ArticlePDF Available

If You Judge, Investigate! Responsibility Reduces Confirmatory Information Processing in Legal Experts


Abstract and Figures

Fair and well justified judicial decisions require that judges evaluate and interpret all relevant facts. However, heuristics and other shortcuts are used here as well. Additionally, it has been demonstrated that experts may be subject to the same decision biases as laypeople. Therefore, we investigated whether and to what extent judicial experts are protected against confirmatory information processing (CIP), a tendency to seek out (selective exposure) and evaluate information more positively (biased assimilation) when it confirms one's own preliminary decision. Results indicate that legal experts (judges, prosecutors, and defense-lawyers) evaluated information supporting their preliminary decision more positively than conflicting information. However, there is a clear expertise effect: domain-specific experts (e.g., criminal-law experts deciding a criminal-law case) showed less CIP than general experts (legal professionals with specializations in other fields than criminal-law), who did not differ from laypeople (pilot study). We further investigated whether either decision-certainty, knowledge, or a feeling of responsibility can be identified as potential underlying mechanism of this expertise effect. Higher decision-certainty or prior knowledge did not correlate with CIP (pilot study). In our main study, general experts significantly reduced their CIP to the same level as domain-specific experts if we induced responsibility. Without this induction general experts again showed more CIP than domain-specific experts. This implies a motivational explanation for the lower CIP in domain-specific experts. The advantage of specialized judges over general judges will be discussed. (PsycINFO Database Record
Content may be subject to copyright.
Published in Psychology, Public Policy, and Law, 22(4), 386-400, ©American Psychological
Association, 2016. This paper is not the copy of record and may not exactly replicate the
authoritative document published in the APA journal. Please do not copy or cite without
author's permission. The final article is available at:
If You Judge, Investigate! Responsibility Reduces
Confirmatory Information Processing in Legal Experts
Susanne M. Schmittat
Johannes Kepler University Linz
Birte Englich
University of Cologne
Author Note
Susanne M. Schmittat, Criminal-Law Department, Johannes Kepler University Linz, Austria
Birte Englich, Psychology Department, University of Cologne, Germany
Correspondence concerning this article should be addressed to Susanne M. Schmittat, Criminal-
Law Department, Johannes Kepler University Linz, Altenberger Strasse 69, 4040 Linz, Austria.
This research was supported by a grant from the Volkswagen Stiftung
Fair and well justified judicial decisions require that judges evaluate and interpret all
relevant facts. However, heuristics and other shortcuts are used here as well. Additionally, it
has been demonstrated that experts may be subject to the same decision biases as laypeople.
Therefore, we investigated whether and to what extent judicial experts are protected against
confirmatory information processing (CIP), a tendency to seek out (selective exposure) and
evaluate information more positively (biased assimilation) when it confirms one’s own
preliminary decision. Results indicate that legal experts (judges, prosecutors, and defense-
lawyers) evaluated information supporting their preliminary decision more positively than
conflicting information. However, there is a clear expertise effect: domain-specific experts
(e.g., criminal-law experts deciding a criminal-law case) showed less CIP than general
experts (legal professionals with specializations in other fields than criminal-law), who did
not differ from laypeople (pilot study). We further investigated if either decision-certainty,
knowledge, or a feeling of responsibility can be identified as potential underlying mechanism
of this expertise effect. Higher decision-certainty or prior knowledge did not correlate with
CIP (pilot study). In our main study, general experts significantly reduced their CIP to the
same level as domain-specific experts if we induced responsibility. Without this induction
general experts again showed more CIP than domain-specific experts. This implies a
motivational explanation for the lower CIP in domain-specific experts. The advantage of
specialized judges over general judges will be discussed.
Keywords: legal decision making, confirmatory information processing, expertise,
responsibility, decision-certainty
If You Judge, Investigate! Responsibility Reduces
Confirmatory Information Processing in Legal Experts
DNA-exoneration cases as well as experimental research have shown that legal
decision making is not free of decision biases and human error (e.g., Danziger, Levav, &
Avnaim-Pesso, 2011; Englich, Mussweiler, & Strack, 2006; Kassin & Gudjonsson, 2004).
However, in order to prevent biases and wrongful judgments, the legal system works with
safeguards at various stages. For instance, legal expertise in form of well-educated and highly
trained judges is supposed to ensure due process when presiding over jury trials, but also in
bench trails or when they pass sentences (Eisenberg & Clermont, 1996; Guthrie, Rachlinski,
& Wistrich, 2001; Rachlinski, Guthrie, & Wistrich, 2007; Wistrich, Guthrie, & Rachlinski,
2005). However, does the experts’ knowledge and experience actually help them to
counteract common “mistakes”? Or may it even be the case that legal expertise facilitates
biases due to the experts’ reliance on intuition (Benner & Tanner, 1987; Dreyfus & Dreyfus,
1986)? If expertise is in fact a protective factor, does it make a difference if judges have a
general legal training or if they are specialized in one area, for instance, criminal-law? With
the aim to qualify expertise as a protective factor we investigated confirmatory information
processing in specialized and general legal experts and laypeople. Furthermore, we
investigated if knowledge, confidence or responsibility constitutes legal expertise as a
protective factor.
Our research aims to shed light on these questions by focusing on one particular bias:
People’s strong affinity to confirm prior beliefs, opinions, attitudes or decisions through
biased information selection and evaluation a tendency called confirmatory information
processing (P. Fischer, Greitemeyer, & Frey, 2008). Confirming a prior belief seems to be a
rather natural behavior (Wason, 1966) compared to the more demanding action of
disconfirming an opinion (Klayman & Ha, 1987). Thus, people intuitively work towards
proving a rule, or maintaining an opinion, instead of refuting it. It is argued that when legal
experts are confronted with a case they instantly form a preliminary decision, which they may
base on their intuition, moral values or develop in accordance to their personality,
background, and political affiliations (e.g., Johnson, 2006; Kulik, Perry, & Pepper, 2003).
The current focus is on what happens afterwards: While checking more carefully for the
adequate sentence does this preliminary decision lead to confirmatory information processing
and thus to potentially biased verdicts?
In the following studies, confirmatory information processing is investigated in actual
judges, prosecutors, and defense lawyers (in our pilot study laypeople were additionally
recruited). To our knowledge, up to now there is no empirical study that addresses
confirmatory information processing in genuine legal experts. Yet, it is important to
understand how legal experts perceive trial information after forming preliminary opinions
and to try to counteract possible biases in order to ensure fair verdicts.
Our studies were conducted with legal experts in an inquisitorial system.
Nevertheless, we believe that our research can also be applied to adversarial systems. Even
though judges in adversarial systems are often not responsible for the decisions, they still
determine the outcome of dispositive motions and settlements, and they preside over bench
trials (Eisenberg & Clermont, 1996; Wistrich et al., 2005). Thus, they are involved in
countless decision making processes, which can all potentially be prone to confirmatory
information processing.
Confirmatory Information Processing
Confirmatory information processing (CIP) subsumes two strongly related tendencies:
selective exposure and biased assimilation (P. Fischer, Greitemeyer, et al., 2008). Selectively
(consciously or unconsciously) choosing only information that supports one’s beliefs and
disregarding conflicting information is known as selective exposure (Festinger, 1957; P.
Fischer, Schulz-Hardt, & Frey, 2008). Biased assimilation on the other hand is the
overestimation of the quality of supporting information (Lord, Ross, & Lepper, 1979). CIP
(also known as confirmation bias or congeniality bias) can range from ignoring conflicting
information to biased interpretation of ambiguous information (Nickerson, 1998) as well as
distorting new information towards a tentative choice (Russo, Carlson, Meloy, & Yong,
CIP is typically investigated with the following paradigm: The participant is presented
with a decision problem with two options. After a (preliminary) decision, more decision-
relevant information in form of arguments is presented to the participant. The amount of
these arguments may vary (P. Fischer, Schulz-Hardt, et al., 2008), however, half of them
always challenge the participants’ preliminary decision and the other half supports it. Next,
participants evaluate each argument on a number of dimensions related to quality (i.e.,
strength, importance, credibility) in order to measure biased assimilation. They are also asked
to select the arguments that they would like to study further, which assesses selective
exposure (e.g., P. Fischer, Jonas, Frey, & Schulz-Hardt, 2005). If supporting arguments
receive a higher quality rating and / or are selected more often than conflicting arguments,
one speaks of confirmatory information processing (P. Fischer, Schulz-Hardt, et al., 2008).
Prior research has identified conditions under which CIP seems to increase, for
instance, when the additional information is presented sequentially rather than simultaneously
(Jonas, Schulz-Hardt, Frey, & Thelen, 2001), when participants are informed that they would
have to describe their attitude towards that topic (Smith, Fabrigar, Powell, & Estrada, 2007),
when the decision problem is framed in terms of gains instead of losses (P. Fischer, Jonas,
Frey, & Kastenmüller, 2008), when self-regulatory resources are depleted (P. Fischer,
Greitemeyer, et al., 2008) or when participants are in a negative mood, whereas participants
in a good mood show no bias at all (Jonas, Graupmann, & Frey, 2006). In a study by J.
Fischer, Fischer, Englich, Aydin, and Frey (2011) the preference for supporting information
was also greater when social power was experimentally increased by embodiment
manipulations. This effect was mediated by decision-certainty, thus, social power led to
increased confidence in the validity of one’s decision, which in turn increased CIP. High
commitment to a choice or attitude, manipulated by having participants write supporting
essays (vs. conflicting essays) about their position, can increase the bias as well (Schwarz,
Frey, & Kumpf, 1980).
On the other hand, irreversible decisions (Frey, 1981), limited information choice or
limited availability of information can lead to the preference of conflicting information (P.
Fischer et al., 2005). A meta-analysis by Hart and colleagues indicates a moderate effect size
for confirmation biases (Cohen’s d = 0.36, based on 91 studies; Hart, Albarracın, et al.,
Confirmatory Information Processing in Criminal Proceedings
CIP is of particular interest in criminal and legal proceedings (for an overview see
Kassin, Dror, & Kukucka, 2013). For instance, within the police force CIP has been studied
in the form of investigator bias, which is a bias towards judgments of deceit (Meissner &
Kassin, 2002), or tunnel view (Findley & Scott, 2006). Once a suspect has been identified,
disconfirming evidence or alternative scenarios are often overlooked or not taken into
consideration (e.g., Rassin, 2010), a mechanism that resembles CIP. Moreover, in two studies
by O’Brien (2009) lay participants reviewed a police file and were asked to articulate a
hypothesis about the identity of the perpetrator when they were either halfway through the
file or after they read the whole file. Afterwards, they took a recognition test and indicated
which leads the police should pursue. Results showed that the early formulation of a
hypothesis led to better recognition of hypothesis-consistent facts. Furthermore, participants
also requested more lines of investigation focused on that first hypothesis. This pattern did
not emerge if participants were asked about their hypothesis after they had read the whole
Furthermore, in a study conducted by Ask and Granhag (2005) police officers
appeared to be blinder to alternative scenarios than students, essentially showing a
confirmation bias. Even independent and naïve crime analysts can be influenced by the prior
interpretation of the detectives, subsequently confirming the detectives’ view (Kerstholt &
Eikelboom, 2007). Likewise, fingerprint experts were also found to be susceptible to context
effects (Dror, Charlton, & Péron, 2006): In a highly realistic setting the participating experts
(not knowing that they were taking part in a study) were informed that the fingerprints, which
they would have to compare, were ostensibly labeled as a non-match by the FBI. They did
not know that they were actually presented with a fingerprint set which they themselves had
previously classified as a match. However, within the new context the belief of a non-match
the experts were more likely to classify the fingerprints also as not matching (Dror et al.,
2006). Similarly, mock jurors also interpreted and even distorted new information according
to their previously held beliefs or towards the currently favored verdict (Carlson & Russo,
2001). Moreover, jurors falsely based their similarity judgment between facial composites of
the offender and the suspect on pre-existing beliefs of guilt (Charman, Gregory, & Carlucci,
Are judges just as biased towards supportive arguments as police detectives and jurors
are? Yet, whereas police detectives’ aim is to find the person who is responsible for a crime
and support it by evidence, the judges’ objective is to determine if the evidence is sufficient
to convict the accused (e.g., Horowitz, 1997). In order to do so, the judge needs to evaluate
all relevant facts.
