Abstract
Recent years have seen a surge of interest in implicit bias. Driving this concern is the thesis, apparently established by tests such as the IAT, that people who hold egalitarian explicit attitudes and beliefs are often influenced by implicit mental processes that operate independently from, and are largely insensitive to, their explicit attitudes. We argue that implicit bias testing in social and empirical psychology does not, and without a fundamental shift in focus could not, establish this startling thesis. We suggest that implicit bias research has been conducted in light of inadequate theories of racism and sexism. As a result, such testing has not sufficiently controlled for subjects’ prejudiced explicit beliefs and emotions, and has not ruled out the possibility that explicit prejudice best explains test subjects’ discriminatory associations and behavior.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
For example, see Oswald et al. (2013), Machery (2017, 2021), and Haslanger (2015) for representatives of each criticism, respectively. Prominent examples in popular press include Singal (2017) and Bartlett (2017). Brownstein, Madva and Gawronski (2020) respond to various criticisms within empirical psychology, the popular press, and philosophy. They respond to an online discussion of ours, but we believe that they misconstrue our argument. We will clarify our argument here.
One might classify “implicit attitudes” in various ways, but these terminological differences will not affect our basic argument. One might, for example, hold that implicit attitudes are just those attitudes that subjects exhibit in implicit attitude testing. On that understanding of “implicit,” our argument could be reformulated as follows: Empirical research on implicit bias has not shown that implicit biases, in this sense, are not straightforward expressions of subjects’ explicit biases. Similarly, if one held that implicit attitudes are just those attitudes that must be measured by implicit attitude tests and could not, in principle, be verbally reported, then our argument could be restated as follows: Empirical research on implicit bias has not demonstrated the existence of biased implicit attitudes because it has failed to provide evidence that the attitudes subjects express in implicit bias testing could not, in principle, be verbally reported. In any case, our fundamental thesis is the same: Empirical research on implicit bias has not demonstrated that subjects’ rapid associations and behavior are not straightforward expressions of their explicit attitudes.
See, for example, (Greenwald et al., 1998, p. 1475).
These questions are taken directly from the IAT as administered by the Project Implicit website, but they are also included in the academic studies that describe the explicit measures used. See, for example, (Sabin et al., 2009).
Some measures of implicit bias, such as the ALPS (Lebrecht, 2009), were not constructed using any explicit measures at all, presumably on the assumption that implicit attitudes can be measured directly and without controlling for explicit attitudes.
In their 1998 paper introducing the IAT, Greenwald et al. employed a more sophisticated measure of explicit bias that we will discuss in the next section. This control was abandoned in many future uses of the IAT, including the version accessible via the Project Implicit Virtual Laboratory, which Greenwald co-founded.
Of course, this assumes that one knows which stereotypes to ask about. Such knowledge would require a good theory of prejudice and insight into the sociopolitical dynamics of the local context.
We include the full content of these measures in the “Appendix”. Where full content was unavailable, we’ve included the “Representative Items” offered by the original authors.
Brownstein et al. (2020), responding to a blog post based on an unpublished version of this manuscript, characterize our skepticism as grounded in the worry that implicit measures are “poor predictors” of prejudiced behavior, and that explicit measures may serve as better predictors. This is a misunderstanding of our point. We are not arguing that implicit bias tests fail to predict behavior. Rather, our point is that empirical research has not shown that the behavior implicit bias tests measure, predict, or explain is at odds with subjects’ explicit attitudes. Our argument does not rest on a doubt that tests such as the IAT measure subjects’ current rapid associations, or that, as Brownstein et al. put it, implicit measures such as the IAT “reflect what is going on in a person's mind in a given moment, which is shaped by complex interactions of person-related and situation-related factors (296).” Again, what we doubt is that what is measured by the IAT is best understood as discrepant with a subject’s explicit attitudes. The IAT may indeed reflect what is going on in a person’s mind at a given moment, but what is going on in her mind at that moment may be an operation of her explicit prejudice.
From “The Modern Racism Scale”; this question is a sample from Greenwald (1998).
We are grateful to an anonymous reviewer for raising this suggestion.
