Personality and Individual Differences 51 (2011) 112–116

Contents lists available at ScienceDirect

Personality and Individual Differences

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

Individual differences in need for cognition and decision making in the Iowa
Gambling Task
Jason L. Harman ⇑
Ohio University, Athens, OH, USA

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

Article history: Differences in decision making between individuals differing in Need for Cognition (NFC) are examined
Received 17 December 2010 using the Iowa Gambling Task (IGT). Previous work using normative one time decisions suggests that
Received in revised form 14 March 2011 individual low in NFC process gains and losses differently than those high in NFC and are more suscep-
Accepted 15 March 2011
tible to decision biases. The IGT is a popular laboratory task that involves making risky decisions from
Available online 13 April 2011
experience involving both gains and losses. In the first experiment, low NFC participants performed sig-
nificantly worse than the high NFC participants. A second experiment designed to examine the nature of
Keywords:
these differences provides evidence that low NFC participants place more importance on gains as
Need for cognition
Risky decision making
opposed to losses when performing the IGT. Results are discussed in light of previous work suggesting
Iowa Gambling Task that low NFC participants place more importance on losses in mixed outcome decisions.
Ó 2011 Elsevier Ltd. All rights reserved.

1. Introduction 1.1. The Iowa Gambling Task

Many of the most important decisions we make in our life are The IGT is a laboratory based card playing paradigm developed
complex, involving risks, trade-offs between possible gains and to study decision making deficits in impaired clinical populations.
losses, and personal experience. Though a great deal of insight onto Recently the IGT has been made available as a clinical measure de-
human decision making performance has come from the study of signed to support diagnosis of brain dysfunction and to assess clin-
simple gambles and one time decisions, few paradigms have been ically relevant decision-making impairment (Bechara, 2007). In the
developed to capture the complex interaction of factors at work in IGT, participants make repeated choices between four decks of
many real world decision making situations. One exception to this cards. After a card is chosen, an amount of money won and some-
is the Iowa Gambling Task (IGT, Bechara, Damasio, Damasio, & times an amount of money lost is displayed and added to a running
Anderson, 1994), a widely used decision making task originally total. The four decks are designed such that two ‘advantageous’
developed to examine real world decision making deficits in indi- decks produce small constant gains with occasional losses (produc-
viduals with lesions of the ventromedial prefrontal cortex. Since its ing net gains over time) while two ‘disadvantageous’ decks pro-
inception, the IGT has been used to discriminate healthy control duce large consistent gains but even larger sporadic losses
participants from multiple populations that display poor decision (producing net losses over time; see Table 1 for all deck contingen-
making in the real world such as: groups with brain damage, anti- cies). To play the game successfully participants must learn these
social personality, drug abuse problems, and Huntington’s disease contingencies over time. The primary dependant variable is the
and incarcerated criminals (Bechara, Tranel, & Damasio, 2000; number of choices from the advantageous decks over the course
Bechara & Damasio, 2002; Bechara, Dolan, & Hindes, 2002; of the game. The IGT is thought to simulate real world decision
Monterosso, Ehrman, Napier, O’Brien, & Childress, 2001; Yechiam making abilities because it involves the integration of multiple
et al., 2008). However, not all control participants perform advan- complex decision making components such as risk, uncertainty, re-
tageously on the IGT and little research has been focused on indi- wards and punishments, ambiguity, and learning from experience
vidual differences that could lead to disadvantageous performance (Buelow & Suhr, 2009). The IGT has successfully discriminated ris-
in healthy controls. The current work uses the IGT to examine deci- ky decision making between healthy control groups and numerous
sion making in individuals who differ in the individual difference clinical populations known to have difficulties in real world deci-
variable need for cognition (NFC; Cacioppo & Petty, 1982). sion making, typically with control groups learning to choose more
from the advantageous decks over time and clinical groups failing
⇑ Address: Ohio University, 200 Porter Hall, Ohio University, Athens, OH 45701, to do so.
USA. Tel.: +1 (740) 707 6944; fax: +1 (740) 593 0579. The IGT is unique as a decision making paradigm in that out-
E-mail address: harmanj1@ohio.edu comes are mixed (both gains and losses can result from the same

0191-8869/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.paid.2011.03.021
 J.L. Harman / Personality and Individual Differences 51 (2011) 112–116 113

