Mimesis Law
22 January 2017

Dueling Judges: Jurors Should Be Free From “Judicial Immunization” For “Implicit Bias”

November 16, 2016 (Fault Lines) — Ed. Note: In light of the concern about the impact of racial bias in the criminal justice system, we have “so ordered” Senior Judge Richard G. Kopf of the District of Nebraska and Senior Judge Mark W. Bennett of the Northern District of Iowa to debate: “Should judges address the issue of ‘implicit bias’ with jurors?” This is Judge Kopf’s argument:

Perhaps the “biggest thing” in social psychology for the last 10 years or so has been research related to “implicit bias.”* That is, (1) testing the thesis that we all have unrecognized or subconscious biases against other people, and (2) if you are clever enough, you can discern the particular bias and then ameliorate it.**

Judge Mark Bennett, my dear friend and colleague, who is also a highly respected scholar, is a true believer in the concept of “implicit bias.” He strongly advocates for active judicial intervention when dealing with jurors so that jurors become aware of their implicit biases and guard against them. See, e.g., Mark W. Bennett, Unraveling the Gordian Knot of Implicit Bias in Jury Selection: The Problems of Judge- Dominated Voir Dire, the Failed Promise of Batson, and Proposed Solutions, Harvard Law & Policy Review, Vol. 4, p. 149 (2010). I respectfully, but strongly, disagree with Judge Bennett. Trial judges should mostly stay out of the “implicit bias” business.

Before addressing my objections, I need to do two things. I should (1) very generally describe the testing method that scholars have used to determine the existence of “implicit bias,” and (2) then describe how Judge Bennett inoculates jurors–in criminal cases–from that “disease.”*** Briefly, very briefly, I next provide some context about these points–after all, this is a blog post!

The academic concept of implicit bias” flows from the Implicit Association Test (IAT) geared to race.**** Here is a description of the race test:

The Implicit Association Test compares millisecond reaction-time differences in a test taker’s responses to varying combinations of stimuli. If the test taker reacts more quickly to some groups of stimuli than to others (for example, pictures of white faces paired with pleasant words versus pictures of black faces paired with the same pleasant words), then the test taker is assumed to have stronger associations with those stimuli or the stimuli are said to be more congruent with the test taker’s unconscious attitudes (for example, a test taker who reacts more quickly to white faces paired with pleasant words would be said to be unconsciously biased in favor of whites and against blacks).*****


Philip E. Tetlock, Gregory Mitchell, and L. Jason Anastasopoulos, Detecting and Punishing Unconscious Bias, The Journal of Legal Studies, University of Chicago (January 2013).

Judge Bennett is convinced that white people, who comprise the great bulk of his jurors and mine, hold unconscious biases against people of color. If you believe the results from administration of the IAT, he is on solid ground. See, e.g., Chris Mooney, Wonk Blog, Across America, whites are biased and they don’t even know it, Washington Post (December 8, 2014).

Judge Bennett is not shy about confronting prospective white jurors.  For example, he may show a video produced by the Urban Alliance on Race Relations. The video shows a black man, while the screen runs a criminal history sheet, and then the video reveals that the man pictured in the video was the arresting officer. See here.

The judge will then engage in his own voir dire that is explicitly designed to be impactful to the jury about race and the video the jury has just seen. After that, he gets very specific about his own concerns. Here is an abbreviated sample of his voir dire script:

After watching the “never judge” Michael Conrad video where I stop it at the end of the “rap sheet” before the rest of the video is played: What do you think of the person pictured in the video? Why? After playing the rest of the video: Were you surprised to find out the image if the person in the video was the arresting officer? [I ask this of jurors whose facial expressions demonstrate obvious surprise] Why? When you think of the term arresting officer what race of an officer to you think of? Why? Can you see how gut feelings can lead to false impressions? Can you see how stereotypes can lead to false impressions? Can you see how generalizations can lead to false impressions? Is it important to keep an open mind until you have heard and seen all of the evidence? Why?

. . .

I want each of you to know that I am worried whether Mr. Rodriguez, as the only Hispanic in the courtroom, is able to receive the fairest trial possible

-what can you tell us to assure that this will happen? Do you have any concerns about Mr. Rodriguez’s ability to receive a fair trial?

