertainty around horizontally differentiated outcomes (i.e., outcomes that differ as a matter of taste) in the case of mysterious consumption focuses consumers on the positive side of uncertainty: the opportunity to be surprised. The preference for uncertainty is not observed when the possible outcomes are vertically differentiated (i.e., outcomes that differ in objective superiority, as is the case in existing demonstrations of uncertainty), or when horizontal uncertainty is reduced to a degree that diminishes the ability to be surprised. The findings reconcile literatures on surprise and uncertainty aversion, and help explain mysterious consumption as a substantive phenomenon in the marketplace.
Article
Résumé Offrir aux clients une expérience de qualité est devenu un enjeu critique pour les entreprises. A cette fin, les chercheurs ont mis en évidence l’opportunité d’utiliser la gamification. Cependant, et malgré la popularité grandissante de cette stratégie, les études rapportent des résultats divergents. Cette recherche vise à réconcilier les résultats antérieurs en identifiant les conditions dans lesquelles la gamification permet d’aboutir à une meilleure expérience. Sur base de la théorie de la résolution de l’incertitude, nous examinons comment l’incertitude de gagner affecte la qualité de l’expérience client. Les résultats de six études, combinant une enquête sur le terrain et des expérimentations, démontrent les bénéfices associés à l’incertitude de gagner en termes de qualité de l’expérience client. Nous montrons que cet effet existe pendant et après l’activité gamifiée, quand les participants sont informés de leur succès ou de leur échec. Par ailleurs, nous mettons en évidence que le fait de perdre le concours renforce l’impact positif de l’incertitude de gagner, ce qui suggère que l’incertitude de gagner permet de contrer l’effet négatif de l’aversion à l’échec lors d’un concours. A l’inverse, nous suggérons que la valeur de la récompense proposée atténue l’impact positif de l’incertitude de gagner en attirant l’attention du client sur l’enjeu plutôt que l’expérience vécue. Nos résultats contribuent à la littérature existante en soulignant le rôle de l’incertitude de gagner comme condition déterminante à la capacité de la gamification à délivrer une expérience de qualité.
Article
Offering customers a high-quality experience has become critical for companies. Scholars have therefore emphasized the opportunity to use gamification. However, despite the increasing popularity of such an approach, prior studies report mixed results. This research aims to reconcile these findings by identifying under which conditions gamification leads to a better experience. Drawing on uncertainty-resolution theory, we examine how uncertainty-to-win affects customer experience quality. The results from six studies, combining a field study and experiments, demonstrate the benefits of uncertainty-to-win for customer experience quality. We find that the uncertainty-to-win effect persists even after people are informed of a win/lose decision. Moreover, we highlight that losing the contest stresses the uncertainty-to-win effect, and consequently counter the harmful effect of loss. Conversely, we suggest that reward value mitigates the positive impact of uncertainty-to-win by focusing customers’ attention on a prize instead of on their experience. Our findings move the literature forward by underscoring how uncertainty-to-win is a condition that explains gamification’s ability to deliver a high-quality experience.
Book
I: Background.- 1. An Introduction.- 2. Conceptualizations of Intrinsic Motivation and Self-Determination.- II: Self-Determination Theory.- 3. Cognitive Evaluation Theory: Perceived Causality and Perceived Competence.- 4. Cognitive Evaluation Theory: Interpersonal Communication and Intrapersonal Regulation.- 5. Toward an Organismic Integration Theory: Motivation and Development.- 6. Causality Orientations Theory: Personality Influences on Motivation.- III: Alternative Approaches.- 7. Operant and Attributional Theories.- 8. Information-Processing Theories.- IV: Applications and Implications.- 9. Education.- 10. Psychotherapy.- 11. Work.- 12. Sports.- References.- Author Index.
Chapter
The main objective of this chapter is to build on Hsee and Zhang's work on the general evaluability theory (GET) and extend it in two respects. First, the chapter elaborates more on some of the key concepts in GET and reviews the most recent findings that provide support for some of the new predictions derived from GET. Second, it reviews more areas to which the evaluability concept applies, more importantly, the research that applies existing findings about evaluation mode and evaluability to changing the choice architecture and nudging people to make decisions that are better for themselves and for society. The chapter clarifies several issues concerning the concept of evaluation mode which may cause or have caused confusion. In substantive areas, research has examined the implications of evaluation mode and evaluability for risk preference and intertemporal preference, prediction-experience consistency and choice-experience consistency, extension neglect, and subjective well-being.
Article
Publisher Summary This chapter introduces regulatory mode theory and reviews evidence for distinguishing between an assessment mode concerned with making comparisons and a locomotion mode concerned with movement from state to state. It considers locomotion and assessment individually and assumes that the nature and the consequences of locomotion versus assessment are to some extent independent of each other. Evidence is presented to show that high achievement performance depends on individuals emphasizing both locomotion and assessment in their goal pursuits. Higher locomotion and higher assessment are shown to have distinct effects on judgment and decision making—including different preferences regarding decision strategies and leadership styles, different emphases in the decision process, and different self-evaluative judgment styles. The chapter highlights the fact that there are trade-offs to each regulatory mode. It describes studies on chronic individual differences and momentary situational differences in regulatory mode, as well as organizational differences that vary in their fit with regulatory mode.
Article
Many consumer promotions involve uncertainty (e.g., purchase incentives offering the chance to receive one of several rewards). Despite retailers' heavy reliance on such promotions, much academic research on uncertainty has demonstrated examples of consumers avoiding and/or disliking uncertainty, implying that promotions involving uncertainty may not be as effective for retailers as promotions offering certain rewards. In an effort to reconcile the prevalence of uncertain promotions with the existing research, this article explores the conditions under which uncertain promotions may be effective for retailers. The article concludes with a discussion of the theoretical and practical implications for these findings.