Table 1 Gaeth, Schreiber, & Lauriola, 2002). Similarly, Chatterjee, Heath,

Mean gains and losses associated with each deck in the Iowa Gambling Task (IGT) and Milberg, and France (2000) found that high NFC participants were
the modified IGT.
resistant to framing effects in the differential discrimination of
Deck ‘A’ Deck ‘B’ Deck ‘C’ Deck ‘D’ losses and gains evaluating gains and losses equally, whereas low
Gains $100 $100 $50 $50 NFC participants were not. Furthermore, Carnevale, Inbar, and
Losses .5 of $250 .1 of $1250 .5 of $50 .1 of $250 Lerner (2010) investigated differences in decision competence be-
Net value (10 trials) $250 $250 $250 $250 tween high and low NFC participants using the Adult Decision-
Modified IGT Making Competence Scale (Bruine de Bruin, Parker, & Fischhoff,
Gains .5 of $250 .1 of $1250 .5 of $50 .1 of $250 2007). Their results showed differences between high and low
Losses $100 $100 $50 $50
Net value (10 trials) $250 $250 $250 $250
NFC participants in susceptibility to biases only in psychophysical
judgment errors (Arkes, 1991) which are types of biases that result
Note: Each deck produces a variable gain (loss) on each trial with some probability from differential sensitivity to gains and losses (e.g. sunk costs and
of a loss (gain) for an average net value over 10 trials.
framing effects). The results discussed above suggest that NFC
could be an important predictor of IGT performance. The first
experiment is designed to explore this possibility using the stan-
choice) and it links two types of decision making, decisions under
dard version of the IGT and separating participants into high and
ambiguity and decisions under risk (Brand, Recknor, Grabenhorst,
low NFC groups.
& Bechara, 2007). In early stages of the IGT participants are making
decisions under ambiguity as they learn the contingent payoffs of
each deck. In later stages the contingent payoffs of each deck are 2. Experiment 1
known and participants are performing a risky decision making
task (Brand et al., 2007; Maia & McClelland, 2004). Successful per- 2.1. Method: participants and procedure
formance in the later stages of the IGT then depend on the partic-
ipant’s appropriate weighting of the sporadic losses compared to Forty-three female and 32 male students at Ohio University
the consistent gains. While most healthy control participants in participated for course credit. Participants completed the IGT and
the above studies perform this task advantageously, up to 25% of a NFC questionnaire in a counterbalanced order. Need for cognition
controls do not (Desmeules, Bechara, & Dube, 2008). This variabil- was assessed using the short form 18 item inventory (Cacioppo
ity in the performance of control participants makes the IGT a et al., 1984).
promising paradigm to study individual differences in decision
making is a task more complex than simple gambles and one time 2.2. Results
decisions.
Participants were grouped into high and low NFC groups via
1.2. Individual differences in the IGT median split. The high NFC group had a mean score of 67.07
(SD = 6.69) and the low NFC group had a mean score of 52.05
Individual differences in IGT performance in the normal popula- (SD = 6.70).1 No gender effects were found in any analyses and are
tion has garnered little study (see Suzuki, Hirota, Takasawa, & omitted from the presented data analyses.
Shigemasu, 2003; Desmeules et al., 2008 for exceptions). Recently The proportion of selections from advantageous decks were
Weller, Levin, and Bechara (2010) found that poor IGT performance separated into five blocks of 20 trials. A 2 (NFC)  5 (block) re-
in non-clinical participants was related to poor performance on a peated measures ANOVA was performed on proportion of advanta-
simpler task that separated choices in the gain domain from geous selections over the five blocks of trials (means are plotted in
choices in the loss domain. Specifically, participants who per- Fig. 1). Results revealed a NFC  block interaction F(4,70) = 2.72,
formed poorly on the IGT performed poorly on risky decisions in p < .05, with the high NFC group choosing more advantageously
the loss domain, showing less discrimination to differences in ex- as trials progress. An expected main effect of trial block was also
pected value, but not in the gain domain. This result suggests that found, F(4,70) = 8.075, p < .01, with all participants on average
susceptibility to framing could lead to disadvantageous perfor- selecting more from the advantageous decks as trials progress.
mance on the IGT underweighting losses in a paradigm where Comparison of the final 40 trials (risky decision making; Brand
losses and gains are mixed and should be weighted equally. et al., 2007) finds the high NFC group outperforming the low NFC
group t(73) = 2.93, p < .01. Repeated measures ANOVA, artificially
1.3. Need for cognition dichotomizing NFC via median split, was chosen to preserve the
dynamic structure of the IGT data. Using regression, collapsing
To test this prediction, we used NFC, an individual difference across IGT data, does not alter the current results.2
capturing people’s tendency to engage in and enjoy effortful
thought (Cacioppo & Petty, 1982). NFC is measured through 18 2.3. Discussion
scale items such as ‘‘I prefer complex to simple problems’’ and
‘‘Thinking is not my idea of fun (reverse scored)’’ (Cacioppo, Petty, The results from experiment 1 show that the high NFC group
& Kao, 1984). In a review of over 100 empirical studies, Cacioppo, outperformed the low NFC group on the IGT. However, due to
Petty, Feinstein, and Jarvis (1996) found consistent reliability the complex nature of the IGT, this could be the result of multiple
(as > .85; test–retest reliability r = .88) and validity for NFC, sup- processes. One possible explanation consistent with previous work
porting a single factor capturing people’s tendency to engage in is that high NFC participants recognized the deeper structure of the
and enjoy effortful cognitive activity, related to reliable differences task choosing a strategy more closely aligned with learned ex-
in information processing. pected values while low NFC participants choose a strategy based
Pertinent to the current study, Smith and Levin (1996) found
1
NFC scores had acceptable reliability in both experiment 1 (a = .83) and
that individuals high in NFC had fewer framing errors than those
experiment 2 (a = .87).
low in NFC, suggesting that high NFC may facilitate discovering 2
NFC is a significant predictor of total IGT performance (advantageous–disadvan-
the deeper structure in different decision problems, though this tageous choices) (F(1,73) = 3.87, p = .05, r2 = .05) and marginally significant when
finding has had mixed support (LeBoeuf & Shafir, 2003; Levin, predicting the final 40 trials (F(1,73) = 2.975, p = .089, r2 = .039).
114 J.L. Harman / Personality and Individual Differences 51 (2011) 112–116

