600886
BBSXXX10.1177/2372732215600886Policy Insights from the Behavioral and Brain SciencesMorewedge et al.
research-article2015
Evaluating and Mitigating Risk
Debiasing Decisions: Improved Decision
Making With a Single Training Intervention
Policy Insights from the
Behavioral and Brain Sciences
2015, Vol. 2(1) 129–140
© The Author(s) 2015
DOI: 10.1177/2372732215600886
bbs.sagepub.com
Carey K. Morewedge1, Haewon Yoon1, Irene Scopelliti2,
Carl W. Symborski3, James H. Korris4, and Karim S. Kassam5
Abstract
From failures of intelligence analysis to misguided beliefs about vaccinations, biased judgment and decision making contributes
to problems in policy, business, medicine, law, education, and private life. Early attempts to reduce decision biases with
training met with little success, leading scientists and policy makers to focus on debiasing by using incentives and changes
in the presentation and elicitation of decisions. We report the results of two longitudinal experiments that found medium
to large effects of one-shot debiasing training interventions. Participants received a single training intervention, played a
computer game or watched an instructional video, which addressed biases critical to intelligence analysis (in Experiment 1:
bias blind spot, confirmation bias, and fundamental attribution error; in Experiment 2: anchoring, representativeness, and
social projection). Both kinds of interventions produced medium to large debiasing effects immediately (games ≥ −31.94%
and videos ≥ −18.60%) that persisted at least 2 months later (games ≥ −23.57% and videos ≥ −19.20%). Games that provided
personalized feedback and practice produced larger effects than did videos. Debiasing effects were domain general: bias
reduction occurred across problems in different contexts, and problem formats that were taught and not taught in the
interventions. The results suggest that a single training intervention can improve decision making. We suggest its use alongside
improved incentives, information presentation, and nudges to reduce costly errors associated with biased judgments and
decisions.
Keywords
debiasing, cognitive bias, judgment and decision making, decisions, training, feedback, nudges, incentives
Tweet
Introduction
As we saw in several instances, when confronted with evidence
that indicated Iraq did not have WMD, analysts tended to
discount such information. Rather than weighing the evidence
independently, analysts accepted information that fit the
prevailing theory and rejected information that contradicted it.
A single training intervention with an instructional game or
video produced large and persistent reductions in decision
bias.
Key Points
•• Biases in judgment and decision making create predictable errors in domains such as intelligence analysis, policy, law, medicine, education, business, and
private life.
•• Debiasing interventions can be effective, inexpensive
methods to improve decision making and reduce the
costly errors that decision biases produce.
•• We found a short, single training intervention (i.e.,
playing a computer game or watching a video) produced persistent reductions in six cognitive biases
critical to intelligence analysis.
•• Training appears to be an effective debiasing intervention to add to existing interventions such as
improvements in incentives, information presentation, and how decisions are elicited (e.g., nudges).
—Report to the President of the United States (Silberman et al.,
2005, p. 169)
Biased judgment and decision making is that which systematically deviates from the prescriptions of objective standards
1
Boston University, MA, USA
City University London, UK
3
Leidos, Reston, VA, USA
4
Creative Technologies Incorporated, Los Angeles, CA, USA
5
Carnegie Mellon University, Pittsburgh, PA, USA
2
Corresponding Author:
Carey K. Morewedge, Associate Professor of Marketing, Questrom
School of Business, Boston University, Rafik B. Hariri Building, 595
Commonwealth Ave., Boston, MA 02215, USA.
Email: morewedg@bu.edu
130
such as facts, rational behavior, statistics, or logic (Tversky
& Kahneman, 1974). Decision bias is not unique to intelligence analysis. It affects the intuitions and calculated decisions of novices and highly trained experts in numerous
domains, including education, business, medicine, and law
(Morewedge & Kahneman, 2010; Payne, Bettman, &
Johnson, 1993) underlying phenomena such as the tendency
to sell winning stocks too quickly and hold on to losing
stocks too long (Shefrin & Statman, 1985), the persistent
belief in falsified evidence linking vaccinations to autism
(Lewandowsky, Ecker, Seifert, Schwarz, & Cook, 2012), and
unintentional discrimination in hiring and promotion practices (Krieger & Fiske, 2006). Biased judgment and decision
making affects people in their private lives. Less biased decision makers have more intact social environments, reduced
risk of alcohol and drug use, lower childhood delinquency
rates, and superior planning and problem solving abilities
(Parker & Fischhoff, 2005).
Decision-making ability varies across persons and within
person across the life span (Bruine de Bruin, Parker, &
Fischhoff, 2007; Dhami, Schlottmann, & Waldmann, 2011;
Peters & de Bruin, 2011), but people are generally unaware
of the extent to which they are biased and have difficulty
debiasing their decision making (Scopelliti et al., 2015;
Wilson & Brekke, 1994). Considerable scientific effort has
been expended developing strategies and methods to improve
novice and expert decision making over the last 50 years (for
reviews, see Fischhoff, 1982; Soll, Milkman, & Payne, in
press). Three general debiasing approaches have been
attempted, each with its pros and cons: changing incentives,
optimizing choice architecture (e.g., improving how decisions are presented and elicited), and improving decision
making ability through training.
Incentives
Changing incentives can substantially improve decision
making. Recalibrating incentives to reward healthy behavior
improves diet (Schwartz et al., 2014), exercise (Charness &
Gneezy, 2009), weight loss (John et al., 2011), medication
adherence (Volpp et al., 2008), and smoking cessation (Volpp
et al., 2009). In one study, during a period in which the price of
fresh fruit was reduced by 50% in suburban and urban school
cafeterias, sales of fresh fruit increased fourfold (French, 2003).
Incentives are not a solution for every bias; bias is prevalent even in high-stake multibillion-dollar decisions (Arkes &
Blumer, 1985). Incentives can also backfire. When incentives
erode intrinsic motivation and change norms from prosociality
to economic exchange, incentives demotivate behavior if they
are insufficient or discontinued (Gneezy, Meier, & Rey-Biel,
2011). Israeli day care facilities that introduced a small fine
when parents picked up their children late, for instance, saw an
increase in the frequency of late pickups. The fine made rude
behavior acceptable, a price to watch the children longer
(Gneezy & Rustichini, 2000). When incentives are too great,
Policy Insights from the Behavioral and Brain Sciences 2(1)
they can make people choke under pressure (Ariely, Gneezy,
Loewenstein, & Mazar, 2009). If people apply inappropriate
decision strategies or correction methods because they do
not know how or the extent to which they are biased, increasing incentives can exacerbate bias rather than mitigate it
(Lerner & Tetlock, 1999). In short, incentives can effectively
improve behavior, but they require careful calibration and
implementation.
Optimizing Choice Architecture
Optimizing the structure of decisions, how choice options are
presented and how choices are elicited, is a second way to
effectively debias decisions. People do make better decisions
when they have the information they need and good options
to choose from. Giving people more information and choices
is not always helpful, particularly when it makes decisions
too complex to comprehend, existing biases encourage good
behavior, or people recognize the choices they need to make
but fail to implement them because they lack self-control
(Bhargava & Loewenstein, 2015; Fox & Sitkin, 2015).
Providing calorie information does not necessarily lead people to make healthier food choices, for instance, and there is
some evidence that smokers actually overestimate the health
risks of smoking—debiasing smokers may actually increase
their health risks (Downs, Loewenstein, & Wisdom, 2009).
