European Journal of Social Psychology, Eur. J. Soc. Psychol.

(2014)
Published online in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/ejsp.2067

Research article
The effect of construal level on risk-taking

EVA LERMER1*, BERNHARD STREICHER1, RAINER SACHS2, MARTINA RAUE1 AND DIETER FREY1
1
Department of Psychology, Ludwig Maximilian University Munich, Munich, Germany; 2Munich Re, Munich,
Germany

Abstract

In a series of studies, we examined the influence of people’s mind-set (construal level (CL): abstract versus concrete) on their risk-
taking behavior. We measured differences in CL (study 1, CL as trait) and manipulated CL (studies 1–5, CL as state) with different
priming methods, which were unrelated to the dependent variable of risk-taking behavior (studies 1, 3, 4, and 5: Balloon Analog
Risk Task; study 2: Angling Risk Task). In all studies, abstract CL resulted in greater risk-taking compared with concrete CL,
which led to lower risk-taking. Risky and safe game strategies mediated the CL effect on risk-taking. A concrete mind-set in-
creased the safe game strategy, whereas an abstract mind-set increased the risky game strategy. Furthermore, different potential
mediators were explored (i.e., focus on payoffs and probabilities, prevention versus promotion focus, attention to pros versus
cons, and mood). A concrete mind-set increased prevention strategies and a negative mood when compared with an abstract
mind-set. In turn, an abstract mind-set increased attention to pros (of an action). Copyright © 2014 John Wiley & Sons, Ltd.

Can differences in mind-set influence risk-taking behavior? Tra- seeking; gain versus loss frames; and situational factors, such
ditionally, risk-taking was assumed to be a stable personality as affect, but also by people’s cognitive mind-set (CL).
trait in that people can be grouped as either risk-seeking versus
avoiding (Eysenck & Eysenck, 1977; Hanoch, Johnson, &
CONSTRUAL LEVEL AND RISK-TAKING
Wilke, 2006; Zuckerman & Kuhlman, 2000). This was then ex-
panded to include other considerations (Johnson, Wilke, &
Weber, 2004). Risk behavior is now more broadly described The main idea of CLT is that the psychological distance (“the
with foci on its multidimensional nature (e.g., Figner & Weber, subjective experience that something is close or far away from
2011). Many research on prospect theory (Kahneman & self, here and now”; Trope & Liberman, 2010, p. 440) to a tar-
Tversky, 1979) shows that gain versus loss frames change the get (e.g., event or object) influences, and is influenced by, the
way people represent events, which in turn influences their level of mental construal. People apply a more abstract, high
preferences. Today, commonly observed influences on risk CL when judging a psychologically distal target. They also,
behavior include subtraits such as sensation seeking (e.g., however, assess more abstract targets to be psychologically
Zuckerman, 2007), self-monitoring (e.g., positive relationship more distal (Liberman, Sagristano, & Trope, 2002). Thus,
between self-monitoring in the form of public performing and the link between the mind-set and psychological distance can
risk-taking behavior; Bell, Schoenrock, & O’Neal, 2000), be described as bidirectional (e.g., Bar-Anan, Liberman, &
constructs such as mood (e.g., De Vries, Holland, & Witteman, Trope, 2006; Henderson, Wakslak, Fujita, & Rohrbach, 2011).
2008; Yuen & Lee, 2003), or visceral influences (e.g., visceral To apply CLT to risk-taking behavior, we focus on the
cues that indicate proximity to desired objects can lead to aspects of CLT that relate to desirability and feasibility
decisions that are less sensitive to risk information; Ditto, concerns about future events (e.g., games of chance), which
Pizarro, Epstein, Jacobson, & MacDonald, 2006). These factors “correspond to the distinction between means and ends”
contribute to the common observation that one sometimes (Liberman & Trope, 1998, p. 7). CLT predicts that abstract
chooses to play it safe when faced with risky situations and other (i.e., high-level) construals promote sensitivity to desirabil-
times not (Freeman & Muraven, 2010). Freeman and Muraven ity considerations (i.e., the value of an end state of an
(2010) raised the question as to what generates people’s action: for instance, amount of money gained) and that
inconsistency in risk-taking behavior. concrete (low-level) construals promote sensitivity to
This question is addressed here using a novel approach, that feasibility considerations (i.e., ease/difficulty of reaching
is, construal level (CL) (Trope & Liberman, 2010). On the basis the end state; Sagristano, Trope, & Liberman, 2002). For
of CL theory (CLT), we assume that risk-taking may be broadly example, desirability would represent getting a high grade
influenced not only by personality traits, such as sensation in a college test, whereas feasibility is the amount of work
*Correspondence to: Eva Lermer, Department of Psychology, Ludwig Maximilian University Munich, Leopoldstr. 13, Munich, Germany 80803, Germany.
E-mail: [email protected]

Copyright © 2014 John Wiley & Sons, Ltd. Received 29 April 2013, Accepted 17 August 2014
Eva Lermer et al.

(e.g., time) one has to invest to obtain a good grade (Liberman influences risk-taking behavior and whether different mecha-
& Trope, 1998). In terms of risk-taking, desirability would nisms would mediate this effect.
reflect the value of the outcome of a game of chance
(e.g., payoff/amount of gain/loss), whereas feasibility is the
probability of winning/loosing. POTENTIAL MEDIATING MECHANISMS OF THE
An important distinction between gain and loss domains in CONSTRUAL LEVEL EFFECT
risk-taking must also be made. In the gain domain, abstract
construals promote risk-taking because abstract construals in-
crease payoff sensitivity. Concrete construals, on the other hand, Despite much research on CLT, the cognitive or affective pro-
cesses responsible for different representations of targets are
promote probability sensitivity. The opposite is true in the loss
still not well understood. However, different parameters can
domain, where abstract construals promote risk aversion and
concrete riskiness. Some of the first research on the influence be linked with CL effects, for example, regulatory focus or
mood. The idea of more risk aversion when the mind-set is
of CL on risk preferences was by Sagristano et al. (2002) who
concrete (versus abstract) is supported by the regulatory focus
investigated the impact of CL on preferences (probabilities
versus payoffs) in different games of chance by manipulating theory (Higgins, 1997). The regulatory focus theory provides a
useful framework for investigating people’s risk tendencies
temporal distance (studies 1–3: same day (near) versus in
(Zou, Scholer, & Higgins, 2014), by distinguishing between
2 months (distant); study 4: next day versus in 2 months). As
predicted by CLT, the results indicated that temporal distance two goal attainment strategies: promotion focus and preven-
tion focus (Crowe & Higgins, 1997). The promotion focus is
increased the influence of outcomes (payoffs and desirability
characterized by eager approach strategies to goal pursuit
concerns) and decreased the influence of winning prob-
abilities (feasibility concerns) on participants’ preferences. (attaining gains). Within the prevention focus, safety (avoiding
losses) is emphasized (Bryant & Dunford, 2008; Förster &
Specifically, the authors showed that participants were willing
Higgins, 2005). Research has indicated that a low-level con-
to take more risks in distant-future gambles than in the near
future. Participants were willing to bid and risk losing more strual processing style fits the prevention focus on safety and
that high-level construal processing fits the promotion focus
money for low-probability (of winning) but high-payoff
on advancement (Förster & Higgins, 2005; Lee, Keller, &
gambles (risky choice), as opposed to high-probability but
low-payoff gambles (nonrisky choice) in the distant future. Sternthal, 2010; Pennington & Roese, 2003). For instance,
participants who were primed with a promotion focus catego-
Furthermore, the authors showed that the inverse was true
rized objects more broadly (and into fewer categories: a typical
for near-future gambles, that is, people are less willing to
take risks in near-future gambles as participants showed a indicator for high-level CL) than participants primed with a
prevention focus (Lee et al., 2010).
higher preference for gambles with high-probabilities but
Furthermore, Förster and Higgins (2005) revealed that the
low-payoff as opposed to gambles with low-probabilities
but high-payoff in near-future gambles. However, two regulatory focus systems are related to either global (abstract)
questions remain unanswered: or local (concrete) processing styles. By measuring reaction
Is the CL effect apparent when people adopt different mind- time, these authors showed that local processing fits a preven-
sets? This can be rephrased in line with the reciprocal charac- tion focus, whereas global processing fits a promotion focus.
ter of CLT (bidirectional relationship of mind-set and psycho- Furthermore, there is substantial empirical evidence that sup-
logical distance, Trope & Liberman, 2010): Will an abstract ports the link between a promotion focus and an increased
mind-set increase risk-taking behavior compared with a con- willingness to take risks compared with a prevention focus
crete mind-set in the gain domain? Previous research revealed (Crowe & Higgins, 1997; Grant & Higgins, 2003; Werth &
that increasing temporal distance is linked to more abstract Förster, 2007). Werth and Förster (2007) summarized that
thinking (Förster, Friedman, & Liberman, 2004). Therefore, people with a promotion focus show a more risky approach
we are particularly interested in whether people who adopt a compared with those with a prevention focus. However, recent
concrete mind-set (after CL manipulation) are less willing to studies limit this assumption to gains frames. Prior research
take risks than abstract thinkers. The second question is: what showed that people have a general tendency to be risk averse
are the underlying psychological mechanisms responsible for in the domain of gains (e.g., Blais & Weber, 2006). More re-
this effect? These are not yet fully clarified. cently, Zou et al. (2014) demonstrated that this tendency is
The influence of CL on risk-taking is of great significance: (i) moderated by the strength of the promotion focus and not by
because people are continuously confronted with risky situations the strength of the prevention focus. Scholer, Zou, Fujita,
and changing mind-sets simultaneously (e.g., driving a car on an Stroessner, and Higgins (2010) revealed that in situations in-
icy road while thinking of the next summer holiday or planning volving loss, it is prevention focus and not promotion focus
what to buy at the grocery store); and (ii) environmental that predicts increased risk-taking when risk-taking appears
circumstances can easily and unexpectedly change the mind- to be the only possibility of returning to the status quo. When
set, for example, when people are confronted with an activity there was no option to return to the status quo, prevention fo-
(e.g., mountaineering), they may have an increased focus on cus predicts risk aversion (Scholer et al., 2010). Zou and col-
the goal (e.g., to get the top; desirability). This is abstract in leagues (2014) showed that promotion motivated participants
terms of CLT, but if CL influences risk-taking, then the focus choose risky options in the domain of gains (when the stock
on details (e.g., the trail and weather; feasibility) would decrease portfolio remained unchanged; i.e., when there was nothing
the risk-taking propensity. Therefore, it is proposed that this ef- to lose but something to win) in no-change or minimal-gain
fect could increase safety. We aimed to investigate whether CL situations, but participants switched to more conservative

