In 2002, Daniel Kahneman was awarded the Nobel Prize for his groundbreaking work in understanding the economics of human decision-making. In the classical economic paradigm, individuals make decisions as rational beings to maximise their utility in circumstances of “perfect information”. Kahneman’s work on behavioural economics demonstrated that this was not so.

Summarising 40 years of research in his book, Thinking, fast and slow,⁶ his key message was that humans are intuitive thinkers, human intuition is imperfect, and these imperfections often result in judgments that deviate from the predictions of classical economic models based on the assumption of rational behaviour. ⁷

According to Kahneman, human judgments can be produced in two ways — a rapid, associative, automatic, and intuitive process (called System 1, or “fast thinking”); and a slower, rule-governed, deliberate, and effortful process (called System 2, or “slow thinking”). The two systems are interconnected, but the division of labour between them is designed to minimise human effort and optimise performance.

Most of our normal decision-making is based on fast thinking (System 1) because the human mind has evolved to allow individuals to reach quick solutions to complex problems. As an evolutionary adaptation, fast thinking generally does very well at modelling familiar situations.⁸ Alas, it can also lead to systemic errors of judgment because it relies on heuristics (mental shortcuts) and biases, with the consequence that some critical information is ignored, while less relevant information receives undue attention.⁹ Intuitive thinking has selective value, but it comes at a price.

Kahneman’s work unseated the rationalist underpinnings of classical economics,¹⁰ but it also connected with parallel themes in the field of cognitive psychology on the nature of human biases.¹¹ It is unsurprising that some people are explicitly racist, sexist, ageist, or ableist. Yet one of the insights from psychology research is that people also have implicit biases (sometimes called unconscious biases) of which they are unaware and over which they have no mindful control. I may pride myself on being egalitarian, but my subconscious (and yours!) is marked by a deep thumbprint of culture and social life that embed preferences and attitudes learned during early childhood and beyond. Through socialisation, we all acquire insistent associations between certain groups and sets of attributes (typically, with a positive or negative valence), and these associations become part of our own value structure.¹²

The direction and strength of these biases can be measured. In 1998, Greenwald et al developed a test of implicit bias by documenting the speed with which respondents were able to associate different concepts when flashed quickly on a computer screen — for example (but greatly simplifying) “White/good”, “White/bad”, “Black/good”, “Black/bad”.¹³ Slower response times were taken as a measure of the cognitive strain (and hence bias) in associating certain concepts with others (eg “Black” and “good”). The Implicit Association Test (IAT) has now been administered millions of times across the globe and it reveals that implicit bias is rife.¹⁴ You and I might not have implicit bias towards the same categories, or to the same degree, but we both have it (even if you are a judicial officer).¹⁵ At a population level, these biases favour culturally dominant and societally valued groups, such as White over Black, rich over poor, young over old, and straight over LGBTIQ. Interestingly, non-dominant groups can share the same implicit biases as dominant groups because of their positive evaluations of a higher status group to which they do not belong.¹⁶ The impact of implicit biases on real-life behaviours is discussed in “Impact of implicit biases on behaviour”, below.

These revelations have importance for judges and magistrates, who make decisions daily on legal matters large and small. Judicial officers take an oath or affirmation upon appointment, binding them to an ideal of neutral decision-making in service to the law. In the terms of the affirmation required of appointees to the High Court, justices solemnly promise to “do right to all manner of people according to law without fear or favour, affection or ill-will”.¹⁷ The core meaning is that a judicial officer will apply the law without “prejudice, partiality or prejudgment”¹⁸ and without concern for the consequences to them. In Lord Mansfield’s immortalised words, “let justice be done, though the heavens fall”.¹⁹

Research on behavioural economics and social psychology throws down the gauntlet to the ideal of judicial impartiality (as does the growing scholarship on law and emotions).²⁰ If all humans have implicit biases in their decision-making, how can judicial officers acquit their duties without partialities towards persons who come before them? These need not be negative, hostile, or adverse assessments of litigants, witnesses, or counsel. Many implicit biases concern the favouritism unconsciously displayed by decision-makers towards others in their “ingroup”; reserving admiration, sympathy, and trust for them, while withholding it from individuals beyond their ingroup.²¹