Studies by Dhami (2003) as well as Dhami and Ayton (2001) give reason to question
the thoroughness of legal decisions. They investigated if bail decisions could be predicted by
a simple heuristic such as the matching heuristic, which searches through a subset of cues
until the critical cue is found that warrants a decision, or by a more elaborate model such as
Franklin’s rule, which weights each cue according to its influence on the decision (Dhami &
Ayton, 2001). Their studies included bail decisions on hypothetical cases made by magistrate
judges (Dhami & Ayton, 2001) as well as observations of real bail decisions made by judges
in the UK (Dhami, 2003). In both studies the matching heuristic was a better predictor than
the more elaborate model, implying that judicial decisions are not necessarily based on all
available cues. Additionally, findings by Danziger, Levav, and Avnaim-Pesso (2011) support
the notion that extraneous variables influence legal judgments by showing that judicial
decisions even depended on the judge’s sensation of hunger. Parole decisions on real cases
were more favorable for the defendant after the judge had taken a break and ate something.
On the other hand, a study by Redding and Reppucci showed that judges were less
biased by their attitudes about social science in their evaluation and admissibility ruling of
social science evidence in a death penalty case than law students (Redding & Reppucci,
1999). Building on these findings, a recent study by Kahan et al. (2016) investigated the
influence of political predispositions on judicial decision making. Their results suggest that
legal experts do overcome identity-protective cognition, which is a similar concept to
confirmatory information processing. Based on participants’ responses to a short version of
the Cultural Cognition Worldview Scales Kahan and colleagues classified their participants
(253 judges, 217 lawyers, 284 law students and 800 general-public members) as hierarchical
individualists, hierarchical communitarians, egalitarian individualists, or egalitarian
communitarians. Participants were then presented with legal cases (statutory interpretation
problems). The researchers created two versions of each case, for instance, in one case the
defendants were either immigrant-aid workers or construction-workers. This manipulation
was supposed to lead to an interpretation of the case that is consistent with the groups’
commitments (i.e., hierarchical individualism etc.), subsequently leading to identity-
protective reasoning. Yet, these case manipulations should in fact have absolutely no bearing
on the decision. Results indicate that only members of the general public showed responses
that reflected the participants’ identities. To some extent even law students showed identity-
protective reasoning, however, judges and lawyers remained uninfluenced, thus unbiased
(Kahan et al., 2016). This study, which has an impressively diverse sample, indicates that
legal training and experience can in fact reduce decision-making biases, and, contradictory to
prior findings about politically-motivated legal decisions by judges (Kulik et al., 2003),
paints a different picture of legal objectivity.
Our current study adds at least two important aspects to this specific line of research:
1) we investigate the effect of preliminary decisions on subsequent information evaluation
instead of analyzing the effect of preexisting political and cultural attitudes on decision
outcomes, and 2) we focus on the underlying mechanism that differentiates levels of legal
expertise, which can either protect judges against or make them even more susceptible to
Experts are defined as being highly knowledgeable within a certain domain and as
possessing superior skills, which they achieve through prolonged periods of experience and
focused practice (for a review on expertise, see Salas, Rosen, & DiazGranados, 2010).
Experts are skilled in distributing their cognitive resources, they process information more
efficiently (Reyna & Lloyd, 2006; Ste-Marie, 1999) and they have superior memory for
meaningful domain-specific information (Brailey et al., 2001; for a review see Vicente &
Wang, 1998). This can be explained by experts’ complex representation of domain
knowledge (Chi, Glaser, & Rees, 1981) and their rich network of connections between
features (Feltovich, Johnson, Moller, & Swason, 1984).
Experts’ superior knowledge can thus potentially act as a protective factor against
decision-making biases. In a study by Krems and Zierer (1994) medical experts with domain-
specific diagnostic skills were more flexible to adjust their first assumption and thus
demonstrated less CIP than medical experts with diagnostic skills in other domains. The
authors argue that the domain-specific knowledge, not the experience in diagnostics, enabled
them to integrate contradicting details into a coherent explanatory model, which made them
more flexible to jump to a different diagnosis. Hence, more knowledge can lead to less CIP in
medical experts.
Another protective factor might be responsibility. In a study by van Dongen and van
Maanen (2013) responsibility predicted reliance on oneself when making a decision: The
more participants felt responsible for the accuracy of a final decision the more they were
willing to accept conflicting yet more reliable advice from a decision aid (van Dongen & van
Maanen, 2013). Similarly, accountability has been shown to increase both time and effort put
into the analysis of information and alternatives (Doney & Armstrong, 1996; Simonson &
Nye, 1992). Therefore, responsibility, which is expected to increase with expertise, may
decrease CIP.
Altogether, experts have both the means and intentions to reach the best possible
decisions. However, research also indicates that judgments of experts are not necessarily
superior or less influenced compared to judgments of laypeople. They can be blinded by their
experience which may result in overconfidence (e.g., Dawes, Faust, & Meehl, 1989; Oskamp,
1965; Vrij, 2004). According to Plous (1993) experts are subject to the same biases as
novices, and they either demonstrate biases to the same degree or to a somewhat reduced
level. For instance, judicial experts also rely on numeric anchors (Englich, Mussweiler, &
Strack, 2005; Englich et al., 2006) or are affected by framing, hindsight bias, base rate
fallacy, or egocentric bias (Eisenberg, 1994; Guthrie et al., 2001). Even specialization does
not mitigate the effects of anchors or frames (Englich & Mussweiler, 2001; Guthrie,
Rachlinski, & Wistrich, 2007). Dan Simon’s work on coherence shifting and its theoretical
application on appellate judges suggest that expert judges as well are likely to unconsciously
use coherence shifts when, for instance, evaluating the harmless error in inadmissible
evidence cases. Coherence shift reasoning describes a bidirectional confirmation bias, in
which the evaluation of one piece of evidence not only influences the evaluation of the
following piece of evidence, but it can also lead to a re-evaluation of already processed
information. Additionally, the most supported conclusion can also influence the evaluation
process, continuously adapting the mental representation of the decision problem. This leads
to excessive confidence since all evaluations end up to be consistent with one another. Based
on this, Simon points out that the design of the trial is based on erroneous assumptions about
how human cognition works (Simon, 2004).
In summary, experts have superior cognitive abilities and high motivation which can
protect them against decision bias, nevertheless, their intense training and experience may
actually facilitate decision biases due to overconfidence or lack of accurate and immediate
performance feedback. The present research aims to clarify the role of legal expertise in
decision biases like CIP.
The Present Research
The main purpose of our studies was to clarify if and to what extent legal experts are
in fact susceptible to CIP while taking into account that expertise can be further differentiated
according to their specific field (domain-specific vs. general). The second aim of the present
research was to identify the underlying process which causes potential expertise differences.
Decision-certainty, knowledge and responsibility were investigated as potential underlying
processes, which led to three different and somewhat conflicting predictions about how legal
experts might be affected by CIP.
Legal Experts
The German legal system is an inquisitorial system and not an adversarial system like
in the US. The most prominent difference between these two legal systems is the role of the
judge. In the adversarial system the judge assumes a rather passive role by guarding the
integrity of the process. Evidence is presented to a lay audience (the jury), who are
responsible for the verdict. In the inquisitorial system the judges are more active. They
question the witnesses, call for expert testimonies and reach the verdict. The central
obligation of the court is to ascertain the truth. Lawyers only ask extra questions and are
allowed to nominate additional witnesses (see Strier, 1992 for a comparison of the two
systems). For the present research it is important to note that German judges are specifically
trained for the bench and are not selected from the ranks of lawyers. Benefiting from this
judicial system, we were able to differentiate between domain-specific and general experts in
our research, since German judges usually have an emphasis on one specific jurisdiction.
This implies that they become specialized in one field (e.g., criminal-law) and mainly gain
experience in that one area, whereas they have general legal expertise in other jurisdictions
(e.g., civil-law), which they have acquired during their studies and basic training. In our
studies, if the specialization of the judge matched with the presented case material (e.g.,
criminal-law judge and criminal-law case) participants of our studies were classified as
domain-specific experts. If it did not match (e.g., criminal-law expert, but labor-law case)
participants were classified as general experts.
Whereas participants in the pilot study were either civil-law or criminal-law experts,
we recruited only criminal-law experts in the main study. Furthermore, since defense lawyers
and prosecutors also solely work on criminal cases, we categorized them as criminal-law
experts as well. If not otherwise reported, it did not have any effect whether the participants
were judges, prosecutors or defense attorneys.
Experts’ high confidence (e.g., Englich et al., 2006; Vrij, 2004) could lead to the
subsequent ignorance of conflicting information, because experts may not see the need for
processing (seemingly) irrelevant information. A positive correlation between decision-
certainty and CIP has already been demonstrated in laypeople samples (J. Fischer et al., 2011;
P. Fischer, Jonas, et al., 2008). The first possible prediction is therefore that due to experts’
generally higher decision-certainty, legal experts may show higher levels of CIP compared to
laypeople, and domain-specific experts show higher levels of CIP than general-experts.
Prior Knowledge
Experts also possess profound knowledge which could influence their information
processing. Applying Krems and Zierer’s findings (1994) about the importance of domain-
specific knowledge for balanced information integration to the legal context, specialized
judges should show low levels or even no CIP at all. Correspondingly, if the case is outside
the judge’s specialty, the lack of domain-specific knowledge should lead to levels of CIP
similar to that of laypeople or even higher. For example, a criminal-law judge should show
lower levels of CIP when he or she adjudicates a criminal-law case but higher levels when
working on a labor-law case. Additionally, this implies that laypeople should show the most
extensive CIP due to their lack of professional legal knowledge. If knowledge is in fact the
underlying causal mechanism for differences in CIP, knowledge and CIP should be
negatively correlated. The second prediction is therefore that domain-specific experts show
lower CIP than both general-experts and laypeople due to specialized knowledge and the
resulting capacity for more conflicting information (Salas et al., 2010).
Besides elevated decision-certainty and profound knowledge, an increased feeling of
responsibility could also have an impact on the information processing of legal experts.
Domain-specific legal experts should feel highly responsible for the consequences that their
decisions can have on the defendant and on society, which in turn could lead to reduced CIP
(van Dongen & van Maanen, 2013). General legal experts on the other hand may feel
(slightly) less responsible, because it is not their field of expertise, whereas laypeople
probably do not feel responsible at all. Low feelings of responsibility in turn should be
associated with more CIP, as it renders making a well-informed decision less important. A
third possible prediction is therefore that domain-specific experts show lower CIP compared
to both general-experts and laypeople due to their high feeling of responsibility.
In summary, three possible predictions can be deduced for domain-specific experts in
the legal context: higher CIP due to overconfidence, lower CIP due to profound knowledge or
lower CIP due to a higher feeling of responsibility. Thus, a decrease in CIP for domain-
specific experts would be consistent with both prior knowledge and responsibility predictions
as these only differ with regard to the assumed mechanism, not with regard to the pattern of
Case Materials
In our studies we used two cases as stimulus material: A labor-law case and a
criminal-law case. In order to demonstrate that the CIP pattern is independent of the verdict
these cases needed to be ambiguous enough to trigger evenly distributed verdicts. Therefore,
both cases involve moral issues and ambiguous legal terminology like “for a compelling
reason” or “it cannot reasonably be expected”. These open terms give judges the necessary
discretionary power (Kepplinger & Zerback, 2009). According to Fiss (1982) “adjudication is
interpretation: adjudication is the process by which a judge comes to understand and express
the meaning of an authoritative legal text and the values embodied in that text” (p. 739). It
stands to reason that everything else being equal not all judges come to the same
interpretation of given information and subsequently arrive at different verdicts, which is
exactly what we tried to achieve with these cases. In order to explore if the judge’s personal
moral evaluation of the deed is reflected in the preliminary decision we measured subjective
moral reprehensibility of the defendants’ behavior and moral evaluation of the decision
options. In daily life, the interpretation of ambiguous situations is influenced by moral
attitudes (Walker, 2000), we therefore expect a similar effect in legal decision-making.
Pilot Study
The main purpose of our pilot study was to investigate if and to what extent legal
experts show CIP at all. Additionally, we wanted to get an initial insight into the effect of the
different expertise levels on CIP and explore decision-certainty and prior knowledge as
potential underlying mechanisms. Last but not least, this pilot study was intended to test the
applicability of the research materials.