Copyright © 2014 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. (2014)
 Construal level and risk

behavior when they had just experienced a large gain. Thus, strategies (e.g., Förster & Higgins, 2005) and that in the
increased risk-taking seems not to be rigidly linked to domain of gains, promotion (versus prevention) motivated
promotion focus. The respective context (domain and amount participants to show increase risk-taking. In particular, a pro-
of gain versus loss) and experience has to be considered. motion focus seems to be linked to abstract thinking and a
Because the current study aims at replicating the findings of prevention focus to concrete thinking. Because these strategies
Sagristano et al. (2002) with a focus on the bidirectional char- are also linked to the willingness to take risks, we investigated
acter of CLT (i.e., focusing on the mind-set: abstract versus whether they mediate the CL effect. Eyal et al. (2004) showed
concrete), we are particularly interested in the impact of pro- a relationship between CL and participants’ focus on the pros
motion and prevention focus on risk-taking in the domain of versus the cons of an action: Participants found more pros and
gains. We predict that a focus on prevention or promotion is fewer cons for activities in the distant future. Finally, Gasper
a potential mediator and that reluctant risk-taking, as an effect and Clore (2002) showed that a happy mood is linked to global
of low CL, should be mediated by a prevention focus. Con- processing and a sad mood to local processing.
versely, intense risk-taking, as a result of high CL, should be We experimentally tested the impact of CL on risk-taking
mediated by a promotion focus, in the domain of gains. behavior in order to corroborate the notion that the level of
Furthermore, people’s focus on the outcome (payoff versus mental construal influences risk-taking. This was carried out
loss) could be another potential mediator. It could be assumed via different priming methods and risk-taking tasks and tests
that a more concrete mind-set, which accompanies detailed of the following potential mediators: attention to probabilities,
thinking, leads to more salience of potentially negative out- outcomes, and pros and cons of actions, as well as promotion
comes (i.e., no gain). This is empirically supported by Eyal, and prevention strategies and mood.
Liberman, Trope, and Walther (2004) who showed that high
construals are related to a focus on considerations in favor of
an action (pros) as opposed to considerations against an action STUDY 1
(cons; Spassova & Lee, 2008). Hence, the focus on pros or
cons could be a potential mediator of the CL effect, in that less
Study 1 investigated whether differences in CL of the mind-set
risk-taking because of low CL is mediated by a focus on the
(concrete versus abstract), either as a trait or as a state, have an
cons, whereas intense risk-taking because of high CL is medi-
impact on risk-taking behavior. Participants’ CL (trait) was
ated by a focus on the pros.
measured, and after being primed in CL (state), they con-
A further potential mediator could be a change in mood.
ducted a risk-taking task.
However, research on the association of mood and CL is
sketchy and ambiguous. Gasper and Clore (2002) showed that
a positive mood is linked to a global (versus local) processing Method
style. Hence, a mood change when the mind-set is concrete
(versus abstract) could cause more risk aversion as people in Fifty-eight university students (54 female; age M = 26.04;
a depressed mood are more conservative in risk-taking than SD = 7.32) were recruited on campus to take part in the study
those in an elated mood (Yuen & Lee, 2003). However, so in exchange for course credit. A German version (Beckmann,
far other attempts to show that differences in mood are related 1993) of the Behavioral Identification Form (BIF; Vallacher &
to people’s CL have failed (Rim, Uleman, & Trope, 2009). Wegner, 1989) was used to assess participants’ tendency to ei-
ther a concrete or abstract mind-set (i.e., CL as trait). The BIF
is a common instrument to measure abstraction (Burgoon,
HYPOTHESES
Henderson, & Markman, 2013) and, therefore, is suitable to
measure CL. The ratio of abstract versus concrete answers
In summary, research on the connection between different (higher values indicate more abstract thinking) was used as
mind-sets (CLs) and the desirability and feasibility of actions, an indicator (cf. Luguri, Napier, & Dovidio, 2012) for abstract
as well as the focus on self-regulatory systems in goal attain- or concrete mind-sets (M = 0.51, SD = 0.17; Cronbach’s
ment, suggests that CL has an impact on risk-taking behavior. α = .796).
This relationship is expressed in the following formal Participants were randomly divided into two CL priming
hypothesis: groups: abstract or concrete. CL manipulation was carried
out via categorization priming (adapted from Fujita, Trope,
Hypothesis 1: Higher CL will result in more risk-taking behav-
Liberman, & Levin-Sagi, 2006), which is a common CL prim-
ior than lower CL.
ing method (Burgoon et al., 2013), where participants are
Assuming that mind-set influences risk-taking behavior, asked to name either a subordinate (concrete CL; e.g., rose)
then one aim is to investigate the potential mechanisms or a superordinate (abstract CL; e.g., plant) category for 30
responsible for the CL effect. Sagristano et al. (2002) showed different items (e.g., flower).
that participants who perceive a greater psychological distance Risk behavior was measured using the Balloon Analog
have an increased preference for outcomes/payoffs over Risk Task (BART; adapted from Lejuez et al., 2002), which
probabilities, compared with participants perceiving less has been shown to be a reliable measurement of risk-taking
psychological distance. Therefore, we aimed to explore propensity (Lejuez, Aklin, Zvolensky, & Pedulla, 2003). Risk-
whether a preference for probabilities or payoffs mediates the iness in the BART is positively correlated with risky behavior
CL impact on risk-taking behavior. Furthermore, previous in real life (e.g., gambling and alcohol and drug use; Chandler
research showed that CL is linked to promotion and prevention & Pronin, 2012; for an overview, see Hunt, Hopko, Bare,

Copyright © 2014 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. (2014)
Eva Lermer et al.

