ArticlePDF Available

Abstract and Figures

People exhibit a bias blind spot: they are less likely to detect bias in themselves than in others. We report the development and validation of an instrument to measure individual differences in the propensity to exhibit the bias blind spot that is unidimensional, internally consistent, has high test-retest reliability, and is discriminated from measures of intelligence, decision-making ability, and personality traits related to self-esteem, self-enhancement, and self-presentation. The scale is predictive of the extent to which people judge their abilities to be better than average for easy tasks and worse than average for difficult tasks, ignore the advice of others, and are responsive to an intervention designed to mitigate a different judgmental bias. These results suggest that the bias blind spot is a distinct metabias resulting from naïve realism rather than other forms of egocentric cognition, and has unique effects on judgment and behavior. This paper was accepted by Yuval Rottenstreich, judgment and decision making.
Content may be subject to copyright.
This article was downloaded by: [4.28.153.60] On: 28 April 2015, At: 06:23
Publisher: Institute for Operations Research and the Management Sciences (INFORMS)
INFORMS is located in Maryland, USA
Management Science
Publication details, including instructions for authors and subscription information:
http://pubsonline.informs.org
Bias Blind Spot: Structure, Measurement, and
Consequences
Irene Scopelliti, Carey K. Morewedge, Erin McCormick, H. Lauren Min, Sophie Lebrecht,
Karim S. Kassam
To cite this article:
Irene Scopelliti, Carey K. Morewedge, Erin McCormick, H. Lauren Min, Sophie Lebrecht, Karim S. Kassam (2015) Bias Blind
Spot: Structure, Measurement, and Consequences. Management Science
Published online in Articles in Advance 24 Apr 2015
. http://dx.doi.org/10.1287/mnsc.2014.2096
Full terms and conditions of use: http://pubsonline.informs.org/page/terms-and-conditions
This article may be used only for the purposes of research, teaching, and/or private study. Commercial use
or systematic downloading (by robots or other automatic processes) is prohibited without explicit Publisher
approval, unless otherwise noted. For more information, contact permissions@informs.org.
The Publisher does not warrant or guarantee the article’s accuracy, completeness, merchantability, fitness
for a particular purpose, or non-infringement. Descriptions of, or references to, products or publications, or
inclusion of an advertisement in this article, neither constitutes nor implies a guarantee, endorsement, or
support of claims made of that product, publication, or service.
Copyright © 2015, INFORMS
Please scroll down for article—it is on subsequent pages
INFORMS is the largest professional society in the world for professionals in the fields of operations research, management
science, and analytics.
For more information on INFORMS, its publications, membership, or meetings visit http://www.informs.org
MANAGEMENT SCIENCE
Articles in Advance, pp. 1–19
ISSN 0025-1909 (print) ISSN 1526-5501 (online) http://dx.doi.org/10.1287/mnsc.2014.2096
© 2015 INFORMS
Bias Blind Spot: Structure,
Measurement, and Consequences
Irene Scopelliti
Cass Business School, City University London, London EC1Y 8TZ, United Kingdom, irene.scopelliti@city.ac.uk
Carey K. Morewedge
Questrom School of Business, Boston University, Boston, Massachusetts 02215, morewedg@bu.edu
Erin McCormick
Dietrich School of Humanities and Social Sciences, Carnegie Mellon University, Pittsburgh, Pennsylvania 15213,
enmccorm@andrew.cmu.edu
H. Lauren Min
Leeds School of Business, University of Colorado, Boulder, Colorado 80309, lauren.min@colorado.edu
Sophie Lebrecht, Karim S. Kassam
Dietrich School of Humanities and Social Sciences, Carnegie Mellon University, Pittsburgh, Pennsylvania 15213
{sophielebrecht@cmu.edu,kskassam@andrew.cmu.edu}
People exhibit a bias blind spot: they are less likely to detect bias in themselves than in others. We report
the development and validation of an instrument to measure individual differences in the propensity to
exhibit the bias blind spot that is unidimensional, internally consistent, has high test-retest reliability, and is
discriminated from measures of intelligence, decision-making ability, and personality traits related to self-esteem,
self-enhancement, and self-presentation. The scale is predictive of the extent to which people judge their abilities
to be better than average for easy tasks and worse than average for difficult tasks, ignore the advice of others,
and are responsive to an intervention designed to mitigate a different judgmental bias. These results suggest
that the bias blind spot is a distinct metabias resulting from naïve realism rather than other forms of egocentric
cognition, and has unique effects on judgment and behavior.
Keywords: bias blind spot; judgment and decision making; metacognition; self-awareness; advice taking;
debiasing
History : Received December 5, 2013; accepted September 13, 2014, by Yuval Rottenstreich, judgment and
decision making. Published online in Articles in Advance.
Introduction
People exhibit numerous systematic biases in judg-
ment (Tversky and Kahneman 1974,Kahneman et al.
1982,Nisbett and Ross 1980), many of which are due
to unconscious processes (Morewedge and Kahneman
2010,Wilson and Brekke 1994). A lack of conscious
access to judgment-forming processes means that
people are often unaware of their own biases (Nisbett
and Wilson 1977) even though they can readily spot
the same biases in the judgments of others. Conse-
quently, most people tend to believe that, on average,
they are less biased in their judgment and behavior
than are their peers. Most people recognize that other
people are likely to be biased when judging an attrac-
tive person, for example, but think that their own
judgment of an attractive person is unaffected by this
type of halo effect. Because the majority of people
cannot be less biased their peers, this phenomenon is
referred to as the bias blind spot (Pronin 2007;Pronin
et al. 2002,2004).
We show that people differ in their propensity
to exhibit the bias blind spot, and develop a reli-
able and valid instrument to measure this individ-
ual difference. Bias blind spot appears to be a unique
construct independent of intelligence and personal-
ity traits related to self-esteem, self-enhancement, and
self-presentation. Bias blind spot appears to be inde-
pendent of general decision-making competence. In
other words, the belief that one is less biased than
one’s peers appears to reflect a biased perception of
the self rather than a more general inferior decision-
making ability, or an accurate reflection of one’s supe-
rior judgment and decision-making ability. We find
that the propensity to believe that one is less biased
than one’s peers has detrimental consequences on
judgments and behaviors that rely on self–other accu-
racy comparisons, including decreasing receptivity to
useful (i.e., debiasing) advice.
Bias Blind Spot
People exhibit a tendency to believe they are less
biased than their peers. College students believe
1
Downloaded from informs.org by [4.28.153.60] on 28 April 2015, at 06:23 . For personal use only, all rights reserved.
Scopelliti et al.: Bias Blind Spot: Structure, Measurement, and Consequences
2Management Science, Articles in Advance, pp. 1–19, © 2015 INFORMS
they are less biased than their classmates, airline
passengers believe they are less biased than other pas-
sengers, and Americans believe they are less biased
than their fellow citizens (Pronin et al. 2002). This com-
mon asymmetry in the assessment of bias in the self
and in others has been observed across a variety of
social and cognitive biases. This bias blind spot occurs
whether people explicitly rate the extent of their bias
relative to their peers, or rate the absolute extent of
their bias and separately rate the extent of bias exhib-
ited by their peers (Pronin et al. 2002,2004;Pronin
and Kugler 2007;Epley and Dunning 2000;West et al.
2012,Van Boven et al. 2000;Wilson et al. 1996).
The bias blind spot has been attributed to the inter-
play of two phenomena: the introspection illusion and
naïve realism (Pronin et al. 2004,Pronin and Kugler
2007). The introspection illusion results from differ-
ences in the availability and perceived diagnostic
value of introspection when assessing oneself and
other people. When people evaluate the extent of
bias in their own behavior, they base their evalua-
tions on introspection (e.g., “I find Rodger attractive,
but I have no memory of that fact influencing how
I judge his speaking ability”). Because introspection
is unlikely to reveal biased thought processes, people
conclude that their judgment and behavior are unbi-
ased. Naïve realism, the belief that one’s perception
reflects the true state of the world, then generates a
false sense that these charitable self-assessments are
genuine rather than positively biased (Pronin et al.
2004;Ross and Ward 1995,1996).
When people evaluate the extent of bias in others,
their assessments rely on behavior rather than on pri-
vate thoughts, because the private thoughts of others
are not accessible. Consequently, the biased behav-
ior of others is not excused by an ostensibly unbi-
ased thought process (e.g., “Joan might not realize
it, but she probably thinks Rodger speaks eloquently
because she is attracted to him”). Because people use
different types of evidence when assessing bias in the
self and in other people, less bias is perceived in the
self than in others. The variety of social and cogni-
tive biases for which this bias blind spot has been
observed (Pronin et al. 2004,West et al. 2012) sug-
gests that susceptibility to the bias blind spot may be
a higher-order latent factor underlying the belief that
one is less likely to exhibit a variety of specific biases
than are one’s peers.
Bias Blind Spot and Decision Making
The propensity to exhibit decision-making biases
appears to vary systematically (Levin et al. 2002,
Stanovich 1999,West and Stanovich 1997). Several
studies have shown that performance across decision-
making tasks tends to have high internal consistency
(Blais et al. 2005;Bornstein and Zickafoose 1999;
Klayman et al. 1999;Stankov and Crawford 1996,
1997;Stanovich and West 2000). Bruine de Bruin
et al. (2007) demonstrated that decision-making com-
petence, measured as composite performance on mul-
tiple decision-making tasks, reflects the outcome of
reliable individual differences and predicts real-world
decision-making ability.
In addition to clarifying the robustness and unique-
ness of the bias blind spot, a major goal of the present
research was to examine its relationship with gen-
eral decision-making ability. If a single higher-order
construct determines the extent to which people rec-
ognize their own bias, three possible relationships
between the bias blind spot and general decision-
making ability emerge.
First, people who are lowest in decision-making
ability may be least aware of their own bias, and
hence may be most likely to exhibit the bias blind
spot. Indeed, the least skilled are often least able to
assess their level of skill (Kruger and Dunning 1999,
Dunning et al. 2003). If bias blind spot is simply one
component of general decision-making ability, suscep-
tibility to the bias blind spot should be negatively
correlated with decision-making competence.
Second, people may accurately assess their degree
of bias relative to their peers. In other words, peo-
ple who believe they are less biased than their peers
(those high in bias blind spot) may be correct. In this
case, bias blind spot may not really reflect suscep-
tibility to a blind spot at all, and there should be
a positive correlation between bias blind spot and
decision-making competence.
Third, susceptibility to the bias blind spot is unre-
lated to general decision-making ability. People who
have not been trained in decision making, and are
therefore unaware of common biases in reasoning
and judgment, show considerable variation in their
decision-making competence (Bruine de Bruin et al.
2007). This suggests that superior decision making
is a function of calibrated intuitions rather than
awareness of appropriate decision-making strategies.
Indeed, research on the limitations of introspection
shows that people have limited insight into the pro-
cesses by which their judgments and decisions are
made (Nisbett and Wilson 1977,Wilson and Dunn
2004). Furthermore, tests comparing bias blind spot
for several specific biases and the commission of
those specific biases have revealed little to no rela-
tion between recognition of bias and its commission
(Stanovich and West 2008,West et al. 2012).
Predictive Value of the Bias Blind Spot Scale
We suggest that the bias blind spot is weakly, if at
all, related to general decision-making ability. More
specifically, that the bias blind spot is a metabias
a bias in the recognition of other biases—that sys-
tematically biases judgment and behavior in unique
Downloaded from informs.org by [4.28.153.60] on 28 April 2015, at 06:23 . For personal use only, all rights reserved.
Scopelliti et al.: Bias Blind Spot: Structure, Measurement, and Consequences
Management Science, Articles in Advance, pp. 1–19, © 2015 INFORMS 3
and predictable ways. Rather than indicate an overly
positive view of the self, it reflects the tendency to
consider different evidence when making self–other
accuracy comparisons. Thus, we predict that it should
(a) serve as an index of bias for self–other com-
parisons that rely on similar cognitive mechanisms,
whether those comparisons are favorable or unfavor-
able to the self. As an index of self–other accuracy
comparisons, we suggest that bias blind spot should
also influence judgments informed by these (social)
comparisons such as (b) the weight placed on advice
from others and (c) receptivity to debiasing training.
Next, we unpack each of these predictions in turn.
A first prediction is that when evaluating their
own abilities, people exhibiting more bias blind spot
should exhibit more naïve realism in their initial
assessments of their self (e.g., “I’m great at using
a computer mouse but terrible at juggling”) and
less often correct those assessment for relevant infor-
mation about others’ abilities (e.g., “000as are most
people”). Rather than simply view the self more pos-
itively, they should be more likely to exhibit biases
reliant on similar cognitive mechanisms. For exam-
ple, they should be more likely to exhibit a better-
than-average effect when evaluating their ability to
perform easy tasks, and a worse-than-average effect
when evaluating their ability to perform difficult
tasks (Kruger 1999).
A second prediction is that commission of bias
blind spot should predict bias in judgments informed
by self–other accuracy comparisons. People who think
they are less vulnerable to bias than their peers, for
example, may attend less to the opinions of others
because they believe that others have a more biased
perception of the world, consistent with the operation
of naïve realism (Ross and Ward 1996,Pronin et al.
2004,Liberman et al. 2012). This should lead them to
place less weight on the advice of others when given
the opportunity to incorporate that advice into their
own judgments.
As another example, people who think they are
less vulnerable to bias than their peers should believe
that their judgment and decision making is in less
need of correction (Wilson and Brekke 1994). Analo-
gous to interventions aiming to curb addiction, where
awareness of the problem is a necessary first step
in facilitating corrective action, awareness of bias
may be an important precursor to bias mitigation
(Kruger and Dunning 1999,Wilson and Brekke 1994).
