Why implicit bias is not (just) in the head, and why that matters
Updated: Dec 5, 2022
Meet Professor P. Professor P seems to be the prototypical egalitarian: she is engaged in anti-racist and feminist activities, explicitly avows egalitarian values, and strives for fair treatment. At the same time, however, her evaluations of Black students’ work are slightly less glowing compared to her evaluations of white students’ assignments, even though they are of the same quality, she displays greater tension and discomfort when interacting with Black students, and is less warm and less patient with them as a result.
The generally accepted explanation for this disparity is so-called implicit attitudes. Even though many of us explicitly avow egalitarianism, we still harbour harmful associations or beliefs towards members of certain social groups. We are not fully aware of having these implicit attitudes and/or the influence they have on our behaviours and decisions. Consequently, one of the main goals in the literature has been to identify the nature of these implicit psychological states. Are they associations, unconscious beliefs, or do we need to introduce a new kind of psychological state in order to explain this phenomenon? As soon as we know the nature of these implicit attitudes, the idea seems to be, we will know how to overcome and limit the harm they currently cause.
Whether this is the right strategy, to focus on the psychological state that supposedly causally explains implicitly biased behaviour, is not a central topic of discussion. This, in my view, is problematic. Psychological research does not measure these implicit psychological states directly, it measures the behaviour. This is clear from the example Professor P: she judges and treats the Black students less favourably, and on the basis of her biased behaviour and the fact that she is not aware of it, an implicit attitude with certain content is ascribed to her. Ascribing the implicit attitude is the next step, after we have settled that there is something problematic about her behaviour. This also applies to the well-known Implicit Association Test (IAT). The IAT, which is supposed to be a measure of the extent to which a person harbours a certain implicit attitude, also measures behaviour: the question is how fast people respond when Black faces and positive words share a key in comparison to when white faces and positive words share a key.
Of course, an important goal of psychological research is to explain how it is possible that people are capable of doing something that does not conform to their explicit beliefs, without being aware of it. Apparently, is the thought, there are things going on in our minds that we are not aware of, otherwise we would not behave as we do. But even if you accept this way of thinking about our psychology, we need to be clear that the goal is to provide a causal explanation of the behaviour, which does not necessarily imply a conceptual relationship as well. Yet, a conceptual relationship is often sneaked in: the assumption is that the implicit attitude not only causes the behaviour, but also characterises the behaviour as implicit and biased. Professor P's behaviour has an implicitly racist nature because it is caused by a racist implicit attitude. But this line of thinking is flawed: in reality, we observe Professor P's biased behaviour, realise that she is unaware of it, and subsequently ascribe an implicit attitude to her. And if that is how we actually proceed, the implicit attitude receives its character from the peculiar behaviour we take it to cause. As a result, we need to pay attention to the implicitly biased behaviour itself. How is it possible that we behave racially without consciously wanting to do so? And on the basis of what do we ascribe an implicit attitude to a person?
A notion that is useful here is ‘acting under a description’. Every time we act, several descriptions of what we are doing can be given. The reason for this is that we can pay attention to different layers and aspects of a person’s behaviour. Imagine a person walking to the supermarket, cooking dinner, or cycling to work. Under these descriptions, the person acts intentionally and consciously. At the same time, however, they do other things unconsciously and in the absence of an intention to do so. The person might step on an ant and kill it, scratch the cutting board while cutting onion, or block someone’s way while taking a left turn. These things they do unintentionally and unconsciously. Such unintentional action is unavoidable. Observers can always come up with descriptions of what we are doing that we did not intend to do, and often we are unaware of acting unintentionally under these descriptions. Researchers could measure the distance the person walks to the supermarket for example, calculate how many pieces of onion they end up with, or count how many left turns the person took while cycling to work. Even though these are not intentional actions, in some sense, the person still does them.
This distinction is useful to capture how we can act and behave in a way that is unfavourable to members of certain social groups, while still being under the impression that we are committed egalitarians. Professor P performs egalitarian intentional actions: she participates in feminist and anti-racist activities for example. The racist things she does do, for example evaluate the papers of Black students less positively, are unintentional actions of hers. In her mind, she is simply critically evaluating the papers and deciding which grade to give to her students. At the same time, if we take all the grades into account, a problematic pattern emerges: equally good papers of white students turn out to receive a slightly higher grade than those of Black students. Professor P is not aware of this pattern, and that is not surprising: it is difficult to recognize patterns in your behaviour.