Lejuez, & Robinson, 2005). The aim of the BART is to inflate Table 1. Construal level (state) effect on risk-taking (studies 1 and 2)
30 computer-simulated balloons by clicking a “pump” button. Mind-set
With each click, a balloon inflates, and the participant receives
a virtual 5¢. The money accumulates until either that balloon Concrete Abstract
bursts (too many pumps), which leads to a total loss of the ac- (study 1: n = 30) (study 1: n = 28)
cumulated money for that balloon, or the participant decides to (study 2: n = 20) (study 2: n = 20)
finish the round and save accumulated money by pressing the M SD M SD
collect button. The aim of the game is to collect as much
money as possible. Because participants can only win or loose Study 1
previously gained money, the BART can be regarded as a Mean pumps 12.67 5.82 16.47 7.72
Bursted balloons 11.03 4.38 13.86 4.75
game in the domain of gains (versus no gains). This is relevant Risk index 0.37 0.73 0.16 0.93
to both replicating the findings of Sagristano et al. (2002) and Study 2
in regard to the aspect of CLT that predicts abstract construals Mean clicks 19.60 7.28 24.49 5.70
(and promotion motivation) increases riskiness in the domain Bad fish 10.90 3.99 13.45 2.66
of gains. Risk index 0.25 1.02 0.44 0.67
Participants were informed in advance that the person with Note: Concrete = low CL mind-set priming group; abstract = high CL mind-set
the highest score will be paid the winnings and that they will group.
receive course credit for participation. Participants did not
know how many pumps would burst a balloon. The number
of pumps required to burst each of the 30 balloons was random SD = 3.30), p = .531, although the way in which it was accu-
but constant across participants. mulated varied. It is important to note that the nature of the
BART task produces similar amounts of gains for both risk-
Results and Discussion seeking and risk-avoiding participants, in that, those who take
more risks gain more money per balloon but have more total
Two different risk-taking indicators of the BART data were cal- losses, whereas those who avoid risk have low gains per bal-
culated in order to determine whether a concrete mind-set leads loon but constantly earn money. Overall, both strategies (types
to more risk aversion than an abstract mind-set. Following the of behavior) lead to similar winnings. However, the current
recommendation of Lejuez et al. (2002), the mean number of study did not focus on the result (total gains) but on the means
pumps per balloon that did not burst and the total number of (risk-taking behavior). Study 2 further assessed the influence
burst balloons were used as indicators of willingness to take of CL as a state on risk-taking behavior by using different
risks. Furthermore, a risk index of the z-standardized means of priming and risk task.
both measures was calculated (cf. Hiemer & Abele, 2012) as
these indicators were sufficiently highly correlated, r = .682,
p < .001. STUDY 2
Regression analysis with independent variables of CL as
state (mind-set contrast coded: abstract [ 1] versus concrete
[1]), CL as trait (BIF, z-standardized), and the product of CL Study 2 builds on study 1 and aims to show that low-state CL
as trait and state were conducted to assess the respective pre- results in less risk-taking by using both a different priming
dictive influence. A significant effect of CL as state on riski- method and a different risk-taking measurement. As in study
ness was evident on the indicators mean clicks, B = 2.10, t 1, participants worked on a risk-taking task after CL priming.
(54) = 2.40, SE = 0.87, p = .019; bursts, B = 1.44, t
(54) = 2.37, SE = 0.60, p = .021; and the risk index, Method
B = 0.28, t(54) = 2.61, SE = 0.11, p = .012, indicating that
abstractly primed participants took more risks than those In study 2, the priming task was a combination of a process
who adopted a concrete mind-set (see Table 1 for means). versus goal simulation task (on the basis of Taylor, Pham,
These results support hypothesis 1, that is, a higher CL leads Rivkin, & Armor, 1998) and a “why/how” priming
to more risk-taking behavior than a lower CL. (adapted from Freitas, Gollwitzer, & Trope, 2004). Partici-
Construal level as trait only showed a significant influence pants were either asked to think about the process (concrete)
on riskiness on the indicator mean clicks, B = 2.05, t or goal (abstract) of an action (i.e., process: preparing for an
(54) = 2.32, SE = 0.88, p = .024, indicating that more abstract exam versus goal: achieving a good university degree) and to
thinking as a trait leads to more riskiness. However, there write down the thoughts that come to mind. Participants sub-
was no significant effect of CL as trait on the other two indica- sequently answered six how (concrete) or why (abstract) ques-
tors, ps > .122. Furthermore, there was no significant interac- tions concerning different activities (e.g., taking part in a
tion of CL as trait and state on riskiness, ps > .390. Thus, we competition). Both procedures (simulation task and, in partic-
conclude that there might be an effect of CL as trait on riski- ular, the why or how task) are established tasks to manipulate
ness, but this effect should be investigated with a larger sample CL (see Burgoon et al., 2013 for a review). Furthermore, the
size in order to obtain convincing evidence. priming was pretested (N = 20). In the pretest, participants
Interestingly, there were no significant differences in the to- were randomly assigned to either concrete or abstract priming
tal amount of money that participants accumulated during the and then asked to give probability judgments of the six best
task (concrete priming: $11.33, SD = 4.61; abstract: $12.00, differentiating items of Wakslak and Trope’s (2009) studies

Copyright © 2014 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. (2014)
 Construal level and risk

1 and 3. Participants rated the probability of occurrences of STUDY 3

fictitious events (i.e., “Scott is deciding whether or not to
join a cooking class. How likely is he to sign up?” p. 54)
In addition to replicating the finding that a concrete mind-set
on a 7-step rating scale (1 = very unlikely to 7 = very likely).
leads to less risk-taking behavior than an abstract mind-set,
In line with previous research, results showed that partici-
we aimed to explore the psychological mechanisms responsi-
pants’ probability judgments within the concrete condition
ble. Sagristano et al. (2002) have shown that temporal dis-
(M = 4.45; SD = 0.49) were significantly higher than those
tance influenced participants’ preferences for outcomes and
of the abstract group (M = 3.84; SD = 0.64), t(18) = 2.275,
probabilities regarding risky choices. With the bidirectional
p = .035, d = 1.08. Results indicated that this priming,
character of CLT in mind, we explored risk-taking by differ-
which is different from the priming used in our study 1,
ent mind-sets. In particular, whether abstract thinking
also led to the adoption of different CL mind-sets.
increases the focus on payoffs (i.e., the desirability of the
Forty university students (26 female; age M = 23.02; SD = 3.03)
outcome of games of chance—amount of gained money)
were randomly assigned to either the abstract or concrete priming
and decreases the focus on probabilities (i.e., feasibility of
group. After priming, participants completed 30 rounds of the
attaining the payoff—likeliness to win) compared with
Angling Risk Task (ART, adapted from Pleskac, 2004). The
concrete thinking. As in the previous studies, participants
ART is a more recent measure of risk propensity and conforms
worked on a risk-taking task after CL priming.
to the general scheme of the BART. The aim of the ART is to
win money by fishing in a virtual pond on a computer screen by
clicking the go fish button (each click catches one fish). In each Method
round, there are 64 fish in the pond: 63 good ones yielding 5¢ each
and one bad fish leading (cf., balloon bursting) to a loss of the Sixty-two university students (32 female; age M = 24.72;
accumulated money in that round. Participants could neither SD = 6.04) were randomly assigned to either the abstract
see the fish nor know the probability of catching a bad fish or concrete priming group. As a priming task, participants
(cloudy condition). Furthermore, they had the opportunity to stop answered the six how (concrete condition) or why (abstract
each round and save the already gained money by clicking the condition) questions of study 2. After priming, participants
collect button. Participants received course credit. completed 20 rounds of the BART. Afterwards, participants
answered short questions about potential mediators. In or-
der to explore whether CL manipulation influences the fo-
Results and Discussion cus on outcomes and probabilities, participants were then
asked the following four self-formulated questions about
As in study 1, two risk behavior indicators were calculated to their focus during the BART task: how much did you focus
test whether a concrete mind-set leads to more risk aversion on the following: (i) gains; (ii) losses; (iii) winning proba-
than an abstract mind-set. The mean number of clicks within bility; and (iv) loosing probability (e.g., during the game I
rounds in which no bad fish were caught was taken as the first predominantly focused on gains). The answer format was a
indicator of willingness to take risk. The second was the total rating scale (ranging from 1 = no agreement to 7 = full
number of bad fish caught. Furthermore, a risk index of the z- agreement).
standardized means of the indicators was calculated as the two The BART screen normally shows an information box,
measures were highly correlated, r = .751, p < .001. which displays the sum of gained money, rounds played,
Regression results showed that participants with an abstract and, if applicable, number of clicks since the last balloon
mind-set took higher risks. This was indicated by a significant burst. To investigate whether the CL effect is evident or
difference in the indicators mean clicks, B = 4.88, t even enhanced when this information is not provided, the
(38) = 2.36, SE = 2.06, p = .023; bad fish, B = 2.55, t info box on the screen was covered by a sheet of paper.
(38) = 2.37, SE = 1.07, p = .023; and the risk index, The expectation was that participants would focus more
B = 0.70, t(38) = 2.55, SE = 0.27, p = .015 (see Table 1 for on their estimated likelihood (the risk) that a balloon will
means). As in study 1, the results of study 2 confirmed burst, rather than assessing the burst risk probability from
hypothesis 1, that is, low-level construals (concrete mind-set) previous rounds. Without the info box, no descriptive in-
lead to more risk aversion compared with high-level construals formation on past events was available except in partici-
(abstract mind-set). pants’ memory.
The difference in gains was marginally significant between
the concrete group (M = $17.57, SD = 3.98) and the abstract
group (M = $19.90, SD = 4.36), t(38) = 1.76, p = .086, Results and Discussion
d = 0.56. As with the BART (study 1), it is inherent to ART
that different types of risk-taking behavior can lead to similar As in study 1, the indicators mean clicks and number of bursts
gains. Although results from study 2 indicate that a higher were used to calculate a risk index (correlation of z-standardized
risk-taking approach within the ART is more advantageous, indicators was sufficient; r = .656, p < .001). Regression results
this is not necessarily the case, because taking higher risks showed that abstract thinkers took more risks than concrete
can in turn also lead to higher losses. However, as in study thinkers (marginally significant on the indicator mean clicks,
1, study 2 did not focus on the total gains but rather on the B = 3.05, t(60) = 1.93, SE = 1.57, p = .057, and significantly
risk-taking behavior. Thus, in the following studies, the on the indicators bursts, B = 1.48, t(60) = 2.02, SE = 0.73,
amount of gained money was not of interest. p = .047, and the risk index, B = 0.49, t(60) = 2.19, SE = 0.22,