Participants and design. Forty-four German judges and prosecutors (16 women, 2
participants did not indicate their gender; MAge = 48.00, SD = 8.94 ranging from 30 to 63
years) and forty-six students from a large German university (31 women; MAge = 23.8, SD =
2.72, ranging from 19 to 29 years) participated in this Study. Judges had an average of 18.68
years of experience on the bench (SD = 9.09, ranging from 1 year and 6 months to 34 years
and 9 months). Since the presented case was a criminal-law case, 11 participants were
categorized as domain-specific experts (8 criminal-law judges, 2 prosecutors, and one former
criminal-law judge). Thirty-one judges had expertise in various other domains (e.g., civil-law,
labor-law, or family-law) and were categorized as general experts. All experts were
approached during a workshop on judicial decision making at the German Judicial Academy
(Deutsche Richterakademie) in Trier, Germany. Laypeople (i.e., students) were recruited
from the university campus and were informed that the study was on judicial decision
making. This pilot study was based on a one-factorial design with three quasi-experimental
conditions of expertise: laypeople vs. general experts vs. domain-specific experts1.
Materials and procedure. A modified version of a classic and well-established CIP
paradigm was employed (e.g., P. Fischer et al., 2005) and adapted to the judicial context (see
main study for a detailed description of the materials). In this adaptation participants had to
imagine that they were a judge presiding over a criminal-law case. The case description was
taken from Englich et al. (2006) who compiled the material in close collaboration with legal
experts. The vignette introduced the case of Mrs. M., a 43 year-old woman who is on trial for
minor theft (total value of 50 Euros). Mrs. M.’s has eleven prior shoplifting charges and has
also pleaded guilty to this one. Last time she was on trial she was sentenced to two years on
probation. Additionally, she received therapy to treat her kleptomania. This psychological
diagnosis was added to the case to increase uncertainty of how this information should be
evaluated and considered in the decision, since we aimed for an ambiguous case. Participants
were asked to decide between a prison sentence and a sentence set out on probation. A
sentence can be set out on probation if it serves “a sufficient warning” (German Criminal
Code, Article 56, paragraph 1). Thus, German law provides sentence guidelines, yet judges
can freely decide which sentence is appropriate according to their interpretation of the law.
After their preliminary decision participants indicated their subjective decision-
certainty on a Likert-type rating item (0 = not at all certain, 10 = absolutely certain)2, and
their subjective competence to judge this case (0 = not at all competent, 10 = very
competent). Additionally, moral reprehensibility of Mrs. M.'s behavior was assessed: “How
morally reprehensible was Mrs. M.’s behavior?” (0 = not at all reprehensible, 10 = very
reprehensible). Additionally, two items assessed moral evaluations of the proposed
sentences: “How morally necessary is it to choose the probationary sentence?” and “How
morally necessary is it to choose the prison sentence?” (0 = not at all necessary, 10 =
absolutely necessary).
In accordance with previous CIP research, participants then received ten arguments
about the case, introduced as arguments that judicial experts had provided and evaluated in
previous studies (see Englich et al., 2006). The arguments were either statements arguing
why a sentence set out on probation with outpatient therapy would be the best option, or
statements arguing in favor of the prison sentence (10 arguments in total, 5 per side).
Depending on the participants’ preliminary decision, the arguments were classified as
either supporting or conflicting. The arguments were presented in a fixed order, alternating
between supporting and conflicting. Biased assimilation (importance, strength evaluations of
the arguments) and selective exposure (selection) items followed (see main study for a more
detailed description of these items). Additionally, as a proxy for prior knowledge, participants
were asked whether they were already familiar with the just presented argument from other
similar cases (yes / no).
Results and Discussion
Manipulation check. An ANOVA revealed a marginal main effect of Expertise on
participants’ subjective competence rating, F(2, 85) = 2.60, p = .080, ηp² = .06. Post-hoc
comparisons, using Games-Howell post-hoc procedures due to unequal variances, revealed
that domain-specific experts felt more competent (M = 6.91, SD = 1.64) than both general
experts (M = 4.81, SD = 2.82), p = .015, Cohen’s d = .84, and laypeople (M = 4.89, SD =
2.98), p = .013, Cohen’s d = .74. There was no difference between general experts and
laypeople, p = .991, Cohen’s d = -.03. Our decision to divide experts into two separate groups
therefore seems warranted.
Preliminary decision. In total 81.8% of the participants chose probation as the
suitable sentence and 18.2% chose the prison sentence. Laypeople, general experts and
domain-specific experts did not show significantly different decision distributions, χ² (2, N =
88) = 2.8, p = .25. Furthermore, a 3 (Expertise) x 2 (Decision) multivariate analysis of
variance (MANOVA) revealed that decision did not interact with expertise on biased
assimilation, F(1,78) = .63, p = .54, or selective exposure, F(1,78) = .54, p = .58. A
systematic effect of the preliminary decision on the main findings is thus unlikely (see P.
Fischer, Jonas, et al., 2008).
As expected, a MANOVA indicated that all morality items corresponded with the
preliminary decision. Participants evaluated the behavior of Mrs. M. as more morally
reprehensible when they chose the prison sentence (M = 6.53, SD = 2.29) compared to the
probationary sentence (M = 4.85, SD = 2.38), F(1,85) = 6.31, p = .014, ηp² = .07.
Furthermore, the necessity of the probationary sentence was evaluated to be higher when
participants chose probation (M = 7.21, SD = 1.94) compared to when they chose the prison
sentence (M = 2.20, SD = 1.32), F(1,85) = 91.18, p < .001, ηp² = .52. Correspondingly, the
prison sentence was evaluated to be more morally necessary if participants chose this
sentence (M = 7.13, SD = 2.77) compared to when they chose the probationary sentence (M =
2.07, SD = 1.97), F(1,85) = 70.76, p < .001, ηp² = .45. Yet, due to the unequal distribution of
the preliminary decision, these results should be interpreted with caution. A separate
MANOVA revealed that laypeople, general experts and domain-specific experts did not
differ in their moral assessments (all ps > .10).
Confirmatory information processing. We computed difference values for
information evaluation (i.e., biased assimilation) and information search (i.e., selective
exposure) by subtracting the corresponding values for conflicting arguments from the values
for supporting arguments (see main study, P. Fischer, Greitemeyer, et al., 2008). One-sample
t-tests revealed that across all conditions biased assimilation (M = 2.61, SD = 2.23) and
selective exposure (M = .62, SD = 1.07) scores were all significantly different from zero,
t(87) = 10.94, p < .001, Cohen’s d = 1.17; t(83) = 5.28, p < .001, Cohen’s d = .58,
respectively; implying that supporting arguments received a higher evaluation and were
selected more often than conflicting arguments.
A MANOVA with biased assimilation and selective exposure as dependent variables
revealed a main effect for expertise on biased assimilation, F(2, 81) = 3.19, p = .046, ηp² =
.07. Post-hoc tests were conducted using Bonferroni adjusted alpha levels to clarify this
effect. Results indicated that biased assimilation was significantly lower in domain-specific
experts (M = 1.44, SD = 1.76) than in general experts (M = 3.16, SD = 2.39), t(40) = -2.188, p
= .035,Cohen’s d = -.78 . Biased assimilation of laypeople (M = 2.51, SD = 2.14) did not
significantly differ from domain-specific experts, t (55) = -1.54, p = .13, Cohen’s d = .53, or
from general experts, t(75) = 1.26, p = .213, Cohen’s d = -.29 (see Figure 1), which replicates
prior research on cognitive biases, where no difference between experts and laypeople was
found (e.g., anchoring effect, Englich et al., 2006; Rachlinski et al., 2007). Yet, the difference
between domain-specific experts and general experts does indicate that level of expertise is
an important moderating factor.
Expertise did not have a significant effect on selective exposure, F(2, 81) = .48, p =
.623. This was probably caused by the generally low number of arguments that participants
selected in total (M = 1.41, SD = 1.62 out of 10), which did not provide enough variance for a
main effect. This floor effect suggests that our participants were generally not interested in
additional reading.
- Insert Figure 1 about here -
Potential underlying processes. There was no effect of expertise on decision-
certainty, F(2, 85) = 1.32, p = .27, ηp² = .03 . Domain-specific experts felt just as certain (M =
7.73 SD = 1.95) as both general experts (M = 6.90, SD = 2.70) and laypeople (M = 7.73, SD =
2.07). Furthermore, decision-certainty neither correlated with biased assimilation (r(86) =
.12, p = .259) nor with selective exposure (r(82) = -.03, p = .809). The positive correlation
between decision-certainty and CIP found by Fischer and his colleagues (J. Fischer et al.,
2011; P. Fischer, Jonas, et al., 2008) was therefore not replicated in our sample.
Consequently, experts’ overconfidence could not account for the differences found in biased
assimilation as a function of expertise group.
A main effect of expertise on prior knowledge (i.e., familiarity with the arguments)
became significant, F (2, 85) = 10.27, p < .001, ηp² = .20. Post-hoc tests with Bonferroni
correction revealed that domain-specific experts (M = 8.36, SD = 2.87) possessed slightly
more prior knowledge compared to general experts (M = 6.03, SD = 4.03), t(24.86) = 2.06, p
= .05, Cohen’s d = .63, and significantly more compared to laypeople (M = 3.28 SD = 3.76),
t(19.17) = 4.94, p < .001, Cohen’s d = 1.43. Furthermore, general experts also indicated more
prior knowledge than laypeople, t(75) = 3.05, p = .003, Cohen’s d = .72. However, prior
knowledge neither correlated with biased assimilation (r(86) = - .01, p = .947) nor with
selective exposure (r(82) = - .17, p = .127) and also does not reflect the found pattern of
biased assimilation. Therefore, prior knowledge does not seem to be a good predictor of CIP,
because it could not explain the expertise differences in biased assimilation.
Main Study
The pilot study provided first insights into the CIP pattern in legal experts. First, CIP
in legal experts seems to depend on the experts’ specialization. Whereas general experts
showed the same degree of biased assimilation as laypeople, domain-specific experts
evaluated arguments in a more balanced way than general experts. Second, decision-certainty
does not seem to be the underlying mechanism for the expertise differences, since all
participants felt equally certain about their choice. Third, as expected, prior knowledge with
the presented arguments did differ between the expertise groups, however, it neither
correlated with biased assimilation nor could it explain the found pattern of biased
The aim of this Study was to replicate the expertise effect of the pilot study with a
larger expert sample. Furthermore, since neither decision-certainty nor prior knowledge could
be supported as underlying mechanisms in the pilot study, a motivational reason for the
difference between domain-specific experts and general experts was investigated: A high
feeling of responsibility for their judicial decisions. Therefore, we hypothesized that a
reminder of this responsibility reduces CIP in general experts, whereas it has no effect on the
domain-specific experts since they already feel highly responsible.
This Study did not include a laypeople condition, mostly because the interesting
difference lies between domain-specific and general experts. Moreover, in the pilot study
laypeople did not significantly differ from general experts in both biased assimilation and
competence ratings. Furthermore, and in order to reduce possible confounding effects that
different specialties may cause, we only recruited criminal-law experts for this study and
manipulated expertise through different case materials. Other limitations of the pilot study
were addressed and improved as well.
Participants and design. One hundred forty-eight legal experts with expertise in
criminal-law were recruited to participate in this study3. Participants were approached on six
different occasions: On five occasions, data was again collected at the German Judicial
Academy, once during a workshop, and four times as an extracurricular activity at criminal-
law conferences, which participants could freely attend. The sixth data collection took place
at a workshop for defense lawyers in Dresden, Germany. One participant had to be excluded
because she did not make a preliminary decision. Furthermore, sixteen experts, who indicated
that they also possessed expertise in labor-law or who left this item unanswered, had to be
excluded due to the nature of our expertise manipulation (see materials below). Additionally,
one person was categorized as an outlier (above 4 SD on the biased assimilation index) and
was also excluded from further analysis. Therefore, one hundred thirty criminal-law experts
were analyzed (30 women; MAge = 44.47, SD = 9.33, ranging from 28 to 77 years).
Participants had an average of 15.50 years of work experience (SD = 8.98, ranging from 1 to
40 years). The sample consisted of fifty-four judges, twenty-four prosecutors and thirty-two
defense lawyers. Twenty participants did not specify their current judicial profession. Judges,
prosecutors and defense attorneys did not differ with regard to age, professional experience
and gender (all Fs < 1; Chi-square < 1, respectively). Furthermore, all main analyses were
checked for effects of current occupation. If not otherwise indicated, occupation had no
Criminal-law experts were randomly assigned to one of four conditions of a 2
(Expertise: domain-specific vs. general) x 2 (Responsibility: control vs. induced) factorial
design with CIP measures as dependent variables.
