Personality and Individual Differences 95 (2016) 40–44

Contents lists available at ScienceDirect

Personality and Individual Differences

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

The relationship between anxiety and risk taking is moderated

by ambiguity
Andrew R. Smith ⁎, Eva E. Ebert, Joshua J. Broman-Fulks
Department of Psychology, Appalachian State University, United States

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

Article history: By definition, risk taking involves uncertainty surrounding potential outcomes. However, risky decisions can vary
Received 12 August 2015 in the amount of ambiguity about the likelihood of each outcome occurring. The current study tested the hypoth-
Received in revised form 5 January 2016 esis that the amount of ambiguity in risky-decisions would moderate the relationship between risk taking and
Accepted 7 February 2016
anxiety. In this study, participants completed individual difference measures and then a version of the Balloon
Available online xxxx
Analogue Risk Task (BART) with either high or low ambiguity about the likelihood of a negative outcome. As
Keywords:
hypothesized, higher levels of anxiety predicted less risk taking in the high ambiguity version of the BART, but
Risk taking anxiety and risk taking were unrelated to one another in the low ambiguity version. This study demonstrates
Anxiety that in order to understand the relationship between anxiety and risk taking, ambiguity level must be taken
Uncertainty into account. Furthermore, this finding provides support for cognitive models of anxiety suggesting that anxious
Ambiguity individuals interpret negative outcomes as more likely to occur than less anxious individuals.
Decision making © 2016 Elsevier Ltd. All rights reserved.
BART

1. Introduction Levin, 2001). One factor that has received a considerable amount of
attention is trait anxiety. Because anxiety is associated with pessimistic
People often make decisions in situations with uncertain outcomes. expectations regarding future events (e.g., Shepperd, Grace, Cole, &
For example, Vanessa, who is running late for work, might choose to Klein, 2005), high anxiety might act as a signal to avoid taking risks. In
drive faster than the speed limit in an attempt to get to work on time. support of this assumption, numerous studies have found that people
Or, while playing poker, Neil might place a large bet knowing that he with higher levels of anxiety tend to be risk-averse (e.g., Giorgetta
will only win the hand if he gets a spade and completes his flush. What et al., 2012; Maner et al., 2007; Maner & Schmidt, 2006). For example,
unites these risky situations is the potential for a negative outcome Maner et al. (2007) found that trait anxiety was negatively correlated
(Vanessa gets a speeding ticket, Neil loses the hand). One important with participants' risk-taking behavior. However, this finding has not
distinction between these situations, however, is that the likelihood of been universal, and studies investigating the relation this relationship
a negative outcome is much less clear for Vanessa than Neil. Neil can have sometimes generated contradictory results. For example, Mitte
precisely compute the likelihood that he will get the needed spade,1 (2007) conducted two similar studies and found the expected relation-
but it is much more difficult for Vanessa. (How likely is it that she will ship between anxiety and risk taking in the first study, but not the sec-
pass a police officer on her way to work?) The current study was de- ond. A number of variables have been investigated in an effort to explain
signed to investigate risky decisions in situations that differed in terms these conflicting results, including both situation- and person-specific
of the ambiguity surrounding the likelihood of the negative outcome. constructs. The domain of the risk, for instance, appears to influence
People's tendency to seek or avoid taking risks is related to a number anxious individuals' risk-taking behaviors, with studies suggesting
of individual differences, including age (Figner, Mackinlay, Wilkening, & that individuals with high anxiety are more likely to take health-
Weber, 2009), gender (Byrnes, Miller, & Schafer, 1999), emotional state related risks but less likely to take risks in most other domains
(Fessler, Pillsworth, & Flamson, 2004), and personality characteristics (e.g., recreation, career, and finance; Nicholson et al., 2005). Similarly,
(Nicholson, Soane, Fenton-O'Creevy, & Willman, 2005; Lauriola & Lauriola, Russo, Lucidi, Violani, and Levin (2005) found that the way
risky health decisions were framed moderated the relationship between
anxiety and risk taking. When making a choice between safe and risky
⁎ Corresponding author: Department of Psychology, Appalachian State University, options that were framed positively, anxiety did not predict partici-
Boone, NC 28608, United States. pants' choices. However, when the options were framed negatively,
E-mail address: smithar3@appstate.edu (A.R. Smith).
1
In Texas Hold'em, if Neil has a flush draw (i.e., he only needs one card to complete his
higher anxiety was related to risk-seeking choices.
flush) on the flop, there is a 34.97% chance he will complete his flush on either the turn One variable that has received relatively little attention in the litera-
(i.e., fourth card) or the river (i.e., fifth card). ture on the relation between anxiety and risk taking is the level of