Fig. 2. Mean number of advantageous choices in the modified IGT. High and low
Fig. 1. Mean number of advantageous choices per 20 trial block. High and low NFC NFC groups performance on the modified IGT is plotted over five blocks of 20 trials
groups performance on the standard IGT is plotted over five blocks of 20 trials with with 10 indicating no preference for advantageous or disadvantageous decks.
10 indicating no preference for advantageous or disadvantageous decks.

on immediate outcomes of the task, giving more weight to large from advantageous decks as trials progressed. There was no differ-
immediate gains. An alternative explanation for the results of ence in modified IGT performance between high and low NFC
study 1 is that both groups perform the IGT in the same manner, groups. Neither the NFC block interaction (F(4,78) = .584,
but with low NFC participants dispositionally more sensitive to p = .675) nor comparison of the final 40 trials (t(81) = 1.065,
gains. p = .29) were significant.3 Fig. 2 shows the mean number of advan-
Experiment 2 further examines the underlying processes lead- tageous choices by NFC group.
ing to differential IGT performance by high and low NFC groups
by using a modified version of the IGT (Bechara et al., 2000) where 3.3. Discussion
the gains and losses are reversed. In this modified IGT, the ‘advan-
tageous’ decks are now disadvantageous, producing constant small Taken together experiments 1 and 2 support the hypothesis
losses with sporadic small gains and a net loss in the long run. The that low NFC participants weight potential gains greater than po-
previously ‘disadvantageous’ decks now produce constant large tential losses in the IGT (see Section 4 for further evidence). These
losses but larger sporadic gains and an overall net gain (see Table results are in part consistent with previous research showing that
1). To perform well on this version of the task participants must low NFC participants are more likely to change choice behavior due
weigh the sporadic large gains enough to overcome the immediate to differences in framing, however, the observed pattern does not
and constant large losses. If participants in the low NFC group follow directly from previous NFC research. While high NFC partic-
choose a strategy based on immediate outcomes of the task, then ipants have been found to use the deeper structure in different
performance on the modified IGT should likewise be poor as ini- decision problems (Carnevale et al., 2010; Smith & Levin, 1996)
tially the advantageous decks produce much larger losses than consistent with advantageous performance on both versions of
the disadvantageous decks. If on the other hand, the low NFC group the IGT, low NFC participants have been associated with more sys-
are more sensitive to gains their performance should be advanta- tematic thought in reference to losses (Chatterjee et al., 2000) but
geous as the sporadic large gains of the advantageous decks will not gains as proposed here. To add more detail to the above anal-
quickly out-weigh the large losses. ysis, data from each experiment was analyzed using the Expec-
tancy Valence Learning model (EVLM, Busemeyer & Stout, 2002),
3. Experiment 2 a cognitive model developed to discriminate the cognitive, motiva-
tional, and learning processes thought to underlie IGT
3.1. Method: participants and procedure performance.