Consequently, greater commission of bias blind spot
should reduce the likelihood of benefitting from bias-
reducing interventions.
Overview of the Studies
We adopted a psychometric approach to the anal-
ysis of the bias blind spot. Our first two studies
report the development of our individual-difference
measure to measure the extent to which a person
believes she is less biased than her peers, the evalu-
ation of its reliability, the verification of its factorial
structure, and the analysis of its discriminant valid-
ity in relationship with potentially related constructs
such as intelligence, cognitive reflection, decision-
making ability, and personality traits related to self-
esteem, self-enhancement, and self-presentation. Our
last three studies report tests of our three judgmental
and behavioral predictions.
Study 1
In Study 1, we generated 27 scale items and then
submitted them to a purification process resulting in
a 14-item bias blind spot scale with good reliabil-
ity and stability. We used an item-generation process
aimed at capturing a broad sense of the construct to
develop the scale. In accordance with the underly-
ing theory (Pronin 2008), we reasoned that the bias
blind spot is a metabias that is exhibited across mul-
tiple self-assessments. Therefore, we treated the bias
blind spot as a latent variable that causes an asymme-
try in the assessment of bias in the self and in others
with respect to several biases, both in the social and
cognitive domains. Accordingly, we approached the
conceptualization and the operationalization of sus-
ceptibility to the bias blind spot as a reflective mea-
surement model (Bollen and Lennox 1991,Edwards
and Bagozzi 2000). That is, a model in which the
direction of causality is from the construct to the
indicators, and in which changes in the underlying
construct are hypothesized to cause changes in the
indicators. According to the proposed conceptualiza-
tion, changes in individual susceptibility to the bias
blind spot (i.e., a latent variable), will cause changes
in the extent to the bias blind spot is observed for
specific biases in judgment and behavior.
Method
Participants. Initial Sample0A total of 172 Ama-
zon Mechanical Turk (AMT) workers (86 women;
Mage =3204 years, SD =1005) accessed and completed
a survey on the Internet and received $4 as com-
pensation. In all studies we restricted participation
to residents of the United States. Participation in any
study reported here resulted in ineligibility for par-
ticipation in subsequent studies (i.e., no person parti-
cipated in more than one study). Participants were
79.7% White, 7% Black, 5.2% multiracial, 4.1% Asian,
and 1.2% Native American; 2.9% did not indicate their
ethnicity. AMT workers have been shown to exhibit
susceptibility to biases in judgment and decision mak-
ing similar to traditional college samples with respect
Downloaded from informs.org by [4.28.153.60] on 28 April 2015, at 06:23 . For personal use only, all rights reserved.
Scopelliti et al.: Bias Blind Spot: Structure, Measurement, and Consequences
4Management Science, Articles in Advance, pp. 1–19, © 2015 INFORMS
to common tasks such as the Asian disease problem,
the conjunction fallacy and outcome bias; to exhibit
similar levels of risk aversion; and to exhibit similar
levels of cooperation in behavioral economics games
such as the prisoner’s dilemma (Berinsky et al. 2012,
Horton et al. 2011,Paolacci et al. 2010).
Follow-Up Sample0A total of 83 participants in the
initial sample completed a follow-up survey consist-
ing of the items in the purified scale for an additional
$4 in compensation (43 women; Mage =3207 years,
SD =1000) yielding a 52% retention rate. Participants
in this subsample were 75.9% White, 9.6% Black,
7.2% Asian, 4.8% multiracial, and 1.2% Native Amer-
ican; 1.2% did not indicate their ethnicity. There were
no differences between the initial sample and the
follow-up sample in terms of age (F=0005, p > 0080),
gender (2=1000, p > 0060), or ethnicity (2=2041,
p > 0070).
Materials. Item Generation0We began developing
scale items by identifying existing indicators of the
construct in the literature. We drew some questions
from Pronin et al. (2002) verbatim, modified questions
from this source, and wrote new questions with the
same structure. Each question was structured so that
it described a bias and asked respondents to indi-
cate the incidence of the bias for two different tar-
gets: the self and the average American (for examples,
see Table 1). Each bias was described as a psycho-
logical tendency or effect, avoiding (when possible)
the use of positively connoted words or negatively
connoted words such as bias or error. After read-
ing a description of each bias, participants rated the
extent to which they exhibit that bias and the extent
to which the average American exhibits that bias on
7-point scales with endpoints, not at all (1) and very
much (7).
Item Scoring0Bias blind spot scores were calculated
by subtracting the perceived self-susceptibility to each
bias from the perceived susceptibility of the average
American to that bias, for each bias, and then by aver-
aging those relative differences. For each participant,
the size of the average difference between self and
average American bias ratings was used as a measure
of the magnitude of her bias blind spot.
Procedure. Participants rated their vulnerability
and the vulnerability of the average American to
commit biases on 27 unique items. Item order was
random. After completing all 27 items, participants
reported their age, gender, and ethnicity. To assess
the test-retest reliability of the instrument, all partic-
ipants in the initial sample were invited by means
of the AMT messaging system to complete the puri-
fied version of the scale eight days after the initial
administration.
Results
Purification0We first reduced the number of items
to improve the psychometric properties of the instru-
ment. We removed items that assessed blind spot by
describing different occurrences of the same underly-
ing judgmental bias. We then computed the correla-
tions between each item and the rest of the scale, and
removed items with item-to-total correlations lower
than 0.40. This purification process led to a final
instrument containing 14 items (see Table 1for the
full list of items).
Reliability0The 14-item scale shows high reliabil-
ity (=0086), well above the acceptable threshold
of 0.70 (Nunnally 1978). All items appeared to be
worth retention. No question eliminations would yield
a higher value of the Cronbach’s alpha coefficient
(-if-item-deletedi< ). Of the 91 pairwise correla-
tions between the items (M=0030), 85 were positive
and significant, 3 were positive and marginally signifi-
cant, and 3 were positive but not significant (ps<0042;
see Table 2). Each item correlated well with the scale,
as signaled by an average item-to-total correlation
equal to 0.52. All further analyses, in this and subse-
quent studies, use this 14-item scale. For all 14 biases,
the majority of participants rated the average Amer-
ican as more susceptible to bias than themselves; all
ts6026, all ps<00001 (for all means, see Table 3).
Exploratory Factor Analysis0We submitted the
14 bias blind spot items to a principal-components
factor analysis (PCA) followed by a parallel anal-
ysis (Horn 1965). The parallel analysis suggests an
underlying single-factor structure, with only the first
eigenvalue observed in the data being higher than
a parallel average random eigenvalue based on the
same sample size and number of variables. In the
single-factor model, all 14 items load onto a single
factor accounting for 35% of the total variance, and
each item has a high correlation with that factor (all
s>0048; see Table 3).
Test-Retest Reliability0An analysis of the second
administration of the scale indicated that the instru-
ment has consistently high reliability (=0088), and
high test-retest reliability (r4815=0080, p < 00001), sig-
naling stability of the bias blind spot scale over time.
Discussion
Study 1 provided evidence for individual differences
in susceptibility to the bias blind spot. High inter-
item correlations and factors loadings suggest that
some people have a higher susceptibility to the bias
blind spot than do others across a variety of judg-
mental biases. The results from this first study sug-
gest that our instrument reliably assessed individual
differences in susceptibility to the bias blind spot. In
addition, the instrument captures a single latent vari-
able as proposed by our construct operationalization.
Furthermore, the high test-retest reliability observed
Downloaded from informs.org by [4.28.153.60] on 28 April 2015, at 06:23 . For personal use only, all rights reserved.
Scopelliti et al.: Bias Blind Spot: Structure, Measurement, and Consequences
Management Science, Articles in Advance, pp. 1–19, © 2015 INFORMS 5
Table 1 Scale Items and Corresponding Biases
Item Description Bias
1 Some people show a tendency to judge a harmful action as worse than an equally harmful inaction. For example,
this tendency leads to thinking it is worse to falsely testify in court that someone is guilty, than not to testify
that someone is innocent.
Action-inaction bias
2 Psychologists have claimed that some people show a tendency to do or believe a thing only because many other
people believe or do that thing, to feel safer or to avoid conflict.
Bandwagon effect
3 Many psychological studies have shown that people react to counterevidence by actually strengthening their
beliefs. For example, when exposed to negative evidence about their favorite political candidate, people tend to
implicitly counterargue against that evidence, therefore strengthening their favorable feelings toward the
candidate.
Confirmation bias
4 Psychologists have claimed that some people show a “disconfirmation” tendency in the way they evaluate
research about potentially dangerous habits. That is, they are more critical and skeptical in evaluating evidence
that an activity is dangerous when they engage in that activity than when they do not.
Disconfirmation tendency
5 Psychologists have identified an effect called “diffusion of responsibility,” where people tend not to help in an
emergency situation when other people are present. This happens because as the number of bystanders
increases, a bystander who sees other people standing around is less likely to interpret the incident as a
problem, and also is less likely to feel individually responsible for taking action.
Diffusion of responsibility
6 Research has found that people will make irrational decisions to justify actions they have already taken. For
example, when two people engage in a bidding war for an object, they can end up paying much more than the
object is worth to justify the initial expenses associated with bidding.
Escalation of commitment
7 Psychologists have claimed that some people show a tendency to make “overly dispositional inferences” in the
way they view victims of assault crimes. That is, they are overly inclined to view the victim’s plight as one he or
she brought on by carelessness, foolishness, misbehavior, or naivetë.
Fundamental attribution error
8 Psychologists have claimed that some people show a “halo” effect in the way they form impressions of attractive
people. For instance, when it comes to assessing how nice, interesting, or able someone is, people tend to
judge an attractive person more positively than he or she deserves.
Halo effect
9 Extensive psychological research has shown that people possess an unconscious, automatic tendency to be less
generous to people of a different race than to people of their race. This tendency has been shown to affect the
behavior of everyone from doctors to taxi drivers.
Ingroup favoritism
10 Psychologists have identified a tendency called the “ostrich effect,” an aversion to learning about potential losses.
For example, people may try to avoid bad news by ignoring it. The name comes from the common (but false)
legend that ostriches bury their heads in the sand to avoid danger.
Ostrich effect
11 Many psychological studies have found that people have the tendency to underestimate the impact or the strength
of another person’s feelings. For example, people who have not been victims of discrimination do not really
understand a victim’s social suffering and the emotional effects of discrimination.
Projection bias
12 Psychologists have claimed that some people show a “self-interest” effect in the way they view political
candidates. That is, people’s assessments of qualifications, and their judgments about the extent to which
particular candidates would pursue policies good for the American people as a whole, are influenced by their
feelings about whether the candidates’ policies would serve their own particular interests.
Self-interest bias
13 Psychologists have claimed that some people show a “self-serving” tendency in the way they view their academic
or job performance. That is, they tend to take credit for success but deny responsibility for failure. They see
their successes as the result of personal qualities, like drive or ability, but their failures as the result of external
factors, like unreasonable work requirements or inadequate instructions.
Self-serving bias
14 Psychologists have argued that gender biases lead people to associate men with technology and women with
housework.
Stereotyping
indicates that the individual difference tapped by our
scale is stable over time.
Study 2
In Study 2, we verified the factorial structure of the
bias blind spot scale that emerged in Study 1 by sub-
mitting the 14 items to a confirmatory factor analysis.
Since the development of a valid and reliable measure-
ment scale introduces the possibility to clarify the rela-
tionship between susceptibility to the bias blind spot
and related constructs that compose its nomological
network (Cronbach and Meehl 1955), we also tested
the discriminant validity of the bias blind spot scale
in relation to a set of established scales and measures
assessing potentially related psychological constructs.
Such comparisons determine whether individual dif-
ferences in bias blind spot reflect individual differ-
ences in more basic or established personality traits.
Specifically, by using five different samples of re-
spondents, we examined the correlations between the
bias blind spot scale and measures of intelligence,
inclination toward cognitive activities, and decision-
making ability (i.e., SAT scores, Need for cognition,
performance on the Cognitive Reflection Test, and
decision-making competence), measures of personal-
ity traits related to self-esteem and self-enhancement
(i.e., self-esteem, superiority, need for uniqueness,
narcissism, and over-claiming tendency), self-presen-
tation (i.e., self-monitoring, self-consciousness, and
Downloaded from informs.org by [4.28.153.60] on 28 April 2015, at 06:23 . For personal use only, all rights reserved.
Scopelliti et al.: Bias Blind Spot: Structure, Measurement, and Consequences
6Management Science, Articles in Advance, pp. 1–19, © 2015 INFORMS
Table 2 Correlations Between the 14 Selected Scale Items
Item 1 2 3 4 5 6 7 8 9 10 11 12 13 14
1 Action-inaction bias
2 Bandwagon effect 00287∗∗
3 Confirmation bias 00326∗∗ 00478∗∗
4 Diffusion of responsibility 00440∗∗ 00347∗∗ 00331∗∗
5 Disconfirmation tendency 0018900236∗∗ 00347∗∗ 00240∗∗
6 Escalation of commitment 00317∗∗ 00413∗∗ 00253∗∗ 00309∗∗ 00267∗∗
7 Fundamental attribution error 00268∗∗ 00222∗∗ 00271∗∗ 00304∗∗ 00061 00216∗∗
8 Halo effect 00356∗∗ 00469∗∗ 00328∗∗ 00273∗∗ 00250∗∗ 00379∗∗ 00331∗∗
9 Ingroup favoritism 00217∗∗ 00321∗∗ 00371∗∗ 00329∗∗ 00245∗∗ 00286∗∗ 00352∗∗ 00398∗∗
10 Ostrich effect 00256∗∗ 00368∗∗ 00313∗∗ 00359∗∗ 00390∗∗ 00329∗∗ 00134 00270∗∗ 00238∗∗
11 Projection bias 00397∗∗ 00244∗∗ 00237∗∗ 00218∗∗ 0013600082 00198∗∗ 00362∗∗ 00281∗∗ 00147
12 Self-interest bias 00336∗∗ 00272∗∗ 00358∗∗ 00323∗∗ 00296∗∗ 00281∗∗ 00249∗∗ 00197∗∗ 00302∗∗ 00261∗∗ 00228∗∗
13 Self-serving bias 00383∗∗ 00403∗∗ 00399∗∗ 00418∗∗ 00357∗∗ 00295∗∗ 00250∗∗ 00447∗∗ 00300∗∗ 00412∗∗ 0017200334∗∗
14 Stereotyping 00280∗∗ 00386∗∗ 00233∗∗ 00253∗∗ 00085 00305∗∗ 00340∗∗ 00426∗∗ 00313∗∗ 0015000378∗∗ 00311∗∗ 00256∗∗
Note. Asterisks indicate significant correlations: ∗∗p < 0001; p < 0005; p < 0010.
social desirability), and more general personality traits
(i.e., the Big Five Inventory (BFI)).