http://dx.doi.org/10.1016/j.paid.2016.02.018
0191-8869/© 2016 Elsevier Ltd. All rights reserved.
 A.R. Smith et al. / Personality and Individual Differences 95 (2016) 40–44 41

ambiguity involved in the likelihood of outcomes. Cognitive models of 2.2.1. Depression, anxiety, and stress
anxiety propose that anxious individuals exhibit biases for threat- Participants completed a computerized version of the 21-item
related information and a propensity to interpret ambiguous stimuli Depression, Anxiety, and Stress Scale (DASS-21; Lovibond &
as more threatening and negative outcomes as more likely to occur Lovibond, 1995). For each item on this scale, participants indicate
than less anxious individuals, which may in turn affect their ability to how often they experienced a situation over the past week using a
process non-threat information and impair decision-making (Butler & 1 (“Did not apply to me at all. NEVER”) to 4 (“Applied to me very
Mathews, 1987; Clark & Wells, 1995). A relatively large and accumulat- much, or most of the time. ALMOST ALWAYS”) point response
ing body of research appears to support these models (e.g., Butler & scale. Example items are “I felt down-hearted and blue” (depres-
Mathews, 1987; Maner & Schmidt, 2006). For example, patients with sion), “I felt I was close to panic” (anxiety), and “I found it difficult
Social Anxiety Disorder exhibit threat interpretation biases toward am- to relax” (stress). In the current sample, internal consistency was
biguous social stimuli on both reaction time and self-report measures relatively good for depression (α = .88) and anxiety (α = .72), and
(Beard & Amir, 2009), and individuals with high levels of trait anxiety acceptable for stress (α = .67).
demonstrate impaired discriminatory fear learning under conditions
of ambiguity (Arnaudova et al., 2013; Lommen, Engelhard, & van den
Hout, 2010). Further, preliminary research suggests that highly anxious 2.2.2. Dispositional optimism
individuals may exhibit impaired decision-making on tasks that involve Participants completed a computerized version of the Life Orienta-
risk with high levels of ambiguity (e.g., the Iowa Gambling Task; IGT), tion Task—Revised (LOT-R; Scheier, Carver, & Bridges, 1994). This 10-
but not low levels of ambiguity (e.g., the Game of Dice Task; Kim et al., item scale (6 critical items and 4 fillers) assesses participants' level of
2015; Zhang et al., 2015). However, these studies have been limited in dispositional optimism. Participants indicate their level of agreement
several respects. For example, these tasks differ in a number of ways with each item on a 1 (“I disagree a lot”) to 5 (“I agree a lot”) point re-
other than their levels of ambiguity. Therefore, it is difficult to know sponse scale. An example item is “In uncertain times, I usually expect
whether the observed differences were due to the level of ambiguity the best”. In the current sample, the scale had relatively good internal
or some other feature of the tasks. In addition, these studies have consistency (α = .78).
often relied on comparisons of relatively small samples of individuals
with a diagnosed anxiety disorder (e.g., OCD; Kim et al., 2015; Zhang
et al., 2015) versus matched controls, despite noting that most of the 2.2.3. Risk taking
clinical patients were taking anxiolytic or antidepressant medications Participants completed a slightly modified version of the Balloon An-
at the time of the assessment, which may have impacted their alogue Risk Task (BART; Lejuez et al., 2002). Risk taking, as measured by
performance. Furthermore, these studies have not examined additional the BART, correlates with a variety of risk taking behaviors, including
constructs (e.g., dispositional optimism) that might partially account for risky sexual behavior (Lejuez, Simmons, Aklin, Daughters, & Dvir,
the relationship between anxiety and risk taking. Thus, additional re- 2004) and alcohol consumption (Fernie, Cole, Goudie, & Field, 2010).
search using tasks that differ only in the ambiguity about the likelihood The BART is a computerized task in which participants pump up numer-