Thirty-three female and 50 male students at Ohio University 4. Supplemental analysis

participated for course credit. Participants completed a NFC ques-
tionnaire and the modified version of the IGT. 4.1. EVLM

3.2. Results According to EVLM, a decision maker integrates the gains and
losses experienced on each trial into a single affective reaction
Participants were grouped into high and low NFC groups via called a valence. An adaptive learning mechanism generates expec-
median split. The high NFC group had a mean score of 65.52 tancies about the valence produced by each deck, which then serve
(SD = 5.05) and the low NFC group had a mean score of 48.55
(SD = 6.70). 3
Regression analysis confirms no significant relationship between NFC and total
Participants performing the modified IGT showed the expected IGT performance (F(1,81) = .076, p = .784, r2 < .01) or the final 40 trials (F(1,81) = .027,
main effect of trial block (F(4,78) = 26.359, p < .01), choosing more p = .87, r2 < .01).
 J.L. Harman / Personality and Individual Differences 51 (2011) 112–116 115

as the inputs into a probabilistic choice mechanism that selects the 5. General discussion
choice on each trial.
EVLM consists of three parameters; attention weight ‘w’, Individual differences in decision-making have received a great
updating rate ‘a’, and sensitivity ‘c’. The weighting parameter deal of attention in recent years. NFC has emerged as an important
‘w’ distinguishes different amounts of weight given to losses as factor in predicting differences in decision competence with high
opposed to gains. The weighting parameter is bound between 0 NFC individuals less susceptible to framing effects and sunk costs
and 1 with values greater than .5 indicating greater weighting when compared to low NFC individuals (Carnevale et al., 2010;
of losses and values below .5 greater weighting of gains. The Chatterjee et al., 2000; Smith & Levin, 1996). These studies have
updating rate parameter ‘a’ (bound between 0 and 1) indicates examined decision competence in terms of normative rule viola-
the influence of previous trials on the current trial. Large updat- tion in one time decisions and judgments. Though this methodol-
ing rates indicate strong recency effects while small rates indicate ogy allows for the subtle examination of component processes, it
weak recency effects. The sensitivity parameter ‘c’ is a learning has been criticized in terms of the external validity of one time lab-
parameter that indicates whether performance changes in line oratory decisions (Gigerenzer, Todd, & the ABC Group, 2000; Klein,
with experienced outcomes with negative values indicating no 1999). Additionally, the use of normative rules as criterion of com-
learning (for a more detailed description of EVLM see Busemeyer petent decision making has been criticized for similar reasons
& Stout, 2002). (Gigerenzer et al., 2000). The current work avoids these limitations
Using EVLM to model IGT data has successfully discriminated by using the IGT, a clinical measurement tool developed specifi-
similar performance on the IGT to different psychological compo- cally to identify individuals with real world decision making defi-
nents between different clinical populations (Busemeyer & Stout, cits. The IGT is complex; involving adaptive decisions from
2002; Yechiam & Busemeyer, 2005; Yechiam, Veinott, Busemeyer, experience, uncertainty, and mixed outcomes in the same task.
& Stout, 2007; Yechiam, Busemeyer, Stout, & Bechara, 2005; Wood, Additionally, the IGT has been extensively validated, identifying
Busemeyer, Koling, Cox, & Davis, 2005). For example Wood et al. decision making deficits in multiple populations (Bechara, 2007)
had younger and older adults perform the IGT and used EVLM to providing strong evidence for the relation between laboratory per-
examine differences in decision making strategies. Though both formance and real world decision making. For these reasons the
groups learned to choose advantageously as the game progressed, IGT can provide an important complimentary method for studying
analysis using EVLM provided evidence that they did so using dif- decision making.
ferent psychological mechanisms. Consistent with previous re- The current work shows that the IGT can be used successfully to
