Legal Techniques for Rationalizing Biased Judicial Decisions: Evidence from Experiments with Real Judges

Date01 September 2019
Published date01 September 2019
DOIhttp://doi.org/10.1111/jels.12229
Journal of Empirical Legal Studies
Volume 16, Issue 3, 630–670, September 2019
Legal Techniques for Rationalizing Biased
Judicial Decisions: Evidence from
Experiments with Real Judges
John Zhuang Liu and Xueyao Li*
Judges rarelyreveal their real reasoningin their opinions when they are influenced by factors
that they know they should not consider. The natural next question is how, when a judge is
improperly influenced, he or she reasons tojustify a biased decision. In a set of experiments
using incumbent Chinese judges, we first replicated the findings of previous studies that
showed judges can be influenced by extra-legal factors. More importantly, we showed that
judges may employ a range of legal techniques to rationalize decision biases: they interpret
legal standards and legal concepts strategically, finesse the applicability of law, infer or deny
causation and foreseeability, anddraw different conclusions from facts.Our findings provide
a more realisticunderstanding of how judges behave,and cast doubt on reasoned elaboration
as a guarantee of judicialtransparency and trustworthiness.
I. Introduction
Every judicial system consists of institutions designed to constrain judicial powers and
ensure accountability. The centrality of reasoned elaboration in judicial decision making
is one of these institutions. Reason-giving limits judicial discretion by ensuring that
written decisions can be read and reviewed by higher courts (Shapiro 1987). It
encourages public monitoring and participation in the judicial decision-making pro-
cess (Eisenberg 1978; Fuller 1978). Giving reasons for decisions forms part and parcel
of the authority of the judiciary, and the use of judicial power unsupported by reasons
detracts from the trustworthiness of the judiciary (e.g., Rawls 1993:231–40; Schauer
1995; Sunstein 2007; Wechsler 1959). Henry Hart and Albert Sacks (1994:147–48)
famously argue that “reasoned elaboration” distinguishes judicial decision making
*Address correspondence to Xueyao Li, Professor, KoGuan Law School, Shanghai Jiang Tong University, Koguan
Law Bldg., 1954 Huashan Rd., Shanghai, China, 200030; email: lixueyao@sjtu.edu.cn. Liu is Assistant Professor of
Law and Economics, School of Management and Economics, the Chinese University of Hong Kong, Shenzhen.
We thank Jianjun Bai, Omri Ben-Shahar, Yun-Chien Chang, Jinhua Cheng, Xin Dai, Yan Ge, William Hubbard,
Holger Spamann, Yingmao Tang, Chao Xi, Wenming Xu, Qun Yang, and Wei Zhang for helpful comments and
discussions. For excellent research assistance, we thank Yanyan Zhang. We are grateful to all of the participating
judges in our studies.
630
from an exercise in “discretionary fiat.” In Law’s Empire, Ronald Dworkin (19 86:94–96)
describes reasoning as a source of integrity in law.
On the other hand, many commentators have noticed that judicial reasoning is
habitually and unusually one-sided, and filled with certainty. Judge Posner (1988) states
that most judicial opinions “imply that even the very toughest case has a right and a wrong
answer and only a fool would doubt that the author of the opinion had hit on the right
one.” He adds that judges “decide cases with greater confidence than the realities of judi-
cial decision making permits, and speak and write with more confidence than they feel”
(Posner 1988). This feature of judicial opinions is particularly perplexing, since many dis-
putes are complex and fraught with conflict and uncertainty (e.g., Schauer 1988; Simon
1998). Karl Llewellyn, among many other jurists, was particularly skeptical of this kind of
certainty. He stated that, for any but the easiest cases, legal certainty has never existed and
never will. Most judicial reasoning is simply the “dressing up” of judicial opinions in the
“garb of certainty” (Llewellyn 1989:73). In other words, reasoning is simply the clothing
judges put on their decisions, rather than the basis on which they make those decisions.
Against this backdrop, there is an impressive collection of studies examining biases
in judicial decision making. These studies examine mechanisms such as “hindsight bias,”
“anchoring,” “probability neglect,” “emotional bias,” “identity protective cognition,” and
the like (e.g., Guthrie et al. 2001, 2007; Kahan 2015), the effects of which bias both prob-
abilistic and analytical reasoning. One strand of studies repeatedly confirms that irrele-
vant information, such as litigant characteristics, can unduly influence judicial decision
making. Wistrich et al. (2015) use several experiments to reveal this general pattern. They
provided judges with vignettes of cases in which they varied the defendants’ characteris-
tics. Although the characteristics were irrelevant to the cases’ contexts and therefore
should not have been considered when rendering a verdict and sentence, they strongly
impacted the judges’ decisions. In an experiment resembling real-world judicial decision
making (presenting judges with raw facts, precedents, and trial judgments and giving
them about an hour to make decisions), Spamann and Klo
¨hn (2016) also show that