Copyright © 2014 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. (2014)
Eva Lermer et al.

Table 2. Construal level (state) effect on risk-taking and focus on detail. Overall, the question as to which psychological mecha-
payoff and probability (study 3) nisms mediate the CL effect on risk-taking remains unclear.
Mind-set

Concrete Abstract
(n = 31) (n = 31) STUDY 4

M SD M SD
An additional study was conducted to investigate the mediat-
Risk-taking ing mechanisms responsible for the CL effect on risk-taking.
Mean pumps 12.57 6.65 15.63 5.72
Bursted balloons 7.35 3.20 8.83 2.51 In this study, the participants’ promotion and prevention foci
Risk index 0.24 0.99 0.24 0.75 and participants’ preferences for explicit safe or risky game
Attention to strategies were examined as potential mediators. As in stud-
Gains 4.48 1.84 4.80 1.37 ies 1 and 3, participants worked on the BART after receiving
Winning probability 5.00 1.36 4.93 1.65 CL priming.
Losses 3.45 1.85 3.38 1.60
Loosing probability 3.67 1.73 4.09 1.79
Method
Note: Concrete = low CL mind-set priming group; abstract = high CL mind-set
group.
Fifty-seven university students (37 female; age M = 23.05;
SD = 3.43) were randomly assigned to either the abstract or
concrete priming group. Participants received course credit
p = .032; see Table 2 for means). Thus, as in studies 1 and 2, hy- for participation. Procedure, CL manipulation, and assessment
pothesis 1 was confirmed. of potential mediators were the same as in study 3, except that
In relation to potential mediating mechanisms, regression re- the info box of the BART was visible again. Participants’ pre-
sults showed no significant CL effect on the focus on gains, vention and promotion strategies were assessed with the regu-
losses, and probability of winning or losing, ps > .354 latory focus questionnaire (adopted from Higgins et al., 2001).
(Table 2). However, participants paying more attention to gains Participants’ preference for safe or risky game strategies were
(z-standardized) generally showed more risk affinity (mean assessed with two items: strategy A: few pumps, constant win-
clicks, B = 1.92, t(60) = 2.46, SE = 0.78, p = .017; bursts, nings of small amounts and little losses, and strategy B: many
B = 0.80, t(60) = 2.17, SE = 0.36, p = .033; and risk index, pumps, high gains but more losses. Participants answered how
B = 0.28, t(60) = 2.57, SE = 0.11, p = .013). Thus, we cannot con- much they applied each strategy on a seven-step rating scale
clude that attention to payoffs or probabilities mediate the CL (1 = not at all to 7 = totally).
effect.
Covering or displaying the info box seemed to have no im-
pact (i.e., ANOVA results of z-standardized mean clicks of stud- Results and Discussion
ies 1 and 3 showed no significant difference, p = .635). The
reason for covering the box was that without any descriptive in- As in studies 1 and 3, the indicators mean clicks and number
formation on previous rounds, participants would have to rely on of bursts were used to calculate a risk index (correlation of
their own estimates of burst probabilities. However, this did not z-standardized indicators was sufficient; r = .727, p < .001) as
seem to be the case. Findings of study 3 showed that abstract an indication of risk propensity. Regression results showed
(versus concrete) thinking and paying more attention to winning that abstract thinkers took more risks than concrete thinkers,
increased risk-taking. on the indicator mean clicks, B = 5.01, t(55) = 2.63, SE = 1.90,
Study 3 did not replicate the findings of Sagristano et al. p = .011, and the risk index, B = 0.50, t(55) = 2.09, SE = 0.23,
(2002) who showed that high CL (high psychological dis- p = .041, but not on the indicator bursts, B = 0.82, t
tance understood as abstract thinking) increases attention to (55) = 1.28, SE = 0.64, p = .205 (Table 3). Thus, hypothesis 1
payoffs and decreases attention to probabilities. One explana- was essentially confirmed.
tion could be the use of different CL manipulations: We used Furthermore, results showed that concrete thinkers scored
a more direct manipulation via priming, whereas Sagristano higher on the prevention focus (z-standardized) compared with
et al. used a more indirect method (i.e., temporal framing). abstract thinkers, B = 0.55, t(55) = 2.15, SE = 0.25, p = .036
Another possible explanation is the potential differences be- (Table 3). However, results of linear regressions showed that
tween real versus intended behavior and past versus future. the prevention focus did not influence risk-taking behavior
We assessed participants’ foci after they had completed a real (mean clicks, bursts, and risk index; all ps > .466). Moreover,
task, and they commented on their own past behavior. there was no difference on the promotion focus by different
Sagristano et al. measured participants’ foci for intended fu- mind-sets, and no significant effect of promotion focus on riski-
ture behavior. In the present research, we used self-reports ness; all ps > .117.
as a measure of participants’ sensitivity to payoffs and prob- To explore the effect of game strategy on risk-taking, the dif-
abilities. Perhaps participants did not have introspective ac- ference value between strategy A and strategy B was calculated
cess to their level of information processing. To summarize, (i.e., strategy A minus strategy B, z-standardized). Regression re-
it remains speculative as to whether, and indeed why, the dif- sults showed a significant difference between CL groups on
ferent methods were responsible for the different outcomes. mean strategy difference, B = 1.01, t(55) = 2.11, SE = 0.48,
Future research is warranted to explore these effects in more p = .039 (Table 3). The difference was significantly higher for

Copyright © 2014 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. (2014)
 Construal level and risk

Table 3. Construal level (state) effect on risk-taking, prevention and Table 4. Regression results for construal level (CL) and strategy A
promotion focus, and strategy in study 4 and on risk-taking, strategy, minus B on risk-taking (studies 4 and 5)
mood, focus on pros and cons, and attention to payoffs and
probabilities in study 5 B t(df) SE p