Changing what and how information is presented can
make choices easier to understand and good options easier to
identify, thus doing more to improve decisions than simply
providing more information. Eligible taxpayers are more
likely to claim their Earned Income Tax Credits, for example,
when benefit information is simplified and prominently displayed (e.g., “ . . . of up to $5,657”; Bhargava & Manoli, in
press). Consumers are better able to recognize that small
reductions in the fuel consumption of inefficient vehicles
saves more fuel than large reductions in the fuel consumption
of efficient vehicles (e.g., improving 16→20 MPG saves
more than improving 34→50MPG) when the same information about vehicle fuel consumption is framed in gallons per
100 miles (GPM) rather than in MPG (Larrick & Soll, 2008).
Both novices and trained experts benefit from the implementation of simple visual representations of risk information,
whether they are evaluating medical treatments or new counterterrorism techniques (Garcia-Retamero & Dhami, 2011,
2013). Moreover, statistical analyses of voting patterns in the
2000 U.S. Presidential Election suggest that had the butterfly
ballots used by Palm Beach County, Florida, been designed in
a manner consistent with basic principles of perception, Al
Gore would have been elected President (Fox & Sitkin, 2015).
Even when people fully understand their options, if one
option is better for them or society but choosing it requires
effort, expertise, or self-control, its selection can be increased
if small nudges in presentation and elicitation methods are
implemented (Thaler & Sunstein, 2008). Nudges take many
forms such as information framing, commitment devices, and
Morewedge et al.
default selection. Voters are more mobilized by message
frames that emphasize a high expected turnout at the polls
(implying voting is normative) than message frames that
emphasize low expected turnouts (implying each vote is
important; Gerber & Rogers, 2009), and consumers prefer
lower fat meat when its fat content is framed as 25% fat than
75% lean (Levin & Gaeth, 1988). Shoppers are willing to
commit to foregoing cash rebates that they currently receive
on healthy foods if they fail to increase the amount of healthy
food that they purchase (Schwartz et al., 2014), and employees substantially increase their contributions to 401k programs when they commit to allocating money from future
raises to their retirement savings before receiving those
raises (Thaler & Benartzi, 2004).
People are more likely to choose an option if it is a default
from which they must opt-out than if it is an option that they
must actively choose (i.e., “opt-in”). In one study, university
employees were 36% more likely to receive a flu shot if
emailed an appointment from which they could opt-out, than
if emailed a link from which they could schedule an appointment (Chapman, Li, Colby, & Yoon, 2010). Organ donation
rates are at least 58% higher in European countries in which
the default is to opt-out of being a donor than in which the
default is to opt-in (Johnson & Goldstein, 2003). Selecting
better default options is not necessarily coercive. It results in
outcomes that decision makers themselves prefer (Goldstein,
Johnson, Herrman, & Heitmann, 2008; Huh, Vosgerau, &
Morewedge, 2014).
The potential applications of optimizing of choice architecture are broad, ranging from increasing retirement savings
and preserving privacy, to reducing the energy, soda, and
junk food that people consume (Acquisti, Brandimarte, &
Loewenstein, 2015; Larrick & Soll, 2008; Schwartz et al.,
2014; Thaler & Benartzi, 2004). Optimizing choice architecture is a cheap way to improve public welfare while
preserving freedom of choice, as it does not exclude options
or change economic incentives (Camerer, Issacharoff,
Loewenstein, O’Donoghue, & Rabin, 2003; Thaler &
Sunstein, 2003, 2008). Critics, however, point out that these
improvements may not do enough. They tend to reduce decision bias in one, not multiple contexts, and do not address the
underlying structural causes of biased decisions such as
poorly calibrated incentives or bad options (Bhargava &
Loewenstein, 2015).
Training
Training interventions to improve decision making, to date,
have met with limited success mostly in specific domains.
Training can be very effective when accuracy requires
experts to recognize patterns and select an appropriate
response, such as in weather forecasting, firefighting, and
chess (Phillips, Klein, & Sieck, 2004). By contrast, even
highly trained professionals are less accurate than very
131
simple mathematical models in other domains such as parole
decisions, personnel evaluations, and clinical psychological
testing (Dawes, Faust, & Meehl, 1989). Whether domainspecific expertise is achievable appears to be contingent on
external factors such as the prevalence of clear feedback, the
frequency of the outcome being judged, and the number and
nature of variables that determine that outcome (Harvey,
2011; Kohler, Brenner, & Griffin, 2002).
Evidence that training effectively improves general decision-making ability is inconclusive at present (Arkes, 1991;
Milkman, Chugh, & Bazerman, 2009; Phillips et al., 2004).
Weather forecasters are well calibrated when predicting the
chance of precipitation (Murphy & Winkler, 1974), for
example, but are overconfident in their answers to general
knowledge questions (Wagenaar & Keren, 1986). Even
within their domain of expertise, experts struggle to apply
their training to new problems. Philosophers trained in logic
exhibit the same preference reversals in similar moral dilemmas as academics without logic training (Schwitzgebel &
Cushman, 2012), and physicians exhibit the same preference
reversals as untrained patients for equivalent medical treatments when those treatments are framed in terms of survival
or mortality rates (McNeil, Pauker, Sox, & Tversky, 1982).
Several studies have shown that people do not apply their
training to unfamiliar and dissimilar domains because they
lack the necessary metacognitive strategies to recognize
underlying problem structure (for reviews, see Barnett &
Ceci, 2002; Reeves & Weisberg, 1994; Willingham, 2008).
Debiasing training methods teaching inferential rules
(e.g., “consider-the-opposite” and “consider-an-alternative”
strategies) that are grounded in two-system models of reasoning hold some promise (e.g., Lilienfeld, Ammirati, &
Landfield, 2009; Milkman et al., 2009; Soll et al., in press).
Two-system models of reasoning assume that people initially
make an automatic intuitive judgment that can be subsequently accepted, corrected, or replaced by more controlled
and effortful thinking: through “System 1” and “System 2”
processes, respectively (Evans, 2003; Morewedge &
Kahneman, 2010; Sloman, 1996). Recognizing that “1,593 ×
1,777” is a math problem and that its answer is a large number, for instance, are automatic outputs of System 1 processes. Deducing the answer to the problem requires the
engagement of effortful System 2 processes.
Effective debiasing training typically encourages the consideration of information that is likely to be underweighted
in intuitive judgment (e.g., Hirt & Markman, 1995), or
teaches people statistical reasoning and normative rules of
which they may be unaware (e.g., Larrick, Morgan, &
Nisbett, 1990). In large doses, debiasing training can be
effective. Coursework in statistical reasoning, and graduate
training in probabilistic sciences such as psychology and
medicine, does appear to increase the use of statistics and
logic when reasoning about everyday problems to which
they apply (Nisbett, Fong, Lehman, & Cheng, 1987).
132
Persistent Debiasing With a Single
Intervention
We tested whether a single debiasing training intervention
could effectively produce immediate and persistent improvements in decision making. In two experiments, we directly
compared the efficacy of two debiasing training interventions,
a video and an interactive serious (i.e., educational) computer
game. Videos and games are scalable training methods that can
be used for efficient teaching of cognitive skills (e.g., Downs,
2014; Haferkamp, Kraemer, Linehan, & Schembri, 2011;
Sliney & Murphy, 2008). The experiments tested whether debiasing training could produce long-term reductions in six cognitive biases affecting all types of intelligence analysis
(Intelligence Advanced Research Projects Activity, 2011).