The risk of implicit bias (whether positive or negative) may be lower in an apex court or in intermediate courts of appeal, which place a premium on weighing arguments, deliberating with care, and producing comprehensive written reasons — activities that reflect System 2 thinking. Yet nine out of 10 matters finalised in Australian courts each year are not made in that milieu but in magistrates’ courts,²² where there is much less distance between the judicial officer and the often-unrepresented defendant. As Roach Anleu and Mack have reported, lower courts in Australia are characterised by “intense time pressure on the presiding magistrate, who is faced, every day, with a large number of matters that appear impossible to complete within the allotted time, and no way of knowing which individual matters will require substantial attention and how long the list will take”.²³ While all judicial decision-making is likely to involve a combination of fast (System 1) and slow thinking (System 2), lower courts place more demand on the former. It is precisely when judicial officers work under conditions of enormous pressure that they “need to be especially on guard against their biases”.²⁴

Let’s assume we accept the claim that everyone has implicit biases of one kind or another. Do they matter? Do they determine our behaviour in the real world, such as by differentially favouring a majority group or disfavouring a minority group?

As it turns out, the relationship between implicit bias and behaviour is not robust. Scholars report that there is only a small to medium correlation between implicit bias scores and explicit (discriminatory) behaviours.²⁵ However, it is argued that even small experimental effects can have large societal consequences in aggregate.²⁶ When individual biases are accumulated across countless settings (education, employment, healthcare, the justice system), across a population, and across time, they can produce “tailwinds and headwinds that profoundly perturb our commitment to giving everyone a fair shot”.²⁷ For this reason, implicit bias is an important social phenomenon even if the impact on an individual’s behaviour is modest.

If individuals have implicit biases in their encounters with the world, and if those biases affect their real-world behaviour, it is natural to ask whether interventions are available to mitigate the impact. Once the Implicit Association Test became widely available in the early 2000s, the corporate world was quick to implement bias testing and anti-bias training for managers, in the optimistic expectation of remediating implicit bias in the workplace. Similar enthusiasm infected researchers who were concerned about implicit bias in the courts, since there is “no legitimate basis for believing that these pervasive implicit biases somehow stop operating in the halls of justice”.²⁸

In a major study in 2012, Kang et al examined “concrete intervention strategies to counter implicit biases for key players in the justice system, such as the judge and jury”.²⁹ One suggested approach was to eliminate or reduce implicit bias by exposing key actors to counter typical associations. Specifically, the authors recommended encouraging intergroup social contact by diversifying the bench, the courtroom, residential neighbourhoods, and friendship circles.³⁰ A second approach was to insulate the biases — that is, accept the existence of implicit biases but alter decision-making so that biases were less likely to translate into adverse behaviour. The debiasing strategies for judges included doubting one’s objectivity; informing oneself about implicit bias; improving decision-making conditions to encourage effortful thinking (System 2) rather than intuitive thinking (System 1); and keeping a tally of decisions to assess patterns of behaviour that might otherwise go unnoticed.³¹

A decade later, Kang reprised his assessment of appropriate interventions, outlining 24 actions to address implicit bias, grouped into four broad categories.³² Many of these actions also appear in Wistrich and Rachlinski’s thoughtful discussion of ways to target implicit bias.³³ It is important to observe that these inventories are largely directed to actions judges can take individually to eliminate or work around their own biases. However, Kahneman had already argued that errors of intuitive thought are often difficult to overcome because System 1 operates automatically and cannot be turned off at will.³⁴ The best we can do, he said, is to recognise situations in which mistakes are likely and try harder to avoid them when the stakes are high.

Reviews of the literature reinforce Kahneman’s conjecture. A meta-analysis of prejudice reduction strategies by Paluck et al showed little support for the hypothesis that “mentalising” successfully acts as a salve for implicit bias.³⁵ According to their analysis, the most important landmark studies showed “remarkably modest effects” from individual interventions. In another systematic review, FitzGerald et al found that common interventions such as training were often ineffective in reducing implicit bias, giving participants and organisations false confidence when in fact the training had no ameliorative effect.³⁶ Moreover, even when implicit biases have been responsive to interventions, Lai et al found that the changes were short-lived (at most a few days), suggesting that biases can be stable over time and resistant to mild interventions.³⁷ This is probably because implicit bias is not only generated but also maintained by exposure to the prevailing culture.³⁸ As Elek and Miller explain, social discrimination is like a virus that can be caught from, and reinforced by, the social environment; and interventions that attempt to change these implicit associations in one’s memory are not consistently effective.³⁹ None of this is good news for judicial officers who seek agency over their hidden preferences.