Materials and procedure. The case used in the pilot study was slightly modified and
another case was compiled. The overall procedure remained almost identical.
Cases. Two cases a labor-law case (layoff without notice) and a criminal-law case
(pilot study) were used to operationalize the two expertise levels within the pure criminal-
law expertise sample. Thus, participants in the domain-specific expertise condition received
the criminal-law case, resulting in a match between case and area of expertise. Participants in
the general expertise condition received a labor-law case, resulting in a mismatch between
case and area of expertise. In total both case files were about 1.5 pages long, which is a
common vignette length when experts are in the sample (e.g., Guthrie et al., 2001; Wistrich,
Guthrie, & Rachlinski, 2005).
Our fictional labor-law case describes a 53 year-old woman, Mrs. B., who used to
work behind the meat counter in a large supermarket. There had been no complaints about
her prior to the following incident: On two consecutive days the branch manager, Mr. K., saw
Mrs. B. taking three slices of ham from the meat counter (allegedly to make a sandwich)
without paying for the ham beforehand. Mr. K. confronted Mrs. B., who did not deny her
actions and apologized instantly. Mrs. B. was unaware of the fact that her behavior could be
interpreted as theft. However, Mr. K. thought that her behavior resulted in a massive breach
of trust and he fired her instantly without notice. Mrs. B. fought against her layoff and as a
result the case was being tried at court. Participants were informed that they were supposed to
adjudicate on this matter. They could decide between “layoff without notice was legitimate”
and “layoff without notice was not legitimate”. This case was based on a number of similar
cases about layoffs for rather trivial reasons (Nöhmaier, 2010). Since case-law is not
commonly practiced in Germany, labor-courts and appeal courts frequently came to different
decisions even if the cases were quite similar (Pressemitteilung Landesarbeitsgericht Baden-
Württemberg, 2010; Pressemittelung Bundesarbeitsgericht, 2010). Therefore, it was expected
that experts have different interpretations of the case leading to an equal decision distribution.
To assist with their decision and to establish a common legal knowledgebase across
all conditions, participants received an extract out of the German Legal Code (BGB, Article
626), which proclaims that an employment contract can be terminated without giving notice,
if there is a compelling reason and if it cannot reasonably be expected from the manager to
still employ the defendant. Furthermore, all circumstances of the individual case need to be
taken into account and the interests of both parties to the contract should be weighted.
Participants’ decision was thus not about whether a layoff is justifiable in general, but only
whether the layoff without notice is - in their eyes - legally acceptable.
In the domain-specific expertise condition, the case was almost identical to the
criminal-law case used in the pilot study. The case was about a woman, who was on trial for
minor theft. Participants were asked to decide between a prison sentence and a suspended
sentence. Participants were provided with Article 56, paragraph 1 from the German Criminal
Code, which states that a sentence can be set out on probation if there are reasons to believe
that the sentence will serve as sufficient warning to the convicted person and that he will
commit no further offences without having to serve the sentence. Also, the character of the
convicted person, his previous history, the circumstances of his offence, his conduct after the
offence, his living circumstances, and the effects to be expected from the suspension all need
to be considered. This paragraph was added to tip the distribution of the verdict more towards
a prison sentence, as the distribution of preliminary decisions was not optimal in the pilot
study (81.8% vs. 18.2%).
Arguments. Participants were given ten arguments, which were provided in a fixed
sequence. An argument for the decision with the lesser consequences for the defendant (i.e.,
probationary sentence or illegitimacy of the layoff without notice) was always presented first.
Furthermore, all arguments were arranged in a specific order starting with weak arguments in
the beginning of the questionnaire followed by stronger arguments. This was done for two
reasons: First, in order to create similar processing conditions across both cases. Second, to
ensure that the first arguments would not be weighted too much as this could bear the risk of
overshadowing all following arguments. The arguments’ strength was determined by a pre-
test in the labor-law case and by the strength-ratings of the pilot study in the criminal-law
One example for an argument in the labor-law case against the legitimacy of the
layoff stated that Mrs. B.’s family depended on her income. The financial restrictions that
would develop after an extraordinary layoff could mean the social decline for the family. An
example of an argument that was in favor of the layoff stated that Mrs. B.’s misbehavior had
far reaching consequences, because she was responsible for the training of new employees
and did not portray a good role model anymore. In the criminal-law case, one argument in
favor of probation, for example, stated that Mrs. M. had a stable and intact family with two
children. A prison sentence would jeopardize her and her children’s social life. An argument
in favor of a prison sentence stated that Mrs. M. is resistant to therapy and that she had
relapsed only three months after her previous therapy. Therefore another therapy does not
seem promising.
Responsibility manipulation. Rozelle and Baxter (1981) successfully induced
responsibility by informing participants that their decision is important and has actual
consequences. We opted for a similar responsibility reminder for our sample of legal experts,
which pointed out that the decision can have tremendous consequences for the defendant.
Therefore, immediately following their preliminary decision, half of the participants
additionally received two short sentences to increase responsibility: “As a legal expert, you
are involved in important decisions with serious consequences for the person concerned.
Please make yourself aware of this societal responsibility before you continue with the next
task.” Right before the evaluation of the arguments participants received a similar reminder
again. Participants who were not in the responsibility conditions did not get these additional
Potential underlying process. Due to demand characteristics we did not add a direct
measurement of responsibility as a manipulation check. Therefore, three items were designed
to measure if responsibility changed participants’ motivation to carefully process all
information: attention (0 = not at all alert, 10 = very alert), importance of a fair decision (0 =
not at all important, 10 = very important), and motivation to consider all arguments (0 = not
at all motivated, 10 = highly motivated) were measured at the end of the questionnaire. Van
Dongen and van Maanen have argued that effort and motivation should be impacted by
responsibility (van Dongen & van Maanen, 2013). Additionally, we measured decision-
certainty again, since results of the pilot study contradicted previous findings (J. Fischer et
al., 2011; P. Fischer, Jonas, et al., 2008).
Morality items. Participants were presented with exploratory morality items. First,
participants were asked to evaluate the defendants’ behavior: “How morally reprehensible
was Mrs. M.’s / Mrs. B.’s behavior?” (0 = not at all reprehensible, 10 = very reprehensible).
Furthermore, moral evaluations of the options presented in the preliminary decision were
assessed. In the general expertise condition the preliminary decision is whether the layoff
without notice was legitimate or not, therefore this item was phrased: “How morally
reprehensible is the layoff without notice” (0 = not at all reprehensible, 10 = very
reprehensible). Since the decision in the domain-expertise condition is to decide between two
different sentence options we assessed the moral evaluation of the available outcome with
two items: “How morally necessary is it to choose the probationary sentence?” and “How
morally necessary is it to choose the prison sentence?” (0 = not at all necessary, 10 =
absolutely necessary).
Procedure. All participants received the description of the case first and subsequently
indicated their preliminary decision. Afterwards, half the participants received the
responsibility reminder. Next, decision-certainty (0 = not at all certain, 10 = very certain),
subjective competence (0 = not at all competent, 10 = very competent), subjective expertise
(0 = no expertise for these types of cases, 10 = high expertise for these types of cases), and
the morality items followed. Arguments were presented next.
Main dependent variables were biased assimilation and selective exposure, assessed
by the evaluation and selection of supporting and conflicting arguments. After reading one
argument, which, for instance, favors the prison sentence, participants rated how important
this argument was for their preliminary decision (0 = not at all important; 10 = very
important) and how strongly this argument favors the prison sentence (0 = not at all strongly,
10 = very strongly). Importance and strength evaluations of supporting arguments were
collapsed into one measure (α = .90, cf. P. Fischer, Jonas, et al., 2008). The same was done
for evaluations of conflicting arguments (α = .64; without one outlier who scored more than 5
SD: α = .74). Selective exposure was measured by having participants choose whether they
would like to read the full version of the argument (yes / no).
When participants had processed all arguments they made a final decision on the case,
again followed by decision-certainty. Afterwards, the three indirect manipulation check items
(attention, motivation, importance) were assessed. The questionnaire ended with a funneled
debriefing, followed by demographic measures. Finally, participants received the information
that, contrary to what was previously stated, they would not receive any additional reading
material. Consistent with nearly all studies on CIP employing this research-paradigm, this
was done in order to keep the duration of the experimental session to a minimum. The study
only focused on the hypothetical selection of information and was not intended to investigate
how participants process the additional information. All participants gave their consent by
handing in their questionnaires, which was entirely their free choice. All participants were
thanked and received a chocolate bar as a token of appreciation. Afterwards, legal experts
were informed about the aim of the Study and they had time to discuss the Study with the
Manipulation check. Subjective competence and subjective expertise were collapsed
into one subjective expertise index (α = .84). A one-way ANOVA revealed that criminal-law
experts working on the criminal-law case felt much more competent (M = 8.21, SD = 1.64)
than criminal-law experts working on the labor-law case (M = 5.20, SD = 1.64), F(1, 128) =
76.82, p < .001, ηp² = .38, supporting our expertise manipulation of domain-specific
(receiving the criminal-law case) and general expertise (receiving the labor-law case) via the
two different cases.
The items motivation, attention and personal importance, which were assessed after
the final decision, were collapsed into a responsibility index (α = .81) and analyzed. An
ANOVA with Expertise and Responsibility as between subject factors revealed a significant
main effect for Expertise, F(1, 126) = 6.10, p = .015, ηp² = .46, indicating that domain-
specific experts felt more responsible (M = 8.26, SD = 1.62) than general experts (M = 7.54,
SD = 1.63). Neither the main effect of the responsibility reminder on the responsibility index
nor the interaction between expertise and responsibility reached significance, F(1,126) =
2.34, p = .129, ηp² = .02, F(1,126) = 2.18, p = .142, ηp² = .02, respectively.
However, due to the hypothesis that the responsibility reminder should only affect
general experts since domain-specific experts already feel highly responsible, planned
contrasts were still analyzed (cf. Tybout et al., 2001). Within the general expertise group
there was a significant effect of the responsibility reminder on the responsibility index,
t(56.32) = 2.17, p = .034, Cohen’s d = .54,, indicating that this index was significantly higher
for the general experts in the responsibility condition (M = 7.98, SD = 1.23) than in the
control condition (M = 7.14, SD = 1.84). Within the domain-specific expertise group the
responsibility index did not differ between the two conditions (control: M = 8.26, SD = 1.67;
responsibility: M = 8.27, SD = 1.61, t < 1). When responsibility was induced, general experts
and domain-specific experts no longer differed with respect to their indirect responsibility (t
< 1), whereas the contrast between the two expert groups without the induction was
significant, t(61) = -2.52, p = .014, Cohen’s d = -.65
Preliminary decision. A Pearson Chi-Square test revealed a marginal difference of
decision distribution across the cases, χ² (1, N = 130) = 2.96, p = .085. Within the criminal-
law case 65.7% (n = 44) of the participants decided for the less severe decision (probation)
and 34.3% (n = 23) for the severe decision (prison sentence), whereas in the labor-law case
the distribution of less severe (illegitimate layoff) vs. severe (legitimate layoff) preliminary
decisions was almost equal (49.2%; n = 31 vs. 50.8%, n = 32, respectively). Only two
participants changed their preliminary decision in the final decision. Consistent with other
CIP research articles the final decision was not subject of further analyses.
Confirmatory information processing. One-Sample t-tests with biased assimilation
and selective exposure as dependent variables (difference scores were again computed by
subtracting the respective conflicting items from the corresponding supporting items) showed
that across all conditions biased assimilation (M = 1.68, SD = 2.27) was significantly
different from zero, t(129) = 8.44, p < .001, Cohen’s d = .74, implying that supporting
arguments again received higher evaluations than conflicting arguments. Selective exposure
(M = .14, SD = 1.30) did not reach significance, t(123) = 1.17, p = .243, Cohen’s d = .11. As
in our pilot study the majority of participants (53.8%) chose not more than two out of ten
arguments in total for further reading resulting in very little variance. Selective exposure
scores will therefore not be subject to further analyses.