Method
Participants. Study 2 made use of five unique sam-
ples that totaled to 661 respondents (344 women;
Mage =3206 years, SD =1102). For each sample, the
content of the questionnaire varied as described next.
All participants were AMT workers residing in the
United States who accessed and completed a survey
on the Internet in exchange for compensation vary-
ing based on the length of the study (Sample 1: $5;
Sample 2: $4; Sample 3: $6; Sample 4: $2; Sample 5:
$4). Participants were 77.0% White, 7.0% Black, 7.0%
Asian, 5.1% multiracial, 1.6% Native American, and
2.4% did not indicate their ethnicity. The five samples
did not differ in terms of bias blind spot scores F < 1.
Table 3 Bias Blind Spot by Scale Item, Factor Loadings from Exploratory Factor Analysis, and Completely Standardized Parameters from
Confirmatory Factor Analysis
Average American Self Factor Completely
loading standardized
Bias M SD M SD t(171) (EFA) parameter (CFA)
1 Action-inaction bias 5012 1033 3096 1069 9086∗∗∗ 00613 00660
2 Bandwagon effect 5073 1006 3060 1068 15007∗∗∗ 00675 00504
3 Confirmation bias 5055 1027 3098 1060 11090∗∗∗ 00642 00552
4 Diffusion of responsibility 5051 1027 4011 1077 10005∗∗∗ 00613 00564
5 Disconfirmation tendency 5031 1017 4048 1052 7031∗∗∗ 00482 00552
6 Escalation of commitment 5036 1013 3057 1067 14028∗∗∗ 00575 00583
7 Fundamental attribution error 4063 1024 3016 1054 12016∗∗∗ 00496 00476
8 Halo effect 5089 1002 4016 1069 12013∗∗∗ 00674 00511
9 Ingroup favoritism 5001 1029 3045 1069 11097∗∗∗ 00598 00599
10 Ostrich effect 5020 1015 4035 1074 6026∗∗∗ 00558 00520
11 Projection bias 5032 1033 4019 1065 7083∗∗∗ 00477 00476
12 Self-interest bias 5075 1024 4075 1062 8010∗∗∗ 00568 00355
13 Self-serving bias 5069 1008 4014 1059 13028∗∗∗ 00681 00677
14 Stereotyping 5046 1015 3062 1081 13032∗∗∗ 00579 00490
Notes. EFA, exploratory factor analysis; CFA, confirmatory factor analysis. Asterisks indicate significant results: ∗∗∗p < 00001.
Materials and Procedure. Sample 1. Participants
(n=260) first completed all 14 items of the bias
blind spot scale in a random order. Participants then
completed a series of personality scales measuring
extant psychological constructs potentially related to
the bias blind spot: the 10-item Rosenberg self-esteem
scale (Rosenberg 1965), the need for uniqueness
scale (32 items; Snyder and Fromkin 1977), the self-
consciousness scale (22 items; Feingstein et al. 1975),
the need for cognition scale (NFC; 18 items; Cacioppo
et al. 1984), the self-monitoring scale (25 items; Snyder
1974), the superiority scale (10 items; Robbins and
Patton 1985), the NPI-16 short measure of narcissism
(16 items; Ames et al. 2006), and the social desir-
ability scale (33 items; Crowne and Marlowe 1960).
Each scale was administered using its original answer
format and coding scheme. Both the order of the
scales and the order of the items were randomized,
Downloaded from informs.org by [4.28.153.60] on 28 April 2015, at 06:23 . For personal use only, all rights reserved.
Scopelliti et al.: Bias Blind Spot: Structure, Measurement, and Consequences
Management Science, Articles in Advance, pp. 1–19, © 2015 INFORMS 7
with scale items nested within scale. Participants also
reported their scores on the math and verbal sections
of the SAT (if they had taken the SAT). To facilitate
the accuracy of participants who might have taken
the test several years before, participants reported
approximate SAT scores on a 12-point scale increasing
in 50-point increments from 200 to 800. Last, partici-
pants reported their age, gender, ethnicity, and high-
est level of education completed.
Sample 2. Participants (n=101) first completed all
14 items of the bias blind spot scale in a random order.
They then completed the original three-item Cogni-
tive Reflection Test (CRT; Frederick 2005) in open-
ended format. Given the diffusion of the three original
CRT on AMT and the ease of retrieving the solution to
those three questions, participants were also admin-
istered nine additional and less common CRT ques-
tions (Frederick, personal communication) resulting
in a total of 12 CRT items. Next, participants reported
their scores on the math and verbal sections of the
SAT if they had taken the test (as did Sample 1).
Finally, participants reported their age, gender, eth-
nicity, and highest level of education completed.
Sample 3. Participants (n=100) first completed
all 14 items of the bias blind spot scale in a ran-
dom order. Then they completed the Adult Decision-
Making Competence inventory (A-DMC; Bruine de
Bruin et al. 2007), a comprehensive instrument assess-
ing general judgment and decision-making ability.
Participants completed the six behavioral decision-
making batteries of measures composing the A-DMC:
resistance to framing (14 paired items), which measures
whether value assessment is affected by irrelevant
variations in problem descriptions; recognizing social
norms, which measures how well people assess peer
social norms (16 items); under/overconfidence, which
measures how well participants recognize the extent
of their own knowledge (34 items); applying decision
rules, which asks participants to indicate, for hypo-
thetical individual consumers using different decision
rules, which products they would buy out of a choice
set (10 items); consistency in risk perception, which mea-
sures the ability to follow probability rules (16 paired
items); and resistance to sunk costs, which measures the
ability to ignore prior investments when making deci-
sions (10 items). These tasks measure different aspects
of the decision-making process and decision-making
skills, such as the ability to assess value and the abil-
ity to integrate information. Both the order of the bat-
teries and the order of the items were randomized,
with scale items nested within battery. Participants
reported their scores on the math and verbal sections
of the SAT if they had taken the test (as did Sam-
ple 1). Finally, participants reported their age, gender,
ethnicity, and highest level of education completed.
Sample 4. Participants (n=102) first completed all
14 items of the bias blind spot scale in a random
order. They then completed 11 Over-Claiming Ques-
tionnaire (OCQ) batteries. In these batteries, respon-
dents indicate their familiarity with a list of items,
some of which exist (e.g., plate tectonics) and some
of which do not actually exist (e.g., plates of paral-
lax). These familiarity ratings serve as the basis for
a claiming response bias index (OCQ bias) measur-
ing self-enhancing tendencies: a statistical estimate of
how strong familiarity with an item has to be for a
respondent to express familiarity with it (Macmillan
and Creelman 1991). The 11 batteries covered topics
including rap artists, rock artists, country artists, hor-
ror movies, comedies, dramas, foreign movies, soap
operas, football, world leaders, and fashion designers
(Paulhus et al. 2003). Each battery included seven real
items and three bogus items. Participants were asked
to rate their familiarity with each item on a 7-point
scale with endpoints, not at all familiar (0) and very
familiar (6). Both the order of the batteries and the
order of the items were randomized, with scale items
nested within battery. Finally, participants reported
their age, gender, ethnicity, and highest level of edu-
cation completed.
Sample 5. Participants (n=98) first completed all
14 items of the bias blind spot scale in a random order.
Next, participants completed a short-form (44-item)
version of McCrae and Costa’s (1987) BFI (John and
Srivastava 1999). Participants reported their scores on
the math and verbal sections of the SAT if they had
taken the test (as did Sample 1). Finally, participants
reported their age, gender, ethnicity, and highest level
of education completed.
Results
Confirmatory Factor Analysis0To test the factorial
structure that emerged in the exploratory factor anal-
ysis in Study 1, we conducted a confirmatory fac-
tor analysis on the data collected from Sample 1. We
first evaluated the assumption of multivariate nor-
mality of the data that is a necessary condition for the
use of maximum likelihood estimation. Toward this
end, we computed Mardia’s (1980) test of multivari-
ate skewness and kurtosis. The test was significant
(2=299051, p < 00001), signaling that the assumption
of multivariate normality of the data was violated. As
a consequence we opted for a robust maximum like-
lihood estimation.
The confirmatory factor analysis suggested that the
single-factor model emerged in the exploratory fac-
tor analysis from Study 1 fit the data in Study 2
well, with a goodness-of-fit index of 0.93, a Bentler’s
comparative fit index of 0.98, a non-normed fit index
of 0.97, and a root mean square error of approxima-
tion (RMSEA) of 0.05, suggesting a good fit between
Downloaded from informs.org by [4.28.153.60] on 28 April 2015, at 06:23 . For personal use only, all rights reserved.
Scopelliti et al.: Bias Blind Spot: Structure, Measurement, and Consequences
8Management Science, Articles in Advance, pp. 1–19, © 2015 INFORMS
Figure 1 Distribution of Bias Blind Spot Scores in Study 2
[–6.5, – 7.0]
[–6.0, – 6.5]
[–5.5, – 6.0]
[–5.5, – 5.0)
[–5.0, – 4.5)
[–4.5, – 4.0)
[–4.0, – 3.5)
[–3.5, – 3.0)
[–3.0, – 2.5)
[–2.5, – 2.0)
[–2.0, – 1.5)
[–1.5, – 1.0)
[–1.0, –0.5)
[–0.5, 0)
[0, 0.5)
[0.5, 1.0)
[1.0, 1.5)
[1.5, 2.0)
[2.0, 2.5)
[2.5, 3.0)
[3.0, 3.5)
[3.5, 4.0)
[4.0, 4.5)
[4.5, 5.0)
[5.0, 5.5)
[5.5, 6.0)
[6.0, 6.5]
[6.5, 7.0]
Bias blind spot score
160
140
120
100
80
60
40
20
0
Frequency (n)
Notes. Scale ranges from 7 to 7. Scores greater than zero indicate bias blind spot. Median column is indicated by gray bar.
the model and the observed data. The factor-loading
estimates are reported in Table 3and were all signifi-
cant; all ts>549, all ps<0001.
Descriptive StatisticsTo examine the distribution of
susceptibility to the bias blind spot, we computed bias
blind spot scores on the five samples following the
same procedure as in Study 1 (=084). Across the
five samples, the distribution of the scores appeared
to be positively skewed with an observed range from
150 to 5.14 (see Figure 1). On average, participants
exhibited a significant bias blind spot (M=148, SD =
095; significantly different from 0; t660=4029, p<
0001). Furthermore, for each of the 14 biases that
composed the bias blind spot scale, participants rated
the average American as more susceptible to that bias
than themselves; all Ms > 090, all ts(659) >1685,
all ps<0001. At the individual level, a large major-
ity (85.2%) of participants exhibited a significant bias
blind spot across the 14 items, with the averages of
their scores being marginally or significantly greater
than 0; all ts(13) >179, all ps<010. Only one par-
ticipant (0.1%) had an average bias blind spot score
that was significantly lower than 0 (t13=−246,
p=003).
We then examined whether demographic variables
accounted for some differences in bias blind spot
scores. Gender appeared unrelated to bias blind spot,
as the scores of male (M=145, SD =095) and
female participants (M=151, SD =095) were not
significantly different; F<1. Similarly, the correlation
between age and bias blind spot scores was not sig-
nificant, r=−004, p=030. Bias blind spot scores
were significantly and negatively affected by level of
education, following approximately a linear pattern
(linear contrast: F1653=1142, p=0001).
Discriminant ValidityWe next examined whether
the bias blind spot scale is distinct from related
constructs: (i) measures of intelligence, inclination
toward cognitive activities, and decision-making abil-
ity (i.e., SAT scores, NFC, CRT, A-DMC); (ii) personal-
ity traits related to self-esteem and self-enhancement
tendencies (i.e., self-esteem, superiority, need for
uniqueness, narcissism, over-claiming tendency);
(iii) personality traits related to self-presentation
(i.e., self-monitoring, self-consciousness, and social
desirability); and (iv) and general personality traits
(i.e., BFI).
For each of the samples we computed bias blind
spot scores (Alphas ranging between 0.82 and 0.85).
For each of the constructs that were measured
using multi-item scales, we computed overall aver-
age scores after reverse scoring appropriate items, and
we examined the correlation of these scores with bias
blind spot scores. Table 4presents these scales and
measures, their Cronbach’s alpha coefficients, means,
standard deviations, and their zero-order correlations
with the bias blind spot scale.