search on aging, older participants relied less on memory (higher study individual differences in decision making, providing both
‘a’ parameter estimates) and more on unbiased weighting of gains support and more detail to previous research while maintaining
and losses (‘w’ estimates close to .5 as opposed to .24 for young relevance to real world outcomes. Results from the two experi-
adults). In this manner, cognitive models such as EVLM can provide ments together provide evidence of decision making differences
a method of analysis more detailed than the standard proportion of between individuals differing in NFC. Experiment 1 showed that
advantageous selections. the high NFC participants performed better on the IGT than the
low NFC group. To determine whether this difference was due to
participants in the low NFC group relying on initial impressions
4.2. Modeling analysis or whether high and low NFC groups differentially weighted gains
and losses, experiment 2 utilized a modified version of the IGT de-
Parameter estimates and fit indices from EVLM for experiments signed to reverse the importance of gains and losses. Following the
1 and 2 are shown in Table 2. Model fitness was assessed using the predictions of differential weighting of gains and losses, the differ-
G2 method introduced by Busemeyer and Stout (2002) and used ences found in Experiment 1 disappeared. Evidence for differential
since as a standard for EVLM evaluation. For experiment 1, where weighting of gains and losses between high and low NFC individu-
low NFC participants performed worse than high NFC participants, als has been found previously (Chatterjee et al., 2000) however
parameter estimates between high and low NFC groups show no these studies have indicated that low NFC individuals discriminate
difference in either the updating parameter ‘a’, t(73) = .651, losses more carefully than gains. The current work suggests that
p = .517, or the sensitivity parameter ‘c’, t(73) = .179, p = .859. The low NFC performance is driven primarily by gains with poor dis-
two groups do differ on the weight parameter ‘w’, t(73) = 2.454, crimination of potential losses. One difference between the current
p < .05. Mean parameter estimates indicate that the low NFC group work and previous studies examining differences in NFC and
weighted gains more (M = .2572) than the high NFC group choice is the complexity of the paradigm. Chatterjee et al. asked
(M = .4137).For experiment 2, where both groups performed participants to rate the happiness of a hypothetical person who
equally well, there were no differences in parameter estimates is faced with a price increase and discount on a multi item pur-
between high and low NFC groups (‘a’, t(81) = 1.207, p = .231; chase. The gains in this paradigm are price discounts. In the IGT
‘w’, t(81) = 1.716, p = .09; ‘c’, t(81) = 1.292, p = .2). participants are making risky decisions from experience with di-
rect gains. Additionally, the occurrence of losses in the IGT is spo-
radic and must be both remembered and integrated. This requires
Table 2
Parameter estimates for all studies. Estimates for the parameters of EVLM (updating
greater processing than integration of consistent gains. The differ-
rate ‘a’, attention to losses ‘w’, and sensitivity ‘c’) are listed for each study separating ent framing of gains between the two paradigms or the sporadic
high and low NFC participants along with G2 fit results. frequency of losses could explain why low NFC individuals focus
on gains in the IGT but not in the multiple outcome judgments
n Parameter estimates +G2
from Chatterjee et al. and should be examined in future work.
a w c
Finally, the current work may help explain important variability
High NFC Low NFC High NFC Low NFC High NFC Low NFC in future IGT studies. The IGT has become an important tool in clin-
Study 1 75 .5338 .4397 .4317 .2572* .6241 .7401 52 ical neuropsychology for studying decision making deficits. In the
Study 2 83 .3783 .4632 .2492 .1407 1.4521 .9549 72 past, analysis of task performance has assumed homogeneity of
*
p < .05. groups. However, some studies have shown that poor performance
Proportion significantly greater than chance, sign test p < .05. by different clinical groups on the IGT could be the result of
116 J.L. Harman / Personality and Individual Differences 51 (2011) 112–116