Experiment 1 targeted three cognitive biases: bias blind
spot (i.e., perceiving oneself to be less biased than one’s
peers; Scopelliti et al., 2015), confirmation bias (i.e., gathering and interpreting evidence in a manner confirming rather
than disconfirming the hypothesis being tested; Nickerson,
1998), and fundamental attribution error (i.e., attributing the
behavior of a person to dispositional rather than to situational
influences; Gilbert, 1998; Jones & Harris, 1967). Experiment
2 targeted three different cognitive biases: anchoring (i.e.,
overweighting the first information primed or considered in
subsequent judgment; Tversky & Kahneman, 1974), bias
induced by overreliance on representativeness (i.e., using the
similarity of an outcome to a prototypical outcome to judge
its probability; Kahneman & Tversky, 1972), and social projection (i.e., assuming others’ emotions, thoughts, and values
are similar to one’s own; Epley, Morewedge, & Keysar,
2004; Robbins & Krueger, 2005).
Many tasks crucial to intelligence analysis are influenced
by these biases (for a review, see Heuer, 1999). Analysts must
assess evidence with uncertain truth value (e.g., anchoring,
bias blind spot, confirmation bias). They must infer cause and
effect when evaluating past, present, and future events (e.g.,
confirmation bias, representativeness), the behavior of persons, and the actions of nations (e.g., fundamental attribution
error, projection). Analysts regularly estimate probabilities
(e.g., anchoring, confirmation bias, projection bias, representativeness), evaluate their own analyses, and evaluate the analyses of others (e.g., anchoring, bias blind spot, confirmation
bias, projection bias). Although each of these cognitive biases
may have its unique influence, multiple biases are likely to act
in concert in any complex assessment (Cooper, 2005).
Attempting to reduce cognitive biases with videos and
games allowed us to administer short, one-shot training
interventions (i.e., approximately 30 and 60 min, respectively) using two different mixes of the four debiasing training procedures proposed by Fischhoff (1982): (1) teaching
people about each bias, (2) teaching people the directional
influence of each bias on judgment, (3) providing feedback,
and (d) providing extended feedback with coaching, intervention, and mitigating strategies. The videos incorporated
debiasing training procedures 1, 2, and mitigating strategies
Policy Insights from the Behavioral and Brain Sciences 2(1)
(i.e., 4 without feedback, intervention, or coaching) in a passive format. The games incorporated all four debiasing training procedures in an interactive format. Each participant
watched one video or played one game, without repetition.
Each video instructed viewers about three cognitive biases,
gave examples of each bias, and provided mitigating strategies
(e.g., consider alternative explanations, anchors, possible outcomes, perspectives, base rates, countervailing evidence, and
potential situational influences on behavior). Each of the interactive computer games elicited the same three cognitive biases
during gameplay by asking players to make in-game decisions
based on limited evidence (e.g., testing a hypothesis, evaluating the behavior of a character in the game, etc.). In an afteraction review (AAR) at the end of each of three levels of each
game, players were given definitions and examples of the
three biases, personalized feedback on the degree to which
they exhibited each bias, and mitigating strategies and practice. Like the video, the mitigating strategies taught in the
game included the following: consider alternative explanations, anchors, possible outcomes, perspectives, base rates,
countervailing evidence, and consider potential situational
influences on behavior. In addition, the games taught formal
rules of logic (e.g., the conjunction of two events can be no
more likely than either event on its own), methods of hypothesis testing (e.g., hold all variables other than the suspected
causal variable constant when testing a hypothesis), and relevant statistical rules (e.g., large samples are more accurate representations than small samples), as well as encouraging
participants to carefully reconsider their initial answers.
Our experiments tested the immediate and persistent
effects of the debiasing interventions by measuring the extent
to which participants committed each bias 3 times: in a pretest
before training, in a posttest immediately after training, and in
follow-up testing 8 or 12 weeks after training (see Figure 1).
The pretest, training, and posttest were conducted in our laboratory and measured immediate debiasing effects of the training interventions. The follow-up was administered online and
measured the persistent debiasing effects of the training interventions over a longer term. Sample sizes were declared in
advance to our government sponsor, and independent thirdparty analyses of the data were performed that confirmed the
accuracy of our results (J. J. Kopecky, J. A. McKneely, & N.
D. Bos, personal communication, June 22, 2015).
Experiment 1: Bias Blind Spot,
Confirmation Bias, and Fundamental
Attribution Error
Method
Participants. Two hundred seventy-eight people in a convenience sample recruited in Pittsburgh, Pennsylvania (132
women; Mage = 24.5, SD = 8.52) received US$30 for completing
a laboratory training session, and an additional US$30 payment
for completing a follow-up test online. Most (80.2%) participants had some college education; 14.3% had graduate or
133
Morewedge et al.
Figure 1. Overview of procedure for training administration and bias assessments (pretest, posttest, follow-up).
Note. Immediate debiasing effects of training interventions (a game or video) were measured by comparing pretest and posttest scores of bias commission
in a laboratory session. Long-term debiasing effects of training interventions were measured in an online follow-up measuring bias commission 8 or 12
weeks later (Experiments 1 and 2, respectively).
professional degrees. A total of 243 participants successfully
completed the laboratory portion of the experiment (game
n = 160; video n = 83); 196 successfully completed the online
follow-up (game n = 130; video n = 66).1
Training interventions
Video. Unbiasing Your Biases is a 30-min unclassified
training video (produced by Intelligence Advanced Research
Projects Activity, 2012). A narrator first defines heuristics
and explains how heuristics can sometimes lead to incorrect inferences. He then defines bias blind spot, confirmation
bias, and fundamental attribution error; presents vignettes in
which actors commit each bias; gives an additional example
of fundamental attribution error and confirmation bias; and
suggests mitigating strategies. The last 2 min of the video is
a comprehensive review of its content.
Game. Missing: The Pursuit of Terry Hughes is a computer game designed to elicit and mitigate bias blind spot,
confirmation bias, and fundamental attribution error (produced by Symborski et al., 2014). It is a first person pointof-view educational game, in which the player searches for
a missing neighbor (i.e., Terry Hughes) and exonerates her
of criminal activity. During interactive gameplay in each of
three levels, players make judgments designed to test the
degree to which they exhibit confirmation bias and the fundamental attribution error. AARs at the end of each level feature experts explaining each bias and narrative examples. To
elicit bias blind spot, players then assess their degree of bias
during each level. Next, participants are given personalized
feedback on the degree of bias they exhibited. Finally, participants perform additional practice judgments of confirmation
bias (five in total) and receive immediate feedback before the
next level begins or the game ends.2
Bias measures. We developed measures of the extent to which
participants committed each of the three cognitive biases: a
Bias Blind Spot scale (BBS), a Fundamental Attribution Error
scale (FAE), and six Confirmation Bias scales (CB). These
were tested to ensure reliability and validity (see Supplemental
Materials). Three interchangeable version of each scale (i.e.,
subscales) were created to measure bias commission at pretest,
posttest, and follow-up. Scoring of each subscale ranged from
0 (no biased answers) to 100 (all answers biased). Confirmation bias scale scores were calculated by averaging the six
CB scales at pretest, posttest, and follow-up. Overall bias
commission scores at pretest, posttest, and follow-up were
calculated by averaging the three bias subscale scores at that
time point (i.e., BBS, FAE, CB).