In order to make the influences of the responsibility reminder visible across both
expertise conditions, biased assimilation was standardized separately for each expertise
condition (i.e., for each case). Biased assimilation (z-scores) was then analyzed with a 2
(Expertise: domain-specific vs. general expertise) x 2 (Responsibility: induced vs. control)
ANOVA. There were no main effects of Expertise, F(1,126) = .15, p = .699, ηp² = .00, or of
Responsibility, F(1,126) = 1.15, p = .285, ηp² = .01. However, the predicted interaction
between Responsibility and Expertise reached significance, F(1, 126) = 8.60, p = .004, ηp² =
.06. Replicating the pilot study, planned contrast analyses revealed that in the control
condition domain-specific experts again showed lower biased assimilation than did general
experts, t(126) = -2.32, p = .022, Cohen’s d = -.67. Essentially, as Figure 2 illustrates, the
responsibility induction significantly reduced biased assimilation in general experts, t(126) =
2.80, p = .006, d = .68, but did not significantly affect biased assimilation of domain-specific
experts, t(126) = 1.33, p = .186, d = .35. Furthermore, general experts did not show biased
assimilation anymore when responsibility was induced, t(29) = .80, p = .429 4.
- insert Figure 2 about here
Decision-certainty was again considered as a possible underlying mechanism. General
and domain-specific experts felt equally certain about their preliminary decision (general
experts: M = 8.17, SD = 1.43; domain-specific experts: M = 8.19, SD = 1.70, F < 1 ). Again,
decision-certainty did not correlate with biased assimilation (r(128) = .03, p = .740).
Additional results. For each expertise condition a MANOVA with preliminary
decision and responsibility as between subject variables and the morality evaluation items as
dependent variables were conducted. As expected, moral evaluations of the sentences
corresponded with participants’ preliminary decision. Within the general expertise condition,
participants who decided for the milder sentence (layoff illegitimate) indicated that they
thought this sentence was less morally reprehensible (M = 3.06, SD = 2.54) compared to
those who chose the severer sentence (layoff legitimate; M = 6.06, SD = 2.61), F(1,59) =
16.80, p < .001, ηp² = .22. A similar pattern became apparent in the domain-specific expertise
condition: Participants who chose the milder sentence (probation) also evaluated this
sentence as more morally necessary (M = 6.55, SD = 2.68) than those who chose the severer
sentence (prison sentence, M = 1.91, SD = 2.22), F(1,60) = 47.81, p < .001, ηp² = .44.
Likewise, participants who chose the prison sentence evaluated this sentence as more morally
necessary (M = 6.36, SD = 2.92) than those participants who chose the probationary sentence
(M = 2.02, SD = 1.65), F(1,60) = 53.84, p < .001, ηp² = .47. In both expertise conditions, the
moral evaluation of the behavior of the defendants remained unaffected by the preliminary
decision (all Fs < 1). Furthermore, the responsibility induction did not affect any of the moral
evaluations, nor did responsibility interact with preliminary decision (all ps > .1).
Additionally, we checked for interfering effects of the preliminary decision, because
decision distribution was not as expected. However, preliminary decision had no effect on
biased assimilation F(1,128) = .423, p = .516).
Again, this Study showed that legal experts are not protected against confirmatory
information processing. Just like laypeople in a number of other studies (e.g., J. Fischer et al.,
2011; P. Fischer, Jonas, et al., 2008; Jonas et al., 2006) legal experts evaluated decision-
supporting arguments more positively than decision-conflicting arguments. Additionally, our
results also show that not all legal experts show the same degree of confirmatory information
processing. Domain-specific experts were less susceptible to CIP than general experts,
replicating our pilot study.
Whereas in the pilot study legal expertise depended on the specialization of the
participant either criminal-law expertise or other expertise, this Study operationalized legal
expertise through the type of case that was assigned to a homogenous group of criminal-law
experts. Participants either worked on a criminal-law case, warranting domain-specific
expertise, or on a labor-law case, for which they (as criminal-law experts) only possessed
general expertise. The manipulation check supported this operationalization: Participants in
the domain-specific expertise condition felt more competent and indicated higher subjective
expertise compared to participants in the general-expertise condition.
Furthermore, the Study’s aim was to investigate whether the reduced biased
assimilation in domain-specific experts was due to a higher feeling of responsibility. It was
hypothesized that a responsibility induction would reduce CIP in general experts, but not in
domain-specific experts, because domain-specific experts by default feel highly responsible
for their actions in their area of expertise. This hypothesis was confirmed for biased
assimilation. When responsibility was induced, general experts demonstrated significantly
less biased assimilation compared to the control condition in which responsibility was not
induced. In fact, after the responsibility induction general experts no longer displayed biased
assimilation at all.
One possible alternative explanation for the expertise effect on biased assimilation is
that general experts might have been in a worse mood than domain-specific experts, because
they could have perceived the task as more difficult and unfamiliar. Negative affect has been
linked to an increase in CIP (Jonas et al., 2006), which would thus correspond with the higher
biased assimilation in general experts. Future studies should therefore include a mood scale.
Due to the ambiguous case materials, we were interested in the experts’ moral
evaluations of the cases. The exploratory morality items revealed that moral evaluations of
the proposed sentences corresponded with participants’ preliminary decision. However, the
moral evaluations of the defendants’ behavior remained unaffected by the preliminary
decisions. Hence, for the legal experts in our study the moral value of the punishment seems
to be the crucial criterion for the preliminary decisions, not their moral evaluation of the
defendants’ behavior.
General Discussion
Is legal expertise a protective factor in legal decision making? If it is, what are the
underlying cognitive mechanisms? Does specialization protect against biases in legal decision
making? We tried to answer these questions with the current studies by focusing on one
specific bias confirmatory information processing (CIP).
Although research on confirmation bias and tunnel vision is quite extensive in
investigative processes, professional judges have rarely been focused on. Yet, in all
inquisitorial systems judges have a fairly active role and actually decide upon further
investigative actions or which expert testimonies are needed (Strier, 1992). Preconceived
opinions can thus extensively impact the outcome for a defendant at multiple stages of the
legal process, if it leads to the neglect of conflicting information. This is also true in
adversarial systems, since judges are still involved in countless decision-making processes
(Eisenberg & Clermont, 1996; Wistrich et al., 2005).
Therefore, research goals of the present studies were to investigate 1) whether legal
experts are susceptible to CIP, 2) whether the extent of the bias depends on the level of legal
expertise, and 3) whether the underlying processes of potential expertise differences could be
identified. The results of the main study as well as of the pilot study support the hypothesis
that legal experts, just like laypeople, show biased assimilation one central aspect of CIP.
Participants in our Study showed surprisingly little interest in reading additional information.
We attribute this to the expertise of our participants, who were not inquisitive enough to read
additional material if it was not absolutely necessary, but time-consuming. It is likely that
they as experts did not expect any new information in the additional material, since the short
summary (the argument) may be sufficient for them to guess the more detailed content.
Therefore, it was not possible to show the expected selective exposure effect in our expert-
samples, independently of the specific expertise-condition.
The data revealed that legal expertise needs to be differentiated into further
categories. In our pilot study, general experts, who possessed basic legal knowledge of the
subject at hand, showed biased assimilation to the same degree as laypeople. However,
domain-specific experts, who had both extensive knowledge and experience in that specific
legal domain, showed significantly less biased assimilation than general experts. The level of
legal expertise seems to be an important moderator. In order to identify the underlying
mechanism of this expertise effect decision-certainty, prior knowledge, and a feeling of
responsibility were investigated.
Results of our main study as well as from the pilot study imply that the degree of
biased assimilation in legal information processing is independent of decision-certainty. This
finding contradicts previous findings (J. Fischer et al., 2011; P. Fischer, Jonas, et al., 2008).
Surprisingly, lay participants (pilot study) felt just as certain about the correctness of their
decision as legal experts did. Legal decision making appears to lead to high decision-certainty
as indicated by the overall high decision-certainty (pilot study): The means of all groups
(domain-specific experts: M = 7.73 SD = 1.95; general experts: M = 6.90, SD = 2.70;
laypeople: M = 7.73, SD = 2.07) were significantly higher than the scale midpoint; all ps <
.01 (scale: 0 = not at all certain, 10 = very certain). In comparison, in studies using extra-
legal materials (J. Fischer et al., 2011; P. Fischer, Jonas, et al., 2008) laypeople’s decision-
certainty measured on an identical scale was much lower (ranging from M = 3.50, SD =
2.16 to M = 5.23, SD = 2.07, control conditions). Even when decision-certainty was
experimentally induced, decision-certainty only reached a mean of 6.29 (SD = 1.78, P.
Fischer, Jonas, et al., 2008). Thus, it is possible that decision-certainty does not correlate with
CIP in legal decision making due to lack of variance. While decision-certainty may play an
influential role in other domains, it seems to be less relevant in legal decision-making
Prior Knowledge
Whereas the decision-certainty of laypeople and legal experts did not significantly
differ, prior knowledge did indeed vary across the three expertise groups (pilot study).
However, it also did not explain the differences in biased assimilation, because the
correlation was almost non-existent (r(86) = - .01, n.s). Furthermore, laypeople and general
experts showed equally strong biased assimilation, although general experts were much more
familiar with the presented arguments.
An increased feeling of responsibility may have been the underlying process for the
lower bias in domain-specific experts, which we consistently find in our pilot study and main
study. If experimentally increased (main study), the responsibility reminder affected the
indirect measures of responsibility in general experts and successfully eliminated biased
assimilation in this group, whereas it did not have an impact on domain-specific experts. This
further supports the hypothesis that domain-specific experts already feel highly responsible
for their decisions. A motivational explanation for different CIP levels in legal experts
therefore seems likely. In their area of specialty experts feel responsible for the outcome of
their decision making. If this feeling is not naturally high, it can (at least temporarily) be
increased by a short reminder. Whether these effects are stable over a longer period of time
needs to be investigated further.
The different case-materials and operationalizations of expertise used in our studies
show that the effects of expertise on CIP were neither case specific nor limited to one type of
legal expertise group. In the pilot study participants were categorized into different expertise
groups according to their professional experience. In the main study the classification into
general- and domain-specific experts depended on the match between profession and type of
case that they had to work on. Across both studies, the same expertise effect was observed,
increasing generalizability of the findings.
Since this is the first data on CIP in experts’ legal decision making, it is certainly still
possible that the current studies did not reveal the true extent of CIP in legal experts. For
instance, the case materials and presented arguments were rather short and presented in a
written format, which can only mimic real-life legal decision making. However, research on
ecological validity of jury simulation studies concluded that it does not seem to make a
difference if the stimulus materials were presented in written form, videotaped or as a mock
trial. Although a mock trial mimics reality the best and clearly has its advantages, it does not
mean that studies employing short trial summaries are less valid (Bornstein, 1999).
There are a few reasons why CIP in legal experts could actually be lower in reality.
For example, consequences of a verdict in a hypothetical case are of course never as severe as
in a real case. Real-life cases could therefore elicit higher motivation to find an adequate
judgment resulting in lower CIP (Hart et al., 2009). Additionally, CIP also seems to be
reduced if the decision is irreversible (Frey, 1981), which many rulings are. On the other
hand, real trials can be much more demanding, which may lead to reduced self-regulatory
resources a factor that has been shown to increase CIP (P. Fischer, Greitemeyer, et al.,
2008). Furthermore, large time gaps between argument presentations facilitate sequential
processing of arguments, another factor that has been linked to higher CIP (Jonas et al.,
2001). Another factor that increases CIP is elevated social power (J. Fischer et al., 2011),
which judges exert as a core aspect of their job. It can therefore be concluded that despite
these differences between experimental studies and real trials, judicial experts are still likely
to exhibit CIP in real life.
Future research should clarify if our results are specific to criminal-law experts, thus,
if this group is especially good at balancing out supporting and conflicting arguments. Yet,
we believe that this is not the case, because it does not explain why criminal-law experts
showed high bias when they worked on the labor-law case. We argue that specialization and
the responsibility that naturally comes with it is the key. It is not something that is inherent to
criminal-law decision making. Therefore, results should be transferable to other jurisdictions
as well.