separate and distinct processes (Yechiam et al., 2005). Likewise, Bruine de Bruin, W., Parker, A. M., & Fischhoff, B. (2007). Individual differences in
adult decision-making competence. Journal of Personality and Social Psychology,
the current results show that non clinical populations, most often
92(5), 938–956.
used as control groups, are themselves not homogenous in their Buelow, M. T., & Suhr, A. (2009). Construct validity of the Iowa Gambling Task.
performance. These differences should be accounted for in future Neuropsychology Review, 19(1), 102–114.
IGT studies, as they may be important sources of variability. For Busemeyer, J. R., & Stout, C. (2002). A contribution of cognitive decision models to
clinical assessment: Decomposing performance on the Bechara gambling task.
example, the mixed findings regarding the use of working memory Psychological Assessment, 14(3), 253–262.
and explicit knowledge in task performance (Maia & McClelland, Cacioppo, J. T., & Petty, E. (1982). The need for cognition. Journal of Personality and
2004; Bechara, Damasio, Tranel, & Damasio, 2005; Brand et al., Social Psychology, 42(1), 116–131.
Cacioppo, J. T., Petty, R. E., Feinstein, J. A., & Jarvis, W. B. G. (1996). Dispositional
2007) could possibly be an artifact of different processing styles differences in cognitive motivation: The life and times of individuals varying in
use by different groups of individuals. In addition to NFC, the cur- need for cognition. Psychological Bulletin, 119(2), 197–253.
rent results point to the possible importance of Regulatory Focus Cacioppo, J. T., Petty, R. E., & Kao, C. F. (1984). The efficient assessment of need for
cognition. Journal of Personality Assessment, 48(3), 306–307.
(RF, Higgins, 1997) in IGT performance. RF is an individual differ- Carnevale, J., Inbar, Y., & Lerner, J. S. (2010). Individual differences in need for
ence that discriminates prevention and promotion focused cognition and decision making competence among leaders. Personality and
self-regulation with prevention focused individuals placing more Individual Differences, doi:10.1016/j.paid.2010.06.023.
Chatterjee, S., Heath, T. B., Milberg, S. J., & France, K. R. (2000). The differential
attention on negative outcomes and promotion focused individuals processing of price in gains and losses: The effects of frame and need for
placing more attention on positive outcomes. Previous work has cognition. Journal of Behavioral Decision Making, 13(1), 61–75.
found links between RF and risky behavior (Hamstra, Bolderdijk, Desmeules, R., Bechara, A., & Dube, L. (2008). Subjective valuation and asymmetrical
motivational systems: Implications of scope insensitivity for decision making.
& Veldstra, 2011; Scholer & Higgins, 2008) making RF a prime can-
Journal of Behavioral Decision Making, 21(2), 211–224.
didate for future work examining individual differences in the IGT. Evans, C. E., Kemish, K., & Turnbull, O. H. (2004). Paradoxical effects of education on
In conclusion, the current work finds clear differences in IGT the Iowa Gambling Task. Brain and Cognition, 54(3), 240–244.
performance between groups differing in NFC. This finding is Gigerenzer, G., & Todd, P. M.the ABC Group. (2000). Simple heuristics that make us
smart. Cary, NC: Oxford University Press.
important in regards to future work using the IGT, in terms of Hamstra, M. R., Bolderdijk, J. W., & Veldstra, J. L. (2011). Everyday risk taking as a
accounting for variability in control populations. Additionally the function of regulatory focus. Journal of Research in Personality, 45(1), 134–137.
current work can provide an important perspective regarding the Higgins, E. T. (1997). Beyond pleasure and pain. American Psychologist, 52(12),
1280–1300.
underlying processes at work in IGT performance. Though support Klein, G. (1999). Sources of power: How people make decisions. Cambridge, MA: MIT
was found for the hypothesis that performance differences be- Press.
tween high and low NFC groups are the result of differential LeBoeuf, R. A., & Shafir, E. (2003). Deep thoughts and shallow frames; on the
susceptibility to framing effects. Journal of Behavioral Decision Making, 16(2),
weighting of gains and losses, other alternative explanations are 77–92.
possible. Though the current data was collected using a homoge- Levin, I. P., Gaeth, G. J., Schreiber, J., & Lauriola, M. (2002). A new look at framing