of the outcomes and assessing a range of constructs is needed to clarify ous balloons, one at a time. Each pump of a balloon earns five points and
the relation between anxiety, risk taking, and ambiguity. participants can collect points at any time. If they collect the points for a
given balloon, that round is over, and they move to the next balloon. If
the balloon explodes before they collect the points, they lose all the
1.1. Current study points for that round. With each pump of the balloon, the likelihood of
the balloon exploding increases. This task requires that participants
Given the inconsistencies observed in previous studies (e.g., Mitte, weigh the likelihood of the negative outcome (i.e., the balloon explodes
2007), indications that the relationship between anxiety and risk taking and all points for that round are lost) with the potential gains
is moderated by various factors (e.g., Nicholson et al., 2005; Lauriola (i.e., getting 5 points for each pump).
et al., 2005), and preliminary evidence suggesting that ambiguity may We included three different balloon colors (blue, purple, and gray),
be particularly relevant to risky decision-making among anxious indi- each with a different initial likelihood of exploding. The blue balloon
viduals (Kim et al., 2015; Zhang et al., 2015), the goal of the present had a 1/10 chance of exploding on the first pump, the purple balloon
study was to investigate whether the level of ambiguity involved in a had a 1/20 chance, and the gray balloon had a 1/40 chance. With each
risky decision would moderate the relationship between anxiety and pump, the chance of explosion increased by decreasing the denomina-
risk taking. Based on previous research, we hypothesized that anxiety tor by 1. For example, the blue balloon had a 1/9 chance of exploding
would predict risk taking under conditions of high ambiguity, but not on the second pump, a 1/8 chance of exploding on the third pump,
under conditions of low ambiguity. and so on.
Most importantly, we created two versions of the BART—one with
2. Method relatively high ambiguity about the likelihood that the balloons would
explode and one with relatively low ambiguity. The high ambiguity
2.1. Participants version closely replicated the classic version of the BART. Participants
were told that the explosion likelihood of the three balloons varied,
One hundred and twenty-four (77.4% women, 22.6% men; but they were not told what the likelihoods were.
Mage = 19.64, SDage = 2.52) undergraduate students from a univer- The low ambiguity version included a visual indicator to let the
sity in the Southeastern region of the United States participated as participants know each balloon's explosion likelihood (see Fig. 1).
partial fulfillment of a course requirement. The visual indicator consisted of an array of balls on the right side
of the screen. The participants were told that the computer picked
a ball at random each time the balloon was pumped. If the computer
2.2. Measures picked a green ball, the balloon did not explode. If the computer
picked the red ball, the balloon exploded. Each time the participant
Participants completed measures of depression, anxiety, stress, opti- pumped up the balloon, a green ball was removed from the array to
mism, and risk taking. We included measures of depression, stress, and show the current explosion likelihood. Aside from the visual indica-
optimism to ensure that the observed relationship between anxiety and tor of the explosion likelihood, the high and low ambiguity versions
risk taking was not driven by another, related construct. of the BART were identical.
42 A.R. Smith et al. / Personality and Individual Differences 95 (2016) 40–44

3. Results

3.1. Relationships between individual differences

For each participant, we calculated the average of their responses for

the depression, stress, anxiety, and optimism scales. As shown in
Table 1, higher levels of anxiety were associated with greater depression
and stress, but lower optimism.