Intelligence, Inclination Toward Cognitive Activi-
ties, and Decision-Making AbilityWe first examined
whether bias blind spot is merely a manifestation of
intelligence and cognitive ability in four ways. First,
we examined SAT scores, as both the verbal and math
scores load highly on psychometric gor general intel-
ligence (Brodnick and Ree 1995,Frey and Detterman
2004,Unsworth and Engle 2007). Of the 559 partic-
ipants in Samples 1, 2, 3, and 5, 248 reported their
SAT test scores. The correlation between bias blind
spot scores and the verbal SAT scores was not sig-
nificant (r247=010, p=013), whereas the corre-
lation between bias blind spot scores and the math
SAT scores (r247=−013, p=005), was significant,
albeit very small.
Second, we examined the correlation between bias
blind spot and NFC, which refers to the extent to
Downloaded from informs.org by [4.28.153.60] on 28 April 2015, at 06:23 . For personal use only, all rights reserved.
Scopelliti et al.: Bias Blind Spot: Structure, Measurement, and Consequences
Management Science, Articles in Advance, pp. 1–19, © 2015 INFORMS 9
Table 4 Scale Descriptive Statistics, Reliabilities, and Correlations with Bias Blind Spot (Study 2)
Construct Sample M SD  N r
Intelligence 1–3, 5
SAT math 605004 111055 248 0013
SAT verbal 640032 93022 — 248 0010
Need for cognition 1 63065 13021 0093 260 0016
Cognitive Reflection Test 2
12 items 3067 3000 0082 101 0023
3 original items 1030 1025 0079 101 0019
9 new items 2038 2002 0071 101 0022
Decision-making competence 3
Resistance to framing 1045 0054 0058 100 0014
Recognizing social norms 0051 0026 0089 100 0007
Under/overconfidence 0091 0007 0089 100 0006
Applying decision rules 0077 0016 0050 100 0002
Consistency in risk perception 0075 0015 0071 100 0009
Resistance to sunk costs 4040 0062 0040 100 0008
Big Five personality traits 5
Neuroticism 24030 5065 0087 98 0006
Extraversion 22084 7042 0088 98 0003
Openness 36062 6081 0085 98 0033∗∗
Agreeableness 31076 5030 0083 98 0010
Conscientiousness 33002 6042 0083 98 0007
Self-esteem 1 19096 6026 0092 260 0015
Superiority 1 35013 7039 0075 260 0002
Need for uniqueness 1 101023 17034 0089 260 0016∗∗
Narcissism 1 20040 3065 0082 260 0004
Over-claiming 4 0050 0026 0095 102 0025
Self-consciousness 1
Public 18062 4058 0082 260 0005
Private 25045 4057 0073 260 0007
Social anxiety 16010 5001 0087 260 0010
Self-monitoring 1 11018 4043 0074 260 0005
Social desirability 1 3030 6080 0097 260 0005
p < 0010; p < 0005; ∗∗ p < 0001.
which people “engage in and enjoy effortful cogni-
tive activities” (Cacioppo and Petty 1982, p. 1). People
who are high in NFC expend more effort processing
information, and perform better on arithmetic prob-
lems, anagrams, trivia tests, and college coursework
(Cacioppo et al. 1996). In line with the results of
West et al. (2012), we observed a small but signifi-
cant and positive correlation between bias blind spot
scores and need for cognition (r42585=0016, p=0001).
This significant correlation suggests that people with
a high NFC tend to believe that they are less suscep-
tible to biases than their peers.
Third, we examined whether participants with
higher bias blind spot scores make more use of intu-
ition than of deliberate reasoning by examining the
correlation between bias blind spot scores and the
CRT (Frederick 2005), a set of questions that each have
an incorrect intuitive response and a correct response
that can be reached through deliberate reasoning. We
computed a composite measure of performance on
the CRT by assigning a score of one to each cor-
rect answer, a score of zero to each incorrect answer,
and summing all items (M=3067, SD =3000). Bias
blind spot scores were significantly and negatively
correlated with CRT scores (r4995= −0022, p=0002).
This pattern of correlations is consistent irrespective
of which CRT items are considered: all 12 items: r=
0023, p=0002; the three “original” items: r= −0019,
p=0006; or the nine “new” items: r= −0022, p=0003.
Participants who believed themselves to be less vul-
nerable to bias than others were more likely to rely on
their intuition and less likely to engage in cognitive
reflection and deliberation (i.e., were less accurate).
Finally, we computed scores for each A-DMC bat-
tery as recommended by Bruine de Bruin et al.1
(2007). Bias blind spot were uncorrelated with perfor-
mance on all of the A-DMC batteries; all rs<0014, all
ps>0018, average r=0008.
Self-Esteem and Self-Enhancement0We next examined
the possibility that the bias blind spot is the manifes-
tation of more basic perceptions of being superior to
others or being different from others (e.g., Taylor and
Brown 1988,Snyder and Fromkin 1980) by examining
1The levels of reliability for each battery reported in Table 4
are generally consistent with those reported by Bruine de Bruin
et al. (2007).
Downloaded from informs.org by [4.28.153.60] on 28 April 2015, at 06:23 . For personal use only, all rights reserved.
Scopelliti et al.: Bias Blind Spot: Structure, Measurement, and Consequences
10 Management Science, Articles in Advance, pp. 1–19, © 2015 INFORMS
the correlations between bias blind spot and superior-
ity, self-esteem, need for uniqueness, narcissism, and
over-claiming. If the domain of these constructs sig-
nificantly overlaps with that of the bias blind spot, we
should observe high correlations between those traits
and bias blind spot scores.
The superiority scale (Robbins and Patton 1985)
measures whether people believe that they are supe-
rior to others. People who believe they are superior
to others may believe that they are less susceptible to
cognitive and social biases than inferior others. How-
ever, perceived self-superiority did not correlate with
bias blind spot scores (r42585= −0002, p=0072).
We then examined the correlation between bias
blind spot and self-esteem (Rosenberg 1965), since
people holding high levels of self-esteem may be
more likely to see themselves as immune from bias
and therefore show higher susceptibility to the bias
blind spot. We found evidence for a significant and
positive correlation between bias blind spot scores
and self-esteem (r42585=0015, p=0002), suggesting
that people who believe they are less vulnerable to
bias than their peers tend to have higher levels of
self-esteem. The small size of the correlation, how-
ever, suggests that the two constructs are theoretically
distinct.
We also examined the possibility that the belief that
one is less vulnerable to bias than one’s peers is a
manifestation of a need for uniqueness (Snyder and
Fromkin 1977), the desire to be dissimilar from oth-
ers. The correlation between bias blind spot scores
and need for uniqueness was positive and significant
(r42585=0016, p < 0001). This small significant correla-
tion suggests that people high in bias blind spot have
a higher desire to distinguish themselves from others,
but that the two constructs are theoretically distinct.
Finally, we examined the relationship between the
bias blind spot and two different operationalizations
of self-enhancement, a narcissism scale (Ames et al.
2006), and over-claiming (Paulhus 1998). Narcissism,
a personality trait that involves a pervasive pattern
of grandiosity, self-focus, and self-importance accom-
panied by preoccupation with success and demands
for admiration (Morf and Rhodewalt 2001), has been
used as a measure of trait self-enhancement (Paulhus
1998). The narcissism scale did not correlate with
bias blind spot scores (r42585=0004, p=0049). Over-
claiming, a measure that can discriminate accuracy
from self-enhancement in response patterns, was
assessed by using the OCQ (Paulhus et al. 2003).
We computed OCQ bias indexes, corresponding to
the mean of the hit rate and the false-alarm rate for
real and false OCQ items (Paulhus et al. 2003), for
each battery of items, and averaged across batter-
ies. Higher values of OCQ bias indicate lower self-
enhancing tendencies, as they indicate that a higher
sense of familiarity needs to be experienced in order
for a respondent to claim familiarity with an item. The
correlation between bias blind spot scores and OCQ
bias was negative and significant, r41015= −0025, p=
0001, suggesting that higher bias blind spot scores
are associated with higher self-enhancing tendencies
as measured by over-claiming, but that the two con-
structs are sufficiently discriminated.
Self-Presentation0We examined whether the bias
blind spot is the manifestation of other personality
traits related to self-presentation. Self-consciousness
is an acute sense of self-awareness, articulated in
three dimensions: private self-awareness (i.e., a ten-
dency to introspect and examine one’s inner self and
feelings), public self-consciousness (i.e., a tendency
to think about how the self is viewed by others),
and social anxiety (i.e., a sense of apprehensiveness
over being evaluated by others in a social context;
Feingstein et al. 1975). We did not observe a signif-
icant correlation between bias blind spot scores and
private self-awareness (r42585=0007, p=0024), public
self-consciousness (r42585= 0005, p=0046), or social
anxiety (r42585= 0010, p=0011). Self-monitoring mea-
sures the extent to which one consciously employs
impression management strategies in social interac-
tions (Snyder 1974). We did not observe a significant
correlation between bias blind spot scores and self-
monitoring (r42585=0005, p=0040). Lastly, we sought
to test whether there was a negative correlation
between bias blind spot and the tendency to answer
in a way that would be viewed favorably by others
(Crowne and Marlowe 1960). The correlation was neg-
ative but not significant (r42585= 0005, p=0041).
Big Five Personality Traits0Finally, we examined
the correlations between bias blind spot and the
Big Five personality traits (McCrae and Costa 1987).
Bias blind spot scores appeared to be uncorrelated
with agreeableness (r4965=0010, p=0032), extraver-
sion (r4965=0003, p=0080), neuroticism (r4965=0006,
p=0057), and conscientiousness (r4965=0007, p=0049),
but moderately correlated with openness to experi-
ence (r4965=0033, p < 0001). Participants who reported
significantly higher openness to experience had higher
bias blind spot scores, although the moderate size of
the coefficient suggests that the two constructs are suf-
ficiently discriminated. This result was not predicted.
Because open-minded thinking has been shown to cor-
relate with superior cognitive ability in a number of
domains (Stanovich and West 2007,West et al. 2008),
it is possible that if people high in openness to expe-
rience have both superior cognitive ability and some
awareness of that ability, they might accurately report
that they are less susceptible to bias than their peers.
Discussion
Bias blind spot appears to be a latent unidimensional
construct that induces people to see themselves as less
Downloaded from informs.org by [4.28.153.60] on 28 April 2015, at 06:23 . For personal use only, all rights reserved.
Scopelliti et al.: Bias Blind Spot: Structure, Measurement, and Consequences
Management Science, Articles in Advance, pp. 1–19, © 2015 INFORMS 11
biased than other people across multiple judgmental
biases (i.e., a meta bias). Bias blind spot scores did
not appear to be correlated with demographic vari-
ables such as gender and age, although bias blind
spot decreased significantly as level of education
increased. The results of Study 2 support the dis-
criminant validity of the bias blind spot by showing
that the construct is not a derivative of several poten-
tially related constructs such as measures of intelli-
gence and cognitive ability, decision-making ability,
self-esteem and self-enhancement, self-presentation,
and general personality traits. Many of the correla-
tions between these traits and bias blind spot were
not significant. Other correlations were in the low to
moderate range (0.17 to 0.33; Cohen 1988), suggesting
that the bias blind spot scale is assessing a distinct
construct.
Most importantly, the perception that one is less
vulnerable to bias than one’s peers truly appears to
constitute a blind spot in self-awareness. The lack of
correlation with general decision-making competence
suggests that susceptibility to bias blind spot (a) does
not reflect generally poorer decision-making ability,
and (b) does not reflect an accurate self-assessment
of decision-making ability. The negative correlation
observed between susceptibility to bias blind spot and
CRT scores could be interpreted as supporting a posi-
tive relation to more general cognitive ability, but also
lends support our suggestion that the bias blind spot
is a distinct meta-bias reflected in the different con-
sideration of evidence when making self and other
assessments. In other words, people who are more
likely to rely on their intuition may more biased, but
also be less likely to correct their initial introspec-
tive self-assessments by considering additional evi-
dence, such as their behavior. A more direct test of
this assumption was performed in Study 3.
Study 3: Bias Blind Spot and
Self-Assessment
In Study 3, we further elucidated the cognitive mech-
anisms underlying the bias blind spot by examin-
ing the relationship between susceptibility to the bias
blind spot and the asymmetric use of evidence in
another set of self–other comparisons: relative perfor-
mance assessments. When evaluating their own skills
and abilities relative to their peers, people tend to
focus on their own capabilities and neglect the capa-
bilities of others. Because they neglect the fact that
easy activities are easy for most people and difficult
activities are difficult for most people, they rate them-
selves as better than the average person when assess-
ing their ability at easy activities, but as worse than
the average person when assessing their ability at dif-
ficult activities (Kruger 1999).
If the bias blind spot is simply due to overly pos-
itive self-perceptions, then people high in bias blind
spot should be more susceptible to the better-than-
average effect, but not to the worse-than-average
effect. That is, they should rate themselves as bet-
ter than average for both easy and difficult activi-
ties. If susceptibility to the bias blind spot is due to
greater naïve realism and confidence in initial self-
assessments (e.g., “I’m great at using a computer
mouse but terrible at juggling”), coupled with a fail-
ure to correct those assessment to account for other’s
abilities (e.g., “000as are most people”), people high in
susceptibility to the bias blind spot should be more
susceptible to both the better-than-average effect and
the worse-than-average effect.
Method
Pretest. In a pretest, 100 AMT workers (47 women;
Mage =2806 years, SD =1004) were presented with a
list of 34 activities in an online survey and com-
pensated $1 for participating. Included in that list
were the activities used by Kruger (1999, Study 2).