Ancillary scales measuring bias knowledge were developed to assess changes in ability to recognize instances of the
three biases and discriminate between them. Bias knowledge
scales were scored on a 0 to 100 scale, with higher scores
indicating greater ability to recognize and discriminate
between the three biases.
Testing procedure. In a laboratory session, each participant
was seated in a private cubicle with a computer. Participants
first completed the pretest measure, consisting of three subscales assessing their commission of each of the three cognitive biases (i.e., BBS, CB, and FAE). Participants also
completed a bias knowledge scale at this time. Next, each
participant was randomly assigned to receive one of the
training interventions, to either play the game or watch the
video, without repetition. Immediately after training, participants completed the posttest measure, consisting of three
subscales assessing their commission of each of the three
cognitive biases posttraining (i.e., BBS, CB, and FAE). Participants also completed a bias knowledge posttest at this
time. To measure the persistence of debiasing training, 8
weeks from the day in which he or she completed the laboratory session, each participant received a personalized link via
email to complete the follow-up measure, consisting of three
subscales assessing his or her commission of each of the
three biases (i.e., BBS, CB, and FAE). He or she had 7 days
to complete the follow-up measure in one sitting. Participants also completed a bias knowledge measure at this time.
The specific bias scales serving as the pretest, posttest, and
follow-up measures of bias commission and bias knowledge
were counterbalanced across participants.
Results
Scale reliability. Subscales were reliable. BBS (Cronbach’s
α): .77pretest, .82 posttest, and .76follow-up. CB: .73pretest, .73 posttest,
and .76follow-up. FAE: .68pretest, .77 posttest, and .78follow-up.
134
Bias commission. Main effects of training on bias commission
overall and for each of the three cognitive biases were analyzed using 2 (training: game vs. video) × 2 (timing: pretest
vs. posttest or pretest vs. follow-up) mixed ANOVAs with
repeated measures on the last factor. To compare the efficacy
of the game and video, between subjects (training: game vs.
video) ANCOVAs were performed to compare the debiasing
effects of the training methods at posttest and follow-up,
controlling for pretest scores. Means of bias commission
scores for overall bias and each of the three biases by training
intervention conditions are presented in Figure 2 (bias
knowledge scores are only reported in the text).
Overall bias. Overall, training effectively reduced cognitive bias immediately and 2 months later, F(1, 241) =
439.23, p < .001, and F(1, 194) = 179.88, p < .001, respectively. Debiasing effect sizes (Rosenthal & Rosnow, 1991)
for overall bias were large for the game (dpre–post = 1.68 and
dpre–follow-up = 1.11) and medium for the video (dpre–post = .69
and dpre–follow-up = .66). The game more effectively debiased
participants than did the video immediately and 2 months
later, F(1, 240) = 68.8, p < .001, and F(1, 193) = 12.69,
p < .001, respectively.
Bias blind spot. Training effectively reduced BBS immediately and 2 months later, F(1, 241) = 151.66, p < .001, and
F(1, 194) = 104.51, p < .001, respectively. Debiasing effect
sizes for BBS were large for the game (dpre–post = .98 and
dpre–follow-up = .89) and medium for the video (dpre–post = .49
and dpre–follow-up = .49). The game more effectively debiased
participants than did the video immediately and 2 months
later, F(1, 240) = 17.31, p < .001, and F(1, 193) = 13.18,
p < .001, respectively.
Fundamental attribution error. Training effectively reduced
FAE immediately and 2 months later, F(1, 241) = 183.74,
p < .001, and F(1, 194) = 85.32, p < .001, respectively. Debiasing effect sizes for FAE were large and medium for the
game (dpre–post = 1.12 and dpre–follow-up = .72) and small and
medium for the video (dpre–post = .38 and dpre–follow-up = .52).
The game more effectively debiased participants than did the
video immediately and 2 months later, F(1, 240) = 50.06,
p < .001, and F(1, 193) = 6.53, p < .05, respectively.
Confirmation bias. Training effectively reduced confirmation bias immediately and 2 months later, F(1, 241) = 181.08,
p < .001, and F(1, 194) = 45.52, p < .001, respectively. Debiasing effect sizes for confirmation bias were large to medium
for the game (dpre–post = 1.09 and dpre–follow-up = .58) and small
for the video (dpre–post = .38 and dpre–follow-up = .26). The game
more effectively debiased participants than did the video
immediately and 2 months later, F(1, 240) = 33.54, p < .001,
and F(1, 193) = 5.17, p < .05, respectively.
Our scales tested six different facets of confirmation bias,
but our game only taught three. This testing structure allowed
Policy Insights from the Behavioral and Brain Sciences 2(1)
us to test the generalization of debiasing training across
trained (Snyder & Swann, 1978; Tschirgi, 1980; Wason,
1960) and untrained facets of confirmation bias (Downs &
Shafir, 1999; Nisbett & Ross, 1980; Wason, 1968). Compared
with their pretest scores, participants exhibited a reduction in
confirmation bias on the trained facets at posttest and followup, t(159) = 9.81, p < .001, d = .78, and t(129) = 2.69, p < .01,
d = .24, respectively. More important, compared with their
pretest scores, participants exhibited reduced confirmation
bias for untrained facets at posttest and follow-up, t(159) =
10.05, p < .001, d = .79, and t(129) = 7.42, p < .001, d = .65,
respectively. Controlling for their pretest scores, participants
performed better on trained than untrained facets of confirmation bias at posttest, t(159) = 2.56, p < .05, d = .20, but
there were no significant differences between trained and
untrained facets at follow-up, t < 1 (for means, see Figure 3).
Bias knowledge. Training also effectively improved bias
knowledge immediately and 2 months later, F(1, 241) =
385.13, p < .001, and F(1, 194) = 64.31, p < .001, respectively. Bias knowledge increased for participants who played
the game (Mpretest = 35.78, Mposttest = 58.54, Mfollow-up = 47.98,
dpre–post = 1.05, and dpre–follow-up = .52) and watched the video
(Mpretest = 35.29, Mposttest = 69.28, Mfollow-up = 50.63, dpre–post =
1.69, and dpre–follow-up = .69). The video more effectively
taught participants to recognize and discriminate bias than
did the game immediately, F(1, 240) = 15.52, p < .001, but
was no more effective 2 months later, F < 1.
Experiment 2: Anchoring, Projection
Bias, and Representativeness
Method
Participants. Two hundred sixty-nine people in a convenience sample recruited in Pittsburgh, Pennsylvania (155
women; Mage = 27.8, SD = 12.01) received US$30 for completing a laboratory training session, and an additional
US$30 payment for completing a follow-up test online. Most
(94.1%) participants had some college education; 19.3% had
graduate or professional degrees. A total of 238 participants
successfully completed the laboratory portion of the experiment (game n = 156; video n = 82); 192 successfully completed the online follow-up (game n = 126; video n = 66).1
Stimuli
Training video. Unbiasing Your Biases 2 (Intelligence
Advanced Research Projects Activity, 2013) had the same
structure as the video in Experiment 1, but addressed anchoring, projection, and representativeness.
Computer game. Missing: The Final Secret is a serious
game designed to elicit and mitigate to anchoring, projection, and representativeness. The game followed a narrative
arc, genre, and structure similar to the game in Experiment 1
135
Morewedge et al.