Mind-set Study 4
CL (IV) → mean clicks (DV) 5.01 2.63 (55) 1.90 .011
Concrete Abstract CL (IV) → bursts (DV) 0.82 1.28 (55) 0.64 .205
(study 4: n = 28) (study 4: n = 29) CL (IV) → risk index (DV) 0.50 2.09 (55) 0.23 .041
(study 5: n = 30) (study 5: n = 30) CL (IV) → SA-SB (mediator) 1.01 2.11 (55) 0.48 .039
CL and SA-SB → mean clicks
M SD M SD CL 3.27 1.82 (54) 1.80 .074
SA-SB 1.71 3.52 (54) 0.48 .001
Study 4 CL and SA-SB → risk index
Risk-taking CL 0.21 1.04 (54) 0.20 .299
Mean pumps 10.67 4.83 15.68 8.86 SA-SB 0.28 4.98 (54) 0.05 .000
Bursted balloons 7.00 2.21 7.82 2.63 Study 5
Risk index 0.25 0.72 0.24 1.04 CL (IV) → mean clicks (DV) 3.25 2.04 (58) 1.95 .046
Promotion focus 3.68 0.61 3.67 0.057 CL (IV) → bursts (DV) 1.33 1.48 (58) 0.90 .144
Prevention focus 3.88 0.52 3.50 0.78 CL (IV) → risk index (DV) 0.44 1.85 (58) 0.24 .068
Strategy difference 0.25 3.26 1.58 3.30 CL (IV) → SA-SB (mediator) 1.00 2.11 (56) .0.47 .039
(z-standardized) CL and SA-SB → mean clicks
Study 5 CL 1.36 0.93 (55) 1.46 .355
Risk-taking SA-SB 1.81 4.58 (55) 0.39 .000
Mean pumps 12.61 4.15 15.86 7.66 CL and SA-SB → risk index
Bursted balloons 12.00 3.18 13.33 3.76 CL 0.13 0.61 (55) 0.21 .539
Risk index 0.22 0.72 0.22 1.09 SA-SB 0.29 5.05 (55) 0.05 .000
Strategy difference 0.53 1.89 0.47 1.72
(z-standardized) Note: CL = construal level (abstract versus concrete); SA-SB = strategy differ-
ence (z-standardized); IV = dependent variable, DV = dependent variable. Me-
Mood diator analyses were only calculated for the variables mean clicks and risk
Positive affect 3.00 0.74 3.13 0.65 index because CL had no significant impact on bursts.
Negative affect 1.33 0.60 1.08 0.14
Attention to pros 4.50 1.54 5.37 1.12
Attention to cons 2.37 1.27 2.53 1.47 only within their behavioral data but also in the self-reports.
Attention to Thus, we conclude that the CL effect might be consciously ac-
Gains 5.17 1.39 5.47 1.07 cessible. This is of great importance to the practical value of
Winning probability 4.93 1.76 5.03 1.42
Losses 3.23 1.54 2.80 1.47 these research findings. The ability of people to be aware of
Loosing probability 3.53 1.65 3.47 1.97 their risk strategy may prove useful in increasing safety. We
discuss this further in the Section on General Discussion.
Note: Concrete = low CL mind-set priming group; abstract = high CL mind-set
group. In summary, study 4 confirmed the effect of CL on risk-
taking behavior by showing that concrete thinkers were more
risk averse than abstract thinkers. Even more important, results
concrete thinkers (indicating less risk preference) than for showed that game strategy can be seen as a kind of mediating
abstract thinkers.1 mechanism for the CL effect on risk-taking. An abstract mind-
A mediation analysis (Baron & Kenny, 1986) was con- set favored a risky game strategy, whereas a concrete mind-set
ducted to test whether the mean strategy difference carries favored a safe strategy. Results further showed that preference
the effect from CL (IV) to risk-taking (DV). Without a poten- for a safe strategy reduced risk-taking, whereas preference for
tial mediator, game strategy regression results showed that the a risky strategy increased risk-taking. Thus, it is proposed that
independent variable CL was significant on mean clicks and CL triggers different game patterns leading to differences in
the risk index but did not influence bursts (Table 4). handling risky situations.
When controlling for the mediator mean strategy differ- Furthermore, results indicated that a concrete (versus ab-
ence, the CL effect on mean clicks was reduced to marginal stract) mind-set led to more preventive strategies. Results for
significance, whereas the regression weight for the mediator both the promotion and prevention foci, however, failed to ex-
mean strategy difference remained significant. Furthermore, plain why concrete thinkers were less risk seeking than ab-
CL no longer predicted the risk index, but the regression stract thinkers. However, results indicated that although
weight of mean strategy difference was still significant. Thus, concrete thinkers paid more attention to prevention than ab-
the variable mean strategy difference worked as a mediator for stract thinkers, they did not differ on promotion focus.
the CL effect on risk-taking behavior. We are well aware that
considering game strategy as a kind of mediator could be
criticized as the used measure was essentially participants’
STUDY 5
self-reports of the behavior they engaged in. Nonetheless, it
is interesting that participants also showed differences, not
Another look at participants’ game strategies is warranted as
1 both study 4 and previous research (Förster & Higgins,
Because separate analysis for strategy A and strategy B showed basically the
same effect, we decided to only report the effect of the difference value of 2005) indicated that a concrete mind-set is linked to preventive
strategy A minus strategy B. (safe) strategies. In study 5, participants’ preferences for

Copyright © 2014 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. (2014)
Eva Lermer et al.