Participants rated the difficulty of each activity on
a 9-point scale (1=not difficult at all; 9=very diffi-
cult). We selected the two activities rated least difficult
and the two activities rated most difficult: operating a
computer mouse (Mdifficulty =1034, SD =0084), copying and
pasting text (Mdifficulty =1043, SD =1013), programming
a computer (Mdifficulty =7058, SD =2058), and juggling
(Mdifficulty =7068, SD =2028). A one-way repeated mea-
sures ANOVA accompanied by planned comparisons
revealed that the two difficult activities were perceived
as significantly more difficult than the two easy activ-
ities (F 411985=11105052, p < 00001), that the two easy
activities were perceived as equally easy, and that the
two difficult activities were perceived as equally diffi-
cult (F s <1).
Participants. A new sample of 156 AMT workers
(90 women; Mage =3200 years, SD =1200) received
$2.50 for participating in Study 3. Participants were
79.5% White, 8.3% Black, 4.5% Asian, 3.2% multira-
cial, and 0.6% Native American; 3.8% did not indicate
their ethnicity.
Materials and Procedure. Participants completed
the 14-item bias blind spot scale in a random order.
Then they rated their comparative ability on the two
easy (operating a computer mouse and copying and past-
ing text) and on the two difficult activities (program-
ming a computer and juggling) on 100-point scales
(0 =I’m at the very bottom; 100 =I’m at the very top).
Activity rating order was random. Last, participants
reported their age, gender, and ethnicity.
Downloaded from informs.org by [4.28.153.60] on 28 April 2015, at 06:23 . For personal use only, all rights reserved.
Scopelliti et al.: Bias Blind Spot: Structure, Measurement, and Consequences
12 Management Science, Articles in Advance, pp. 1–19, © 2015 INFORMS
Results and Discussion
Bias blind spot scores were computed following the
procedure described in Study 1 (=0083). A mixed
model analysis was used to estimate the effect of sus-
ceptibility to the bias blind spot, type of activity (easy
versus difficult), and their interaction on compara-
tive ability ratings. Since each participant assessed her
ability on both the easy and the difficult activities, we
let the intercept vary randomly to take into account
the lack of independence between the observations.
We estimated the following model:
CAij =0+1BBSi+2TAj+3BBSi×TAj+Ui+ij
where CA refers to comparative ability, index irefers
to participants and index jrefers to the type of
activity (easy versus difficult). The dependent vari-
able was thus the average comparative ability rat-
ing expressed by each participant on easy or difficult
activities. The explanatory variables were each partic-
ipant’s bias blind spot score (BBSi5, the type of activ-
ity (TAj, dummy coded, with 0 =easy, 1 =difficult)
and the interaction between these two variables, and
Uiindicated each participant’s random effect. Both
bias blind spot scores and comparative ability ratings
were standardized. Results are based on a total of
624 observations, where each observation is a compar-
ative ability rating on an activity provided by a par-
ticipant. The results revealed a significant interaction
between susceptibility to the bias blind spot and type
of activity (easy vs. difficult) on comparative ability
assessments (b=0027, SE =0006, t= −4084, p < 00001).
Table 5reports the results of this estimation and the
comparison of this factorial model with an alternative
model that includes only the main effects.
The interaction between bias blind spot and the
type of activity was further explored by conducting a
simple slope analysis. Bias blind spot was associated
with higher comparative ability ratings on easy activ-
ities (=0013, t=3019, p=00002), and lower compar-
ative ability ratings on difficult activities (= −0014,
Table 5 Regression Results for the Effects of Bias Blind Spot and Type of Activity (Easy vs. Difficult) on (Standardized) Comparative Ability
Ratings (Study 3)
Main effect model Factorial model
Variable baSE t p baSE t p
Intercept 0072 0004 19067 <00001 0072 0004 19053 <00001
Type Activity b1044 0006 24021 <00001 1044 0006 25035 <00001
BBSc0004 0003 1022 0022 0013 0004 3055 <00001
Type Activity b×BBS c0027 0006 4084 <00001
2 Log-likelihood 1,286.19 1,263.85
Degrees of freedom 6 7
Akaike information criterion 1,298.19 1,277.85
Bayesian information criterion 1,324.81 1,308.90
Likelihood ratio test 22.34 (1), p < 00001
aUnstandardized coefficients.
bDummy coded.
cStandardized.
t= −3051, p < 00001). These results suggest that people
high in susceptibility to the bias blind spot are more
susceptible to both the better-than-average effect and
the worse-than-average effect. These results provide
further evidence that the bias blind spot is a form of
egocentric cognition driven by an asymmetric consid-
eration of evidence when evaluating the self and oth-
ers, rather than by a general self-enhancement motive
that induces a person to see herself as superior to
her peers.
Study 4: Bias Blind Spot and
Advice Taking
In Study 4, we began to examine our last set of
predictions, that bias blind spot affects judgments
informed by self–other accuracy comparisons. Specif-
ically, we examined the effect of the bias blind spot on
the weight given to advice from other people (Yaniv
and Kleinberger 2000). If bias blind spot does influ-
ence bias in judgments reliant on self–other accu-
racy comparisons, people who are more susceptible
to the bias blind spot should view their judgments
to be more accurate than the judgments of other
people, based on the belief that they are based on
more objective perceptions of the world (Pronin et al.
2004). Consequently, people high in bias blind spot
should give less weight to advice received from oth-
ers. They should be less likely to correct their initial
(self-informed) judgments when subsequently given
the opportunity to incorporate others’ judgments into
their own.
Method
Participants. A total of 178 AMT workers (109
women; Mage =3308 years, SD =1202) received $8 for
completing Study 4. Participants were 75.3% White,
6.7% Black, 6.7% multiracial, 6.2% Asian, and
1.7% Native American; 3.4% did not indicate their
ethnicity.
Downloaded from informs.org by [4.28.153.60] on 28 April 2015, at 06:23 . For personal use only, all rights reserved.
Scopelliti et al.: Bias Blind Spot: Structure, Measurement, and Consequences
Management Science, Articles in Advance, pp. 1–19, © 2015 INFORMS 13
Materials and Procedure. Participants completed
the 14-item bias blind spot scale in a random order.
Next, they estimated the weight of 30 household
objects (e.g., a trash can, a statue, a doll house), one at
a time. For each object they were shown a color pic-
ture and were provided with the object’s dimensions
(length, height, and width). After making each initial
weight estimate, participants were given advice in the
form of the guess of another “randomly selected par-
ticipant” in the study (i.e., the average estimate of two
pilot participants), and were then given the opportu-
nity to estimate the weight of the object again. Thus,
participants had the option to revise their original
estimate based on the advice that they received from
the other participant. The order in which objects were
judged was random. Finally, participants reported
their age, gender, and ethnicity.
Results and Discussion
We computed bias blind spot scores following the pro-
cedure described in Study 1 (=0082). We computed
the weight on advice (WOA) relative to each object,
by computing the difference between the original esti-
mate and the estimate made after receiving the advice
of the other participant (as in Gino and Moore 2007,
Gino 2008,Harvey and Fischer 1997,Yaniv 2004).
WOA reflects how much a participant uses the advice
she receives (Yaniv 2004) and is defined as
WOA =final estimate initial estimate
advice initial estimate 0
WOA is equal to 0 when participants entirely dis-
count the advice. In such a case, final estimates
are equal to initial estimates, meaning that partici-
pants did not revise their estimates after receiving
the advice. WOA is equal to 1 when participants’
final estimate is equal to the advice received. In
this case participants give maximum weight to the
advice received. Finally, WOA equals a value between
0 and 1, when participants weigh both their initial
estimate and the received advice positively. Follow-
ing the procedure used in prior research (Gino and
Moore 2007,Gino 2008,Yaniv 2004), cases in which
the advice equaled the initial estimate were excluded
from the analysis, since WOA in those cases equaled a
number divided by 0, making it impossible to quan-
tify how much a participant did or did not use the
advice. For cases in which the final estimate did not
fall between the initial estimate and the advice (e.g.,
participants provide revised estimates that are fur-
ther away from the advice than initial estimates),
and WOA was thus greater than 1, values above 1
were adjusted to 1.2One participant whose aggregate
2We estimated the same model on WOA scores that were not sub-
ject to this recommended adjustment, and the effect of susceptibil-
ity to the bias blind spot is even stronger: b1= −0005, SE =0001,
t=3065, p < 00001.
WOA score value was more than five standard devia-
tions above the mean WOA score was excluded from
all subsequent analyses. No other participants were
excluded. Results are based on a total of 5,049 obser-
vations, each observation being the weight of an
object estimated by a participant.
A mixed model analysis was used to estimate the
effect of susceptibility to the bias blind spot on WOA.
Because each participant evaluated the weight of mul-
tiple objects, we let the intercept vary randomly to
take into account the lack of independence between
the observations. We estimated the following model,
WOAij =0+b1BBSi+Ui+ij1
where index ireferred to participants and index j
refers to objects. The dependent variable was thus
the value for WOA for each participant and for each
object. The explanatory variable was each partici-
pant’s bias blind spot score (BBSi5, and Uiindicated
each participant’s random effect. The results revealed
a significant and negative effect of susceptibility to
the bias blind spot on WOA (b1= −0003, SE =0001, t=
3023, p=00001). On average, participants with bias
blind spot scores one standard deviation below the
mean placed six percent more weight on the advice
of others than participants with bias blind spot scores
one standard deviation above the mean (WOA =00223
vs. WOA =00164).
This analysis was complemented by the estimation
of a logit mixed model that examined the relationship
between bias blind spot scores and the rate at which
participants completely ignored advice (WOA =0).
We estimated the following model:
NO_WOAij =0+b1BBSi+Ui+ij1
where index ireferred to participants and index j
refers to objects. The dependent variable was a
dummy variable that indicates the complete igno-
rance of advice (i.e., D=1 when WOA =0; D=0 oth-
erwise) for each participant and for each object. The
explanatory variable was each participant’s bias blind
spot score (BBSi), and Uiindicated each participant’s
random effect. Consistent with the previous model,
the results revealed a significant and positive effect of
susceptibility to the bias blind spot on the likelihood
of completely ignoring advice (b1=00104, SE =00044,
z=2034, p=0002). Susceptibility to the bias blind spot
increased participants’ likelihood of confirming their
initial estimate and completely neglecting advice.
In short, participants who exhibited high levels of
bias blind spot were more likely to ignore the advice
provided, whereas participants who exhibited low
levels of bias blind spot weighed the advice more
heavily. These results are consistent with the results
of Liberman et al. (2012), who showed that naïve
Downloaded from informs.org by [4.28.153.60] on 28 April 2015, at 06:23 . For personal use only, all rights reserved.
Scopelliti et al.: Bias Blind Spot: Structure, Measurement, and Consequences
14 Management Science, Articles in Advance, pp. 1–19, © 2015 INFORMS
realism, by making people believe that their own
perceptions of the world are objective, leads them
to underweight the input of their peers. Our results
extend those by Liberman et al. (2012) by showing
how individual differences in susceptibility to the bias
blind spot significantly affect the extent to which peo-
ple incorporate others’ advice in their judgments. Fur-
thermore, our results provide additional evidence in
support of the role of naïve realism as the process
underlying the bias blind spot (Pronin et al. 2004).
Study 5: Bias Blind Spot and Bias
Reduction
In Study 5, we expanded our analysis of bias blind
spot to see whether people characterized by higher
susceptibility to the bias blind spot are more resis-
tant to debiasing procedures. There is little evidence
that people who are more aware of their own biases
are better able to overcome them. West et al. (2012)
examined the effect of the bias blind spot on the inci-
dence of six cognitive biases, and observed no evi-
dence that people characterized by lower bias blind
spot scores were less likely to commit bias. A rela-
tionship between bias blind spot and decision-making
ability might emerge, however, after an intervention
aimed at reducing bias.
We predicted that the perception that one is less
vulnerable to bias than others may constitute a bar-
rier to the activation of corrective strategies aimed at
avoiding bias (Wilson and Brekke 1994). This barrier
may prevent people who believe that others are more
susceptible to bias from actively responding to train-
ing procedures designed to correct biased judgments.
In order to test this prediction, we had participants
read an article designed to increase awareness of the
occurrence of a specific bias, the fundamental attribu-
tion error (FAE) that also suggested how to mitigate
that bias. We predicted that susceptibility to the bias
blind spot would moderate the effect of this debiasing
training on subsequent commission of the FAE.
Method
Participants. A total of 297 AMT workers (170
women; Mage =3107 years, SD =1007) completed an
online survey and received $2 for their participa-
tion. Participants were 76.4% White, 7.4% multiracial,
6.4% Black, 5.7% Asian, and 1.0% Native American;
3.0% did not indicate their ethnicity.
Procedure. Participants completed the 14-item bias
blind spot scale with items in random order. Next,
participants were randomly assigned to one of two
conditions: debiasing training or control. Participants
assigned to the debiasing training condition read an
explanatory article about the FAE, a tendency to over-
estimate the impact of individual dispositions on
behaviors whereas underestimating the impact of sit-
uational variables (Jones and Harris 1967). The article
described the existence of the FAE, provided exam-
ples of its occurrence, and explained how one might
correct the bias (see the section “debiasing training”
in the appendix). Participants assigned to the con-
trol condition read an article of equal length report-
ing the results of a research study on trust (see the
section “control training” in the appendix). All par-
ticipants then answered nine questions designed to
test the occurrence of the FAE in judgment inspired
by existing paradigms, in a random order (Jones and
Harris 1967,Snyder and Frankel 1976,Scopelliti et al.
2015). Finally, participants reported their age, gender,
and ethnicity.
Results and Discussion
Bias blind spot scores were computed following the
procedure described in Study 1 (=0080). Answers
to the nine FAE questions were highly correlated with
one another (=0080). We summed the number of
questions in which participants exhibited the FAE that
served as our primary dependent variable (M=4017,
SD =2073). The FAE scores were neither normally
distributed, nor could be normalized by log transfor-
mation. Therefore we estimated a Poisson regression,
appropriate to model count dependent variables that
have only non-negative integer values.