Figure 2. Bias commission by training intervention in Experiments 1 and 2.
Note. Left and right columns illustrate the mitigating effects of training on bias commission overall and for each of the three cognitive biases in
Experiments 1 and 2, respectively. Scales range from 0 to 100; higher scores indicate more biased answers (95% CI). Both training interventions
effectively debiased participants. Overall, the game more effectively debiased participants than did the video in Experiments 1 and 2. Symbols indicate
statistically significant and marginally significant differences between game and video conditions at posttest and follow-up. CI = confidence interval.
†
p < .10. *p < .05. **p < .01. ***p < .001.
(see Barton et al., 2015). Players exonerate their employer
of a criminal charge and uncover the criminal activity of her
accusers, while making decisions testing their commission
of each of the cognitive biases during gameplay. Experiment
2 introduced adaptive training in the AARs. When players
gave biased answers to practice questions, they received
additional practice questions (up to 16 in total) and feedback.2
136
Policy Insights from the Behavioral and Brain Sciences 2(1)
Debiasing effect sizes for overall bias were large for both the
game (dpre–post = 1.74 and dpre–follow-up = 1.16) and video (dpre–
post = 1.75 and dpre–follow-up = 1.07). However, the game more
effectively debiased participants than did the video immediately, F(1, 235) = 13.44, p < .001, and marginally 3 months
later, F(1, 189) = 3.66, p = .057.
Anchoring. Training effectively reduced anchoring immediately and 3 months later, F(1, 236) = 127.94, p < .001, and F(1,
190) = 78.42, p < .001, respectively. Debiasing effect sizes for
anchoring were medium for the game (dpre–post = .70 and dpre–
follow-up = .63) and large to medium for the video (dpre–post = .80
and dpre–follow-up = .66). The game and video were equally effective immediately and 3 months later, Fs < 1, ps > .62.
Figure 3. Debiasing effects of the game were observed for both
trained and untrained facets of confirmation bias in Experiment
1, suggesting that debiasing effects of training generalized across
domains.
Note. Scales range from 0 to 100, higher scores indicate more bias (95%
CI). Asterisk indicates significant difference between trained and untrained
facets of confirmation bias at posttest, controlling for pretest scores.
CI = confidence interval.
*p < .05.
Scale development. Scales measuring commission of anchoring, projection, and representativeness, and scales measuring
bias knowledge were developed and scored following a procedure similar to that used in Experiment 1 (see Supplemental Materials).
Testing procedure. The experiment adhered to the same testing procedure as described in Experiment 1, with the exception that the follow-up was administered 12 weeks after
participants completed their laboratory session.
Results
Scale reliability. Subscales were reliable: Anchoring (Cronbach’s
α): .60pretest, .52 posttest, and .62follow-up. Projection Bias: .63pretest,
.78 posttest, and .77follow-up. Representativeness: .86pretest, .87 posttest,
and .93follow-up.
Bias commission. The same analyses were performed as in
Experiment 1. All bias commission scale means are presented in Figure 2.
Overall bias. Overall, training effectively reduced cognitive bias immediately and 3 months later, F(1, 236) = 719.58,
p < .001, and F(1, 190) = 246.17, p < .001, respectively.
Projection. Training effectively reduced projection immediately and 3 months later, F(1, 236) = 197.29, p < .001, and
F(1, 190) = 34.52, p < .001, respectively. Debiasing effect
sizes for projection were large to medium for the game (dpre–
post = 1.11 and dpre–follow-up = .54) and medium to small for the
video (dpre–post = .49 and dpre–follow-up = .14). The game more
effectively debiased participants than did the video immediately and 3 months later, F(1, 235) = 34.42, p < .001, and
F(1, 189) = 13.49, p < .001, respectively.
Representativeness. Training effectively reduced bias due
to overreliance on representativeness immediately and 3
months later, F(1, 236) = 599.55, p < .001, and F(1, 190) =
216.36, p < .001, respectively. Debiasing effect sizes for representativeness were large for both the game (dpre–post = 1.51
and dpre–follow-up = 1.05) and video (dpre–post = 1.80 and dpre–
follow-up = 1.09). The game more effectively debiased participants than did the video immediately, F(1, 235) = 10.85, p <
.01, but was no more effective 3 months later, F < 1, p = .37.
Bias knowledge. Training effectively improved bias knowledge immediately and 3 months later, F(1, 236) = 506.52,
p < .001, and F(1, 190) = 216.36, p < .001, respectively. Bias
knowledge increased for participants who played the game
(Mpretest = 35.89, Mposttest = 63.16, Mfollow-up = 50.65, dpre–post =
1.42, and dpre–follow-up = 1.05) and watched the video (Mpretest =
39.03, Mposttest = 74.11, Mfollow-up = 52.04, dpre–post = 1.53, and
dpre–follow-up = 1.09). The video more effectively taught participants to recognize and discriminate bias than did the
game immediately, F(1, 235) = 11.07, p < .001, but was no
more effective 3 months later, F < 1.
Conclusions and Recommendations
People generally intend to make decisions that are in their
own and society’s best interest, but biases in judgment and
decision making often lead them to make costly errors. More
than 40 years of judgment and decision-making research
suggests feasible interventions to debias and improve decision making (Bhargava & Loewenstein, 2015; Fischhoff,
137
Morewedge et al.
1982; Fox & Sitkin, 2015; Soll et al., in press). This research
and its methods can be used to align incentives, present
information, elicit choices, and educate people so they are
better able to make good decisions.
Debiasing interventions are not, by default, coercive.
Presenting information in a manner in which options are
easier to evaluate generally improves decisions by making
people better able to evaluate those options along the dimensions that are important to them. Commuting ranks among
the most unpleasant daily experiences (Kahneman, Krueger,
Schkade, Schwarz, & Stone, 2004), for instance, but people
are relatively insensitive to the duration of a prospective
commute unless they are provided with a familiar comparison standard (Morewedge, Kassam, Hsee, & Caruso, 2009).
Decisions usually have some underlying structure that
biases the decision making process or its outcome. For some
decisions such as whether to be an organ donor, one option
must be specified as the default even if one defers the decision. Selecting a default option that is beneficial for the decision maker or society can improve the public good while
preserving freedom of choice (Camerer et al., 2003; Thaler
& Sunstein, 2003, 2008). Furthermore, people actively seek
out many kinds of debiasing interventions such as timesaving recommendation systems (Goldstein et al., 2008) and
commitment devices to give them the willpower to make
choices that are unappealing in the present but will benefit
them more in the long-term (e.g., Schwartz et al., 2014;
Thaler & Benartzi, 2004).
Debiasing interventions are not, by default, more costly
than the status quo. New incentives do not have to impose a
financial cost to taxpayers or decision makers. Social influence is an underutilized but powerful nonpecuniary motive
for positive behavior change, for instance, that can produce
significant reductions in environmental waste and energy
consumption (Cialdini, 2003; Schultz, Nolan, Cialdini,
Goldstein, & Griskevicius, 2007). Moreover, existing incentives are only effective if they motivate behavior as they
were intended. If incentives are misaligned, misinterpreted,
or poorly framed, they may be costly and ineffective or
counterproductive.