explicit safe versus risky game strategies were explored again. risks than concrete thinkers. However, this effect was only
Furthermore, participants’ foci on pros and cons of outcomes significant for mean clicks, B = 3.25, t(58) = 2.04, SE = 1.59,
were also questioned in an attempt to identify mediating p = .046, and marginally significant for the risk index,
mechanisms of the CL effect. Eyal et al. (2004) showed that B = 0.44, t(58) = 1.85, SE = 0.24, p = .068. The bursts indica-
psychological distance (i.e., temporal distance) is linked to a tor showed no significant difference, B = 1.33, t(58) = 1.48,
focus on pros (of an action) over cons. Accordingly, abstract SE = 0.90, p = .144 (see Table 3 for means). Thus, hypothesis
thinkers should focus more on the positive aspects of a risky 1 was essentially confirmed.
task than concrete thinkers do. As in study 4, a mediation analysis (Baron & Kenny, 1986)
Additionally, the impact of CL on participants’ mood was was calculated to test whether mean strategy difference medi-
assessed as another potential mediator. Gasper and Clore ated between CL (IV) and risk-taking (DV). As in study 4, re-
(2002) showed that a positive (versus sad) mood is linked to gression results showed that game strategy mediated the CL
a global (versus local) processing style (which is assumed to effect on risk-taking (Table 4).2 Next, we investigated whether
be equivalent to an abstract mind-set). Furthermore, Yuen a focus on pros or cons mediated the CL effect on risk-taking.
and Lee (2003) showed that a negative mood can be linked Regression results showed that abstract thinkers focused more
to less riskiness. Therefore, the effect of CL manipulation on on pros (i.e., positive aspects of the game) than concrete
mood changes and, in turn, risk-taking behavior was also ex- thinkers, B = 0.86, t(58) = 2.47, SE = 0.35, p = .016. However,
plored. Furthermore, as the results of study 3 do not agree with further regression analyses revealed that a focus on pros did
the findings of Sagristano et al. (2002), the effect of CL manip- not predict risk-taking propensity (mean clicks: B = 0.66, t
ulation on participants’ foci on outcomes versus probabilities (58) = 1.46, SE = 0.58, p = .256; risk index: B = 0.08, t
was readdressed with the aim of replicating (or otherwise) (58) = 0.99, SE = 0.08, p = .323). The participants’ focus on
the results of study 3. In doing so, we hoped to shed some light cons (i.e., negative aspects of the game) indicated no differ-
on this issue. ence between abstract and concrete thinkers. There was no re-
In order to explore the outlined questions, participants’ lationship between a focus on cons and risk-taking; all
mind-sets were primed before carrying out the BART. After- ps > .526. Thus, we could find no support for the assumption
wards, they were asked to what extent they pursued a safe or that attention to pros or cons mediates the CL effect.
risky game strategy during the task. In addition, their focus In further search of mediating mechanisms, we also ex-
on positive and negative aspects of the game, their current plored the influence of participants’ mood. Regression results
mood, and their focus on payoffs (i.e., gains or losses) and showed no CL impact on the PA, p > .472. However, abstract
on probabilities (i.e., winning or losing) was assessed. Again thinkers scored significantly less on the NA dimension
participants received course credit for participation. than concrete thinkers, B = 0.25, t(58) = 2.24, SE = 0.11,
p = .029, but neither NA nor PA predicted risk-taking (for
Method mean clicks and risk index, all ps > .430). Thus, the assump-
tion as to whether mood is a mediator of the CL effect could
Sixty university students (52 female; age M = 22.75, SD = 5.94) not be supported.
were randomly assigned to either the abstract or concrete prim- Finally, we reexamined whether focus on payoffs (i.e., gains
ing group. The same priming task as in study 4 served as a CL or losses) and probabilities (i.e., loosing or winning) mediated
manipulation. For measuring risk-taking behavior, participants the CL effect on risk-taking behavior. Regression results
played 30 rounds of the BART with an uncovered info box. showed no effect of CL on attention to gains, winning
Participants’ preference for safe and risky game strategies was probability, losses, and loosing probability, ps > .271. Thus,
measured as in study 4. Participants were also asked how much again, the findings of Sagristano et al. (2002) that temporal dis-
they agreed with the following two statements: (i) I predomi- tance is associated with intended risk behavior could not be rep-
nantly focused on positive aspects during the game and (ii) I licated in the context of mind-set manipulation on risk-taking
predominantly focused on negative aspects during the game, behavior.
so as to assess the influence of CL on focusing on pros over In summary, study 5 supported the findings of studies 1–4 in
cons. The answers format was a rating scale (from 1 = no agree- that CL affects risk-taking behavior. Abstract thinkers took
ment to 7 = full agreement). As a measure for mood, participants more risks than concrete thinkers. Furthermore, study 5 repli-
then answered the German version of the positive affect nega- cated the findings of study 4 in that game strategy mediates
tive affect schedule (Grühn, Kotter-Grühn, & Röcke, 2010). the CL effect on risk-taking behavior. Abstract thinkers pursued
The answer format was a rating scale (1 = very little or no inten- a more risky game strategy, whereas concrete thinkers preferred
sity to 5 = very strong). Finally, participants answered the a safe game strategy. These strategies resulted in either more
questions concerning focus on outcomes and probabilities used (abstract thinkers) or less (concrete thinkers) risky behavior.
in study 3. The interesting point here is that the CL manipulation not only
influenced risk-taking but participants were also aware of the
Results and Discussion kind of strategy they pursued. These results also revealed that
abstract thinkers focused more on the pros of an action than
Again, the mean clicks and number of bursts indicators were concrete thinkers. These findings are in line with results from
used as dependent variables for risk-taking to calculate a risk
2
index (correlation of z-standardized indicators was sufficient; As in study 4, separate analysis of strategy A and strategy B showed ba-
sically the same (mediating) effect. Therefore, we again decided to only
r = .803, p < .001) as a further indicator for risk propensity. report the mean difference value. Both results are provided in the supple-
Regression results indicated that abstract thinkers took more mentary material.

Copyright © 2014 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. (2014)
 Construal level and risk

Eyal et al. (2004) who showed that psychological distance is mediated the effect of CL on riskiness, as abstract thinkers
linked to a focus on pros over cons (i.e., the more distant future had a higher preference for risky strategy and a lower
is associated with pros). However, there was no CL effect on preference for safe strategy than concrete thinkers. Risky
participants’ focus on the cons of the game. One explanation strategy was linked to more risk-taking and the safe strategy
for this might lie in the operationalization of assessing partici- to less risk-taking. Finally, the present research indicates that
pants’ foci on pros and cons. In contrast to Eyal et al. (2004), CL considered as a trait (study 1) has an effect on risk-taking.
we assessed participants’ foci with two proximal questions However, the influence of CL as a trait could only be found
concerning their explicit attention to positive and negative as- on one out of three risk-taking indicators and should be
pects. Perhaps there is a ceiling effect for the focus on cons be- further researched with larger sample sizes in order to make
cause of social desirability. However, this is speculative and empirically satisfactory conclusions.
should be clarified in future research. This research adds some important theoretical implica-
Furthermore, participants’ focus on pros had no effect on tions for CLT. First, it was shown that CLT is an important
risk-taking. Accordingly, a focus on pros did not mediate the perspective in studying risk-taking. This is significant as
CL effect on risk-taking. In relation to participants’ mood, re- CLT is not traditionally considered in risk research. Al-
sults showed that there was no CL effect on PA, but abstract though risk has become a very popular area of research,
thinkers showed a less NA than concrete thinkers. Gasper the reasons why people show inconsistencies in their han-
and Clore (2002) found that participants in a sad mood were dling of risk are not yet clear. It is, however, generally
less likely to focus on abstract targets and more likely to focus agreed that “different whos react differently to different
on concrete targets compared with participants in a happy whens” (Figner & Weber, 2011, p. 211). Consequently, a
mood. Rim et al. (2009) found no link between mood and particular focus on people’s mind-sets is essential for a bet-
CL. Our results add to these ambiguous findings by showing ter understanding of how people handle risk. Thus, we be-
a relationship between CL and mood by using an inverse ap- lieve that consideration of CL in future research would
proach (i.e., influence of CL on mood) to that used by Gasper make a decisive contribution to risk research and, in doing
and Clore (2002). Interestingly, however, mood had no effect so, enhance existing insights. However, a better under-
on riskiness and thus did not mediate the effect of CL on standing of the mediating mechanisms responsible for CL
risk-taking. effects is also necessary. We present this research as a con-
tribution to this enterprise.
Second, our results give further empirical support to the
GENERAL DISCUSSION
reciprocal (bidirectional) character of CLT. CL effects are
not limited to variation of psychological distance (i.e., tempo-
The present research (studies 1–5) reveals a consistent pat- ral distance). They can also be found after a mind-set manip-
tern of the influence of the mind-set (abstract versus con- ulation and on the behavioral level. Our research shows that
crete) on risk-taking behavior. All studies showed that mind-set manipulation influences people’s risk propensity.
concrete thinking decreases risk propensity compared with However, further research is needed to combine present and
abstract thinking. Although, we could not directly replicate previous research findings into an integrated framework for
the findings of Sagristano et al. (2002), we could show that a better understanding of risk-taking. For instance, future re-
the bidirectional character of CLT also holds for risk-taking search should investigate whether the bidirectionality of CLT
behavior. Sagristano et al. (2002) have shown that, by ma- also holds for risk estimation, in particular, whether the CL
nipulating time frames, low-level construals lead to less effect on real risk-taking is mediated by changes in risk
riskiness compared with high-level construals. We ad- perception.
dressed this issue by manipulating participants’ mind-sets Furthermore, some practical implications emerge from our
(i.e., abstract versus concrete) and could consistently show research. The results support the suggestion that a concrete
that a concrete mind-set leads to less riskiness than an ab- thinking leads to more risk aversion relative to abstract think-
stract mind-set. ing. Therefore, it is assumed that CL is a very important aspect
However, despite replicating some findings from previ- in safety considerations. In line with Chandler and Pronin
ous research, most of our attempts to discover applicable (2012), who showed that thought speed influences risk-taking,
mediating mechanisms failed. First, our CL manipulation we propose that interventions designed to boost concrete
did not influence participants’ foci on payoffs and probabil- thinking would decrease people’s tendency to put themselves
ities (studies 3 and 5). Second, results showed that abstract at risk. These findings emphasize the trainability of risk behav-
thinkers scored lower on the prevention focus dimension ior skills, especially in contexts where less riskiness is more
than concrete thinkers (study 4), but there was no differ- advantageous. This is, for example, of particular relevance to
ence in the promotion focus between CL groups. Third, motorcyclists, and road traffic in general; sports (e.g., moun-
our results showed that abstract thinkers focused more on taineering); or decision-making in private and business con-
pros (of an action) than concrete thinkers (study 5), and texts (e.g., taking out insurance). These findings could, for
there was no difference in the focus on cons between the instance, be incorporated into road safety trainings with the
CL groups. Fourth, concrete thinkers showed higher values aim of increasing safety by teaching road users how they can
on the NA dimension than abstract thinkers (study 5), but activate a concrete mind-set while driving. It would also be
there was no difference on the PA dimension between useful in high-risk sports to know whether one can increase
abstract and concrete thinkers. Fifth, only game strategy safety by activating a concrete mind-set. As a concrete exam-
(measured as a self-report of the participants’ behavior) ple, mountain guides could be given a guideline prompting