We predicted that the debiasing training would
reduce the incidence of bias more for participants
who were less susceptible to the bias blind spot. This
prediction was tested using the procedures outlined
by Aiken and West (1991) to decompose the pre-
dicted interaction using regression analysis. First, bias
blind spot scores were mean-centered by subtract-
ing the mean bias blind spot score from all obser-
vations. Second, we created the interaction term of
(dummy-coded) training factor (debiasing vs. control)
by (mean-centered) bias blind spot score. Next, the
FAE scores were regressed on the training factor (debi-
asing vs. control), bias blind spot scores, and the inter-
action between these two variables.
The analysis revealed a significant interaction
between training and susceptibility to the bias blind
spot (2415=4073, p=0003). To shed light on the
nature of this interaction, we conducted a floodlight
analysis (Spiller et al. 2013) estimating the marginal
effect of training on commission of the FAE at all levels
of susceptibility to the bias blind spot. Table 6reports
the results of this estimation and the comparison of
this factorial model with an alternative model that
includes only the main effects. We used the Johnson-
Neyman technique to identify the range of bias blind
spot scores for which the effect of the debiasing train-
ing was significant. This analysis revealed that there
was a significant positive effect of debiasing training
Downloaded from informs.org by [4.28.153.60] on 28 April 2015, at 06:23 . For personal use only, all rights reserved.
Scopelliti et al.: Bias Blind Spot: Structure, Measurement, and Consequences
Management Science, Articles in Advance, pp. 1–19, © 2015 INFORMS 15
Table 6 Poisson Regression Results for the Effects of Bias Blind Spot and Type of Training on Fundamental Attribution Error Commission (Study 5)
Main effect model Factorial model
Variable baSE Wald 2p baSE Wald 2p
Intercept 1071 0006 819041 <00001 1081 0007 605062 <00001
Type of Training b0036 0006 39027 <00001 0057 0011 26020 <00001
BBSc0008 0003 5087 0002 0015 0005 10048 <00001
Type Trainingb×BBS c0015 0007 4073 0003
2 Log-likelihood 1,471.535 1,466.813
Degrees of freedom 4.000 3
Akaike information criterion 1,477.535 1,474.813
Bayesian information criterion 1,488.617 1,489.588
Likelihood ratio test 4.72 (1), p=0003
aUnstandardized coefficients.
bDummy coded.
cMean centered.
on reducing FAE commission for bias blind spot scores
lower than 2.55 (original scale) (2415=3087, p=0005),
but not for bias blind spot scores higher than 2.55.
The effect of a training procedure designed to
reduce commission of the FAE was stronger for par-
ticipants with a lower susceptibility to the bias blind
spot than for participants with a higher susceptibility
to the bias blind spot. The results suggest that high
susceptibility to the bias blind spot may constitute a
barrier to bias reduction.
General Discussion
The foregoing studies provide an in-depth portrait of
the bias blind spot. We found that the tendency for
people to be better able to identify bias in the judg-
ments and behaviors of others than in their own judg-
ment and behavior is captured by a single latent fac-
tor, and that there is substantial individual variation in
susceptibility to this meta-bias. Using a psychometric
approach, we developed and validated an instrument
to measure individual differences in susceptibility to
the bias blind spot. We found that the bias blind spot
is discriminated from intelligence, cognitive ability,
cognitive reflection, and personality traits related to
self-esteem, self-enhancement, and self-presentation.
Moreover, the bias blind spot does not appear to be a
facet of general decision-making ability or an accurate
perception of the extent to which one is biased. Rather,
susceptibility to the bias blind spot appears to be a
true blind spot in self-awareness, because of asymmet-
ric consideration of evidence when assessing the self
and other people.
Susceptibility to bias blind spot appears to predict
at least two kinds of bias in social judgment. First,
bias blind spot predicts the extent to which people
exhibit biases reliant on similar cognitive mechanisms,
namely the consideration of different evidence in self–
other comparisons. Participants higher in susceptibil-
ity to bias blind spot were more likely to ignore the
abilities of others when judging their own abilities,
resulting in a greater propensity to believe that they
were better than average on easy tasks and worse than
average on difficult tasks than their peers (Study 3).
Second, bias blind spot predicts bias in judgments
reliant on self–other accuracy comparisons. It pre-
dicted the extent to which people ignored the advice
of others, as participants high in bias blind spot placed
less weight on advice than did participants low in bias
blind spot (Study 4). Susceptibility to the bias blind
spot was also associated with reduced responsiveness
to training designed to reduce the incidence of a dif-
ferent judgmental bias (Study 5).
Consequences and Implications of the Bias
Blind Spot
The identification of individual differences in suscepti-
bility to the bias blind spot has important implications
for the incidence of bias and the improvement of judg-
ment and decision making. As we have demonstrated,
awareness of one’s vulnerability to bias is an impor-
tant antecedent of openness to advice that in turn
affects decision quality. Research maintains that inte-
grating advice from external sources improves deci-
sion making (Larrick and Soll 2006,Johnson et al. 2001,
Budescu and Rantilla 2000,Yaniv 2004; see Bonaccio
and Dalal 2006 for a review). People high in bias blind
spot, for example, may be particularly likely to ignore
the advice of peers or experts when engaging in finan-
cial or medical decision making and require alterna-
tive forms of guidance to improve the quality of their
decisions.
Second, awareness of one’s susceptibility to bias
blind spot appears to be an important indicator of
receptivity to efforts to improve one’s decision mak-
ing. Indeed, Wilson and Brekke (1994) suggest that a
critical impediment to the correction of the contami-
nating influence of biases on judgment is people’s lack
of humility about their vulnerability to bias. When
people are unaware of their bias, they are unlikely
Downloaded from informs.org by [4.28.153.60] on 28 April 2015, at 06:23 . For personal use only, all rights reserved.
Scopelliti et al.: Bias Blind Spot: Structure, Measurement, and Consequences
16 Management Science, Articles in Advance, pp. 1–19, © 2015 INFORMS
to adopt corrective strategies to avoid the sources of
bias that influence their judgment. Consequently, peo-
ple who are more susceptible to bias blind spot are
less prone to improve their decision making by engag-
ing in bias reduction strategies, responding to training,
and taking advice. Consequently they may need more
persuasion or evidence that their decision making is
subject to error to acknowledge their potential for bias.
Our results corroborate this view by showing that
people who believe that they are relatively immune to
bias are less likely to enact corrective strategies, even
when correcting strategies are explained and explic-
itly suggested. In the work place, the results suggest
that employees high in bias blind spot may be less
receptive to training designed to improve their deci-
sion making, and may need more or different kinds of
training. At home, the results suggest that people high
in bias blind spot may need more or different forms
of guidance to improve their ability to make conse-
quential decisions affecting their personal life. It has
been proposed that decision making may be a teach-
able skill (Baron and Brown 1991,Bruine de Bruin et al.
2007,Fischhoff 1982,Larrick 2004), and correlational
evidence has suggested that people who have received
formal training in decision making may obtain bet-
ter life outcomes (Larrick et al. 1993). If so, then the
bias blind spot represents an obstacle to improving
the quality of both work and life because it bolsters
resistance to debiasing training aimed at improving
decision-making ability. This influence is not irrevo-
cable. We have found that propensity to exhibit bias
blind spot can be reduced by as much as 39% in a
related research program consisting of scenarios in
which participants could exhibit bias blind spot and
were then provided with critical feedback and training
(Symborski et al. 2014).
Third, some have argued that the bias blind spot is
a critical determinant of misunderstanding, mistrust,
and pessimism when attempting to reach agreements
in interpersonal, political and professional relation-
ships as it obstructs the resolution of conflicts (Pronin
et al. 2004), and have highlighted the importance of
mitigating bias blind spot to promote dialogue, diplo-
macy, and peace processes (Pronin et al. 2004). Pronin
et al. (2006), for example, elucidate the consequences
of perceiving others as more biased than oneself in the
context of terrorism. When terrorists were depicted as
biased and irrational rather than objective and ratio-
nal, people indicated a greater preference for a mili-
tary over a diplomatic resolution to a conflict, which is
likely to result in a spiral of conflict escalation. Assess-
ing individual differences in bias blind spot may help
predict the likelihood of interpersonal conflict, misun-
derstanding, and the need for dispute resolution. For
example, high susceptibility to the bias blind spot may
exacerbate the perception of the gap between antago-
nists in a controversy (i.e., false polarization, Robinson
et al. 1995,Pronin et al. 2002). People high in bias blind
spot, being more likely to believe that their view is
correct, may be less likely to take the necessary steps
to arrive at an agreement. More generally, susceptibil-
ity to the bias blind spot is likely to predict problem-
atic interpersonal interactions when those interactions
require one to acknowledge the potential for bias or
error in one’s own thinking and reasoning.
Limitations and Directions for Future Research
Of course, the present research is not without its own
limitations. We relied upon samples drawn from AMT
that are not representative samples. Fortunately, AMT
samples have been shown to exhibit vulnerability to
biases in judgment and decision making similar to tra-
ditional college samples (Berinsky et al. 2012,Horton
et al. 2011,Paolacci et al. 2010). We thus expect that
the results should generalize to other samples.
In the discriminant validity testing conducted in
Study 2, we collected measures to test multiple rela-
tionships within the same sample (Sample 1). Most of
the correlation coefficients computed on this sample
were not significant, but we did observe three small
significant correlations. Considering the fact that these
significant correlations were obtained within a larger
set of multiple comparisons, it is possible that their
significance level is inflated. We have refrained from
drawing strong inferences from those results.
Although our discriminant validity test shows how
the bias blind spot is distinct from and independent
of several other constructs, in Studies 3–5 we did not
pit our scale against other potential predictors of the
judgments and behaviors investigated. Our choice not
to include measures for which we would expect to
find null results was driven by the tradeoff between
ruling out a potential alternative and using a length-
ier study likely to induce participant fatigue and at an
additional cost. Future research may compare the pre-
dictive power of the bias blind spot scale with that of
other traits and biases.
Conclusion
We find that bias blind spot is a latent factor in
self-assessments of relative vulnerability to bias. This
meta-bias affected the majority of participants in our
samples, but exhibited considerable variance across
participants. We present a concise, reliable, and valid
measure of individual differences in bias blind spot
that has the ability to predict related biases in self-
assessment, advice taking, and responsiveness to bias
reduction training. Given the influence of bias blind
spot on consequential judgments and decisions, as
well as receptivity to training, this measure may prove
Downloaded from informs.org by [4.28.153.60] on 28 April 2015, at 06:23 . For personal use only, all rights reserved.
Scopelliti et al.: Bias Blind Spot: Structure, Measurement, and Consequences
Management Science, Articles in Advance, pp. 1–19, © 2015 INFORMS 17
useful across a broad range of domains such as per-
sonnel assessment, information analysis, negotiation,
consumer decision making, and education.
Acknowledgments
The authors gratefully recognize the research support of a
contract from the Intelligence Advanced Research Projects
Activity (IARPA) via the Air Force Research Laboratory
(AFRL) to Carey K. Morewedge and Karim S. Kassam. The
U.S. Government is authorized to reproduce and distribute
reprints for Governmental purposes notwithstanding any
copyright annotation thereon. 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 IARPA,
AFRL, or the U.S. Government. The authors thank Nino
Miceli for advice on the data analysis.
Appendix. Fundamental Attribution Error
Training
Debiasing Training
Please read the following:
A mistake in judgment that people often make is under-
estimating how much a situation can determine someone’s
behavior. This error occurs when we think another person’s
action tell us something meaningful about their personal-
ity when instead anyone placed in their shoes would have
acted the same way. For example, one might see people
in anxiety-inducing situations (e.g., bungee jumping for the
first time), and infer that they are anxious people in general,
rather than infer that they are average people in a nerve-
racking situation. Similarly, one may think that an author
paid to write in favor of a political leader has actually a
favorable opinion of that political leader.
To combat this error, we have to acknowledge the impor-
tance of situations in determining behavior.
Control Training
Please read the following:
Our decisions to trust people with our money are based
more on how they look then how they behave, according
to a new study. Researchers found people are more likely
to invest money in someone whose face is generally per-
ceived as trustworthy, even when they are given negative
information about this person’s reputation.
The researchers found that 13 out of 15 participants
invested more, on average, in the trustworthy identities.
In a second experiment, the researchers gave the volun-
teers information about whether the trustees had good or
bad histories. Even with this inside information, the average
amount invested in those who looked “trustworthy” was
6% higher.
References
Aiken LS, West SG (1991) Multiple Regression: Testing and Interpret-
ing Interactions (Sage, Newbury Park).
Ames DR, Rose P, Anderson CP (2006) The NPI-16 as a short
measure of narcissism. J. Res. Personality 40:440–450.
Baron J, Brown RV (1991) Teaching Decision Making to Adolescents
(Erlbaum, Hillsdale, NJ).
Berinsky AJ, Huber GA, Lenz GS (2012) Evaluating online labor
markets for experimental research: Amazon.com’s mechanical
Turk. Political Anal. 20:351–368.
Blais AR, Thompson MM, Baranski JV (2005) Individual differences
in decision processing and confidence judgments in compar-
ative judgment tasks: The role of cognitive styles. Personality
and Individual Differences 38:1701–1713.
Bollen K, Lennox R (1991) Conventional wisdom on measurement:
A structural equation perspective. Psych. Bull. 110:305–314.
Bonaccio S, Dalal RS (2006) Advice taking and decision-making:
An integrative literature review, and implications for the orga-
nizational sciences. Organ. Behav. Human Decision Processes 101:
127–151.