It might be seen as one limitation of our main study that participants did not revise
their final decision (only two participants adapted their preliminary decision), which implies
that biased assimilation does not seem to affect the final verdict. However, this missing effect
on the final decision is consistent with previous research. Final decisions in other studies on
CIP have remained largely unrevised as well (e.g., P. Fischer, Greitemeyer, et al., 2008; Jonas
et al., 2006). Often, final decisions were even completely omitted from analysis (J. Fischer et
al., 2011; P. Fischer, Jonas, et al., 2008; P. Fischer et al., 2005; Jonas et al., 2001;
Kastenmüller, Jonas, Fischer, Frey, & Fischer, 2013). Further, the experimental setting of our
Study may have caused participants to not change their preliminary decision. Psychological
studies with the aim to understand legal decision making might be interpreted by legal
experts as an attempt to identify flaws in legal decision making. Decision consistency may
thus reflect the effort to protect oneself against negative assessment. Yet, it would have been
interesting to see if a more balanced information evaluation would have increased the chance
of participants changing their minds.
Due to the fact that genuine legal experts were recruited most of whom had many
years of work experience the studies allow to draw valid implications about the information
processing pattern in judicial decision making. Even though we used stimulus material that
was fitted towards the inquisitorial legal system, our results can also be applied to the
adversarial system. In the adversarial system, even if it is rarely upon judges to make guilty /
not-guilty verdicts, it is very likely that judges still form preliminary decisions. This may then
unconsciously influence the judge’s decisions on evidence admissibility or which objections
are sustained or overruled. Thus, a form of confirmatory information processing may be at
work here as well. Additionally, 1/3 of the civil trials in the United States are actually bench
trials (Eisenberg & Clermont, 1996) and U.S. judges are involved in seven times as many
cases as juries (Wistrich et al., 2005). At any of these decision processes CIP may take place.
Three practical implications can be deduced. First of all, preliminary decisions seem
to set the stage for further information processing in legal decision making. Consequently,
first impressions and the first confrontation with the case seem to determine the final outcome
through the subsequent unbalanced argument evaluation. O’Brien (2009) demonstrated that
the formulation of a hypothesis about the identity of the perpetrator in the middle of the case
file was already enough to bias further investigations towards that hypothesis. This was not
the case if participants were only asked to identify the perpetrator after they had read the
complete case file. Therefore, judges should refrain from making a preliminary decision as
long as they can, because biased information processing may lead to unjust verdicts if certain
arguments may not receive the needed attention. Judges and juries should be instructed to
avoid forming preliminary decisions, for instance, by not talking about the case with
colleagues or other jurors, which might help to minimize the chance of hypothesis
formulation (O’Brien, 2009).
Second, results from the present research indicate that it is not just expertise, but
domain-specific expertise that can protect against biased decision making. Therefore, if
possible, law specialties should be encouraged. Fortunately, the number of specialized judges
is growing (Rachlinski et al., 2007).
Third, if it is not possible to modify the system, it might be a possibility to try to
increase responsibility in general judges in order to reduce CIP, which in turn might facilitate
fair trial proceedings. Whereas the presently used induction may seem artificial, it was
sufficient to show that this feeling of responsibility may be the underlying mechanism, which
was the first and foremost purpose of the studies. Field research could reveal if this or an
adapted responsibility reminder would also work in real life. However, it has not yet been
investigated how long the effects of such manipulations last or if the reminder loses its
effectiveness after multiple presentations.
It could also be speculated whether the responsibility induction also works in a sample
of laypeople, thus potential jurors. If it is truly not a lack of knowledge that drives the effect
but motivation, then laypeople might profit just as much from a responsibility induction as
general experts. Additionally, general experts and laypeople showed the same level of biased
assimilation in the pilot study which also implies that the responsibility induction could have
the same effect in laypeople. Future research should explore this, as the practical implications
for potential jurors are high and the implementation is rather simple: A separate and strongly
empathized responsibility reminder could just be added to the standard juror instructions,
which usually just focus on the importance to not convict if reasonable doubt exists
(Horowitz, 1997).
We would like to thank the Volkswagen Stiftung for providing the funds for this
research. Furthermore, we would like to thank the German Judicial Academy and the Defense
Attorneys’ Union for giving us the opportunity to conduct research in their facilities, as well
as our research assistants: Miriam Schwarzenthal, Jasmin Schmitz, Martin Schafföner, Tobias
Wingen, Elisabeth Jackson, Janna Krahe, Jana Berkessel and Benedikt Ball. Furthermore, we
thank Helmut Appel, Barbara Stoberock, Pascal Burgmer, and three anonymous reviewers for
their helpful comments.
Declaration of Conflicting Interests
The author(s) declared no conflicts of interest with respect to the authorship and/or
publication of this article.
Ask, K., & Granhag, P. A. (2005). Motivational Sources of Confirmation Bias in Criminal
Investigations: The Need for Cognitive Closure. Journal of Investigative Psychology
and Offender Profiling, 2(1), 4363.
Benner, P., & Tanner, C. (1987). Clinical Judgment: How Expert Nurses Use Intuition. The
American Journal of Nursing, 87(1), 23.
Bornstein, B. H. (1999). The ecological validity of jury simulations: Is the jury still out? Law
and Human Behavior, 23(1), 7591.
Brailey, K., Vasterling, J. J., Franks, J. J., The, S., Journal, A., & Spring, N. (2001). Memory
of Psychodiagnostic Information: Biases and Effects of Expertise. The American
Journal of Psychology, 114(1), 55.
Carlson, K. A., & Russo, J. E. (2001). Biased interpretation of evidence by mock jurors.
Journal of Experimental Psychology: Applied, 7(2), 91103.
Charman, S. D., Gregory, A. H., & Carlucci, M. (2009). Exploring the diagnostic utility of
facial composites: beliefs of guilt can bias perceived similarity between composite and
suspect. Journal of Experimental Psychology. Applied, 15(1), 7690.
Chi, M. T. H., Glaser, R., & Rees, E. (1981). Expertise in problem solving. In R. J. Sternberg
(Ed.), Advances in the Psychology of Human Intelligence (pp. 175). Hillsdale, NJ, NJ:
Danziger, S., Levav, J., & Avnaim-Pesso, L. (2011). Extraneous factors in judicial decisions.
Proceedings of the National Academy of Sciences of the United States of America,
108(17), 688992.
Dawes, R., Faust, D., & Meehl, P. (1989). Clinical versus actuarial judgment. Science,
243(4899), 16681674.
Dhami, M. K. (2003). Psychological models of professional decision making. Psychological
Science, 14(2), 175180.
Dhami, M. K., & Ayton, P. (2001). Bailing and jailing the fast and frugal way. Journal of
Behavioral Decision Making, 14, 141168.
Doney, P. M., & Armstrong, G. M. (1996). Effects of accountability on symbolic information
search and information analysis by organizational buyers. Journal of the Academy of
Marketing Science, 24(1), 5765.
Dreyfus, H., & Dreyfus, S. E. (1986). Mind Over Machine. New York: Free Press.
Dror, I. E., Charlton, D., & Péron, A. E. (2006). Contextual information renders experts
vulnerable to making erroneous identifications. Forensic Science International, 156(1),
Eisenberg, T. (1994). Differing Perceptions of Attorney Fees in Bankruptcy Cases.
Washington University Law Review, 72(3), 979995.
Eisenberg, T., & Clermont, K. M. (1996). Trial by Jury or Judge: Which Is Speedier?
Judicature, 79, 176180,199.
Englich, B., & Mussweiler, T. (2001). Sentencing Under Uncertainty: Anchoring Effects in
the Courtroom1. Journal of Applied Social Psychology, 31(7), 15351551.
Englich, B., Mussweiler, T., & Strack, F. (2005). The Last Word in CourtA Hidden
Disadvantage for the Defense. Law and Human Behavior, 29(6), 705722.
Englich, B., Mussweiler, T., & Strack, F. (2006). Playing Dice With Criminal Sentences: The
Influence of Irrelevant Anchors on Experts’ Judicial Decision Making. Personality and
Social Psychology Bulletin, 32(2), 188200.
Feltovich, P. J., Johnson, P. E., Moller, J. H., & Swason, D. B. (1984). LCS: The role and
development of medical knowledge in diagnostic expertise. In W. J. Clancey & E. H.
Shortliffe (Eds.), Readings in medical artificial intelligence: The first decade (pp. 275
319). Reading, MA, MA: Addison Wesley.
Festinger, L. (1957). A theory of cognitive dissonance. Stanford CA: Stanford University
Findley, K. A., & Scott, M. S. (2006). The Multiple Dimensions of Tunnel Vision in
Criminal Cases. Wisconsin Law Review, 291.
Fischer, J., Fischer, P., Englich, B., Aydin, N., & Frey, D. (2011). Empower my decisions:
The effects of power gestures on confirmatory information processing. Journal of
Experimental Social Psychology, 47(6), 11461154.
Fischer, P., Greitemeyer, T., & Frey, D. (2008). Self-regulation and selective exposure: the
impact of depleted self-regulation resources on confirmatory information processing.
Journal of Personality and Social Psychology, 94(3), 38295.
Fischer, P., Jonas, E., Frey, D., & Kastenmüller, A. (2008). Selective exposure and decision
framing: The impact of gain and loss framing on confirmatory information search after
decisions. Journal of Experimental Social Psychology, 44(2), 312320.
Fischer, P., Jonas, E., Frey, D., & Schulz-Hardt, S. (2005). Selective Exposure to
Information: The Impact of Information Limits. European Journal of Social Psychology,
35(4), 469492.
Fischer, P., Schulz-Hardt, S., & Frey, D. (2008). Selective exposure and information
quantity: how different information quantities moderate decision makers’ preference for
consistent and inconsistent information. Journal of Personality and Social Psychology,
94(2), 23144.
Fiss, O. (1982). Objectivity and interpretation. Stanford Law Review, 34(4), 739763.
Frey, D. (1981). Reversible and Irreversible Decisions: Preference for Consonant Information
as a Function of Attractiveness of Decision Alternatives. Personality and Social
Psychology Bulletin, 7(4), 621626.
Guthrie, C., Rachlinski, J. J., & Wistrich, A. J. (2001). Inside the judicial mind. Cornell Law
Review, 86(4), 777.
Guthrie, C., Rachlinski, J. J., & Wistrich, A. J. (2007). Blinking on the bench: How judges
decide cases. Cornell Law Review, 93(1), 143.
Hart, W., Albarracın, D., Eagly, A. H., Brechan, I., Lindberg, M. J., & Merrill, L. (2009).
Feeling validated versus being correct: A meta-analysis of selective exposure to
information. Psychological Bulletin, 135(4), 555588.
Horowitz, I. A. (1997). Reasonable doubt instructions: Commonsense justice and standard of
proof. Psychology, Public Policy, and Law, 3(2-3), 285302.
Johnson, B. D. (2006). The Multilevel Context of Criminal Sentencing: Integrating Judge-
and County-Level Influences. Criminology, 44(2), 259298.
Jonas, E., Graupmann, V., & Frey, D. (2006). The influence of mood on the search for
supporting versus conflicting information: dissonance reduction as a means of mood
regulation? Personality & Social Psychology Bulletin, 32(1), 315.
Jonas, E., Schulz-Hardt, S., Frey, D., & Thelen, N. (2001). Confirmation bias in sequential
information search after preliminary decisions: An expansion of dissonance theoretical
research on selective exposure to information. Journal of Personality and Social
Psychology, 80(4), 557571.
Kahan, D. M., Hoffman, D. A., Evans, D., Devins, N., Lucci, E. A., & Cheng, K. (2016).
“Ideology” or “Situation Sense”? An Experimental Investigation of Motivated
Reasoning and Professional Judgment. University of Pennsylvania Law Review, 64,
Working Paper No. 201526. Retrieved from
Kassin, S. M., Dror, I. E., & Kukucka, J. (2013). The forensic confirmation bias: Problems,
perspectives, and proposed solutions. Journal of Applied Research in Memory and
Cognition, 2(1), 4252.
Kassin, S. M., & Gudjonsson, G. H. H. (2004). The Psychology of Confessions: A Review of
the Literature and Issues. Psychological Science in the Public Interest, 5(2), 3367.
Kastenmüller, A., Jonas, E., Fischer, P., Frey, D., & Fischer, J. (2013). Selective exposure
among advisors: the impact of impression motivation. Journal of Applied Social
Psychology, 43(1), 138146.
Kepplinger, H. M., & Zerback, T. (2009). Der Einfluss der Medien auf Richter und
Staatsanwälte. Publizistik, 54(2), 216239.
Kerstholt, J. H., & Eikelboom, A. R. (2007). Effects of prior interpretation on situation
assessment in crime analysis. Journal of Behavioral Decision Making, 20(5), 455465.
Klayman, J., & Ha, Y. (1987). Confirmation, disconfirmation, and information in hypothesis
testing. Psychological Review, 94(2), 211228.