3.2. Risk taking

Participants' adjusted pumps for each balloon color (see Lejuez

et al., 2002) were calculated by computing the average number of
times the participant pumped up the balloon for all the rounds
when the balloon did not explode. Therefore, the adjusted pumps is
the average number of pumps for the rounds when the participant
made the choice to stop pumping up the balloon (rather than the choice
being made for him/her because the balloon exploded).2 Higher
adjusted pump scores indicate greater risk taking.
A 3 (balloon color: blue, purple, and gray) × 2 (ambiguity version:
low or high) analysis of variance on participants' adjusted pump
scores revealed a main effect of balloon color, F(2, 121) = 188.62,
p b .001, η2p = .76. Participants' adjusted pump scores were higher
for the balloons that exploded less often. There was also a main effect
of ambiguity version, F(1, 122) = 41.85, p b .001, η 2p = .26.
Participants who went through the low ambiguity version of the
BART exhibited higher levels of risk taking (i.e., had higher adjusted
pump scores) than participants who went through the high ambigu-
ity version. Finally, the analysis revealed a balloon color × ambiguity
version interaction, F(2, 121) = 64.06, p b .001, η2p = .51. As shown in
Fig. 2, the adjusted pump scores for the participants who went
through the low-ambiguity version varied across the three balloon
colors to a greater extent than participants who went through the
high-ambiguity version.
In short, participants were sensitive to the explosion rates of the
Fig. 1. Screenshots of the low ambiguity version of the Balloon Analogue Risk Task before
three balloons, and participants who went through the low ambiguity
the first pump (top) and after the 6th pump (bottom). version were more sensitive to these differences than participants
who went through the high ambiguity version of the BART. These
results suggest that the two versions of the BART varied with regards
to the amount of ambiguity about the likelihoods of the balloons
2.3. Procedure
exploding.

After providing their consent, participants completed the DASS and

LOT-R. Participants were then randomly assigned to complete either 3.3. Relationship between risk taking and anxiety
the high or low ambiguity version of the BART. To motivate the
participants in their task, they were told that performance on the In order to test our hypothesis that the relationship between anxiety
task is related to real-world decision-making skills. They were also and risk taking is moderated by ambiguity, we combined participants'
told that they would see their score and the average person's score average adjusted pump scores for all three balloons.3 We used the
at the end of the study. Furthermore, the participants were told PROCESS SPSS macro (Hayes, 2013) to conduct a regression analysis
that their task was to score as many points as possible because the with participants' anxiety score as the predictor, the ambiguity version
more points they earned, the more reward (candy) they would as the moderator, and the participants' adjusted pump score as the out-
receive. come variable. In this analysis, the variables were mean centered and
The participants received instructions about the BART and went we added participants' gender, age, and levels of depression, stress,
through 5 practice rounds. During the practice rounds the balloon was
yellow in order to limit learning the explosion rates. Aside from the
second practice round, the explosion likelihoods of the practice balloons
were 1/15 on the first pump. On the second practice round, the balloon 2
Participants' adjusted pump scores and average (unadjusted) pumps per round were
exploded on the first pump. This ensured that every participant highly correlated, r(122) = .98, p b .001. Therefore, analyses conducted on average pumps
rather than adjusted pump scores were virtually identical.
experienced an explosion during the practice rounds. After the 3
In addition to examining participants' overall adjusted pump scores, we also conduct-
practice rounds, the participants completed 60 rounds. The rounds ed separate analyses for each of the three balloons. A similar, but not identical pattern of
were presented in two blocks of 30 rounds—10 per balloon results was observed for the three balloons. For all three balloons, anxiety did not predict
color—with the order of the balloons randomized within each risk taking (ps N .50). For all three balloons, ambiguity predicted risk taking (all ps b .001).
block. After completing the 60 rounds, the participants were asked With the purple and gray balloon, participants' adjusted pump scores were higher in the
low ambiguity condition. However, with the blue balloon, scores were higher in the high
their age and gender, were shown their score and the average ambiguity condition. Finally, there was the predicted interaction between anxiety level
participant's score (from a pilot study), and given their earned reward and ambiguity version for the purple (p = .03) and gray (p = .005) balloons, but not
based on their performance. the blue balloon (p = .20).
 A.R. Smith et al. / Personality and Individual Differences 95 (2016) 40–44 43

Table 1
Descriptive statistics and relationships among participant characteristics.