Perhaps the most unusual aspect of the judge’s approach to “implicit bias” is his requirement that each juror certify the following on the verdict form:


By signing below, each juror certifies the following:

(1) that consideration of the defendant’s race, color, religious beliefs, national origin, or sex was not involved in reaching the juror’s individual decision, and

(2) that the individual juror would have returned the same verdict for or against the defendant on the charged offense regardless of the defendant’s race, color, religious beliefs, national origin, or sex. [Bold in original]

I now turn to my criticisms. To keep this post to a readable length, I will limit myself to just a few of my objections. And, I intend to be brief.

First, the premise of Mark’s efforts, while obviously well-intentioned, is based upon questionable science. The IAT has been heavily criticized by serious scholars:

Our primary concern is, however, with the way a small number of studies are being used to make strong claims in applied settings, including courtrooms (see Feuss & Sosna, 2007).  Although one might quibble over a particular reanalysis or the implications of a specific outlier, the broad picture that emerges from our reanalysis is that the published results are likely to be conditional and fragile and do not permit broad conclusions about the prevalence of discriminatory tendencies in American society.  Given the paucity of studies showing strong links between IAT scores and behavior, given our inability to gain access to published data sets, and given the weakness of the data that we did obtain, psychologists and legal scholars do not have evidentiary warrant to claim that the race IAT can accurately or reliably diagnose anyone’s likelihood of engaging in discriminatory behavior, less still that there is substantial evidence of such linkages (contra Greenwald & Krieger, 2006). [Underlines by Kopf.]

Hart Blanton, Jonathan Klick, Gregory Mitchell,  James Jaccard, Barbara Mellers, Barbara Mellers, Philip E. Tetlock, Strong Claims and Weak Evidence: Reassessing the Predictive Validity of the IAT, Journal of Applied Psychology (2009). See also Oswald FL, Mitchell G, Blanton H, Jaccard J, Tetlock PE, Predicting ethnic and racial discrimination: a meta-analysis of IAT criterion studies, J Pers Soc Psychol (August, 2013) (“IATs were poor predictors of every criterion category other than brain activity, and the IATs performed no better than simple explicit measures. These results have important implications for the construct validity of IATs, for competing theories of prejudice and attitude-behavior relations, and for measuring and modeling prejudice and discrimination.”).

Second, I have no doubt that evolution has taught us to discriminate between friend and foe based upon instantly assessing otherwise innocent cues (skin color for example******). But there is very little evidence for trials, which are inherently methodical, tedious and relatively long, that those ancient shortcuts motivate decisions based upon hours or days of deliberation by real jurors based upon jury instructions dedicated to prodding the rational mind.  That is, if we look at actual behaviors such as serving as an actual juror in a real case, there is precious little reason to believe that “implied  bias” is at work.

Third, who is to say when an “implied bias” warrants overt actions like those taken by Judge Bennett?  According to one study of the results of the administration of the  IAT, both black women and white women are “biased” against fat people. See here. Should fat defendants get special handling? This “implicit bias” business puts judges into a swamp from which judges may never get out once they walk into that primordial muck.

Let’s take, for example, a counterterrorism case. Research has shown that “liberals supported use of the technology to detect unconscious racism but not unconscious anti-Americanism, whereas conservatives showed the reverse pattern . . .”. Detecting and Punishing Unconscious Bias, supra, at 1. In our hypothetical counterterrorism case, shouldn’t the intellectually honest but active “implicit bias” judge demand a certification from jurors that their verdict of acquittal was not driven by a dislike of the government?

Finally, I am completely creeped out by a United States District Judge consciously manipulating the supposedly unconscious minds of jurors. I am truly frightened by such authoritarianism dressed up in the guise of justice. But, I concede that may just be me. After all, I took the race IAT and here is what the computer told me: “Your data suggest a moderate automatic preference for White people over Black people.”


* In this context, the word “bias” properly understood is not a “bias” at all. Social psychologists who care about being precise prefer to describe what they are researching as an “association,” which is “implicit” meaning a subconscious association. Thus, I prefer “implicit association” rather than “implicit bias.” The distinction is important both as an academic matter and a policy matter. However, for this post, I will not quibble with wording.