Krems, J. F., & Zierer, C. (1994). Sind Experten gegen kognitive Täuschungen gefeit? Zur
Abhängigkeit des confirmation bias von Fachwissen. Zeitschrift für experimentelle und
angewandte Psychologie, 41(1), 98115.
Kulik, C. T., Perry, E. L., & Pepper, M. B. (2003). Here comes the judge: The influence of
judge personal characteristics on federal sexual harassment case outcomes. Law and
Human Behavior, 27(1), 6986.
Lord, C. G., Ross, L., & Lepper, M. R. (1979). Biased assimilation and attitude polarization:
The effects of prior theories on subsequently considered evidence. Journal of
Personality and Social Psychology, 37(11), 20982109.
Meissner, C. A., & Kassin, S. M. (2002). “He’s guilty!”: Investigator bias in judgments of
truth and deception. Law and Human Behavior, 26(5), 469480.
Nickerson, R. S. (1998). Confirmation bias: A ubiquitous phenomenon in many guises.
Review of General Psychology, 2(2), 175220.
Nöhmaier, N. (2010). Arbeitsrecht: Kündigung wegen Kokolores. Focus Online. Retrieved
O’Brien, B. (2009). Prime suspect: An examination of factors that aggravate and counteract
confirmation bias in criminal investigations. Psychology, Public Policy, and Law, 15(4),
Oskamp, S. (1965). Overconfidence in case-study judgments. Journal of Consulting
Psychology, 29(3), 261265.
Plous, S. (1993). The psychology of judgment and decision making. McGraw-Hill series in
social psychology. New York, NY England: Mcgraw-Hill Book Company.
Pressemitteilung Landesarbeitsgericht Baden-Württemberg. (2010, March 30).
Widerruflicher Vergleich im “Maultaschenfall.” [Press Release]. Retrieved from
Pressemittelung Bundesarbeitsgericht. (2010). Fristlose Kündigung - Interessenabwägung -
Abmahnung - Fall “Emmely.” [Press Release]. Retrieved from
Rachlinski, J. J., Guthrie, C., & Wistrich, A. J. (2007). Heuristics and Biases in Bankruptcy
Judges. Journal of Institutional and Theoretical Economics JITE, 163(1), 167186.
Rassin, E. (2010). Blindness to alternative scenarios in evidence evaluation. Journal of
Investigative Psychology and Offender Profiling, 7(2), 153163.
Redding, R. E., & Reppucci, N. D. (1999). Effects of lawyers’ socio-political attitudes on
their judgments of social science in legal decision making. Law and Human Behavior,
23(1), 3154.
Reyna, V. F., & Lloyd, F. J. (2006). Physician decision making and cardiac risk: effects of
knowledge, risk perception, risk tolerance, and fuzzy processing. Journal of
Experimental Psychology. Applied, 12(3), 17995.
Rozelle, R. M., & Baxter, J. C. (1981). Influence of role pressures on the perceiver:
Judgments of videotaped interviews varying judge accountability and responsibility.
Journal of Applied Psychology.
Russo, J. E., Carlson, K. A., Meloy, M. G., & Yong, K. (2008). The goal of consistency as a
cause of information distortion. Journal of Experimental Psychology. General, 137(3),
Salas, E., Rosen, M. A., & DiazGranados, D. (2010). Expertise-Based Intuition and Decision
Making in Organizations. Journal of Management, 36(4), 941.
Schwarz, N., Frey, D., & Kumpf, M. (1980). Interactive effects of writing and reading a
persuasive essay on attitude change and selective exposure. Journal of Experimental
Social Psychology, 16(1), 117.
Simon, D. (2004). A Third View of the Black Box : Cognitive Coherence in Legal Decision
Making Cognitive Coherence in Legal Decision Making. University of Chicago Law
Review, 71(2), 511586.
Simonson, I., & Nye, P. (1992). The effect of accountability on susceptibility to decision
errors. Organizational Behavior and Human Decision Processes, 51(3), 416446.
Smith, S. M., Fabrigar, L. R., Powell, D. M., & Estrada, M.-J. (2007). The role of
information-processing capacity and goals in attitude-congruent selective exposure
effects. Personality & Social Psychology Bulletin, 33(7), 94860.
Ste-Marie, D. M. (1999). Expert-Novice Differences in Gymnastic Judging: An Information-
processing Perspective. Applied Cognitive Psychology, 13(3), 269281.<269::AID-ACP567>3.0.CO;2-Y
Strier, F. (1992). What Can the American Adversary System Learn from an Inquisitional
System of Justice. Judicature, 76(3), 109111.
Tybout, A., Sternthal, B., Keppel, G., Verducci, J., Meyers-Levy, J., Barnes, J., …
Steenkamp, J.-B. (2001). Analysis of Variance. Journal of Consumer Psychology, 10(1-
2), 535.
van Dongen, K., & van Maanen, P. P. (2013). A framework for explaining reliance on
decision aids. International Journal of Human Computer Studies.
Vicente, K. J., & Wang, J. H. (1998). An ecological theory of expertise effects in memory
recall. Psychological Review, 105(1), 3357.
Vrij, A. (2004). Why professionals fail to catch liars and how they can improve. Legal and
Criminological Psychology, 9(2), 159181.
Walker, J. S. (2000). Choosing biases, using power and practicing resistance: Moral
development in a world without certainty. Human Development, 43(3), 135156.
Wason, P. (1966). Reasoning. In B. M. Foss (Ed.), New horizons in Psychology (Vol 1., Vol.
1, pp. 135151). Harmondsworth: Penguin Books.
Wistrich, A. J., Guthrie, C., & Rachlinski, J. J. (2005). Can Judges Ignore Inadmissible
Information? The Difficulty of Deliberately Disregarding. University of Pennsylvania
Law Review, 153(4), 12511345.
1 This study was run as the second part in a combined investigation package of judicial
decision making. The questionnaire prior to the critical case included a color priming (the
defendant either wore a black or a white t-shirt). A 3 (expertise) x 2 (color) ANOVA revealed
no significant main effects or interactions for color on confirmatory information processing
(all ps > .1).
2 Participants’ prediction of how many percent of the other participants made the same
preliminary decision as they did (0% - 100% in increments of 10%) was also assessed.
Several judicial experts in our study consistently indicated to be uncertain as to who exactly
was referred to by other participants. This variable will therefore be excluded from all
subsequent analyses.
3 Although we tried to only collect the data of criminal-law experts, recruitment
circumstances like shared conference facilities led to the participation of other legal-experts
as well.
4 We additionally checked for interfering effects: A 3 x 2 x 2 ANOVA with Occupation
(judge vs. prosecutor vs. defense-attorney), Expertise (domain-specific vs. general) and
Responsibility (induced vs. control) revealed that occupation had no effect on biased
assimilation, F(2, 98) = .26, p = .775, nor did it interact with expertise, F(2, 98) = 1.74, p =
.181, or with responsibility, F(2, 98) = .61, p = .534. Since defense-lawyers differ the most
from criminal-law judges and prosecutors with regard to training and professionals objectives
(establishment of the truth vs. defending a client), we temporarily excluded defense-lawyers
from the analysis. The pattern for biased assimilation remained unchanged (Responsibility x
Expertise: F(1, 75) = 3.98, p = .050, ηp² = .05).
Figure 1. Biased assimilation as a function of expertise group. A high value points towards a
large difference between the evaluation of supporting and conflicting arguments, indicating
confirmatory information processing. Error bars represent standard errors.
Figure 2. Expertise as a function of responsibility. The displayed scores are the standardized
scores of biased assimilation. Error bars represent standard errors.
... Also, Swedish prosecutors displayed a guiltconfirming mindset once they have charged a suspect (Lidén et al., 2019). Since the structure of the law of criminal procedure triggers psychological mechanisms in prosecutors and judges, thereby shaping the decisions of both, the initial preference for the incriminating story may not be compensated for in the trial (e.g., Englich et al., 2006;Rassin, 2020;Schmittat & Englich, 2016). Collectively, this stresses the importance of balanced and well-founded decisions of prosecutors. ...
... This approach allows us to test psychological mechanisms extensively with adequately large samples, before replicating the effect with an expert sample (prosecutors) in follow-up studies. Besides, research has demonstrated that influences on experts' decision making might be highly similar to those on lay people's decisions Miller, 2019;Schmittat & Englich, 2016), supporting the proposed approach. ...
... occasionally on professional judges (Eisenberg, 1994;Englich, 2006;Englich et al., 2005;2006;Miller, 2019;Rachlinski, et al., 2007;Schmittat & Englich, 2016). However, very few empirical studies have focused on prosecutors' decision-making. ...
In many legal systems, preliminary proceedings and the prosecutor’s charging decision predict the outcome of the trial, as reflected in high conviction rates. This charging decision is mostly based on the final police report, which is primarily aimed to substantiate the given suspicion and may lead to a preponderance of incriminating evidence. We investigated if a written statement by the defense (an alternative story) can balance out this information disparity and can alter the charging decision. Building on classical debiasing methods, the Story Model and narrative persuasion, we hypothesized that an alternative story reduces the likelihood of being charged, because the police report becomes less unique. Participants of three experiments (criminal-law students, N = 684) received either only the police report, or the police report plus an alternative story (vs. a statement merely claiming innocence) and were asked to make charging decisions. The presence of prior conviction evidence (PCE) was varied. Results show that 1) an alternative story reduces (and PCE increases) the likelihood of being charged, 2) a statement merely claiming innocence is less but still effective, and 3) PCE effects were small and inconsistent. Underlying mechanisms, implications for criminal procedural law, and applicability to adversarial systems are discussed.
... Theoretical grounds of coherence effects, also known as information distortion (e.g., Carlson & Russo, 2001;Russo, Medvec, & Meloy, 1996) and biased assimilation (Schmittat & Englich, 2016), emphasize its unconscious nature (Simon, 2004) and are in line with the theory of explanatory coherence (Thagard, 1989;. Instead of selectively exposing oneself to specific information (i.e., information search), the mental representation of the decision task shifts towards a state of internal coherence (Glöckner & Betsch, 2008). ...
Full-text available
Legal judgments require one to make sense of a complex set of typically contradicting pieces of information that can easily be interpreted in a biased manner. This systematic judgment bias can be caused by biased information search (i.e., confirmatory search) as well as biased information processing (i.e., coherence effects) in which the interpretation of information is changed to fit the emerging favored option. In four studies, we investigate the complex interplay between both kinds of influences. Participants completed three legal cases in which they could freely search for information. We manipulated between subjects whether systematic search was possible or not and measured the assessment of each selected piece of information. In line with previous studies, we observe strong coherence effects in each study, in that the evidence interpretation strongly depended on the current tendency towards acquittal or conviction. In contrast to our expectation, however, people searched for information that was contrary to their current belief in the given case (i.e., disconfirmatory information search). We also observed a trend towards an interaction between both factors, in that coherence effects were slightly stronger for neutral and pro-guilty evidence when systematic information search was possible. Our results underline an unconscious striving for coherence when making complex judgments that is not easily corrected.
... Furthermore, the training and experience of judges may cause them to respond differently than law students. That would be in line with the findings by Schmittat and Englich (2016), who found that criminal law experts showed less preference for confirming information in a criminal law case than did experts in other areas of law and laypeople. In addition to the need to determine whether law students and defence lawyers have a preference for alternative scenarios, the question of whether and how the consideration of clearly exonerating evidence differs between the different parties at trial also warrants investigation. ...
Full-text available
Order of evidence presentation affects the evaluation and the integration of evidence in mock criminal cases. In this study, we aimed to determine whether the order in which incriminating and exonerating evidence is presented influences cognitive dissonance and subsequent display of confirmation bias. Law students (N = 407) were presented with a murder case vignette, followed by incriminating and exonerating evidence in various orders. Contrary to a predicted primacy effect (i.e. early evidence being most influential), a recency effect (i.e. late evidence being most influential) was observed in ratings of likelihood of the suspect’s guilt. The cognitive dissonance ratings and conviction rates were not affected by the order of evidence presentation. The effects of evidence presentation order may be limited to specific aspects of legal decisions. However, there is a need to replicate the results using procedures and samples that are more representative of real-life criminal law trials.