neous sample in regards to age and education, two variables effects: Distribution of effect sizes, individual differences, and independence of
types of effects. Organizational Behavior and Human Decision Processes, 88(1),
known to effect IGT performance (Wood et al., 2005; Evans, 411–429.
Kemish, & Turnbull, 2004), future work is needed to examine Maia, T. V., & McClelland, L. (2004). The somatic marker hypothesis: Still many
other possible third variable explanations such as IQ and memory questions but no answers. Trends in Cognitive Sciences, 9(4), 162–164.
Monterosso, J., Ehrman, R., Napier, K. L., O’Brien, C. P., & Childress, A. R. (2001). Three
capacity.
decision-making tasks in cocaine-dependent patients: Do they measure the
same construct? Addiction, 96(12), 1825–1837.
Scholer, A. A., & Higgins, E. T. (2008). Distinguishing levels of approach and
avoidance: An analysis using regulatory focus theory. In A. J. Elliot (Ed.),
References Handbook of approach and avoidance motivation (pp. 489–504). New York:
Psychology Press.
Arkes, H. R. (1991). Costs and benefits of judgment errors: Implications for Smith, S. M., & Levin, P. (1996). Need for cognition and choice framing effects.
debiasing. Psychological Bulletin, 110(3), 486–498. Journal of Behavioral Decision Making, 9(4), 283–290.
Bechara, A. (2007). Iowa Gambling Task Professional Manual. Boca Raton, Fl: Suzuki, A., Hirota, A., Takasawa, N., & Shigemasu, K. (2003). Application of the
Psychological Assessment Resources. somatic marker hypothesis to individual differences in decision making.
Bechara, A., & Damasio, H. (2002). Decision-making and addiction (part I): Impaired Biological Psychology, 65(1), 81–88.
activation of somatic states in substance dependent individuals when Weller, J. A., Levin, I. P., & Bechara, A. (2010). Do individual differences in Iowa
pondering decisions with negative future consequences. Neuropsychologia, Gambling Task performance predict adaptive decision making for risky gains
40(10), 1675–1689. and losses? Journal of Clinical and Experimental Neuropsychology, 32(2), 141–150.
Bechara, A., Damasio, A. R., Damasio, H., & Anderson, S. W. (1994). Insensitivity to Wood, S., Busemeyer, J., Koling, A., Cox, C. R., & Davis, H. (2005). Older adults as
future consequences following damage to human prefrontal cortex. Cognition, adaptive decision makers: evidence from the Iowa gambling task. Psychology
50(1–3), 7–15. and Aging, 20(2), 220–225.
Bechara, A., Damasio, H., Tranel, D., & Damasio, A. R. (2005). The Iowa Gambling Yechiam, E., & Busemeyer, R. (2005). Comparison of basic assumptions embedded in
Task and the somatic marker hypothesis: Some questions and answers. Trends learning models for experience-based decision making. Psychonomic Bulletin &
in Cognitive Sciences, 9(4), 159–162. Review, 12(3), 387–402.
Bechara, A., Dolan, S., & Hindes, A. (2002). Decision-making and addiction (part II): Yechiam, E., Busemeyer, J. R., Stout, J. C., & Bechara, A. (2005). Using cognitive
Myopia for the future or hypersensitivity to reward? Neuropsychologia, 40(10), models to map relations between neuropsychological disorders and human
1690–1705. decision-making deficits. Psychological Science, 16(12), 973–978.
Bechara, A., Tranel, D., & Damasio, H. (2000). Characterization of the decision- Yechiam, E., Kanz, J. E., Bechara, A., Stout, J. C., Busemeyer, J. R., Altmaier, E. M., et al.
making deficit of patients with ventromedial prefrontal cortex lesions. Brain: A (2008). Neurocognitive deficits related to poor decision making in people
Journal of Neurology, 123(11), 2189–2202. behind bars. Psychonomic Bulletin & Review, 15(1), 44–51.
Brand, M., Recknor, E. C., Grabenhorst, F., & Bechara, A. (2007). Decisions under Yechiam, E., Veinott, E. S., Busemeyer, J. R., & Stout, J. C. (2007). Cognitive models for
ambiguity and decisions under risk: Correlations with executive functions and evaluating basic decision processes in clinical populations. In Advances in
comparisons of two different gambling tasks with implicit and explicit rules. clinical cognitive science: Formal modeling of processes and symptoms
Journal of Clinical and Experimental Neuropsychology, 29(1), 86–99. (pp. 81–111). Washington, DC, US: American Psychological Association.