Mean (SD) Depression Anxiety Stress Optimism Age Gender

Depression 1.53 (0.49) – .43⁎⁎⁎ .31⁎⁎⁎ −.57⁎⁎⁎ −.01 0.07

Anxiety 1.47 (0.41) – .42⁎⁎⁎ −.47⁎⁎⁎ .00 −0.09
Stress 1.91 (0.41) – −.33⁎⁎⁎ −.03 −0.18
Optimism 3.39 (0.72) – −.03 −0.23
Age 19.64 (2.52) – 0.55⁎⁎
77% women
Gender –
23% men

Note: Depression, anxiety, and stress coded on a 1–4 point scale; optimism coded on a 1–5 point scale; values listed for relationships between depression, anxiety, stress, optimism, and age
are r-values. Values listed for relationship with gender are d-values with positive numbers indicating higher values for men and negative numbers indicating higher values for women.
⁎ p b .05.
⁎⁎ p b .01.
⁎⁎⁎ p b .001.

and optimism as covariates.4 This analysis revealed that, overall, anxiety Although previous research has revealed high levels of uncertainty
did not predict participants' adjusted pump scores, t = − 0.74, to be associated with reduced risk taking, the current study suggests
b = − 0.39, p = .46, 95% CI [− 1.44, 0.66]. The ambiguity version did that the relationship between anxiety and risk taking may not be driven
predict risk taking, t = 5.63, b = 2.01, p b .001, 95% CI [1.30, 2.71]. solely by the uncertainty involved in risky-decisions. Both the high and
Most importantly, there was a significant anxiety × ambiguity condition low ambiguity versions of the BART involved uncertainty in that each
interaction, t = 2.45, b = 2.17, p = .016, 95% CI [0.42, 3.93]. This time the balloon was pumped, there was a chance it would explode.
interaction explained a significant proportion of the variance, R2 = .04, Rather, what separated the two versions was that one version had out-
F(1, 115) = 6.00, p = .016. Fig. 3 plots the relationship at one SD above comes that were relatively ambiguous and the other had outcomes that
and one SD below the mean of participants' anxiety score. Simple effects were relatively unambiguous. Thus, these results are consistent with the
analyses revealed that for the participants who went through the high finding that intolerance of uncertainty and intolerance of ambiguity,
ambiguity version of the BART, higher anxiety was associated with less while related, appear to be distinct constructs (Rosen, Ivanova, &
risk taking, t = − 2.23, b = −1.50, p = .03, 95% CI [−2.82, − 0.17]. Knäuper, 2014; but see Lauriola, Levin, & Hart, 2007), and it is important
However, in the low ambiguity version, anxiety and risk taking were un- for research to consider both constructs with regard to the decision-
related to one another, t = 0.97, b = 0.68, p = .34, 95% CI [−0.73, 2.08]. making strategies of anxious individuals.
This analysis also revealed that gender predicted participants' While our study adds to the literature regarding anxiety and risk tak-
adjusted pump scores, t = −2.25, b = −0.99, p = .03, 95% CI [−1.86, ing, some limitations warrant acknowledgment. Perhaps the most nota-
− 0.12]—a result consistent with previous studies findings that men ble limitation was the use of a non-clinical, undergraduate student
generally take more risks than women (Byrnes et al., 1999). None of sample. Although the pattern of results in the present study was largely
the other covariates significantly predicted risk taking (all ps N .27). consistent with previous research utilizing clinical samples, additional
research is needed to determine whether these findings will extend to
samples with clinical levels of anxiety. In addition, the present study
4. Discussion only employed one measure of risk-taking behavior, and it is not clear
whether these findings will extend to other behavioral risk-taking mea-
The current study investigated whether the amount of ambiguity in sures (e.g., the IGT, Bechara, Damasio, Damasio, & Anderson, 1994; the
the likelihood of a negative outcome moderates the relationship be- Cups Task, Levin & Hart, 2003). Previous research has suggested that
tween anxiety and risk taking. Consistent with our prediction, when performance on one task is not always correlated with performance
the likelihood of the negative outcome was ambiguous, participants on another task (Bishara et al., 2009), which may be at least partially re-
with higher anxiety exhibited less risk taking. This finding replicates nu- lated to the level of ambiguity surrounding the outcomes. For example,
merous other studies demonstrating that anxiety and risk taking are the BART and IGT require participants to learn the probabilities while
often negatively related (e.g., Giorgetta et al., 2012; Maner & Schmidt, going through the task, whereas the likelihoods in the Cups Task are
2006). Most notably, Maner et al. (2007, Study 2) found that anxious clearly observable by the participants. Future research could investigate
individuals took fewer risks as measured by the BART—the same risk- the relationship between anxiety and risk taking across different
taking measure used in the current study. While the results in the
high ambiguity condition replicated previous studies, the pattern was
different in the low ambiguity condition. When the likelihood of the
negative outcome was relatively unambiguous, there was not a signifi-
cant relationship between anxiety and risk taking. Anxious individuals
were, on average, no more risk-seeking or risk-avoidant than their less
anxious counterparts. This finding appears consistent with cognitive
models of anxiety and suggests that, in the absence of information re-
garding the probability of positive and negative outcomes, the tendency
of anxious persons to perceive increased threat and higher probability
of negative outcomes may lead them to being less willing to engage in
risk taking behaviors.