Bornstein BH, Zickafoose DJ (1999) “I know I know it, I know I
saw it”: The stability of the confidence–accuracy relationship
across domains. J. Experiment. Psych.: Appl. 5:76–88.
Brodnick RJ, Ree MJ (1995) A structural model of academic perfor-
mance, socio-economic status, and Spearman’s g.Educational
Psych. Measurement 55:594–605.
Bruine de Bruin W, Parker AM, Fischhoff B (2007) Individual dif-
ferences in Adult Decision-Making Competence. J. Personality
Soc. Psych. 92:938–956.
Budescu DV, Rantilla AK (2000) Confidence in aggregation of
expert opinions. Acta Psychologica 104:371–398.
Cacioppo JT, Petty RE (1982) The need for cognition. J. Personality
Soc. Psych. 42:116–131.
Cacioppo JT, Petty RE, Kao CE (1984) The efficient assessment of
need for cognition. J. Personality Assessment 48:306–307.
Cacioppo JT, Petty RE, Feinstein JA, Jarvis WBG (1996) Disposi-
tional differences in cognitive motivation: The life and times
of individuals varying in need for cognition. Psych. Bull. 119:
197–253.
Cohen J (1988) Statistical Power Analysis for the Behavioral Sciences,
2nd ed. (Lawrence Erlbaum Associates, Hillsdale, NJ).
Crowne DP, Marlowe D (1960) A new scale of social desirabil-
ity independent of psychopathology. J. Consulting Psych. 24:
349–354.
Cronbach LJ, Meehl PE (1955) Construct validity in psychological
tests. Psych. Bull. 52:281–302.
Dunning D, Johnson K, Ehrlinger J, Kruger J (2003) Why people
fail to recognize their own incompetence. Current Directions
Psych. Sci. 12:83–87.
Edwards JR, Bagozzi RP (2000) On the nature and direction of the
relationship between constructs and measures. Psych. Methods
5:155–174.
Epley N, Dunning D (2000) Feeling “holier than thou”: Are self-
serving assessments produced by errors in self- or social pre-
diction? J. Personality Soc. Psych. 79:861–875.
Feingstein A, Scheier M, Buss A (1975) The self-consciousness scale.
J. Consulting and Clinical Psych. 43:522–527.
Fischhoff B (1982) Debiasing. Kahneman D, Slovic P, Tversky A,
eds. Judgment Under Uncertainty: Heuristics and Biases (Cam-
bridge University Press, Cambridge), 422–444.
Frederick S (2005) Cognitive reflection and decision making.
J. Econom. Perspect. 19:24–42.
Frey MC, Detterman DK (2004) Scholastic assessment or g? The
relationship between the scholastic assessment test and general
cognitive ability. Psych. Sci. 15:373–378.
Gino F (2008) Do we listen to advice just because we paid for
it? The impact of advice cost on its use. Organ. Behav. Human
Decision Processes 107:234–245.
Gino F, Moore DA (2007) Effects of task difficulty on use of advice.
J. Behav. Decision Making 20:21–35.
Harvey N, Fischer I (1997) Taking advice: Accepting help, improv-
ing judgment and sharing responsibility. Organ. Behav. Human
Decision Processes 70:117–133.
Horn JL (1965) A rationale and test for the number of factors in
factor analysis. Psychometrika 32:179–185.
Horton JJ, Rand DG, Zeckhauser RJ (2011) The online laboratory:
Conducting experiments in a real labor market. Experiment.
Econom. 14:399–425.
Downloaded from informs.org by [4.28.153.60] on 28 April 2015, at 06:23 . For personal use only, all rights reserved.
Scopelliti et al.: Bias Blind Spot: Structure, Measurement, and Consequences
18 Management Science, Articles in Advance, pp. 1–19, © 2015 INFORMS
John OP, Srivastava S (1999) The Big Five trait taxonomy: History,
measurement and theoretical perspectives. Pervin LA, John
OP, eds. Handbook of Personality: Theory and Research (Guilford,
New York), 102–138.
Johnson TR, Budescu DV, Wallsten TS (2001) Averaging probabil-
ity judgments: Monte Carlo analyses of asymptotic diagnostic
value. J. Behavioral Decision Making 14:123–140.
Jones EE, Harris VA (1967) The attribution of attitudes. J. Experi-
ment. Soc. Psych. 3:1–24.
Kahneman D, Slovic P, Tversky A (1982) Judgment Under Uncer-
tainty: Heuristics and Biases (Cambridge University Press,
New York).
Klayman J, Soll JB, Gonzalez-Vallejo C, Barlas S (1999) Overconfi-
dence: It depends on how, what and whom you ask. Organ.
Behav. Human Decision Processes 79:216–247.
Kruger J (1999) Lake Wobegon be gone! The “below-average effect”
and the egocentric nature of comparative ability judgments. J.
Personality Soc. Psych. 77:221–232.
Kruger J (2003) Return of the ego—Self-referent information as a
filter for social prediction: Comment on Karniol (2003). Psych.
Rev. 110:585–590.
Kruger J, Dunning D (1999) Unskilled and unaware of it: How
difficulties in recognizing one’s own incompetence lead to
inflated self-assessments. J. Personality Soc. Psych. 77:1121–1134.
Larrick RP (2004) Debiasing. Koehler DJ, Harvey N, eds. Blackwell
Handbook of Judgment and Decision Making (Blackwell Publish-
ers, Oxford, UK), 316–337.
Larrick RP, Soll JB (2006) Intuitions about combining opinions:
Misappreciation of the averaging principle. Management Sci.
52:111–127.
Larrick RP, Nisbett RE, Morgan JN (1993) Who uses the cost-
benefit rules of choice? Implications for the normative status of
microeconomic theory. Organ. Behav. Human Decision Processes
56:331–347.
Levin IP, Gaeth GJ, Schreiber J, Lauriola M (2002) A new look
at framing effects: Distribution of effect sizes, individual dif-
ferences, and independence of types of effects. Organ. Behav.
Human Decision Processes 88:411–429.
Liberman V, Minson JA, Bryan CJ, Ross L (2012) Naïve realism and
capturing the “wisdom of dyads.” J. Experiment. Soc. Psych.
48:507–512.
Macmillan NA, Creelman CD (1991) Detection Theory: A User’s
Guide (Cambridge, New York).
Mardia KV (1980) 9 Tests of unvariate and multivariate normal-
ity. Rao CR, ed. Handbook of Statistics, Vol. 1 (Elsevier B.V.,
Amsterdam), 279–320.
McCrae RR, Costa PT (1987) Validation of the five-factor model
of personality across instruments and observers. J. Personality
Soc. Psych. 52:81–90.
Morewedge CK, Kahneman D (2010) Associative processes in intu-
itive judgment. Trends in Cognitive Sci. 14:435–440.
Morf CC, Rhodewalt F (2001) Unraveling the paradoxes of nar-
cissism: A dynamic self-regulatory processing model. Psych.
Inquiry 12:177–196.
Nisbett RE, Ross L (1980) Human Inference: Strategies and Shortcom-
ings of Social Judgment (Prentice-Hall, Englewood Cliffs, NJ).
Nisbett RE, Wilson TD (1977) Telling more than we can know:
Verbal reports on mental processes. Psych. Rev. 84:231–259.
Nunnally JC (1978) Psychometric Theory, 2nd ed. (McGraw-Hill,
New York).
Paolacci G, Chandler J, Ipeirotis PG (2010) Running experiments
using Amazon mechanical Turk. Judgment and Decision Making
5:411–419.
Paulhus DL (1998) Interpersonal and intrapsychic adaptiveness of
trait self-enhancement: A mixed blessing? J. Personality Soc.
Psych. 74:1197–1208.
Paulhus DL, Harms PD, Bruce MN, Lysy DC (2003) The over-
claiming technique: Measuring self-enhancement independent
of ability. J. Personality Soc. Psych. 84:890–904.
Pronin E (2007) Perception and misperception of bias in human
judgment. Trends in Cognitive Sci. 11:37–43.
Pronin E (2008) How we see ourselves and how we see others.
Science 320:1177–1180.
Pronin E, Kugler MB (2007) Valuing thoughts, ignoring behavior:
The introspection illusion as a source of the bias blind spot. J.
Experiment. Soc. Psych. 43:565–578.
Pronin E, Lin DY, Ross L (2002) The bias blind spot: Perceptions
of bias in self versus others. Personality Soc. Psych. Bull. 28:
369–381.
Pronin E, Gilovich TD, Ross L (2004) Objectivity in the eye of the
beholder: Divergent perceptions of bias in self versus others.
Psych. Rev. 111:781–799.
Pronin E, Kennedy K, Butsch S (2006) Bombing versus negotiating:
How preferences for combating terrorism are affected by per-
ceived terrorist rationality. Basic Appl. Social Psych. 28:385–392.
Robbins SB, Patton MJ (1985) Self-psychology and career devel-
opment: Construction of the superiority and goal instability
scales. J. Counseling Psych. 32:221–231.
Robinson RJ, Keltner D, Ward A, Ross L (1995) Actual versus
assumed differences in construal: “Naive realism” in inter-
group perception and conflict. J. Personality Soc. Psych. 68:
404–417.
Rosenberg M (1965) Society and the Adolescent Self-Image (Princeton
University Press, Princeton, NJ).
Ross L, Ward A (1995) Psychological barriers to dispute resolution.
Adv. Experiment. Soc. Psych. 27:255–304.
Ross L, Ward A (1996) Naiïve realism in everyday life: Implications
for social conflict and misunderstanding. Reed ES, Turiel E,
Brown T, eds. Values and Knowledge. The Jean Piaget Symposium
Series (Lawrence Erlbaum Associates, Hillsdale, NJ), 103–135.
Scopelliti I, Morewedge CK, Min HL, McCormick E, Kassam KS
(2015) Neglect of external demands (NED) scale: Measuring
the coherence, consequences, and correction of correspondent
inference making. Working paper, City Univeristy London,
London.
Symborski C, Barton M, Quinn M, Morewedge CK, Kassam K,
Korris J (2014) Missing: A serious game for the mitigation
of cognitive bias. Interservice/Industry Training, Simulation Ed.
Conf., Orlando, FL, December.
Snyder M (1974) Self-monitoring of expressive behavior. J. Person-
ality Soc. Psych. 30:526–537.
Snyder ML, Frankel A (1976) Observer bias: A stringent test
of behavior engulfing the field. J. Personality Soc. Psych. 34:
857–864.
Snyder CR, Fromkin HL (1977) Abnormality as a positive char-
acteristic: Development and validation of a scale measuring
need for uniqueness. J. Abnormal Psych. 86:518–527.
Snyder CR, Fromkin HL (1980) Uniqueness (Plenum, New York).
Spiller SA, Fitzsimons GJ, Lynch JG Jr, McClelland GH (2013)
Spotlights, floodlights, and the magic number zero: Simple
effects tests in moderated regression. J. Marketing Res. 50:
277–288.
Stankov L, Crawford JD (1996) Confidence judgments in studies
of individual differences. Personality and Individual Differences
21:971–986.
Stankov L, Crawford JD (1997) Self-confidence and performance
on tests of cognitive abilities. Intelligence 25:93–109.
Stanovich KE (1999) Who Is Rational? Studies of Individual Differences
in Reasoning (Psychology Press, Mahwah, NJ).
Stanovich KE, West RF (2000) Individual differences in reasoning:
Implications for the rationality debate? Behavioral Brain Sci.
23:645–665.
Stanovich KE, West RF (2007) Natural myside bias is independent
of cognitive ability. Thinking and Reasoning 13:225–247.
Stanovich KE, West RF (2008) On the relative independence of
thinking biases and cognitive ability. J. Personality Soc. Psych.
94:672–695.
Taylor SE, Brown JD (1988) Illusion and well-being: A social
psychological perspective on mental health. Psych. Bull. 103:
193–210.
Tversky A, Kahneman D (1974) Judgment under uncertainty:
Heuristics and biases. Science 185:1124–1130.
Downloaded from informs.org by [4.28.153.60] on 28 April 2015, at 06:23 . For personal use only, all rights reserved.
Scopelliti et al.: Bias Blind Spot: Structure, Measurement, and Consequences
Management Science, Articles in Advance, pp. 1–19, © 2015 INFORMS 19
Unsworth N, Engle RW (2007) The nature of individual differences
in working memory capacity: Active maintenance in primary
memory and controlled search from secondary memory. Psych.
Rev. 114:104–132.
Van Boven L, Dunning D, Loewenstein G (2000) Egocentric empa-
thy gaps between owners and buyers: Misperceptions of the
endowment effect. J. Personality Soc. Psych. 79:66–76.
West RF, Stanovich KE (1997) The domain specificity and gener-
ality of overconfidence: Individual differences in performance
estimation bias. Psychonomic Bull. Rev. 4:387–392.
West RF, Toplak ME, Stanovich KE (2008) Heuristics and biases as
measures of critical thinking: Associations with cognitive abil-
ity and thinking dispositions. J. Educational Psych. 100: 930–941.
West RF, Meserve RJ, Stanovich KE (2012) Cognitive sophistication
does not attenuate the bias blind spot. J. Personality Soc. Psych.
103:506–513.
Wilson TD, Brekke N (1994) Mental contamination and mental cor-
rection: Unwanted influences on judgments and evaluations.
Psych. Bull. 116(1):117–142.
Wilson TD, Dunn EW (2004) Self-knowledge: Its limits, value, and
potential for improvement. Ann. Rev. Psych. 55:493–518.
Wilson TD, Houston CE, Etling KM, Brekke N (1996) A new look at
anchoring effects: Basic anchoring and its antecedents. J. Exper-
iment. Psych.: General 125:387–402.