Thus, though we are sympathetic with Eduoard Machery’s (2017) suggestion that implicit bias research has attempted to construct theories “on quicksand,” we nonetheless think that our particular criticism may be addressable with the right kind of philosophical intervention. Machery (2016) argues that the distinction between “implicit” and “explicit” attitudes is meaningless, because bias should be understood as a character trait. It thus makes no sense to draw a distinction between implicit and explicit bias. It is consistent with our argument, however, that it may make sense to think of prejudice as a trait that could, in principle, be partly constituted by either explicit or implicit attitudes. And this leaves open the possibility that empirical research may be attempting to show the surprising and morally significant thesis that people who have egalitarian explicit attitudes may nonetheless possess prejudiced implicit attitudes. Thus, unlike Machery (2016), we are not arguing that empirical psychologists have been attempting to demonstrate a thesis that relies on a conceptual confusion. Rather, we are attempting to show that they have not demonstrated the truth of that thesis. Our thanks to a reviewer for suggesting this clarification. For a recent presentation of methodological criticisms of implicit bias research that are distinct from the critique we press here, see Machery (2021).
As an example of the “hyping” of implicit bias research, Brownstein et al. (2020) offer Nicholas Kristof’s claim that, “It’s sobering to discover that whatever you believe intellectually, you are biased about race, gender, age, or disability.” They suggest that this conclusion is not informed by the actual science, which suggests that “explicit beliefs about social concepts are, in fact, strong moderators of implicit attitudes about those concepts” (Brownstein et al., 2020, p. 298, fn. 15).
For example, consider interpretations such as Madva and Brownstein's (2018) proposal that implicit attitudes are clusters of semantic-affective associations that can be trained to change over time, but tend to be unresponsive to the semantic content of our other mental states, Gendler's proposal that, like phobias, implicit attitudes are explained by “aliefs” (Gendler, 2008, 2011), Levy's proposal that they are “patchy endorsements” that resist rational correction, and the view that they are the result of “system 1” processes on a dual system theory of mind (Gawronski & Bodenhausen, 2006).
Strawson (1962).
As Zheng then elaborates in a footnote, “Here, with respect to this particular trait, we adopt the Strawsonian ‘objective’ attitude, the attitude we take towards non-human animals, young children, and beings that are not fully moral agents” (79, fn. 30).
For example, Kawakami et al., (2005, 2007) have tested the effects of “counterstereotype training,” which involves having subjects respond “Yes” to images of Black people or women, or having subjects nod (Wennekers, 2012) or pull a joystick toward themselves when prompted with Black or Arab-Muslim faces and push it away from themselves in response to White faces. Forbes and Schmader (2010) tested the effects of training subjects to associate the phrase “women are good at” with math terms, as well as interventions that aim to reduce the influence of one's implicit biases on one's actions without intervening on one's psychology (Beauclac and Kenyon, 2014). For a defense of using these forms of de-biasing techniques outside the laboratory, see Madva (2017).
Even philosophers who argue for metaphysical conceptions of implicit prejudice that challenge common conceptions of implicit bias assume that the empirical research has at least demonstrated a discrepancy between subjects’ implicit and explicit attitudes. Consider, for example, Carruthers’ (2018) proposal that the same mental structures can underlie and explain both “explicit” prejudice (what a subject would report about herself when prompted) and “implicit” associations and behavior. Carruthers may be right that we need not develop a special ontology to explain implicit bias test results. But even he grants that the empirical research has shown a discrepancy between what subjects would be willing to report if asked and their rapid associations. It is this basic assumption that we are critical of. Our thanks to an anonymous reviewer for encouraging us to distinguish our argument from Carruthers’.
To be clear, we do not think that moral philosophers should simply ignore the findings of empirical psychologists, either. For an elaboration on the relationship between normative moral psychology and empirical psychology, see Wolf (2007).
We are grateful for feedback we received from Susan Wolf, Douglas MacLean, Ram Neta, Robert Smithson, Charles Siewart, Calvin Lai, Elizabeth Reis, Matthew Dennis, Pamela Reis, Chris Hakkenberg, and the anonymous referees who reviewed the manuscript. We have also benefited from discussion with audiences at the Rocky Mountain Ethics Congress and at a meeting of the American Philosophical Association, where Elís Miller Larson commented on the paper. We would also like to thank Brad Cokelet and David Shoemaker for hosting an online discussion of the paper on the PEA Soup blog.
References
Antony, L. (2016). Bias: Friend or Foe? Reflections on Saulish Skepticism. In J. Saul & M. Brownstein (Eds.), Implicit bias and philosophy, volume I: Metaphysics and epistemology. Oxford University Press.
