One of the major recommendations made by many groups at MIT over the past several years is that all community members be offered information about unconscious bias, which is the automatic reliance on a stereotype while making judgments or decisions. Psychologists and neuroscientists have for many years distinguished between automatic and deliberative mental processes; Daniel Kahneman won the Nobel Memorial Prize in Economic Sciences in 2002 for his application of these ideas to economics. Psychologists prefer the term implicit bias, but I will use the more common term unconscious bias.
Last fall I wrote a summary of methods and provides of unconscious bias education for a faculty committee charged by the Provost to evaluate training options for faculty. This blog entry is adapted from that summary, which was divided into brief sections intended to inform faculty leaders considering whether and how to implement a program of unconscious bias education. I’ve removed most of the references.
1. What is your goal? Unconscious bias education is not the same as diversity training, sensitivity training, harassment training, cultural competence, etc.
Why are you considering unconscious bias education for your organization? Is it to change behavior? To improve the climate for underrepresented groups? To help diversify the leadership? You may be considering the wrong approach, with a significant chance of making things worse.
Unconscious bias education aims to increase awareness of how our minds work, so as to help people and organizations adopt practices that improve decision-making. All evaluative processes by people about people — such as formation of teams, student teaching evaluations, employee recruitment, annual performance review, customer service, faculty promotion and tenure — are affected by unconscious bias. They may also be affected by explicit or intentional bias, but here we are concerned with “the hidden biases of good people” (Ref. 1). The consequence of unchecked bias is persistent inequity of representation, compensation, and experience.
Whenever possible, it is preferable to change the system or process to eliminate the possibility for bias to work: changing the rules is the surest way to change outcomes (Ref. 2). For example, implementing blind orchestra auditions dramatically increased the number of women in professional orchestras. However, such blind processes are unrealistic for many selection processes such as faculty recruitment. Becoming aware of unconscious bias is a first step in addressing these issues, but it is insufficient to guarantee equitable outcomes. Research shows that changing unconscious bias does not necessarily lead to changes in explicit bias or behavior (Ref. 3 online).
If your goals go beyond education to creating and sustaining effective change, it is best consider a systemic, data-driven approach. In higher education, these include the NSF ADVANCE program in the US and Athena SWAN in the UK. Although many academic sources focus on gender inequity as the outcome of unconscious bias, systemic interventions can be tailored to address inequity across race, age, and other dimensions of diversity. A promising example from the corporate world is the Bias Interrupters Model.
The next several sections summarize different models for unconscious bias education.
2. Self-paced electronic delivery
Many organizations have developed self-paced online modules in response to California AB 1825 or, more recently, Title IX and the federal Violence Against Women Act. Many of these are motivated by legal compliance to reduce sexual assault or harassment, and are not mainly about unconscious bias education. For example, MIT requires all incoming students to take Everfi’s Haven online module. Some vendors have a range of modules, e.g. Aperian’s Globesmart, which includes cultural competence and is being tried in the MIT Sloan School. The only online module I know that focuses on unconscious bias education is Google’s video of an in-person presentation. Google employees are invited (but not required) to watch this video as the first half of their corporate unconscious bias education and mitigation strategy. I am aware of no peer-reviewed published research on the efficacy of such methods, however at its excellent Rework site Google has published some of their results showing that that retention of learning was as good with the video as it was following face-to-face workshops. For this reason, Google has abandoned in-person lecture-based unconscious bias education. More than half of Google’s employees have watched the video. At MIT, faculty search chairs in one school found the video helpful and its use is spreading.
3. In-person lecture-based workshops
The lecture or workshop presented by an expert is the default practice for any kind of teaching in academia, and the same is true for unconscious bias education. The efficacy of this method is generally unknown. At MIT, over the past year we’ve had many lectures and workshops in this category, several of them from Project Implicit, an organization whose mission is to develop and deliver methods for investigating and applying phenomena of implicit social cognition, including especially phenomena of implicit bias based on age, race, gender or other factors.
The expense and lack of assessment are shortcomings of this approach. The cost of outside consultants, including academics doing this work, makes it prohibitive for scaling up to all of MIT. An alternative is to partner with these experts to develop an in-house version of the workshop. The University of Wisconsin-Madison team (WISELI) gave a workshop on gender bias to MIT’s Academic Council on April 5, 2016, and to a broader audience that afternoon. WISELI helps organizations tailor the workshop, and have provided MIT with a set of materials that could be used to implement their model at MIT. WISELI has also published research results on the efficacy of their methods for faculty learning along with an analysis of social psychological factors in the implementation of such workshops (Ref. 6 online, Ref. 7 online).
4. Interactive theater
Interactive theater is designed not to inform, but to persuade — to change attitudes and behaviors. In this format, trained actors present situations where unconscious bias exists and micro-inequities play out, usually repeating the scene with different strategies, and halting the scene for actors to speak their inner thoughts. The method is designed to increase empathy, and is popular with audiences.
This genre was pioneered in the 1970s by Augusto Boal through his Theatre of the Oppressed in Brazil and Europe. In Boal’s version, audience members can intervene with suggestions or even participation in the scene.
The genre was utilized as a tool for equity and inclusion in higher education by the University of Michigan CRLT players starting in 2000. Other groups using this method include CSW Associates, the Harvard Bok Players, the Cornell Interactive Theatre Ensemble, and the Berkeley Interactive Theater Program. The efficacy of interactive theater in changing behavior is difficult to establish, but a growing number of universities are using it in a portfolio of activities designed to improve teaching and campus climate. At MIT, Lecturer Daena Giardella uses interactive theater with the SPLC SpeakUp! Curriculum to teach bystander intervention skills at the Sloan School.