Copyright © 2014 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. (2014)
Eva Lermer et al.

them to carefully reconsider their situation in detail at certain less risk-taking is favorable or desired, then a concrete mind-
intervals during the trip. set should be adopted. In contrast, an abstract mind-set is fa-
vored for more risk-taking. Given that different game strategies
Limitations (safe versus risky) mediated the CL effect on risk-taking, fur-
ther research on cognitive influences could help understand
Some limitations of the study have to be mentioned. First, and improve our behavior in risky situations.
we want to emphasize that our results of the CL effect on
risk-taking behavior are limited to the domain of gains
(versus no gains). Although the risk measures used in the
present research correlate positively with risky behavior in REFERENCES
real life (Hunt et al., 2005), it has to be noted that both
the BART and the ART are games in the domain of gains Bar-Anan, Y., Liberman, N., & Trope, Y. (2006). The association between
(because participants can only loose what they have previ- psychological distance and construal level: Evidence from an implicit asso-
ously won). Second, we did not succeed in defining mech- ciation test. Journal of Experimental Psychology: General, 135, 609–622.
doi: 10.1037/0096-3445.135.4.609
anisms to explain the CL effect on risk-taking. The found Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable dis-
mediator game strategy is not ideal because our measure tinction in social psychological research: Conceptual, strategic and statisti-
was solely a self-report of the behavior participants had en- cal considerations. Journal of Personality and Social Psychology, 51,
1173–1182. doi: 10.1037/0022-3514.51.6.1173
gaged in. It is interesting that participants are consciously Beckmann, J. (1993). Verhaltensidentifikations-Fragebogen. Eine deutsche
aware of their chosen risk strategy, but game strategy is Übersetzung der Behavioral Identification Form von Vallacher und Wegner
no real mediator in the proper sense. Third, we could not [Behavior-identification-questionnaire. A German translation of the Behav-
ioral Identification Form by Vallacher and Wegner]. Unpublished manu-
replicate the findings of Sagristano et al. (2002) regarding script. Max Planck Institute for Psychological Research, Munich, Germany.
participants’ focus on payoffs and probabilities. This may Bell, N. J., Schoenrock, C. J., & O’Neal, K. K. (2000). Self-monitoring
well be down to the fact that we used a different measure and the propensity for risk. European Journal of Personality, 14,
107–119. doi: 10.1002/(SICI)1099-0984(200003/04)14:2<107::AID-
(i.e., self-reports) and that participants had not introspective PER365>3.0.CO;2-T
access to their level of information processing. Fourth, for Blais, A. R., & Weber, E. U. (2006). A domain-specific risk-taking
future research (especially regarding investigation of CL (DOSPERT) scale for adult populations. Judgment and Decision Making,
1, 33–47. doi: 10.1037/t13084-000
as trait), larger sample sizes are recommended in order to Bryant, P., & Dunford, R. (2008). The influence of regulatory focus on
increase study power. risky decision-making. Applied Psychology, 57, 335–359. doi: 10.1111/
Although we could only show that game strategy medi- j.1464-0597.2007.00319.x
Burgoon, E. M., Henderson, M. D., & Markman, A. B. (2013). There are
ated the CL effect, we suggest that CL influences various many ways to see the forest for the trees: A tour guide for abstraction. Per-
variables, which separately or in combination result in spectives on Psychological Science, 8(5), 501–520. doi: 10.1177/
more willingness to take risks (e.g., the interaction of more 1745691613497964
Chandler, J. J., & Pronin, E. (2012). Fast thought speed induces risk taking.
negative mood, higher need for prevention, and less focus Psychological Science, 23, 370–374. doi: 10.1177/0956797611431464
on pros of an action for concrete thinkers than for abstract). Crowe, E., & Higgins, E. T. (1997). Regulatory focus and strategic inclina-
This presents another promising topic for future risk re- tions: Promotion and prevention in decision-making. Organizational Be-
havior and Human Decision Processes, 69, 117–132. doi: 10.1006/
search, viz investigation of influences such as consideration obhd.1996.2675
of future consequences (Strathman, Gleicher, Boninger, & De Vries, M., Holland, R. W., & Witteman, C. L. M. (2008). In the winning
Edwards, 1994), or learned carelessness (Frey & Schulz- mood: Affect in the Iowa gambling task. Judgment and Decision Making,
Hardt, 1996), which appear to be promising constructs for 3, 42–50.
Ditto, P. H., Pizarro, D. A., Epstein, E. B., Jacobson, J. A., & MacDonald, T.
a better understanding of how people handle risk. It is K. (2006). Visceral influences on risk-taking behavior. Behavioral Decision
speculated that different mind-sets are related to, for in- Making, 19, 99–113. doi: 10.1002/bdm.520
stance, different levels of consideration of future conse- Eyal, T., Liberman, N., Trope, Y., & Walther, E. (2004). The pros and cons of
temporally near and distant action. Journal of Personality and Social Psy-
quences and thus, influence risk propensity. chology, 86, 781–795. doi: 10.1037/0022-3514.86.6.781
Furthermore, it would also be interesting to explore which Eysenck, S. B., & Eysenck, H. J. (1977). The place of impulsiveness in a di-
daily activities or situational variables activate concrete or ab- mensional system of personality description. British Journal of Social and
Clinical Psychology, 16, 57–68. doi: 10.1111/j.2044-8260.1977.tb01003.x
stract mind-sets. Moreover, although the chosen instruments Figner, B., & Weber, E. U. (2011). Who takes risks when and why? Determi-
(i.e., BART and ART) seem to be reliable for measuring nants of risk taking. Current Directions in Psychological Science, 20,
risk-taking propensity, the CL effect should also be investi- 211–216. doi: 10.1177/0963721411415790
Förster, J., Friedman, R. S., & Liberman, N. (2004). Temporal construal ef-
gated in real-life settings. Another argument for real-life set- fects on abstract and concrete thinking: consequences for insight and crea-
tings in future studies is the desire to assess the mediators tive cognition. Journal of Personality and Social Psychology, 87, 177. doi:
between the priming task and participants’ behavior. We 10.1037/0022-3514.87.2.177
Förster, J., & Higgins, T. (2005). How global versus local perception fits
refrained from this favored method because of our specific ex- regulatory focus. Psychological Science, 16, 631–636. doi: 10.1111/
perimental setting, in that participants were not familiar with j.1467-9280.2005.01586.x
the risk task (i.e., BART and ART). It would be interesting Freeman, N., & Muraven, M. (2010). Self-control depletion leads to increased
to investigate our findings with a natural target group that is risk taking. Social Psychological and Personality Science, 1, 175–181. doi:
10.1177/1948550609360421
confronted with real risks (e.g., with drivers during road safety Freitas, A. L., Gollwitzer, P., & Trope, Y. (2004). The influence of abstract
training). and concrete mindsets on anticipating and guiding others’ self-regulation
In conclusion, our results shed some light on the ques- efforts. Journal of Experimental Social Psychology, 40, 739–752. doi:
10.1016/j.jesp.2004.04.003
tion as to why people differ in risk-taking. Apparently, dif- Frey, D., & Schulz-Hardt, S. (1996). Eine Theorie der gelernten Sorglosigkeit.
ferences in mind-set can influence risk-taking behavior. If [A theory of learned carelessness]. In H. Mandel (Ed.), Bericht über den 40.