Bar-Anan, Y., Nosek, B., & Vionello, M. (2009). The sorting paired features task: A measure of association strengths. Experimental Psychology, 65(5), 329–343.
Bartlett, T. (2017). Can we really measure implicit bias? Maybe not. The Chronicle of Higher Education. http://www.chronicle.com/article/Can-We-Really-Measure-Implicit/238807
Beaman, L., Chattopadhyay, R., Duflo, E., Rande, R., & Topalova, P. (2009). Powerful women: Does exposure reduce bias? The Quarterly Journal of Economics, 124, 1497–1540.
Beauclac, G., & Kenyon, T. (2014). Critical thinking education and debiasing. Informal Logic, 34(4), 341–363.
Brownstein, M., Madva, A., & Gawronski, B. (2020). “Understanding implicit bias: Putting criticism into perspective. Pacific Philosophy Quarterly, 10, 276–307.
Brownstein, M., & Saul, J. (2016). Introduction. In J. Saul & M. Brownstein (Eds.), Implicit bias and philosophy, volume I: Metaphysics and epistemology. Oxford: Oxford University Press.
Carruthers, P. (2018). Implicit versus explicit attitudes: Differing manifestations of the same representational structures? Review of Philosophy and Psychology., 9(1), 52–72.
Forbes, C. E., & Schmader, T. (2010). Retraining attitudes and stereotypes to affect motivation and cognitive capacity under stereotype threat. Journal of Personality and Social Psychology, 99(5), 740–754.
Gawronski, B., & Bodenhausen, G. (2006). Associative and propositional processes in evaluation: An integrative review of implicit and explicit attitude change. Psychological Bulletin, 132(2006), 692–731.
Gendler, T. (2008). Alief and belief. The Journal of Philosophy, 105(10), 634–663.
Gendler, T. (2011). On the epistemic costs of implicit bias. Philosophical Studies, 156, 33–63.
Glick, P., & Fiske, S. T. (1996). The ambivalent sexism inventory: Differentiating hostile and benevolent sexism. Journal of Personality and Social Psychology, 70(3), 491–512.
Greenwald, A., McGhee, D., & Schwartz, J. (1998). Measuring individual differences in implicit cognition: The implicit association test. Journal of Personality and Social Psychology, 74, 1464–1480.
Haslanger, S. (2015). Social structure, narrative, and explanation. Canadian Journal of Philosophy, 45(1), 1–15.
Henry, P. J., & Sears, D. O. (2002). The symbolic racism 2000 scale. Political Psychology, 23, 2.
Kawakami, K., Dovidio, J. F., & van Kamp, S. (2005). Kicking the habit: Effects of nonstereotypic association training and correction processes on hiring decisions. Journal of Experimental Social Psychology, 41(1), 68–75.
Kawakami, K., Dovidio, J. F., & van Kamp, S. (2007). The impact of counter-stereotypic training and related correction processes on the application of stereotypes. Group Processes and Intergroup Relations, 10(2), 139–156.
Lebrecht, S., Pierce, L., Tarr, M., & Tanaka, J. (2009). Perceptual other-race training reduces implicit racial bias. PLoS ONE, 4(1), 4215.
Levy, N. (2015). Neither fish nor fowl: Implicit attitudes as patchy endorsements. Nous, 49(4), 800–823.
Levy, N. (2017). Am i a Racist? Implicit bias and the ascription of racism. The Philosophical Quarterly, 67, 268.
Machery, E. (2017). Should we throw the IAT on the scrap heap of indirect measures?’ Comment on the Brains Blog, January 17. http://philosophyofbrains.com/2017/01/17/how-can-we-measure-implicit-bias-a-brains-blog-roundtable.aspx
Machery, E. (2016). De-freuding implicit attitudes. In J. Saul & M. Brownstein (Eds.), Implicit bias and philosophy, volume I: Metaphysics and epistemology (p. 2016). Oxford: Oxford University Press.
Machery, E. (2021). Anomalies in implicit attitudes research. Wires Cognitive Science. Early View: https://doi.org/10.1002/wcs.1569
Madva, A. (2017). Biased against debiasing: On the role of (institutionally sponsored) self-transformation in the struggle against prejudice. Ergo, 4(6), 145–179.
Madva, A., & Brownstein, M. (2018). Stereotypies, prejudice, and the taxonomy of the implicit social mind. Nous, 52(3), 611–644.