Small changes in message framing and choice elicitation
can produce debiasing effects for little additional cost. In two
laboratory studies, simply framing an economic stimulus as
a “bonus” rather than a “rebate” more than doubled how
much of that stimulus was spent (Epley, Mak, & Idson,
2006). In a field study run in the United Kingdom, adding a
single sentence to late tax notices that truthfully stated the
majority of U.K. citizens pay their taxes on time increased
the clearance rate of late payers to 86% (£560 million out of
£650 million owed), compared with a clearance rate of 57%
the previous year (£290 million out of £510 million owed;
Cialdini, Martin, & Goldstein, 2015).
Training interventions have an upfront production cost, but
the marginal financial and temporal costs of training many
additional people are minimal. The results of our experiments
suggest that even a single training intervention, such as the
games and videos we tested in this article, can have significant debiasing effects that persist across a variety of contexts
affected by the same bias. Participants who played our games
exhibited large reductions in cognitive bias immediately
(−46.25% and −31.94%), which persisted at least 2 or 3
months later (−34.76% and −23.57%) in Experiments 1 and
2, respectively. Participants who watched the videos exhibited medium and large reductions immediately (−18.60%
and −25.70%), which persisted at least 2 or 3 months later
(−20.10% and −19.20%) in Experiments 1 and 2, respectively. The greater efficacy of the games than the videos suggest that personal feedback and practice increase the
debiasing effects of training, but more research is needed to
determine precisely why it was more effective. Most important, these results suggest that despite its rocky start
(Fischhoff, 1982), training is a promising avenue through
which to develop future debiasing interventions.
Decision research is in an exciting phase of expansion,
from important basic research that identifies and elucidates
biases to the development and testing of practical interventions. Laboratory experiments provide a safe and inexpensive microcosm in which to uncover new biases, develop
new theories, and test new interventions. Many researchers
now test successful laboratory interventions and their extensions in larger field experiments, such as randomized controlled trials, to determine which biases and interventions are
most influential in particular contexts (Haynes, Service,
Goldacre, & Torgerson, 2012). This work extends outside the
ivory tower. Researchers have produced numerous successful collaborations with government and industry partners
that have reduced waste and improved the health and finances
of the public (e.g., Chapman et al., 2010; Mellers et al., 2014;
Schultz et al., 2007; Schwartz et al., 2014; Thaler & Benartzi,
2004). Ad hoc collaborations and targeted programs, such as
the development and testing of training inventions that we
report, have been very successful (see also Mellers et al.,
2014). Several countries have even established panels of
behavioral scientists to develop interventions from within
government, such the Social and Behavioral Sciences Team
in the United States.
Decision making is pervasive in professional and everyday life. Its study and improvement can contribute much to
the public good.
Notes
1.
Participants were excluded before analyses in Experiment 1
because they played early game prototypes (n = 20), experienced game crashes (n = 3) and server errors during scale
administration (n = 6), or were unable to finish the laboratory
session in 4 hr (n = 6). In addition, those who did not complete
the follow-up test within 7 days of receiving notification were
not included in follow-up analyses (n = 47). Participants were
excluded before analyses in Experiment 2 because of game
crashes (n = 1), experimenter or participant error (n = 3), or
138
2.
Policy Insights from the Behavioral and Brain Sciences 2(1)
failed attention checks (n = 27). In addition, those who did not
complete the follow-up within 7 days of receiving notification
were not included in follow-up analyses (n = 45).
There were four variants of the Experiment 1 game, including
whether a game score or narrative examples were included or
excluded in the AARs. Moreover, half of the participants in the
game condition played the full game, and half played only the
first round. We did not observe a significant difference across
these game and methodological variations in their reduction of
overall bias at posttest and follow-up, Fs ≤ 2.29, ps ≥ .13. In
Experiment 2, all players completed the whole game, but there
were four variants including whether hints or game scores
were provided. We did not observe a significant difference
across these variants in their reduction of overall bias at posttest and follow-up, all ts ≤ 1.77, ps ≥ .08. In both experiments,
we report the results collapsed across these variations.
Acknowledgment
The authors thank Marguerite Barton, Abigail Dawson, Sophie
LeBrecht, Erin McCormick, Peter Mans, H. Lauren Min, Taylor
Turrisi, and Shane Schweitzer for their assistance with the execution of this research.
Authors’ Note
The views and conclusions contained herein are those of the authors
and should not be interpreted as necessarily representing the official
policies or endorsements, either expressed or implied, of the
Intelligence Advanced Research Projects Activity (IARPA), the Air
Force Research Laboratory (AFRL), or the U.S. Government.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect
to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support
for the research, authorship, and/or publication of this article: This
work was supported by the Intelligence Advanced Research
Projects Activity via the Air Force Research Laboratory Contract
Number FA8650-11-C-7175.
References
Acquisti, A., Brandimarte, L., & Loewenstein, G. (2015). Privacy
and human behavior in the age of information. Science, 347,
509-514.
Ariely, D., Gneezy, U., Loewenstein, G., & Mazar, N. (2009). Large
stakes and big mistakes. The Review of Economic Studies, 76,
451-469.
Arkes, H. R. (1991). Costs and benefits of judgment errors:
Implications for debiasing. Psychological Bulletin, 110, 486-498.
Arkes, H. R., & Blumer, C. (1985). The psychology of sunk cost.
Organizational Behavior and Human Decision Processes, 35,
124-140.
Barnett, S. M., & Ceci, S. J. (2002). When and where do we apply
what we learn? A taxonomy for far transfer. Psychological
Bulletin, 128, 612-637.
Barton, M., Symborski, C., Quinn, M., Morewedge, C. K., Kassam,
K. S., & Korris, J. H. (2015, May). The use of theory in designing a serious game for the reduction of cognitive biases. Digital
Games Research Association Conference, Lüneberg, Germany.
Bhargava, S., & Loewenstein, G. (2015). Behavioral economics
and public policy 102: Beyond nudging. American Economic
Review, 105, 396-401.
Bhargava, S., & Manoli, D. (in press). Psychological frictions and
incomplete take-up of social benefits: Evidence from an IRS
field experiment. American Economic Review.
Bruine de Bruin, W., Parker, A. M., & Fischhoff, B. (2007).
Individual differences in adult decision-making competence.
Journal of Personality and Social Psychology, 92, 938-956.
Camerer, C., Issacharoff, S., Loewenstein, G., O’donoghue, T., &
Rabin, M. (2003). Regulation for conservatives: Behavioral
economics and the case “for asymmetric paternalism.”
University of Pennsylvania Law Review, 151, 1211-1254.
Chapman, G. B., Li, M., Colby, H., & Yoon, H. (2010). Opting in
vs opting out of influenza vaccination. Journal of the American
Medical Association, 304, 43-44.
Charness, G., & Gneezy, U. (2009). Incentives to exercise.
Econometrica, 77, 909-931.
Cialdini, R. B. (2003). Crafting normative messages to protect the
environment. Current Directions in Psychological Science, 12,
105-109.
Cialdini, R. B., Martin, S. J., & Goldstein, N. J. (2015). Small
behavioral science-informed changes can produce large policyrelevant effects. Behavioral Science & Policy, 1, 21-27.
Cooper, J. R. (2005). Curing analytic pathologies: Pathways to
improved intelligence analysis. Washington, DC: Center for
the Study of Intelligence, Central Intelligence Agency.
Dawes, R. M., Faust, D., & Meehl, P. E. (1989). Clinical versus
actuarial judgment. Science, 243, 1668-1674.
Dhami, M. K., Schlottmann, A., & Waldmann, M. R. (2011).