Yaniv I (2004) Receiving other people’s advice: Influence and ben-
efit. Organ. Behav. Human Decision Processes 93:1–13.
Yaniv I, Kleinberger E (2000) Advice taking in decision mak-
ing: Egocentric discounting and reputation formation. Organ.
Behav. Human Decision Processes 83:260–281.
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
You are free to download this work and share with others for any purpose, even commercially if you
distribute your contributions under the same license as the original, and you must attribute this work
as “Management Science. Copyright 2015 INFORMS. http://dx.doi.org/10.1287/mnsc.2014.2096, used
under a Creative Commons Attribution License: http://creativecommons.org/licenses/by-sa/4.0/.”
Downloaded from informs.org by [4.28.153.60] on 28 April 2015, at 06:23 . For personal use only, all rights reserved.
... That is, an average person thinks they are less biased than an average person, a belief that is clearly incorrect. A caveat of the bias blind spot is that people overestimate their accuracy mostly for easy tasks, but also tend to underestimate their decision-making abilities in difficult tasks (Scopelliti et al., 2015). This bias has far-reaching consequences for everyday decision-making-for example, we tend to overestimate the accuracy of our beliefs and neglect the opinions of others. ...
... Apart from the use of numerous biases uncovered by psychological research, we also use a series of biases that have been uncovered by economists. We used 13 psychological biases from previous research on the bias blind spot (Scopelliti et al., 2015), and 7 economic biases that we selected for this study. These 20 biases were presented in a random order to every participant. ...
... This study was pre-registered at: https://aspredicted.org/e4cg8.pdf. We made one deviation from the preregistration: due to a coding error, we have not collected data for one bias (the ostrich effect) in the L2 condition, and thus investigated 13 psychological biases instead of the 14 biases used in Scopelliti et al. (2015). Data and materials are available at: https://osf.io/j43df/. ...
Article
Aims and objectives/purpose/research questions Extant research suggests that processing information in a second language (L2) affects decision-making, possibly by affecting metacognition. We hypothesized that processing in L2 will reduce the bias blind spot effect, whereby people (on average) erroneously think that they are less susceptible to biases than others. Design/methodology/approach In Experiment 1, participants assessed their susceptibility and the susceptibility of others to 13 psychological and 7 economic biases, in either L1 (Polish) or L2 (English). In Experiment 2, participants assessed the 7 most severe bias blind spots from Experiment 1. Data and analysis We recruited 500 participants for each experiment via Prolific (832 overall, after exclusions). The main hypothesis and moderators were tested via mixed-model regressions. Findings/conclusions In Experiment 1, participants showed an overall bias blind spot, which decreased in the L2 condition, but only for psychological biases. In Experiment 2, we replicated the L2-bias blind spot attenuation effect. An exploratory analysis suggests that the effect of L2 is the result of both lower ratings of other-susceptibility and higher ratings of self-susceptibility. Originality Our study provides unique insights into how L2 affects metacognition. We are the first to study how the use of L2 can attenuate the bias blind spot. Our findings provide rare support for the psychological distancing (“birds-eye view”) explanation for the foreign language effect. Significance/implications Bilinguals using L2 showed some resilience to the bias blind spot, suggesting metacognition is language dependent.
... Bias can lead to poor judgments and decisions in domains such as intelligence analysis (Dhami et al., 2019;Morewedge et al., 2015), medicine (Arkes, 2013;Bornstein & Emler, 2001) and forensics (Kassim et al., 2013) to name a few. This is compounded by the fact that individuals, including children (Elashi & Mills, 2015;Hagá et al., 2018), often report that they are less susceptible to biases than others-a phenomenon known as the bias blind spot (BBS; Pronin et al., 2002;Pronin & Kugler, 2007;Scopelliti et al., 2015;West et al., 2012). ...
... The positive relation between the BBS and cognitive sophistication also has not been consistently observed. For instance, Scopelliti et al. (2015) found no relation between scores on their BBS scale and self-reported verbal SAT scores and a small negative correlation with self-reported math SAT scores. However, since West et al. (2012) only used cognitive bias items and Scopelliti et al. (2015) only used social/motivational items, it is possible that the discrepancy is attributable to the different domains. ...
... For instance, Scopelliti et al. (2015) found no relation between scores on their BBS scale and self-reported verbal SAT scores and a small negative correlation with self-reported math SAT scores. However, since West et al. (2012) only used cognitive bias items and Scopelliti et al. (2015) only used social/motivational items, it is possible that the discrepancy is attributable to the different domains. Another limitation of all previous studies examining the link between the BBS and cognitive sophistication is that the primary measurecognitive ability-was always based on self-reported SAT scores. ...
Preprint
Full-text available
Individuals often assess themselves as being less susceptible to common biases compared to others. This bias blind spot (BBS) is thought to represent a metacognitive error. In this research, we tested three explanations for the effect: The cognitive sophistication hypothesis posits that individuals who display the BBS more strongly are actually less biased than others. The introspection bias hypothesis posits that the BBS occurs because people rely on introspection more when assessing themselves compared to others. The conversational processes hypothesis posits that the effect is largely a consequence of the pragmatic aspects of the experimental situation rather than true metacognitive error. In two experiments (N = 1057) examining 18 social/motivational and cognitive biases, there was strong evidence of the BBS. Among the three hypotheses examined, the conversational processes hypothesis attracted the greatest support, thus raising questions about the extent to which the BBB is a metacognitive effect.
... In design thinking, a practical approach has been to address this by involving users and colleagues and, more generally, by adopting an empathic lens in the design process (e.g., Wright & McCarthy, 2008;Kouprie & Visser 2009;Köppen & Meinel, 2015). Such attempts, however, face a critical challenge: UX practitioners, just as everyone else, might be subject to a bias blind spot where "they are less likely to detect bias in themselves than in others" (Scopelliti et al., 2015(Scopelliti et al., , p. 2468. The result is a tendency to ignore where and how one's own types of bias could be addressed despite the potentially harmful effects of bias-induced errors during design-based decision-making. ...
... The bias blind spot is a systematic information-processing asymmetry where people are less likely to detect bias in themselves than in others (Scopelliti et al., 2015). Several other biases are thought to contribute to the bias blind spot. ...
... Although some have questioned their effectiveness (Oliver et al., 2017), most studies suggest a positive effect of debiasing interventions (Arkes et al., 1988;Shaffer et al., 2016;Ludolph & Schulz, 2018), for instance as video games and direct warning (Lewandowsky et al., 2012). The challenge with such interventions is that people might be oblivious (blind) to their own susceptibility to cognitive bias (Scopelliti et al., 2015). ...
... If no form of data was available in a paper, we contacted the corresponding and/or first author via email. Two attempts were made to contact authors and effects were excluded when we received no response or were informed that data were no longer available (i.e., Sciandra, 2019;Scopelliti et al., 2015;See et al., 2011;Study 4;Sniezek et al., 2004;Tzioti et al., 2014;Yaniv, 2004a;Yaniv & Kleinberger, 2000, Studies 2 to 4;Wanzel et al., 2017). ...
Article
Full-text available
The degree to which people take advice, and the factors that influence advice-taking, are of broad interest to laypersons, professionals, and policy-makers. This meta-analysis on 346 effect sizes from 129 independent datasets ( N = 17, 296) assessed the weight of advice in the judge-advisor system paradigm, as well as the influence of sample and task characteristics. Information about the advisor(s) that is suggestive of advice quality was the only unique predictor of the overall pooled weight of advice. Individuals adjusted estimates by 32%, 37%, and 48% in response to advisors described in ways that suggest low, neutral, or high quality advice, respectively. This indicates that the benefits of compromise and averaging may be lost if accurate advice is perceived to be low quality, or too much weight is given to inaccurate advice that is perceived to be high quality. When examining the three levels of perceived quality separately, advice-taking was greater for subjective and uncertain estimates, relative to objective estimates, when information about the advisor was neutral in terms of advice quality. Sample characteristics had no effect on advice-taking, thus providing no evidence that age, gender, or individualism influence the weight of advice. The findings contribute to current theoretical debates and provide direction for future research.
Chapter
Beliefs play a central role in our lives. They lie at the heart of what makes us human, they shape the organization and functioning of our minds, they define the boundaries of our culture, and they guide our motivation and behavior. Given their central importance, researchers across a number of disciplines have studied beliefs, leading to results and literatures that do not always interact. The Cognitive Science of Belief aims to integrate these disconnected lines of research to start a broader dialogue on the nature, role, and consequences of beliefs. It tackles timeless questions, as well as applications of beliefs that speak to current social issues. This multidisciplinary approach to beliefs will benefit graduate students and researchers in cognitive science, psychology, philosophy, political science, economics, and religious studies.
Article
Beliefs play a central role in our lives. They lie at the heart of what makes us human, they shape the organization and functioning of our minds, they define the boundaries of our culture, and they guide our motivation and behavior. Given their central importance, researchers across a number of disciplines have studied beliefs, leading to results and literatures that do not always interact. The Cognitive Science of Belief aims to integrate these disconnected lines of research to start a broader dialogue on the nature, role, and consequences of beliefs. It tackles timeless questions, as well as applications of beliefs that speak to current social issues. This multidisciplinary approach to beliefs will benefit graduate students and researchers in cognitive science, psychology, philosophy, political science, economics, and religious studies.
Article
Full-text available
A crisis requires the affected population, governments or non-profit organizations, as well as crisis experts, to make urgent and sometimes life-critical decisions. With the urgency and uncertainty they create, crises are particularly amenable to inducing cognitive biases that influence decision-making. However, there is limited empirical evidence regarding the impact of cognitive biases on estimation, judgment, and decision-making tasks in crises. Possible biases occurring in crises are: (1) to be influenced by how information is framed (i.e., framing effect), (2) to overly rely on information that confirms rather than opposes preliminary assumptions (i.e., confirmation bias), (3) to rely heavily on a skewed informational cue when making estimations (i.e., anchoring bias), and (4) to see the own decision-making as less biased than decision-making of others (i.e., bias blind spot). We investigate these four cognitive biases using three online survey experiments targeting crisis-affected people of the general public (n = 460, mTurk workers), governmental and non-profit workers (n = 50, mTurk workers), and crisis experts (n = 21, purposefully sampled). Our findings show that crisis experts are the least biased group but are still significantly affected by anchoring, framing, and bias blind spot. Crisis-affected people from the general public showed the strongest susceptibility to all four biases studied. The findings have implications for future research on crisis information systems (IS) design. As crisis response is increasingly facilitated through IS, we propose debiasing functions that account for biased user behavior in crises.
Article
When teaching statistics, educators sometimes overestimate their students' knowledge and abilities. This is due to the curse of knowledge, a cognitive bias that causes people—especially experts—to overestimate how likely others are to know and understand the same things as them. This can lead to various issues, including struggling to communicate with students, and making students feel less comfortable in the classroom. To address this, educators should first identify situations where this bias can affect their teaching. In doing so, they should consider relevant risk factors, and potentially also solicit feedback from relevant individuals. Then, educators can reduce this bias and its impact on their teaching by using techniques such as keeping the curse of knowledge and their audience in mind, assessing students' knowledge, assuming lack of knowledge unless there is strong evidence to the contrary, and avoiding saying that things are obvious.
Article
Full-text available
Reactions to trait self-enhancers were investigated in 2 longitudinal studies of person.perception in discussion groups. Groups of 4-6 participants met 7 times for 20 rain. After Meetings 1 and 7, group members rated their perceptions of one another. In Study 1, trait self-enhancement was indexed by measures of narcissism and self-deceptive enhancement. At the first meeting, self-enhancers made positive impressions: They were seen as agreeable, well adjusted, and competent. After 7 weeks, however, they were rated negatively and gave self-evaluations discrepant with peer evaluations they received. In Study 2, an independent sample of observers (close acquaintances) enabled a pretest index of discrepancy self-enhancement: It predicted the same deteriorating pattern of interpersonal perceptions as the other three trait measures. Nonetheless, all self-enhancement measures correlated positively with self-esteem.
Article
Evidence is reviewed which suggests that there may be little or no direct introspective access to higher order cognitive processes. Subjects are sometimes (a) unaware of the existence of a stimulus that importantly influenced a response, (b) unaware of the existence of the response, and (c) unaware that the stimulus has affected the response. It is proposed that when people attempt to report on their cognitive processes, that is, on the processes mediating the effects of a stimulus on a response, they do not do so on the basis of any true introspection. Instead, their reports are based on a priori, implicit causal theories, or judgments about the extent to which a particular stimulus is a plausible cause of a given response. This suggests that though people may not be able to observe directly their cognitive processes, they will sometimes be able to report accurately about them. Accurate reports will occur when influential stimuli are salient and are plausible causes of the responses they produce, and will not occur when stimuli are not salient or are not plausible causes.
Article
Many decisions are based on beliefs concerning the likelihood of uncertain events such as the outcome of an election, the guilt of a defendant, or the future value of the dollar. Occasionally, beliefs concerning uncertain events are expressed in numerical form as odds or subjective probabilities. In general, the heuristics are quite useful, but sometimes they lead to severe and systematic errors. The subjective assessment of probability resembles the subjective assessment of physical quantities such as distance or size. These judgments are all based on data of limited validity, which are processed according to heuristic rules. However, the reliance on this rule leads to systematic errors in the estimation of distance. This chapter describes three heuristics that are employed in making judgments under uncertainty. The first is representativeness, which is usually employed when people are asked to judge the probability that an object or event belongs to a class or event. The second is the availability of instances or scenarios, which is often employed when people are asked to assess the frequency of a class or the plausibility of a particular development, and the third is adjustment from an anchor, which is usually employed in numerical prediction when a relevant value is available.