Mandelbaum, E. (2016). Attitude, inference, association: on the propositional structure of implicit bias. Nous, 50(3), 629–658.
McConohay, J. (1982). Self-interest versus racial attitudes as correlates of anti-busing attitudes in louisville: Is it the buses or the blacks? The Journal of Politics, 44(3), 692–720.
McConohay, J., Hardee, J. B., & Batts, V. (1981). Has racism declined in America? It depends on who is asking and what is asked. Journal of Conflict Resolution, 25(4), 563–579.
Nosek, B., & Banaji, M. (2001). The Go/No-Go Association task. Social Cognition, 19(6), 625–666.
Oswald, F. L., Mitchell, G., Blanton, H., & Jaccard, J. (2013). Predicting ethnic and racial discrimination: A meta-analysis of IAT criterion studies. Attitudes and Social Cognition, 105(2), 171–192.
Payne, B. (2009). Attitude misattribution: Implications for attitude measurement and the implicit-explicit relationship. In A. Black, W. Prokasy, R. Petty, R. Fazio, & P. Brinol (Eds.), Attitudes: Insights from the new wave of implicit measures. Erlbaum.
Rachlinski, J., Johnson, S. L., Wistrich, A. J., & Guthrie, C. (2009). Does unconscious racial bias affect trial judges? Notre Dame Law Review, 84, 1195.
Sabin, J., Nosek, B., Greenwald, A., & Rivara, F. P. (2009). Physicians’ implicit and explicit attitudes about race by MD race, ethnicity and gender. Journal of Health Care Poor Underserved, 20(3), 896.
Saul, J. (2012). Scepticism and implicit bias. Disputatio, Lecture 2012.
Sie, M., & van Voorst Vader-Bours, N. (2016). “Stereotypes and prejudices: Whose responsibility? Indirect personal responsibility for implicit biases. In M. Brownstein & J. Saul (Eds.), Philosophy and implicit bias, volume II: Moral responsibility, structural injustice, and ethics. Oxford University Press.
Singal, J. (2017). Psychology’s favorite tool for measuring racism isn’t up to the job. New York Magazine. https://www.thecut.com/2017/01/psychologys-racism-measuring-tool-isnt-up-to-the-job.html
Sriram, N., & Greenwald, A. (2009). Brief implicit association test. Experimental Psychology, 56(4), 283–294.
Strawson, P. F. (1962). Freedom and resentment. In Proceedings of the British Academy (p. 48).
Swim, J. K., Aikin, K. J., Hall, W. S., & Hunter, B. A. (1995). Sexism and racism: Old fashioned and modern prejudices. Journal of Personality and Social Psychology, 68(2), 199–214.
Washington, N., & Kelly, D. (2016). “Who’s responsible for this? Moral responsibility, externalism, and knowledge about implicit bias. In M. Brownstein & J. Saul (Eds.), Philosophy and implicit bias, volume II: Moral responsibility, structural injustice, and ethics. Oxford University Press.
Wennekers, A. M., Holland, R. W., Wigboldus, D. H., & van Knippenberg, A. (2012). First see, then nod: The role of temporal contiguity in embodied evaluative conditioning of social attitudes. Social Psychological and Personality Science, 3(4), 455–461.
Wittenbreck, B., Judd, C. M., & Park, B. (1997). Evidence for racial prejudice at the implicit level and its relationship with questionnaire measures. Journal of Personality and Social Psychology, 72(2), 262.
Wolf, S. (2007). Moral psychology and the unity of the virtues. Ratio, 20(2), 145–167.
Zheng, R. (2016). Attributability, accountability, and implicit bias. In M. Brownstein & J. Saul (Eds.), Philosophy and implicit bias, volume II: Moral responsibility, structural injustice, and ethics. Oxford University Press.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendix
Appendix
Where the full content was not made available by authors, we include the “Representative Items” listed.
Glick, Peter and Susan T. Fiske. “The Ambivalent Sexism Inventory: Differentiating Hostile and Benevolent Sexism.” Journal of Personality and Social Psychology, 70(3), (1996).
Greenwald, A., D. McGhee, and J. Schwartz. “Measuring Individual Differences in Implicit Cognition: The Implicit Association Test”. Journal of Personality and Social Psychology 74 (1998): 1464–80.
Henry, P.J. and David O. Sears, “The Symbolic Racism 2000 Scale,” Political Psychology, 23:2 (2002).