** For an in-depth introduction, see Teige-Mocigemba, SKlauer, KC Sherman, JW, A practical guide to Implicit Association Tests and related tasks, University of California Davis (June 1, 2016).

*** Judge Bennett was kind enough to share with me some of his practices, procedures and forms. I trust Judge Bennett will correct me if my description is too far off base.

****You can take the various (see below) tests here.

*****There are numerous tests for different subject matter areas: Gender – Science. This IAT often reveals a relative link between liberal arts and females and between science and males; Race IAT; Race (‘Black – White’ IAT) This IAT requires the ability to distinguish faces of European and African origin. It indicates that most Americans have an automatic preference for white over black; Skin-tone IAT; Skin-tone (‘Light Skin – Dark Skin’ IAT) This IAT requires the ability to recognize light and dark-skinned faces. It often reveals an automatic preference for light-skin relative to dark-skin; Age IAT; Age (‘Young – Old’ IAT). This IAT requires the ability to distinguish old from young faces. This test often indicates that Americans have automatic preference for young over old; Weight IAT Weight (‘Fat – Thin’ IAT). This IAT requires the ability to distinguish faces of people who are obese and people who are thin. It often reveals an automatic preference for thin people relative to fat people; Presidents IAT Presidents (‘Presidential Popularity’ IAT). This IAT requires the ability to recognize photos of Barack Obama and one or more previous presidents; Arab-Muslim IAT Arab-Muslim (‘Arab Muslim – Other People’ IAT). This IAT requires the ability to distinguish names that are likely to belong to Arab-Muslims versus people of other nationalities or religions; Disability IAT Disability (‘Disabled – Abled’ IAT). This IAT requires the ability to recognize symbols representing abled and disabled individuals. Asian IAT Asian American (‘Asian – European American’ IAT). This IAT requires the ability to recognize White and Asian-American faces, and images of places that are either American or Foreign in origin; Weapons IAT Weapons (‘Weapons – Harmless Objects’ IAT). This IAT requires the ability to recognize White and Black faces, and images of weapons or harmless objects; Sexuality IAT Sexuality (‘Gay – Straight’ IAT). This IAT requires the ability to distinguish words and symbols representing gay and straight people. It often reveals an automatic preference for straight relative to gay people; Gender-Career IAT Gender – Career. This IAT often reveals a relative link between family and females and between career and males; Native IAT Native American (‘Native – White American’ IAT). This IAT requires the ability to recognize White and Native American faces in either classic or modern dress, and the names of places that are either American or Foreign in origin.

****** I agree with Judge Bennett that we all are “hard-wired” to instantaneously distinguish between “us” and “them.” My hunch is that this truism is an artifact of evolution. See also Karin Brulliard, Eek, a snake! Humans may be hard-wired to spot serpents — and fast, Washington Post (November 10, 2016). But my hunch is also that humans are “hard wired” to be empathetic. See, e.g., Line Goguen-Hughes, Survival of the Kindest, Mindful (December 23, 2010) (“In 1871, eleven years before his death, Charles Darwin published what has been called his ‘greatest unread book, ‘The Descent of Man and Selection in Relation to Sex’. . .. In the fourth chapter, entitled ‘Comparison of the Mental Powers of Man and the Lower Animals,’ Darwin explained the origin of what he called ‘sympathy” [which today would be termed empathy, altruism, or compassion], describing how humans and other animals come to the aid of others in distress.”).

15 Comments on this post.

Leave a Reply



  • MOK
    16 November 2016 at 9:31 am - Reply

    Having not given this as much thought as I should, my cynical initial impression of Judge Bennett’s technique of juror “indoctrination” (yes, that is what it is), is that it is an insult to their intelligence. At present, I think I would refuse to sign that silly certificate. Would he hold me in contempt?

    Hell, let’s also make sure we don’t have any hidden Huskers or Hawkeyes on those juries in Sioux City!

  • Bernie King
    16 November 2016 at 9:50 am - Reply

    How far would meddling against implicit bias go? Why stop at defendants? Should judges actively try to neutralize implicit bias against black witnesses? Black police officers? Black lawyers,?