legally irrelevant characteristics of a defendant influence judges’ verdicts. Judges are
more likely to convict an unrepentant defendant than one who is repentant.
More germane to our study, research also shows that judges rarely reveal the real
reasons for their decisions when they have been influenced by factors they know they
should not have taken into account. For example, Spamann and Klo
¨hn (2016) required
their experimental subjects to write reasons after they made decisions. Judges’ written
reasons, however, never mention defendant characteristics at all. Liu (2018) found a simi-
lar pattern in his experiments with judges in China, showing that at least some judges are
influenced by irrelevant defendant character evidence, but none of them mention it in
their reasoning.
Our current study is a natural next step from this previous literature. Judges are
expected, and often required, to provide reasons for their decisions. However, when
judges are unduly influenced, how do they justify their biased decisions; in other words,
what legal techniques do they employ to rationalize their biased decisions?
Theories and preliminary evidence suggest that judges may use a wide range of
legal techniques to do so. For example, critics have long speculated that the judicial
Techniques for Rationalizing Biased Judicial Decisions 631
interpretations that judges employ in their opinions are simply window dressing for their
ideologies. When a judge uses originalism to interpret the Constitution, she does so not
because she is a true adherent to that strategy and finds its result determinative, but
because it is a means to a conservative end (Post & Siegal 2006; Chilton & Posner 2015).
This speculation, however, can hardly be verified by rigorous empirical evidence, due to
a conflation of legal and ideological preferences. In criminal law, Nadler and McDonnell
(2011) found that actors’ bad motives or bad moral character could not only increase
perceived blame and responsibility, but also perceived causal influence and intentional-
ity. This finding implies that laypeople may use causation and intention to rationalize
their decisions, which have readily been influenced by irrelevant defendant characteris-
tics. Sood and Darley (2012) found a similar pattern with respect to harm and criminal
liability. To impose criminal responsibility that has been driven by irrelevant moral intui-
tion, people impute harm to conduct where harm was not previously reported. Both
Nadler and McDonnell (2011) and Sood and Darley (2012) focus on the psychological
mechanism at work in certain legal contexts. They do not intend to document a more
general pattern in the law.
In this article, we provide an empirical analysis lying at the intersection of two
themes: biased judicial decisions and reasoned elaboration. We try to provide an outline
of the legal techniques judges can employ to rationalize biased decisions. We report three
experiments that studied this area, in the domains of contract law (Study 1), criminal law
(Study 2), and torts (Study 3), and we studied an array of techniques, spanning both legal
and factual matters. Our experiments used more than 100 real, incumbent Chinese
judges as subjects, and employed a standard between-group design. Judges were ran-
domly assigned to different conditions. They read cases vignettes that were adapted from
real cases.
We first induced the judges to make biased decisions, using stimuli that are, under
the law applicable to the precise question at hand, strictly irrelevant (e.g., the defendant’s
moral character in a determination of guilt). Study 1 was a lease contract case involving
liquidated damages and the court’s discretion in lowering such damages. In the treat-
ment condition, the defendant was described as maintaining an extramarital relationship
with a corrupt government official. In Study 2, a defendant who bred and sold parrots at
home was charged with “illegal sale of precious and endangered wild animals and their
related products.” He was described as indulging in gambling in one experiment condi-
tion and as a good father in the other condition. In Study 3, the defendant had stored
oxygen tanks in his courtyard, which led to a large fire, and he was being sued for tort lia-
bility by a plaintiff who voluntarily participated in firefighting but fell on the road and
hurt himself. In one condition, the defendant was storing oxygen to produce crystal
meth; in the other, he did so to better care for his mother’s lung disease (this case was
inspired by Nadler and McDonnell [2011]). Using these designs, we successfully repli-
cated the findings of prior experimental studies showing that judges can be influenced
by extra-legal factors (Wistrich et al. 2015; Spamann & Klo
¨hn 2016; Liu 2018).
Second, as the focus of our study, we asked the judges to provide reasons for their
decisions. A few reasons that are most relevant to the case issue were given to them in the
form of options in a multiple-choice question, and the stimulus factor was also included
632 Liu and Li

To continue reading

Request your trial

VLEX uses login cookies to provide you with a better browsing experience. If you click on 'Accept' or continue browsing this site we consider that you accept our cookie policy. ACCEPT