McConohay, John, Betty B. Hardee and Valerie Batts, “Has Racism Declined in America? It Depends on Who is Asking and What is Asked,” The Journal of Conflict Resolution 25(4), 1981.
The “Modern Racism” Scale:
Payne, B. (2009). “Attitude Misattribution: Implications for Attitude Measurement and the Implicit-Explicit Relationship.” In A. Black and W. Prokasy (Eds.) R. Petty, R. Fazio, and P. Brinol (Eds.), Attitudes: Insights from the new wave of implicit measures. Hillsdale, NJ: Erlbaum.
1.1 Sample 1 (ANES Panel Study)
Feelings: Do you feel warm, cold, or neither warm nor cold toward blacks? [Response options: Warm, cold, neither warm nor cold].
If warm: Do you feel a little warm, moderately warm, or extremely warm toward blacks?
If cold: Do you feel a little cold, moderately cold, or extremely cold toward blacks?
Sympathy: How often have you felt sympathy for blacks? [Always, most of the time, about half the time, once in a while, or never].
Admiration: How often have you felt admiration for blacks? [Always, most of the time, about half the time, once in a while, or never].
Influence: Would you say that blacks have too much influence in American politics, just about the right amount of influence in American politics, or too little influence in American politics? [Response options: Too much influence, Just about the right amount of influence, Too little influence].
1.2 Sample 2 (ANES Time Series)
Feelings: Do you feel warm, cold, or neither warm nor cold toward blacks? [Response options: Warm, cold, neither warm not cold].
If warm: Do you feel a little warm, moderately warm, or extremely warm toward blacks?
If cold: Do you feel a little cold, moderately cold, or extremely cold toward blacks?
Sympathy: How often have you felt sympathy for blacks? [Always, most of the time, about half the time, once in a while, or never].
Admiration: How often have you felt admiration for blacks? [Always, most of the time, about half the time, once in a while, or never].
Stereotypes:
Where would you rate BLACKS on this scale? [ 1 = Hardworking; 7 = Lazy].
Where would you rate BLACKS on this scale? [ 1 = Intelligent; 7 = Unintelligent].
Symbolic Racism:
-
1.
Irish, Italians, Jewish, and other minorities overcame prejudice and worked their way up, blacks should do the same without special favors” [Disagree, Somewhat disagree, Neither agree nor disagree, Somewhat agree, Strongly agree].
-
2.
Generations of slavery have created conditions that make it difficult for blacks to work their way out of the lower class” [Disagree, Somewhat disagree, Neither agree nor disagree, Somewhat agree, Strongly agree].
-
3.
It’s really a matter of some people just not trying hard enough; if blacks would only try harder they could be just as well of as whites. [Disagree, Somewhat disagree, Neither agree nor disagree, Somewhat agree, Strongly agree].
-
4.
Over the past few years, blacks have gotten less than they deserve.” [Disagree, Somewhat disagree, Neither agree nor disagree, Somewhat agree, Strongly agree]
1.3 Sample 3 (Associated Press/Yahoo! News/Stanford University study)
Liking: “How much do you like or dislike each of the following groups? Whites … Blacks …” [Response options: dislike a great deal, dislike a moderate amount, dislike a little, Neither like nor dislike, like a little amount, like a moderate amount, like a great deal].
Admiration: “How often have you felt admiration for blacks?” [Extremely often, Very often, Moderately often, Rarely, Never].
Sympathy: “How often have you felt sympathy for blacks?” [Extremely often, Very often, Moderately often, Rarely, Never].
Stereotypes: Respondents were asked “How well does each of these words describe most blacks?” and were shown a list of 14 adjectives (Friendly, Determined to succeed, Law abiding, Hard-working, Intelligent at school, Smart at everyday things, Good neighbors, Dependable, Keep up their property, Violent, Boastful, Complaining, Lazy, Irresponsible). [Extremely well, very well, moderately well, slightly well, not well at all].
Wittenbreck B., CM Judd, and B Park. “Evidence for racial prejudice at the implicit level and its relationship with questionnaire measures.” Journal of Personality and Social Psychology. 72(2):1997.
Rights and permissions
About this article
Cite this article
Reis-Dennis, S., Yao, V. ‘I love women’: an explicit explanation of implicit bias test results. Synthese 199, 13861–13882 (2021). https://doi.org/10.1007/s11229-021-03401-3
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11229-021-03401-3