Jerry Kang, Nilanjana Dasgupta, Kumar Yogeeswaran, & Gary Blasi, Are Ideal Litigators White? Measuring the Myth of Colorblindness, 7 J. Empirical Leg. Studies 886-915 (2010) .
This study examined whether explicit and implicit biases in favor of Whites and against Asian Americans would alter mock jurors’ evaluation of a litigator’s deposition. We found evidence of both explicit bias as measured by self-reports, and implicit bias as measured by two Implicit Association Tests. In particular, explicit stereotypes that the ideal litigator was White predicted worse evaluation of the Asian American litigator (outgroup derogation); by contrast, implicit stereotypes predicted preferential evaluation of the White litigator (ingroup favoritism). In sum, participants were not colorblind, at least implicitly, towards even a “model minority,” and these biases produced racial discrimination. This study provides further evidence of the predictive and ecological validity of the Implicit Association Test.
Executive Summary
Introduction. “Racial discrimination.” Today, few terms generate greater anxiety, concern, resentment, and passion in American society. But such moral and emotional intensity doesn’t shed much light on what “racial discrimination” actually is. Even if we define racial discrimination narrowly—to cover only disparate treatment of a specific individual because of that individual’s race—there remains substantial empirical complexity about what “because of” actually means.
The empirical complexity arises, in part, from the operation of implicit social cognitions (“ISCs”). In the past decade, scientists working across the boundaries of neuroscience, cognitive psychology, social psychology, and behavioral economics have demonstrated the existence of implicit social cognitions generally, including ISCs about racial groups. These ISCs turn out not to be randomly oriented; instead, they are biased in predictable directions in favor of groups higher on the social hierarchy. More recently, scientists have been documenting evidence of “predictive validity”—namely, that ISCs predict decisions, choices, and behavior in realistic settings. Such findings convert esoteric mind science into a real-world problem.
If ISCs based on race predict worse treatment in the real world, then we have identified a new stream of “race discrimination” even when defined narrowly. Of course, it is less offensive than the kind of racism embraced by racial supremacists. But the fact that Bull Connor and his dogs are so much worse does not mean that race discrimination caused by ISCs is necessarily de minimis. If nothing else, we should be more skeptical about easy assurances that today’s racial disparities are caused only by objective differences in “merit” across racial groups.
To respond thoughtfully to the problem of racial discrimination, we need less opinion and more data. As our contribution, we study the link between ISCs and behavior within the legal domain, about an understudied minority group. Specifically, we ask: When individuals imagine the ideal litigator, does a White man (as compared to an Asian American man) come to mind? More important, do such implicit stereotypes influence evaluation of the litigator?
I. Psychological Theory. Part I provides a brief introduction to Implicit Social Cognitions, and how they might be measured through reaction time instruments such as the Implicit Association Test (IAT). In addition, we describe Alice Eagly’s Role Congruity Theory, which explains how the perceived “lack of fit” between professional roles and social groups can undermine professional success, and extend her theory to race discrimination.
II. The Experiment. Part II describes our study and reports our results. We gathered 68 participants from a non-student, community pool.
IATs. They completed two computerized Implicit Association Tests (IATs) that assessed their (1) implicit stereotypes linking the ideal litigator with particular racial groups and (2) implicit racial attitudes toward Asians relative to Whites.
Depositions. Participants were then told that they would hear two depositions from two unrelated cases. At the beginning of each deposition, participants were shown for five seconds a picture of the litigator on a computer screen accompanied by his name. We manipulated the race of the litigator by varying his name and photograph to be prototypically White (“William Cole”) or Asian (“Sung Chang”). Participants then listened to the deposition through headphones and, at the same time, read the script of the deposition presented on a computer screen. The transcript identified who was speaking, which meant that participants saw labels such as “Attorney Cole” or “Attorney Chang”.
At the end of the deposition, participants were asked to evaluate the litigator’s competence (6 items), warmth (6 items), and their willingness to hire him or recommend him to family and friends (2 items). Next, participants saw a picture of the second litigator, then listened to the second deposition and evaluated the second litigator on the same dimensions.