4
A similar analysis predicting participants' adjusted pump scores while not controlling
for gender, age, depression, stress, and optimism yielded similar results. Specifically, anx-
iety did not predict participants' adjusted pump scores, t = 0.09, b = 0.04, p = .93, 95% CI
[−0.83, 0.91]. The ambiguity version did predict risk taking, t = 5.73, b = 2.05, p b .001,
95% CI [1.34, 2.76]. Most importantly, there was a significant anxiety × ambiguity condi- Fig. 2. Participants' average adjusted pump scores as a function of balloon color and BART
tion interaction, t = 2.68, b = 2.36, p = .008, 95% CI [0.62, 4.11]. version. Error bars represent ±1 SE.
44 A.R. Smith et al. / Personality and Individual Differences 95 (2016) 40–44

Clark, D. M., & Wells, A. (1995). A cognitive model of social phobia. In R. G. Heimberg, M.
R. Liebowitz, D. A. Hope, & F. R. Schneier (Eds.), Social phobia: Diagnosis, assessment,
and treatment (pp. 69–93). New York, NY: Guilford Press.
Fernie, G., Cole, J. C., Goudie, A. J., & Field, M. (2010). Risk-taking but not response inhibi-
tion or delay discounting predict alcohol consumption in social drinkers. Drug and
Alcohol Dependence, 112(1–2), 54–61. http://dx.doi.org/10.1016/j.drugalcdep.2010.
05.011.
Ferrey, A. E., & Mishra, S. (2014). Compensation method affects risk-taking in the Balloon
Analogue Risk Task. Personality and Individual Differences, 64, 111–114. http://dx.doi.
org/10.1016/j.paid.2014.02.008.
Fessler, D. T., Pillsworth, E. G., & Flamson, T. J. (2004). Angry men and disgusted women:
An evolutionary approach to the influence of emotions on risk taking. Organizational
Behavior and Human Decision Processes, 95(1), 107–123. http://dx.doi.org/10.1016/j.
obhdp.2004.06.006.
Figner, B., Mackinlay, R. J., Wilkening, F., & Weber, E. U. (2009). Affective and deliberative
processes in risky choice: Age differences in risk taking in the Columbia Card Task.
Journal of Experimental Psychology: Learning, Memory, and Cognition, 35(3),
709–730. http://dx.doi.org/10.1037/a0014983.
Giorgetta, C., Grecucci, A., Zuanon, S., Perini, L., Balestrieri, M., Bonini, N., ... Brambilla, P.
(2012). Reduced risk-taking behavior as a trait feature of anxiety. Emotion, 12(6),
1373–1383. http://dx.doi.org/10.1037/a0029119.
Hayes, A. F. (2013). Introduction to mediation, moderation, and conditional process analysis:
Fig. 3. Participants average adjusted pump scores as a function of their level of anxiety and A regression-based approach. New York, NY, US: Guilford Press.
BART version. Kim, H. W., Kang, J. I., Namkoong, K., Jhung, K., Ha, R. Y., & Kim, S. J. (2015). Further evi-
dence of a dissociation between decision-making under ambiguity and decision-
making under risk in obsessive–compulsive disorder. Journal of Affective Disorders,
176, 118–124.
behavioral tasks that vary in respect to the amount of ambiguity regard- Lauriola, M., & Levin, I. P. (2001). Personality traits and risky decision-making in a con-
ing the likelihood of the outcomes. trolled experimental task: An exploratory study. Personality and Individual
Finally, a possible limitation in the current study is that the partici- Differences, 31(2), 215–226. http://dx.doi.org/10.1016/S0191-8869(00)00130-6.
Lauriola, M., Levin, I. P., & Hart, S. S. (2007). Common and distinct factors in decision mak-