The preceding empirical findings are not necessarily a cause for despondency. As FitzGerald et al conclude in their systematic review, “the fact that there is scarce evidence for particular bias-reducing techniques does not weaken the case for implementing widespread structural and institutional changes that are likely to reduce implicit biases”.⁴⁰ The central question for this article is whether greater judicial diversity is one such mechanism for mitigating the impact of implicit bias in resolving legal disputes.

The first point to note is that diversity is a structural issue, and hence may be a suitable candidate for a remedy. Although people sometimes use the term “diversity appointment” to describe an individual judicial officer from a non-traditional background, this is loose talk. Diversity is not an attribute of a person but of a population.⁴¹ At a point in time, individuals generally have only one sex, one religion, one ethnicity (even if mixed), and so on. Some attributes are immutable, while others may change over a person’s lifetime. The combination of attributes that can potentially coexist in a single individual is very large, and they give each person a unique array of characteristics. But this is not what we mean when speaking of diversity. Judicial diversity focusses on single attributes (and occasionally on binary combinations, where intersectionality is in issue) and examines the variability of that attribute across the relevant population or subpopulation. Thus we may ask, is the Bench largely male, largely Christian, and largely white? Diversity is a structural property of the corpus of judges and magistrates, not a property of any one of them.

To assess whether greater judicial diversity is an effective structural remedy for implicit bias in judicial decision-making, it is helpful to revisit Kahneman’s two ways of thinking — the rapid, associative process (System 1), and the slower, effortful process (System 2). In principle, the impact of System 1 thinking, with its errors and biases, could be reduced through three different, but cumulative, approaches:

This article began by drawing together two strands of overlapping scholarship on the way humans make decisions — Kahneman’s Nobel prize-winning insights into behavioural economics, and the psychology literature on implicit basis. Common to both fields is the realisation that humans are often intuitive decision-makers, and those intuitions are conditioned by the explicit and implicit norms of the society in which we grow up and live. For judges and magistrates committed to an ethic of impartiality, the existence of unconscious attitudes and preferences is a professional challenge, whether these favour the judicial officer’s “ingroup” or disfavour an “outgroup”. That challenge is likely to be greatest for those who work in busy lower courts with unremitting caseloads.

Much work has been done to consider how judicial officers might individually ameliorate their implicit biases when exercising judicial power, and many courts and judicial education bodies now run programs for that purpose. Yet, there is conflicting evidence about whether these programs are effective, with recent systematic reviews suggesting the impact of personal interventions is often modest and short lived.

Taking an alternative approach, this article has investigated whether implicit bias can be addressed by structural or institutional change, and specifically whether greater diversity in the judiciary can mitigate these biases. I suggested two ways this could be done: (a) by appointing more judicial officers whose non-traditional backgrounds enliven different System 1 intuitions from those of past appointees; and (b) by using counter-stereotypes to alter the formation of implicit biases at their inception and their preservation thereafter. A judiciary that looks more like Australian society (at least in respect of personal or social attributes that are considered to matter) may, in the long run, have that effect.

The focus on individual action to counter implicit bias may give organisations the sense of a quick cure, but it can also divert attention from deeper structural, institutional, and historical causes.⁶⁷ It needs to be acknowledged that structural change is often challenging. Bringing greater diversity to the Bench is a lengthy process because judicial officers are guaranteed long tenure, and change therefore comes “one retirement at a time”. Greater diversity may also require reform of existing institutions, such as moving from a model of appointment based on “virtually unfettered executive discretion”⁶⁸ to one that gives a formal role to a judicial appointments commission, with greater transparency about the value of diversity in selection.

On its own, diversity will never be a complete cure for the ills of implicit bias among those who hold office as judges or magistrates. The quantitative representation of different groups on the Bench has value, but it does not capture how much each group is heard or how much influence they have.⁶⁹ Diversity therefore needs to be supplemented by institutional change to support greater inclusion. When greater judicial diversity is administered alongside other structural remedies, it may mitigate the adverse effects of a complex phenomenon so that judicial officers can do their best to “do right to all manner of people according to law”.