  • CLS
    16 November 2016 at 9:58 am - Reply

    Judge Kopf:

    Thank you and Judge Bennett for slugging this issue out for our readers.
    I think this is the best Fault Lines debate yet.

  • Anon
    16 November 2016 at 11:56 am - Reply
    • Richard G. Kopf
      16 November 2016 at 12:50 pm - Reply


      I look like the kid, but I sure as hell have none of his talents.

      You made me smile. I loved the movie and this scene in particular.

      You must be old, but that’s OK. Burt Reynolds still lives!

      All the best.


  • Mark Bennett
    16 November 2016 at 3:34 pm - Reply

    BK, of course in making credibility determinations of witnesses when we are called upon to do so we should be aware of our implicit biases. And MOC, great idea I am working on a Husker/Hawkeye/Cyclone IAT. 😃

    • MOK
      17 November 2016 at 9:19 am - Reply

      I plead guilty for forgetting the Cyclones. Dare I say, better include the Panthers, too? They are in the Northern District. 🙂

  • Brian
    16 November 2016 at 6:03 pm - Reply

    Judge, I’d like to take the liberty of posing at least partial answers to some of your questions/concerns.

    First, however, I feel the need to note that, at the same time you begin to state your concerns, you switch from calling it “implicit bias” to “implied bias”. Since both terms are nearly synonymous, normally this would be a nonissue, but in this case it may be of interest, even if the difference is largely academic. If something is implied it is *expressed* indirectly, but if something is implicit it has been *accepted* indirectly. Thus you have, indirectly (heh), answered many of your own questions. Remember that the whole point of voir dire is to remove jurors with potentially troublesome biases.

    “That is, if we look at actual behaviors such as serving as an actual juror in a real case, there is precious little reason to believe that “implied bias” is at work.”

    I believe you may have been oversimplifying things when you wrote this. Jurors are supposed to look at the whole of the case and decide based on everything they hear. A problem arises when the jury is told of something that isn’t admissible in court, because human beings are incapable of completely un-seeing or un-hearing something once it’s been called to their attention. The opposite, however, is also possible – a highly prejudiced juror may discount a witness’ entire testimony (while it is being made) for one reason or another, and then when evidence to support that testimony arrives it makes no impact because the testimony is (in their mind) already false. This often happens to those involved in cults and conspiracy theories, but isn’t limited to them.

    “Third, who is to say when an “implied bias” warrants overt actions like those taken by Judge Bennett?”

    I propose the following:
    1. The bias must be prevalent enough in our society at large that it is likely at least half the jurors would have it.
    2. The bias must be implicit.
    3. The bias must be one that is difficult for mere questioning to expose.
    4. The bias must directly* prejudice jurors against the defendant’s case. The prosecution can take care of themselves.
    * One thing you did not address is the specific nature of the bias. See below.

    “Should fat defendants get special handling?”

    Usually no. An implicit dislike of fat people doesn’t necessarily translate to ignoring their testimony or feeling they should rot in jail. However, an implicit belief that black people are criminals, or that cops (when arguing for the prosecution) are never corrupt, usually does. We see those types of cases all the time here on Fault Lines.

    “In our hypothetical counterterrorism case, shouldn’t the intellectually honest but active “implicit bias” judge demand a certification from jurors that their verdict of acquittal was not driven by a dislike of the government?”

    Hell no. For starters, jury nullification (which is what you suggest might be happening in that case) isn’t very common. In addition, dislike of the government is rarely an implicit bias on its own – generally we are more tolerant, if not outright supportive, of those in power. Usually it is caused by something more than just “they look different”, and that something can be found by attorneys. In this case, you are applying a solution in a setting where the problem doesn’t exist. Most importantly, however, the prosecution can take care of themselves in this regard. The defendant is supposed to be guaranteed the right to a fair trial. Not the government.

    Ultimately, I think Judge Bennett has the right of it. Anyone with a relevant obvious bias is going to be excluded during voir dire already. What he’s trying to do is take the others, shake up their preconceptions a little, make them consciously confront themselves, and see what falls out. Judge Bennett’s given anecdote (paragraph 20) is likely representative of what actually ends up happening once you require a reasonable person to reconcile unconscious attitudes with a desire to do the right thing. To be honest, I don’t see what’s so terribly authoritarian about that when the same person is about to be given the power to decide someone’s fate for the forseeable future.