Copyright © 2014 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. (2014)
 Construal level and risk

Kongress der Deutschen Gesellschaft für Psychologie (pp. 604–611). Luguri, J. B., Napier, J. L., & Dovidio, J. F. (2012). Reconstruing intolerance:
Göttingen: Hogrefe. Abstract thinking reduces conservatives prejudice against nonnormative
Fujita, K., Trope, Y., Liberman, N., & Levin-Sagi, M. (2006). Construal levels groups. Psychological Science, 23, 756–763. doi: 10.1177/
and self-control. Journal of Personality and Social Psychology, 90, 0956797611433877
351–367. doi: 10.1037/0022-3514.90.3.351 Pennington, G. L., & Roese, N. J. (2003). Regulatory focus and temporal dis-
Gasper, K., & Clore, G. L. (2002). Attending to the big picture: Mood and tance. Journal of Experimental Social Psychology, 39, 563–576. doi:
global versus local processing of visual information. Psychological Science, 10.1016/S0022-1031(03)00058-1
13, 34–40. doi: 10.1111/1467-9280.00406 Pleskac, T. J. (2004). Evaluating cognitive sequential risk-taking models: Ma-
Grant, H., & Higgins, E. T. (2003). Optimism, promotion pride, and preven- nipulations of the stochastic process. Doctoral dissertation, Graduate School
tion pride as predictors of quality of life. Personality and Social Psychology of the University of Maryland. Retrieved March 18, 2014 from: http://drum.
Bulletin, 29(12), 1521–1532. doi: 10.1177/0146167203256919 lib.umd.edu/bitstream/1903/1854/1/umi-umd-1849.pdf
Grühn, D., Kotter-Grühn, D., & Röcke, C. (2010). Discrete affects across the Rim, S., Uleman, J. S., & Trope, Y. (2009). Spontaneous trait inference and
adult lifespan: Evidence for multidimensionality and multidirectionality of construal level theory: Psychological distance increases nonconscious trait
affective experiences in young, middle-aged and older adults. Journal of thinking. Journal of Experimental Social Psychology, 45, 1088–1097.
Research in Personality, 44, 492–500. doi: 10.1016/j.jrp.2010.06.003 doi: 10.1016/j.jesp.2009.06.015
Hanoch, Y., Johnson, J., & Wilke, A. (2006). Domain specificity in exper- Sagristano, M. D., Trope, Y., & Liberman, N. (2002). Time-dependent gam-
imental measures and participant recruitment: An application to risk- bling: Odds now, money later. Journal of Experimental Psychology: Gen-
taking behavior. Psychological Science, 17, 300–304. doi: 10.1111/ eral, 131, 364–376. doi: 10.1037/0096-3445.131.3.364
j.1467-9280.2006.01702.x Scholer, A. A., Zou, X., Fujita, K., Stroessner, S. J., & Higgins, E. T.
Henderson, M. D., Wakslak, C. J., Fujita, K., & Rohrbach, J. (2011). Con- (2010). When risk seeking becomes a motivational necessity. Journal
strual level theory and spatial distance: Implications for mental representa- of Personality and Social Psychology, 99, 215–231. doi: 10.1037/
tion, judgment, and behavior. Social Psychology, 42, 165–173. doi: a0019715
10.1027/1864-9335/a000060 Spassova, G., & Lee, A. Y. (2008). Self-construal and temporal distance. Ad-
Hiemer, J., & Abele, A. E. (2012). High power = motivation? Low power = sit- vances in Consumer Research, 35, 841–842.
uation? The impact of power, power stability and power motivation on risk- Strathman, A., Gleicher, F., Boninger, D. S., & Edwards, C. S. (1994). The
taking. Personality and Individual Differences, 53, 486–490. doi: 10.1016/ consideration of future consequences: Weighing immediate and distant out-
j.paid.2012.04.008 comes of behavior. Journal of Personality and Social Psychology, 66,
Higgins, E. T. (1997). Beyond pleasure and pain. American Psychologist, 52, 742–752. doi: 10.1037/0022-3514.66.4.742
1280–1300. doi: 10.1037/0003-066X.52.12.1280 Taylor, S. E., Pham, L. B., Rivkin, I. D., & Armor, D. A. (1998). Harnessing
Higgins, E. T., Friedman, R. S., Harlow, R. E., Idson, L. C., Ayduk, O. N., & the imagination: Mental simulation, self-regulation and coping. American
Taylor, A. (2001). Achievement orientations from subjective histories of Psychologist, 53, 429–439. doi: 10.1037/0003-066X.53.4.429
success: Promotion pride versus prevention pride. European Journal of So- Trope, Y., & Liberman, N. (2010). Construal-level theory of psychological
cial Psychology, 31, 3–23. doi: 10.1002/ejsp.27 distance. Psychological Review, 117, 440–463. doi: 10.1037/a0018963
Hunt, M. K., Hopko, D. R., Bare, R., Lejuez, C. W., & Robinson, E. V. Vallacher, R. R., & Wegner, D. M. (1989). Levels of personal agency:
(2005). Construct validity of the Balloon Analogue Risk Task (BART): As- Individual variation of action identification. Journal of Personality
sociations with psychopathy and impulsivity. Assessment, 12, 416–428. and Social Psychology, 57, 660–671. doi: 10.1037/0022-3514.57.4.660
doi: 10.1177/1073191105278740 Wakslak, C., & Trope, Y. (2009). The effect of construal level on subjective
Johnson, J. G., Wilke, A., & Weber, E. U. (2004). Beyond a trait view of risk- probability estimates. Psychological Science, 20, 52–58. doi: 10.1111/
taking: A domain-specific scale measuring risk perceptions, expected bene- j.1467-9280.2008.02250.x
fits, and perceived-risk attitude in German-speaking populations. Polish Werth, L., & Förster, J. (2007). How regulatory focus influences consumer be-
Psychological Bulletin, 35, 153–163. havior. European Journal of Social Psychology, 37, 33–51. doi: 10.1002/
Kahneman, D., & Tversky, A. (1979). Prospect theory: An analysis of decision ejsp.343
under risk. Econometrica, 47, 263–291. doi: 10.2307/1914185 Yuen, K. S. L., & Lee, T. M. C. (2003). Could mood state affect risk-taking
Lee, A. Y., Keller, P. A., & Sternthal, B. (2010). Value from regulatory con- decisions? Journal of Affective Disorders, 75, 11–18. doi: 10.1016/
strual fit: The persuasive impact of fit between consumer goals and message S0165-0327(02)00022-8
concreteness. Journal of Consumer Research, 36(5), 735–747. doi: Zou, X., Scholer, A. A., & Higgins, E. T. (2014). In pursuit of progress:
10.1086/605591 Promotion motivation and risk preference in the domain of gains. Jour-
Lejuez, C. W., Aklin, W. M., Zvolensky, M. J., & Pedulla, C. M. (2003). Eval- nal of Personality and Social Psychology, 106(2), 183. doi: 10.1037/
uating the Balloon Analogue Risk Task (BART) as a predictor of adoles- a0035391
cent real-world risk-taking behaviors. Journal of Adolescence, 26, Zuckerman, M. (2007). Sensation seeking and risky behavior. Washington
475–479. doi: 10.1016/S0140-1971(03)00036-8 DC: American Psychological Association. doi: 10.1037/11555-000
Lejuez, C. W., Read, J. P., Kahler, C. W., Richards, J. B., Ramsey, S. E., Stu- Zuckerman, M., & Kuhlman, D. M. (2000). Personality and risk-taking: Com-
art, G. L., … Brown, R. A. (2002). Evaluation of a behavioral measure of mon biosocial factors. Journal of Personality, 68, 999–1029. doi: 10.1111/
risk taking: The Balloon Analogue Risk Task (BART). Journal of Experi- 1467-6494.00124
mental Psychology: Applied, 8, 75–84. doi: 10.1037/1076-898X.8.2.75
Liberman, N., Sagristano, M. D., & Trope, Y. (2002). The effect of temporal
distance on level of mental construal. Journal of Experimental Social Psy-
chology, 38, 523–534. doi: 10.1016/S0022-1031(02)00013-6 SUPPORTING INFORMATION
Liberman, N., & Trope, Y. (1998). The role of feasibility and desirability con-
siderations in near and distant future decisions: A test of temporal construal
theory. Journal of Personality and Social Psychology, 75, 5–18. doi: Additional supporting information may be found in the online
10.1037/0022-3514.75.1.5 version of this article at the publisher’s web-site.

Copyright © 2014 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. (2014)