Explicit measures. Finally, we measured the degree to which participants personally endorsed the stereotype linking ideal litigators’ personality to race by asking them to judge how well each of the five litigator traits described Asian Americans as a group and White Americans as a group. In addition to measuring personal stereotypes, we asked about societal stereotypes—what “most Americans” believed.
Results. We found that both explicit and implicit stereotypes of ideal litigators as more White than Asian predicted more favorable evaluations of the White litigator over the Asian American one. Focusing on the implicit bias findings, we discovered:
- the existence of an implicit attitude favoring Whites over Asian Americans (IAT D = 0.62);
- the existence of an implicit stereotype associating “litigator” with “White” (IAT D = 0.45);
- correlations between these implicit stereotypes with higher evaluations of the White lawyer (ingroup favoritism) (competence r = 0.32; likeability r = 0.31; willingness to hire r = 0.26). These results were confirmed through hierarchical regressions.
III. Law and Policy Implications. Part III briefly explores policy implications of these findings and responds to various objections.
First, we have added to the growing body of evidence that we are not as colorblind as we think. This goes for even a “model minority.”
Second, we have contributed to the predictive validity literature, by showing that implicit biases predict behavior. We add that our participants were not undergraduates, but jury-eligible adult residents drawn from the community. Also, the dependent variables were not only written vignettes. Instead, there was a multimedia component, in which participants actually listened to a full five-minute long deposition, which provided a richer set of materials for participants to interpret and judge.
Since correlations are often too abstract, we provide concrete examples of the effect sizes in the paper. Consider the following comparison between two hypothetical participants, “James” who is implicitly colorblind (IAT D score of zero) and “Greg” who has an IAT D score of 1. According to our hierarchical regressions, Greg would likely evaluate Attorney Cole, the White lawyer, very favorably as 6.01 on the 1-7 scale on competence, 5.57 on likeability, and 5.65 in terms of recommending his services to friends and family. By contrast, James would evaluate Attorney Cole as 5.06 in terms of competence, 4.36 in terms of likeability, and 4.30 in terms of recommending his services to others.
Table 4: Differences in White Litigator Evaluations by Hypothetical James v. Greg
| White Litigator Evaluation | James (IAT D = 0) |
Greg (IAT D = 1) |
Difference |
| Competence | 5.06 | 6.01 | 0.95 |
| Likeability | 4.36 | 5.57 | 1.21 |
| Hireability (all on 1-7 scale) |
4.30 | 5.65 | 1.35 |
Conclusion. People who decry the play of the “race card” believe that we already compete in something like a meritocratic tournament, in which individuals are evaluated based on their performance only. Differences in evaluation are presumed to come only from differences in actual merit, which is independent of social categories, such as race. If this is what’s going on, then most claims of racial discrimination can be seen as self-interested whining by those who lost in a fair game.
But do we really live in such a world? Or in less Manichean terms, how much does race continue to influence our merit evaluations? And by this, we don’t mean to go back decades in an individual’s life to trace how race might have affected her trajectory of human capital development. We mean, instead: “Does race influence merit evaluations right now, when the performance is objectively indistinguishable?”
Our study demonstrates that explicit and implicit stereotypes about litigators and Whiteness alter how we evaluate identical lawyering, simply because of the race of the litigator. The race was only primed by a five-second picture and the last name of the lawyer shown on the transcript. Nonetheless, race was sufficiently salient as to predict different evaluations of the litigator’s deposition. Explicit and implicit bias produced net racial discrimination against a “model” minority either by elevating Whites or by putting Asians down.
Many folks resent “affirmative action” programs and instead say that everyone should be colorblind. Appeals to an only partially redeemed history are rejected. It’s as if some statute-of-limitations has passed on claims of justice for past wrongs. But if we are sincere and accurate about our own colorblindness, then the race of the litigator should not cause one iota of difference in how we evaluate a garden-variety deposition. But our data show otherwise—that race still does matter. We need more evidence on how and why; more important, we need to start studying what we might do about it.
- draft text @ SSRN (pdf) :: please e-mail me for a final e-copy (per copyright restrictions)
- Videostream of talk at Conference of Empirical Legal Studies Nov. 2009
- I’m the 2nd speaker; starting at 38 minutes
- (since I used Mindmanager to present, the screen is not shown–instead it’s the last slide used by the prior speaker).