    • Richard Kopf
      17 November 2016 at 7:31 am - Reply


      Thank you for your thoughtful comments. Several quick points in response:

      1. You don’t address the questionable scientific basis for applying “implicit bias” theory in the courtroom. By the way, “it can’t hurt” is not an adequate answer because it just might hurt. That’s why doctors are so conservative when treating people–first do harm they are taught. We should never forget that caution.

      2. Your list about when to address or not address implicit biases–while sounding neutral–is merely an enumeration of your value judgments about what unconscious biases should and should not be probed by a trial judge. And, the notion that the “prosecution can take care of itself” is far too flip.

      3. Don’t like the counterterrorism example, substitute a white cop charged with shooting a black motorist and the jury panel includes several black jurors? Shouldn’t the white cop get the benefit of a certification that the guilty verdict was not premised upon an implicit bias against white cops?

      4. If you don’t see authoritarianism in a government official (in this case a judge) shaping or attempting to shape the unconscious mind of an innocent citizen (prospective juror) who has been compelled by law to appear before that official, then you and I must respectfully agree to disagree.

      All the best.


    • Richard Kopf
      17 November 2016 at 7:35 am - Reply


      Thanks for pointing out my misuse of “implied” before “bias” at one point in my post. That was my “unconscious” error.I always meant to use the word “implicit.” I apologize.

      All the best.


  • Peter Gerdes
    16 November 2016 at 10:12 pm - Reply

    The Certification is a bit troubling as in many cases it will literally be false. If skin color is a critical detail in a witness report obviously the jury can’t certify they would have returned the same verdict if the defendant had been a different color or that consideration of color was not involved.

    Even if we try and read this requirement charitably it is still disturbing as the requirement either collapses into triviality (don’t use race/color/etc.. when it isn’t an appropriate grounds to reach a decision) or risks barring the use of accurate beliefs about these categories.

    For instance can one consider particular religious beliefs as a possible motive or lack thereof? Can one use the paucity of white people in the area to estimate the likelihood the white defendant is the white suspect? Surely that must be ok. But if that is ok can I use facts about the relative frequency of Quakers carrying guns? What about the relative frequency of whites and blacks carrying guns?

    We only want to exclude unjustified use of these factors but as your recidivism posts bring out people disagree violently about what that is.

  • traderprofit
    16 November 2016 at 11:03 pm - Reply

    You are putting too much thinking into this, and definitely too much writing. It’s making my head hurt.
    Please just stop it.
    This isn’t as difficult as you judges seem to make it.

  • Mark Bennett
    17 November 2016 at 5:49 am - Reply

    Peter, you raise some interesting points, thank you. I have always been open the the fact that the statutory language of the certification may not be appropriate under certain fact patterns- easy to postulate but none have occurred in any cases I have tried. Because I instruct before opening statements in all civil and criminal cases, and the certification is part of the verdict form at the conclusion of the instructions, and I get that to the lawyers usually 7-10 days before trial, there is ample opportunity to raise issues relating to the certification well in advance of trial. No lawyer has ever done so but I would welcome them to.

  • Mark Bennett
    17 November 2016 at 8:58 am - Reply

    Rich, I am no more shaping jurors unconscious minds than you do in jury selection I am simply informing them of how implicit bias works and that it may effect decision making and asking them to focus on facts not gut instincts or stereotypes. Labeling that authoritarianism is authoritarianism. 😄

    • Anon
      17 November 2016 at 3:44 pm - Reply

      I don’t think informing jurors on how implicit bias works is “authoritarian,” but it does come off somewhat condescending. Most people in the privacy of their own minds are aware of their own biases. My experience with jurors informs me that most people when sworn in rise to the occasion. That is, I think they do their honest best to evaluate the case on the merits, and make a sincere effort to check their prejudices. To the extent that someone is unable to do so, I don’t think any instruction a judge can give will assist them any better. It strikes me that it may have the opposite effect. I think I might be a little indignant if I had to sign a form saying that I would agree not to be prejudiced in making my decision as a juror. Again, it comes off in a condescending way that could irritate jurors. As a CDL, I wouldn’t want the instruction, or the signed certificate for that reason.