This shows that a crucial aspect of bias, that it is about treating a person less favourably than another, in virtue of their membership of a certain social group, contributes to it being implicit. Bias is about making distinctions between members of different social groups. Mostly, we do not intentionally differentiate as such – we just grade a paper, sit down somewhere in the waiting room, or choose a candidate as the next police chief. We are not aware that if we would take our actions (and those of others) together, we end up with a biased pattern. Professor P’s racist behaviour is implicit, then, because she does not recognize the racist pattern that grows out of her, individual and intentional, paper evaluations. Similarly, she does not realize a problematic pattern in how warm she is to the different students, and that her discomfort is related to the race of the respective student.
Of course, we can discover psychological explanations for this pattern, although I suspect that here, too, it would be good to pay more attention to the circumstances in which a person decides. We should reluctant to simply attribute implicit attitudes to a person. But my main concern is this: it follows that whether someone is implicitly biased is not easy to determine. And that increases the harm for the members of marginalized groups.
Imagine being one of Professor P’s Black students. You have the impression that you understand the topic, have offered interesting contributions to the discussions, and that your paper is insightful, original, and the argumentation is clear and comprehensive. Yet, your grade is the same as your white classmate’s, even though they did not prepare the classes very well and made quite obvious points in the paper, with arguments that were already extensively discussed during the classes. You know that implicit biases exist, and you have read about studies that show that Black students are often treated unfavourably, even though teachers do not intend to do so. You also realize, however, that experimental findings cannot be generalized to all cases. You know that, even if studies show a racist bias, not all subjects, even those that gave a bad grade to a paper of a Black student, are necessarily implicitly biased. After all, sometimes Black students write bad papers. Only in the experimental setting, wherein the circumstances are completely the same, it can be concluded with absolute certainty that Black students are treated differently because of their race. As a result, you do not feel you have sufficient ground to criticize Professor P’s practices. Perhaps you were mistaken, and your paper was really not as good as you thought it was. Perhaps the discomfort you and your Black fellow students feel is just in your heads.
In other words, the case of Professor P. suggests that implicit bias can be identified easily, but in real life things are not that simple. For victims of implicit bias, it is typically difficult to know that they were treated unfavourably, and this places extra burdens on them to try interpret situations and maintain their confidence in face of these experiences. In reality, assignments, interviews, and other interactions are never exactly the same, and it isn’t entirely clear where the difference in assessment and treatment comes from. We cannot simply settle the matter by attributing an implicit attitude to Professor P and get things over with. Whether we have a case of implicit bias is, unlike in experimental research, not clear-cut. Importantly, even though these patterns are often easiest to identify by persons being a regular target of biased behaviour, i.e., members of a marginalized social groups, they also have a harder time being taken seriously by others and themselves. As a result, implicit bias is even more problematic than is sometimes acknowledged. These persons are hit with a double whammy: not only are they treated unfavourably, they also have to go through the struggle of making sense of this experience and communicating it to others. If we want to be truly egalitarian, then, we must at least ensure that talking about implicit prejudice becomes a path of less resistance.
Dovidio, J. F., Gaertner, S. E., Kawakami, K., & Hodson, G. (2002). Why can't we just get
along? Interpersonal biases and interracial distrust. Cultural Diversity and Ethnic
Minority Psychology, 8(2), 88. https://doi.org/10.1037/1099-9809.8.2.88
Fricker, M. (2007). Epistemic injustice: Power and the ethics of knowing. Oxford University
Glover, D., Pallais, A., & Pariente, W. (2017). Discrimination as a self-fulfilling prophecy:
Evidence from French grocery stores. The Quarterly Journal of Economics, 132(3),
Greenwald, A.G., McGhee, D. E., & Schwartz, J.L. (1998). Measuring individual
differences in implicit cognition: the implicit association test. Journal of Personality and Social Psychology. https://doi.org/10.1037/0022-35220.127.116.114
Holroyd, J., & Puddifoot, K. (2020). Epistemic Injustice and Implicit Bias. In E. Beeghly and
A. Madva (Eds. An Introduction to Implicit Bias: Knowledge, Justice, and the Social
Mind (pp. 116-133). Routledge.
Holroyd, J., Scaife, R. & Stafford, T. (2017). Responsibility for implicit bias. Philosophical
Compass, 12, e12410. https://doi.org/10.1111/phc3.12410.
Uhlmann, E.L. & Cohen, G.L. (2005). Constructed criteria: Redefining merit to justify
discrimination. Psychological Science.
Wood, M., Hales, J., Purdon, S., Sejersen, T., & Hayllar, O. (2009). A test for racial
discrimination in recruitment practice in British cities: research report no 607.
London: Department for Work and Pensions.
Photo by Matheus Viana