Judgment and decision making as a skill: Learning, development and evolution. Cambridge, UK: Cambridge University
Press.
Downs, J. S. (2014). Prescriptive scientific narratives for communicating usable science. Proceedings of the National Academy of
Sciences, 111, 13627-13633.
Downs, J. S., Loewenstein, G., & Wisdom, J. (2009). Strategies
for promoting healthier food choices. The American Economic
Review, 99, 159-164.
Downs, J. S., & Shafir, E. (1999). Why some are perceived as more
confident and more insecure, more reckless and more cautious, more trusting and more suspicious, than others: Enriched
and impoverished options in social judgment. Psychonomic
Bulletin & Review, 6, 598-610.
Epley, N., Mak, D., & Idson, L. C. (2006). Bonus of rebate? The
impact of income framing on spending and saving. Journal of
Behavioral Decision Making, 19, 213-227.
Epley, N., Morewedge, C. K., & Keysar, B. (2004). Perspective taking in children and adults: Equivalent egocentrism but differential correction. Journal of Experimental Social Psychology,
40, 760-768.
Evans, J. S. B. (2003). In two minds: Dual-process accounts of reasoning. Trends in Cognitive Sciences, 7, 454-459.
Fischhoff, B. (1982). Debiasing. In D. Kahneman, P. Slovic, & A.
Tversky (Eds.), Judgment under uncertainty: Heuristics and
Morewedge et al.
biases (pp. 422-444). Cambridge, UK: Cambridge University
Press.
Fox, C. R., & Sitkin, S. B. (2015). Bridging the divide between
behavioral science and policy. Behavioral Science & Policy,
1, 1-12.
French, S. A. (2003). Pricing effects on food choices. The Journal
of Nutrition, 133, 841S-843S.
Garcia-Retamero, R., & Dhami, M. K. (2011). Pictures speak
louder than numbers: On communicating medical risks to
immigrants with limited non-native language proficiency.
Health Expectations, 14, 46-57.
Garcia-Retamero, R., & Dhami, M. K. (2013). On avoiding framing
effects in experienced decision makers. The Quarterly Journal
of Experimental Psychology, 66, 829-842.
Gerber, A. S., & Rogers, T. (2009). Descriptive social norms and
motivation to vote: Everybody’s voting and so should you. The
Journal of Politics, 71, 178-191.
Gilbert, D. T. (1998). Ordinary personology. In D. T. Gilbert, S. T.
Fiske, & G. Lindsey (Eds.), The handbook of social psychology
(Vol. 2, pp. 89-150). Hoboken, NJ: John Wiley & Sons, Inc.
Gneezy, U., Meier, S., & Rey-Biel, P. (2011). When and why
incentives (don’t) work to modify behavior. The Journal of
Economic Perspectives, 25, 191-209.
Gneezy, U., & Rustichini, A. (2000). A fine is a price. Journal of
Legal Studies, 29, 1-17.
Goldstein, D. G., Johnson, E. J., Herrmann, A., & Heitmann, M.
(2008). Nudge your customers toward better choices. Harvard
Business Review, 86, 99-105.
Haferkamp, N., Kraemer, N. C., Linehan, C., & Schembri, M.
(2011). Training disaster communication by means of serious
games in virtual environments. Entertainment Computing, 2,
81-88.
Harvey, N. (2011). Learning judgment and decision making from
feedback. In M. K. Dhami, A. Schlottmann, & M. R. Waldmann
(Eds.), Judgment and decision making as a skill: Learning,
development, and evolution (pp. 199-226). Cambridge, UK:
Cambridge University Press.
Haynes, L., Service, O., Goldacre, B., & Torgerson, D. (2012).
Test, learn, adapt: Developing public policy with randomized
controlled trials. London: Cabinet Office Behavioural Insights
Team.
Heuer, R. J. (1999). Psychology of intelligence analysis.
Washington, DC: Center for the Study of Intelligence, Central
Intelligence Agency.
Hirt, E. R., & Markman, K. D. (1995). Multiple explanation: A consider-an-alternative strategy for debiasing judgments. Journal
of Personality and Social Psychology, 69, 1069-1086.
Huh, Y. E., Vosgerau, J., & Morewedge, C. K. (2014). Social
defaults: Observed choices become choice defaults. Journal of
Consumer Research, 41, 746-760.
Intelligence Advanced Research Projects Activity. (2011). Sirius Broad
Agency Announcement, IARPA-BAA-11-03. Retrieved from http://
www.iarpa.gov/index.php/research-programs/sirius/baa
Intelligence Advanced Research Projects Activity. (2012).
Unbiasing your biases I. Alexandria, VA: 522 Productions.
Intelligence Advanced Research Projects Activity. (2013).
Unbiasing your biases II. Alexandria, VA: 522 Productions.
John, L. K., Loewenstein, G., Troxel, A. B., Norton, L., Fassbender,
J. E., & Volpp, K. G. (2011). Financial incentives for extended
139
weight loss: A randomized, controlled trial. Journal of General
Internal Medicine, 26, 621-626.
Johnson, E. J., & Goldstein, D. G. (2003). Do defaults save lives?
Science, 302, 1338-1339.
Jones, E. E., & Harris, V. A. (1967). The attribution of attitudes.
Journal of Experimental Social Psychology, 3, 1-24.
Kahneman, D., Krueger, A. B., Schkade, D. A., Schwarz, N., &
Stone, A. A. (2004). A survey method for characterizing daily
life experience: The day reconstruction method. Science, 306,
1776-1780.
Kahneman, D., & Tversky, A. (1972). Subjective probability: A
judgement of representativeness. Cognitive Psychology, 3,
430-454.
Kohler, D. J., Brenner, L., & Griffin, D. (2002). The calibration of
expert judgment: Heuristics and biases beyond the laboratory.
In T. Gilovich, D. Griffin, & D. Kahneman (Eds.), Heuristics
and biases: The psychology of intuitive judgment (pp. 686715). New York, NY: Cambridge University Press.
Krieger, L. H., & Fiske, S. T. (2006). Behavioral realism in employment discrimination law: Implicit bias and disparate treatment.
California Law Review, 94, 997-1062.
Larrick, R. P., Morgan, J. N., & Nisbett, R. E. (1990). Teaching the
use of cost-benefit reasoning in everyday life. Psychological
Science, 1, 362-370.
Larrick, R. P., & Soll, J. B. (2008). The MPG illusion. Science, 320,
1593-1594.
Lerner, J. S., & Tetlock, P. E. (1999). Accounting for the effects of
accountability. Psychological Bulletin, 125, 255-275.
Levin, I. P., & Gaeth, G. J. (1988). How consumers are affected by
the framing of attribute information before and after consuming the product. Journal of Consumer Research, 15, 374-378.
Lewandowsky, S., Ecker, U. K., Seifert, C. M., Schwarz, N., &
Cook, J. (2012). Misinformation and its correction continued
influence and successful debiasing. Psychological Science in
the Public Interest, 13, 106-131.
Lilienfeld, S. O., Ammirati, R., & Landfield, K. (2009). Giving
debiasing away: Can psychological research on correcting
cognitive errors promote human welfare? Perspectives on
Psychological Science, 4, 390-398.
McNeil, B. J., Pauker, S. G., Sox, H. C., Jr., & Tversky, A. (1982).
On the elicitation of preferences for alternative therapies. New
England Journal of Medicine, 306, 1259-1262.