pants were not given monetary incentives for their performance on
ing under ambiguity and risk: A psychometric study of individual differences.
the BART; instead, they received points so they could get a non- Organizational Behavior and Human Decision Processes, 104(2), 130–149. http://dx.
monetary reward (i.e., candy). Ferrey and Mishra (2014) recently doi.org/10.1016/j.obhdp.2007.04.001.
Lauriola, M., Russo, P. M., Lucidi, F., Violani, C., & Levin, I. P. (2005). The role of personality
demonstrated that the compensation method used in the BART may
in positively and negatively framed risky health decisions. Personality and Individual
affect participants' risk taking, with participants who were paid for Differences, 38(1), 45–59. http://dx.doi.org/10.1016/j.paid.2004.03.020.
their participation in addition to being paid based on their performance Lejuez, C. W., Read, J. P., Kahler, C. W., Richards, J. B., Ramsey, S. E., Stuart, G. L., ... Brown, R.
tending to take more risks than participants who were only paid for A. (2002). Evaluation of a behavioral measure of risk taking: The Balloon Analogue
Risk Task (BART). Journal of Experimental Psychology: Applied, 8(2), 75–84. http://dx.
their performance or not paid at all. While our compensation method doi.org/10.1037/1076-898X.8.2.75.
might have influenced participants' risk taking, it seems unlikely that Lejuez, C. W., Simmons, B. L., Aklin, W. M., Daughters, S. B., & Dvir, S. (2004). Risk-taking
this can account for the interactive effect of ambiguity and anxiety on propensity and risky sexual behavior of individuals in residential substance use treat-
ment. Addictive Behaviors, 29(8), 1643–1647. http://dx.doi.org/10.1016/j.addbeh.
risk-taking behavior. Future research is needed to assess whether 2004.02.035.
alternative compensation methods may impact the relation between Levin, I. P., & Hart, S. S. (2003). Risk preferences in young children: Early evidence of in-
ambiguity, risk-taking, and anxiety. dividual differences in reaction to potential gains and losses. Journal of Behavioral
Decision Making, 16(5), 397–413. http://dx.doi.org/10.1002/bdm.453.
Despite the need for further research, the current study demon- Lommen, M. J. J., Engelhard, L. M., & van den Hout, M. A. (2010). Neuroticism and avoid-
strates that in order to understand the relationship between anxiety ance of ambiguous stimuli: Better safe than sorry? Personality and Individual
and risk taking, the ambiguity level must be taken into account. This is Differences, 49, 1001–1006. http://dx.doi.org/10.1016/j.paid.2010.08.012.
Lovibond, P. F., & Lovibond, S. H. (1995). The structure of negative emotional states: Com-
especially important because many real-world decisions vary in terms parison of the Depression Anxiety Stress Scales (DASS) with the Beck Depression and
of the ambiguity surrounding the likelihood of the outcomes. Further- Anxiety Inventories. Behaviour Research and Therapy, 33(3), 335–343. http://dx.doi.
more, this finding provides additional support for cognitive models of org/10.1016/0005-7967(94)00075-U.
Maner, J. K., & Schmidt, N. B. (2006). The role of risk avoidance in anxiety. Behavior
anxiety suggesting that anxious individuals interpret ambiguous stimuli
Therapy, 37, 181–189. http://dx.doi.org/10.1016/j.beth.2005.11.003.