... For instance, prior expertise enhances the influence of ideology on judicial decision-making, but accumulated experience does not (Miller and Curry, 2009). Legal experts with domain-specific expertise in criminal law show less sensitivity to confirmatory bias than legal professionals without this expertise (with specializations in other fields than criminal law) (Schmittat and Englich, 2016). Thus, our findings and previous evidence suggest that specific knowledge, background, and technical skills in criminal law have a relevant role in overriding cognitive and emotional biases which can influence decisionmaking. ...
Full-text available
Traditional and mainstream legal frameworks conceive law primarily as a purely rational practice, free from affect or intuition. However, substantial evidence indicates that human decision-making depends upon diverse biases. We explored the manifestation of these biases through comparisons among 45 criminal judges, 60 criminal attorneys, and 64 controls. We examined whether these groups’ decision-making patterns were influenced by (a) the information on the transgressor’s mental state, (b) the use of gruesome language in harm descriptions, and (c) ongoing physiological states. Judges and attorneys were similar to controls in that they overestimated the damage caused by intentional harm relative to accidental harm. However, judges and attorneys were less biased towards punishments and harm severity ratings to accidental harms. Similarly, they were less influenced in their decisions by either language manipulations or physiological arousal. Our findings suggest that specific expertise developed in legal settings can attenuate some pervasive biases in moral decision processes.
... Furthermore, the counterfactual arguments used in court may not be equally persuasive for all judges and juries. Research on judicial decision-making and its underlying processes has investigated several factors influencing judges' and jurors' evaluation of cases, arguments, and supporting evidence (Danziger, Levav, & Avnaim-Pesso, 2011;Dhami, 2003;Ellison & Munro, 2009;Posner, 2008;Schmittat & Englich, 2016). Judges' expertise has emerged as a potential key factor in the quality of judicial decisions, as more expert judges may be able to make more rigorous and accurate evaluations than less experienced colleagues, based on their greater familiarity with the application of the law to real-world cases, and greater ability to detect potential sources of bias (Catellani, 1992). ...
Full-text available
Past research has shown that counterfactual (“If…then…”) thoughts influence causal and responsibility attribution in the judicial context. However, little is known on whether and how the use of counterfactuals in communication affects lay jurors’ and judges’ evaluations. In two studies, we asked mock lay jurors (Study 1) and actual judges (Study 2) to read a medical malpractice case followed by an expert witness report which included counterfactuals focused on either the physician, the patient, or external factors. Results showed that counterfactual focus had a strong direct effect on both lay jurors’ and judges’ causal and responsibility attributions. Counterfactual focus also moderated the effect of outcome foreseeability on responsibility attribution. Discussion focuses on how counterfactual communication can direct causal and responsibility attribution and reduce the importance of other factors known to influence judicial decision‐making. The potential implications of these findings in training programs and debiasing interventions are also discussed. This article is protected by copyright. All rights reserved.
... For starters, confirmation bias may lead investigators to select information that is incriminating for the primary suspect, while leaving out exonerating information (e.g. Kassin, Dror, & Kukucka, 2013;Kassin, Goldstein, & Savitsky, 2003;Schmittat & Englich, 2016). Ultimately, this biased selection may lead to a skewed view on the probability that the suspect is guilty. ...
Full-text available
Purpose Fact finding is an important part of the job of criminal trial judges and juries. In the literature, several potential pitfalls hindering fact finding have been identified, such as context effects (i.e. an unintended effect of non‐probative information on conviction) and confirmation bias (i.e. a skewed selection of and overreliance on guilt‐confirming evidence and neglect of exonerating information). In the present study, the effect of irrelevant contextual information on conviction and subsequent confirmation bias was tested. Method A sample of Dutch professional criminal trial judges (N = 105) studied a case file and decided on their conviction of the suspect’s guilt, and subsequent investigation endeavours. There were two versions of the file, differing in non‐probative details that might affect conviction, such as crime severity and facial appearance of the suspect. Results Findings suggest that context information indeed affected conviction, and the subsequent preference for guilt‐confirming investigation endeavours. Conclusion Professional judges may be susceptible to bias threatening the objectivity of legal decision‐making.
Full-text available
This thesis studies the pretrial screening stage of the criminal process. It does so by, first, explaining its regulation in specific procedures in three different countries: Spain, Colombia, and the United States. And second, by evaluating such procedures by reference to values and criteria which are proposed as guidelines. It is my argument that one cannot find an effective substantial screening of the case at this stage in any of the studied procedures. As a consequence, the pretrial screening becomes ineffective as a means to avoid unnecessary trials and the abuse of guilty pleas. Furthermore, I propose that an acceptable configuration of the pretrial screening stage requires (1) the presentation of a verifiable legal argument that allows for confrontation and control; (2) the respect of the defendant's guarantees; (3) the consideration of the victim's interests; and (4) the avoidance of unreasonable delays in the solution of cases. Esta tesis estudia la fase intermedia del proceso penal como momento de control del caso previo al juicio. Lo hace a través de la explicación de cómo se configura en procedimientos específicos de diferentes países (España, Colombia y Estados Unidos), y también a través de su evaluación con referencia a los valores y criterios que se considera deben guiar su regulación. Se defiende que en ninguno de los procedimientos estudiados se realiza un eficaz control sustancial del caso en esta fase y esto genera que sea ineficaz como mecanismo para evitar la apertura de juicios innecesarios y para racionalizar el uso de las aceptaciones tempranas de culpabilidad. Se postula, además, que para su configuración aceptable se debe: (1) exigir la presentación de un argumento jurídico verificable frente al que se permita contradicción y control; (2) respetar las garantías del procesado; (3) considerar los intereses de la víctima; y, (4) no generar demoras irrazonables en la solución de los casos.
Purpose The purpose of this study was to investigate the influence of sexual objectification on the attribution processes of the guilt of a defendant – and also on the level of guilt. It was also hypothesized that legal expertise could be a protective factor in countering the influence of sexual objectification. Design/methodology/approach Sexual objectification can be defined as the perspective in which a person is evaluated solely in terms of his or her body parts or sexual function. As yet, no studies have assessed the influence of sexual objectification on guilt assessment in the legal system; this paper aims to explore whether sexual objectification has an influence on the attribution processes of a defendant's guilt. Findings The statistical analysis revealed that the sexually objectified defendant received a guilty verdict more often than a non-sexually objectified defendant; additionally, legal experts were more likely to identify the defendant as not guilty than non-legal experts. The findings support the hypothesis that sexual objectification is indeed one of the common stereotypes that lead to discrimination. Originality/value The present study provides novel findings regarding sexual objectification in the forensic context in which the defendant is viewed and evaluated.
Full-text available
Background: Bias in reasoning rather than knowledge gaps has been identified as the origin of most diagnostic errors. However, the role of knowledge in counteracting bias is unclear. Objective: To examine whether knowledge of discriminating features (findings that discriminate between look-alike diseases) predicts susceptibility to bias. Design: Three-phase randomized experiment. Phase 1 (bias-inducing): Participants were exposed to a set of clinical cases (either hepatitis-IBD or AMI-encephalopathy). Phase 2 (diagnosis): All participants diagnosed the same cases; 4 resembled hepatitis-IBD, 4 AMI-encephalopathy (but all with different diagnoses). Availability bias was expected in the 4 cases similar to those encountered in phase 1. Phase 3 (knowledge evaluation): For each disease, participants decided (max. 2 s) which of 24 findings was associated with the disease. Accuracy of decisions on discriminating features, taken as a measure of knowledge, was expected to predict susceptibility to bias. Participants: Internal medicine residents at Erasmus MC, Netherlands. Main measures: The frequency with which higher-knowledge and lower-knowledge physicians gave biased diagnoses based on phase 1 exposure (range 0-4). Time to diagnose was also measured. Key results: Sixty-two physicians participated. Higher-knowledge physicians yielded to availability bias less often than lower-knowledge physicians (0.35 vs 0.97; p = 0.001; difference, 0.62 [95% CI, 0.28-0.95]). Whereas lower-knowledge physicians tended to make more of these errors on subjected-to-bias than on not-subjected-to-bias cases (p = 0.06; difference, 0.35 [CI, - 0.02-0.73]), higher-knowledge physicians resisted the bias (p = 0.28). Both groups spent more time to diagnose subjected-to-bias than not-subjected-to-bias cases (p = 0.04), without differences between groups. Conclusions: Knowledge of features that discriminate between look-alike diseases reduced susceptibility to bias in a simulated setting. Reflecting further may be required to overcome bias, but succeeding depends on having the appropriate knowledge. Future research should examine whether the findings apply to real practice and to more experienced physicians.
One collateral consequence of a criminal conviction for parents with young children is the loss of custody, which turns on the subjective “best interests of the child” standard. This research explored whether criminal conviction and substance abuse history influenced custody decisions. Experiment 1 presented community participants with a vignette describing a parent with combinations of manipulated stigmatized characteristics (i.e., gender, race, offender status, and substance abuse). Participants completed a custody determination scale, which showed that mothers with an offense received more favorable custody decisions than fathers with an offense, as did ex‐offenders without substance abuse. Experiment 2 added a positive or negative psychological fitness evaluation of the ex‐offender. It found main effects of the professional parenting evaluation and replicated the parent's substance abuse findings from Experiment 1, but not the offense status result. Most importantly, these results were significant after controlling for the participants’ ratings of the best interests of the child in question and mediation analyses revealed that the child's best interests only partially explained the relationships between substance abuse, the parental evaluation, and the custody determination. This suggests that participants made custody decisions based on factors extraneous to the current legal standard, namely, the best interests of the child.
Full-text available
Recently, in a number of high-profile cases, defendants who were prosecuted, convicted, and sentenced on the basis of false confessions have been exonerated through DNA evidence. As a historical matter, confession has played a prominent role in religion, in psychotherapy, and in criminal law-where it is a prosecutor's most potent weapon. In recent years, psychologists from the clinical, personality, developmental, cognitive, and social areas have brought their theories and research methods to bear on an analysis of confession evidence, how it is obtained, and what impact it has on judges, juries, and other people.
Full-text available
Confirmation bias, as the term is typically used in the psychological literature, connotes the seeking or interpreting of evidence in ways that are partial to existing beliefs, expectations, or a hypothesis in hand. The author reviews evidence of such a bias in a variety of guises and gives examples of its operation in several practical contexts. Possible explanations are considered, and the question of its utility or disutility is discussed.
In a criminal trial, the standard of proof is successfully met when the prosecution presents facts that convince the jury "beyond a reasonable doubt." Recent research suggests that most definitions of beyond reasonable doubt that have passed constitutional muster elicit subjective standards of certainty of guilt that are lower than previously observed. These results intimate that citizens may have altered the balance between due process and crime control values and may be willing to convict on a lesser showing than due process requires. This article discusses the possibility that, rather than the well-established "leniency bias" (acquittal bias) in criminal trials, juries may now exhibit a conviction bias in certain types of trials.
This Article reports the results of a study on whether political predispositions influence judicial decisionmaking. The study was designed to overcome the two principal limitations on existing empirical studies that purport to find such an influence: the use of nonexperimental methods to assess the decisions of actual judges; and the failure to use actual judges in ideologically-biased-reasoning experiments. The study involved a sample of sitting judges (n = 253), who, like members of a general public sample (n = 800), were culturally polarized on climate change.
Research on juridical decision making has demonstrated that largely disparate sentences are often given for identical crimes. This may be the case because judges' sentencing decisions are influenced by a recommended or demanded sentence. Building on research on judgmental anchoring (Tversky & Kahneman, 1974), the present investigation examines whether a sentencing demand has a direct influence on a given sentence. Using criminal trial judges as participants, Study 1 demonstrates that such a direct influence does, in fact, exist. Sentencing decisions are assimilated to the sentence demanded by the prosecutor. Study 2 further reveals that this influence is independent of the perceived relevance of the sentencing demand. Study 3 demonstrates that this influence is also independent of judges' experience.
Audience confirmation bias (ACB) refers to the extent to which people prefer information supporting (vs. conflicting with) their audience's views. In two studies, we showed that advisors shifted their ACB toward the needs of their advisees (i.e., audience): When advisors were led to believe that their advisees wanted to defend their views, the ACB was higher compared with when advisees were open minded for critique. Study 2 indicated that this pattern occurred because advisors wanted to have a pleasant interaction with their advisees (impression motivation): Whereas impression‐motivated advisors exhibited a stronger ACB when they were asked to give advice to a defensive (vs. open‐minded) advisee, accuracy‐motivated advisors showed a balanced ACB, regardless of their advisee's needs.