Mellers, B., Ungar, L., Baron, J., Ramos, J., Gurcay, B., Fincher,
K., . . . Tetlock, P. E. (2014). Psychological strategies for
winning a geopolitical forecasting tournament. Psychological
Science, 25, 1106-1115.
Milkman, K. L., Chugh, D., & Bazerman, M. H. (2009). How can
decision making be improved? Perspectives on Psychological
Science, 4, 379-383.
Morewedge, C. K., & Kahneman, D. (2010). Associative processes in intuitive judgment. Trends in Cognitive Sciences, 14,
435-440.
Morewedge, C. K., Kassam, K. S., Hsee, C. K., & Caruso, E. M.
(2009). Duration sensitivity depends on stimulus familiarity.
Journal of Experimental Psychology: General, 138, 177-186.
Murphy, A. H., & Winkler, R. L. (1974). Subjective probability
forecasting experiments in meteorology: Some preliminary
results. Bulletin of the American Meteorological Society, 55,
1206-1216.
140
Nickerson, R. S. (1998). Confirmation bias: A ubiquitous phenomenon in many guises. Review of General Psychology, 2,
175-220.
Nisbett, R. E., Fong, G. T., Lehman, D. R., & Cheng, P. W. (1987).
Teaching reasoning. Science, 238, 625-631.
Nisbett, R. E., & Ross, L. (1980). Human inference: Strategies
and shortcomings of social judgment. Englewood Cliffs, NJ:
Prentice Hall.
Parker, A. M., & Fischhoff, B. (2005). Decision-making competence: External validation through an individual-differences
approach. Journal of Behavioral Decision Making, 18, 1-27.
Payne, J. W., Bettman, J. R., & Johnson, E. J. (1993). The adaptive
decision maker. Cambridge, UK: Cambridge University Press.
Peters, E., & de Bruin, W. B. (2011). Aging and decision skills.
In M. K. Dhami, A. Schlottmann, & M. R. Waldmann (Eds.),
Judgment and decision making as a skill: Learning, development, and evolution (pp. 113-140). Cambridge, UK: Cambridge
University Press.
Phillips, J. K., Klein, G., & Sieck, W. R. (2004). Expertise in judgment and decision making: A case for training intuitive decision skills. In D. J. Koehler & N. Harvey (Eds.), Blackwell
handbook of judgment and decision making (pp. 297-315).
Malden, MA: Blackwell Publishing, Ltd.
Reeves, L., & Weisberg, R. W. (1994). The role of content and
abstract information in analogical transfer. Psychological
Bulletin, 115, 381-400.
Robbins, J. M., & Krueger, J. I. (2005). Social projection to ingroups
and outgroups: A review and meta-analysis. Personality and
Social Psychology Review, 9, 32-47.
Rosenthal, R., & Rosnow, R. L. (1991). Essentials of behavioral research: Methods and data analysis. New York, NY:
McGraw-Hill.
Schultz, P. W., Nolan, J. M., Cialdini, R. B., Goldstein, N. J., &
Griskevicius, V. (2007). The constructive, destructive, and
reconstructive power of social norms. Psychological Science,
18, 429-434.
Schwartz, J., Mochon, D., Wyper, L., Maroba, J., Patel, D., &
Ariely, D. (2014). Healthier by precommitment. Psychological
Science, 25, 538-546.
Schwitzgebel, E., & Cushman, F. (2012). Expertise in moral reasoning? Order effects on moral judgment in professional philosophers and non-philosophers. Mind & Language, 27, 135-153.
Scopelliti, I., Morewedge, C. K., McCormick, E., Min, H. L.,
Lebrecht, S., & Kassam, K. S. (2015). Bias blind spot:
Structure, measurement, and consequences. Management
Science. doi:10.1287/mnsc.2014.2096
Shefrin, H., & Statman, M. (1985). The disposition to sell winners too early and ride losers too long: Theory and evidence.
Journal of Finance, 40, 777-790.
Silberman, L. H., Robb, C. S., Levin, R. C., McCain, J., Rowan,
H. S., Slocombe, W. B., . . . Cutler, L. (2005, March 31).
Report to the President of the United States. Washington, DC:
Policy Insights from the Behavioral and Brain Sciences 2(1)
Commision on Intelligence Capabilities of the United States
Regarding Weapons of Mass Destruction.
Sliney, A., & Murphy, D. (2008, February). JDoc: A serious game
for medical learning. Proceedings of the First International
Conference on Advances in Computer-Human Interaction (ACHI2008), Sainte Luce, Martinique. doi:10.1109/ACHI.2008.50
Sloman, S. A. (1996). The empirical case for two systems of reasoning. Psychological Bulletin, 119, 3-22.
Snyder, M., & Swann, W. B. (1978). Hypothesis-testing processes in social interaction. Journal of Personality and Social
Psychology, 36, 1202-1212.
Soll, J., Milkman, K., & Payne, J. (in press). A user’s guide to debiasing. In G. Keren & G. Wu (Eds.), Wiley-Blackwell handbook
of judgment and decision making. New York, NY: Blackwell.
Symborski, C., Barton, M., Quinn, M., Morewedge, C. K., Kassam,
K., & Korris, J. (2014, December). Missing: A serious game for
the mitigation of cognitive bias. Interservice/Industry Training,
Simulation and Education Conference, Orlando, FL.
Thaler, R. H., & Benartzi, S. (2004). Save more tomorrow™: Using
behavioral economics to increase employee saving. Journal of
Political Economy, 112, S164-S187.
Thaler, R. H., & Sunstein, C. R. (2003). Libertarian paternalism.
American Economic Review, 93, 175-179.
Thaler, R. H., & Sunstein, C. R. (2008). Nudge. New Haven, CT:
Yale University Press.
Tschirgi, J. E. (1980). Sensible reasoning: A hypothesis about
hypotheses. Child Development, 51, 1-10.
Tversky, A., & Kahneman, D. (1974). Judgment under uncertainty:
Heuristics and biases. Science, 185, 1124-1131.
Volpp, K. G., Loewenstein, G., Troxel, A. B., Doshi, J., Price, M.,
Laskin, M., & Kimmel, S. E. (2008). A test of financial incentives to improve warfarin adherence. BMC Health Services
Research, 8, Article 272.
Volpp, K. G., Troxel, A. B., Pauly, M. V., Glick, H. A., Puig, A.,
Asch, D. A., . . . Audrain-McGovern, J. (2009). A randomized,
controlled trial of financial incentives for smoking cessation.
New England Journal of Medicine, 360, 699-709.
Wagenaar, W. A., & Keren, G. B. (1986). Does the expert know?
The reliability of predictions and confidence ratings of experts.
In E. Hollnagel, G. Mancini, & D. D. Woods (eds.) Intelligent
decision support in process environments (pp. 87-103). Berlin,
Germany: Springer.
Wason, P. C. (1960). On the failure to eliminate hypotheses in a conceptual task. Quarterly Journal of Experimental Psychology,
12, 129-140.
Wason, P. C. (1968). Reasoning about a rule. The Quarterly Journal
of Experimental Psychology, 20, 273-281.
Willingham, D. T. (2008). Critical thinking: Why is it so hard to
teach? Arts Education Policy Review, 109, 21-32.
Wilson, T. D., & Brekke, N. (1994). Mental contamination and
mental correction: Unwanted influences on judgments and
evaluations. Psychological Bulletin, 116, 117-142.