as more threatening and negative outcomes as more likely to occur than Maner, J. K., Richey, J. A., Cromer, K., Mallott, M., Lejuez, C. W., Joiner, T. E., & Schmidt, N. B.
less anxious individuals. (2007). Dispositional anxiety and risk-avoidant decision-making. Personality and
Individual Differences, 42(4), 665–675. http://dx.doi.org/10.1016/j.paid.2006.08.016.
Mitte, K. (2007). Anxiety and risky decision-making: The role of subjective probability
and subjective costs of negative events. Personality and Individual Differences, 43(2),
References 243–253. http://dx.doi.org/10.1016/j.paid.2006.11.028.
Nicholson, N., Soane, E., Fenton-O'Creevy, M., & Willman, P. (2005). Personality and
Arnaudova, I., Krypotos, A. -M., Effting, M., Boddez, Y., Kindt, M., & Beckers, T. (2013).
domain-specific risk taking. Journal of Risk Research, 8(2), 157–176. http://dx.doi.
Individual differences in discriminatory fear learning under conditions of ambiguity:
org/10.1080/1366987032000123856.
A vulnerability factor for anxiety disorders? Frontiers in Psychology, 4, 1–9. http://dx.
Rosen, N. O., Ivanova, E., & Knäuper, B. (2014). Differentiating intolerance of uncertainty
doi.org/10.3389/fpsyg.2013.00298.
from three related but distinct constructs. Anxiety, Stress & Coping: An International
Beard, C., & Amir, N. (2009). Interpretation in social anxiety: When meaning precedes
Journal, 27(1), 55–73. http://dx.doi.org/10.1080/10615806.2013.815743.
ambiguity. Cognitive Therapy and Research, 33(4), 406–415. http://dx.doi.org/10.
Scheier, M. F., Carver, C. S., & Bridges, M. W. (1994). Distinguishing optimism from neurot-
1007/s10608-009-9235-0.
icism (and trait anxiety, self-mastery, and self-esteem): A reevaluation of the Life Ori-
Bechara, A., Damasio, A. R., Damasio, H., & Anderson, S. W. (1994). Insensitivity to future
entation Test. Journal of Personality and Social Psychology, 67(6), 1063–1078. http://
consequences following damage to human prefrontal cortex. Cognition, 50(1–3),
dx.doi.org/10.1037/0022-3514.67.6.1063.
7–15. http://dx.doi.org/10.1016/0010-0277(94)90018-3.
Shepperd, J. A., Grace, J., Cole, L. J., & Klein, C. (2005). Anxiety and outcome predictions.
Bishara, A. J., Pleskac, T. J., Fridberg, D. J., Yechiam, E., Lucas, J., Busemeyer, J. R., ... Stout, J. C.
Personality and Social Psychology Bulletin, 31, 267–275. http://dx.doi.org/10.1177/
(2009). Similar processes despite divergent behavior in two commonly used
0146167204271322.
measures of risky decision making. Journal of Behavioral Decision Making, 22(4),
Zhang, L., Dong, Y., Yifu, J., Zhu, C., Yu, H., Chen, X., & Wang, K. (2015). Dissociation of
435–454. http://dx.doi.org/10.1002/bdm.641.
decision making under ambiguity and decision making under risk: A neurocognitive
Butler, G., & Mathews, A. (1987). Anticipatory anxiety and risk perception. Cognitive
endophenotype candidate for obsessive–compulsive disorder. Progress in Neuro-
Therapy and Research, 11(5), 551–565. http://dx.doi.org/10.1007/BF01183858.
Psychopharmacology and Biological Psychiatry, 57, 60–68. http://dx.doi.org/10.1016/j.
Byrnes, J. P., Miller, D. C., & Schafer, W. D. (1999). Gender differences in risk taking: A
pnpbp.2014.09.005.
meta-analysis. Psychological Bulletin, 125, 367–383. http://dx.doi.org/10.1037/0033-
2909.125.3.367.