5. In-person interactive workshops and their assessment
Education research has established that active-learning methods, such as those used in the MIT TEAL classroom, are more effective than lecture-based pedagogy. This is also true for adult learners. This is why experts favor the use of active-learning methods in evidence-based diversity interventions (Ref. 8 online).
A variety of methods are used in such workshops; what they have in common is the combination of theory and practice using active-learning techniques. At MIT, I used a search committee simulation exercise as part of a 2-hour workshop in March, 2016. Google’s Bias Busters workshop utilizes similar methodology. Carnegie Mellon University has partnered with Google to adapt their workshop to faculty, and more than 2/3 of the senior faculty in Engineering at CMU have participated in this workshop, which is given after participants have viewed the Google Unconscious Bias video (online and here).
How effective are these methods? Peer-reviewed research is beginning to provide answers. Project Implicit held a research contest to assess the efficacy of interactive methods for reducing the expression of implicit racial prejudice. The Implicit Association Test and self-reported racial attitudes showed a reduction of prejudice for several methods that involved counterstereotypic imaging and strategies to override biases (Ref. 10 online). A group at Montana State University report high efficacy in hiring of more women STEM faculty based on a three-step intervention involving search committee training, a guidebook, and a faculty advocate for work-life and family issues (Ref. 11 online). Similarly, a study conducted at four Midwestern research universities found some shift in attitudes by men toward women in STEM following participation in a diversity training session (Ref. 12 online).
Pushback is a serious barrier to the efficacy of such programs. Many of the studies cited previously comment on this. As noted in a recent article in PNAS (Ref. 13 online),
“Results across experiments showed that men evaluate the gender-bias research less favorably than women, and, of concern, this gender difference was especially prominent among STEM faculty…. These results suggest a relative reluctance among men, especially faculty men within STEM, to accept evidence of gender biases in STEM.”
For these reasons, experts recommend using multiple strategies to encourage learning and change (Ref. 14 online and Bias Interrupters). Mandatory efforts often generate resistance; voluntary efforts are more successful (Ref. 15 online).
6. Concluding thoughts
Learning from the experience of others is an important step to avoid the numerous pitfalls facing institutions that implement unconscious bias workshops or other diversity interventions. There is a growing social science literature in this area, much of it directed to gender equity in STEM. The self-reported experiences of women in other fields are similar to those in STEM, as shown by the 2002 Report on the Status of Women Faculty at MIT.
Several groups at MIT representing faculty and staff are following the example of Carnegie Mellon University (Ref. 9) to adapt the Google Bias Busters workshop model, tailoring it to their MIT audiences. A combination approach that utilizes a research overview, the implicit association tests, and practicing roleplay scenarios incorporates elements that social psychologists have found effective in shifting attitudes. This is a good beginning.
One-time teaching interventions do not create permanent change. In the late 1980s and early 1990s, with great effort MIT implemented a series of diversity workshops for faculty aimed at informing faculty how to respond to complaints. This effort was not sustained and shows little impact on current personnel practices. For these reasons, I encourage the development of an in-house capability for giving workshops periodically, incorporating elements into new faculty orientation, search committee training, leadership and new department head orientations, etc. Combining such workshops with a review of procedures to de-bias hiring and other practices could help MIT achieve its goal of advancing a respectful and caring community that embraces diversity and empowers everyone to learn and do their best.
1. M. R. Banaji and A. G. Greenwald, Blindspot: Hidden Biases of Good People (Bantam, 2016)
2. I. Bohnet, What Works: Gender Equality by Design (Belknap Press, 2016)
3. P. S. Forscher et al, “A Meta-Analysis of Change in Implicit Bias,” preprint available at https://www.researchgate.net/publication/308926636_A_Meta-Analysis_of_Change_in_Implicit_Bias (2016)
4. Kirwan Institute, Ohio State University, State of the Science: Implicit Bias Review 2014, http://kirwaninstitute.osu.edu/wp-content/uploads/2014/03/2014-implicit-bias.pdf
5. Committee on Women in Science, Engineering, and Medicine, National Academies, Resources on Implicit Bias, http://sites.nationalacademies.org/PGA/cwsem/PGA_161607
6. M. Carnes et al., “The Effect of an Intervention to Break the Gender Bias Habit for Faculty at One Institution: A Cluster Randomized, Controlled Trial,” Acad. Med. 90(2):221 (2015); doi: 10.10971/ACM.0000000000000552
7. C. Isaac, L. B. Manwell, P. G. Devine, C. Ford, & J. T. Sheridan, “Difficult Dialogues: Faculty Responses to a Gender Bias Literacy Training Program,” The Qualitative Report 21(7):1243 (2016); doi: 10.1111/j.1471-.
8. C. A. Moss-Racusin et al, “Scientific Diversity Interventions,” Science 343:615 (2014); doi: 10.1126/science.1245936; http://science.sciencemag.org/content/sci/343/6171/615.full.pdf
9. https://engineering.cmu.edu/about/events/2016/02_12_implicit_bias.html; https://scs4all.cs.cmu.edu/biasbusters/
10. C. K. Lai et al, “Reducing Implicit Racial Preferences: I. A Comparative Investigation of 17 Interventions,” J. Exp. Psych. Gen. 143(4):1765 (2014) ; doi: 10.1037/a0036260
11. J. L. Smith, I. M. Handley, A. V. Zale, S. Rushing, & M. A. Potvin, “Now Hiring! Empirically Testing a Three-Step Intervention to Increase Faculty Gender Diversity in STEM,”BioScience 65(11):1084 (2015); doi: 10.1093/biosci/biv138/-/DC1
12. S. M. Jackson, A. L. Hillard, & T. R. Schneider, “Using implicit bias training to improve attitudes toward women in STEM,” Soc. Psychol. Educ. 17:419 (2014); doi: 10.1007/s11218-014-9259-5
13. I. M. Handley, E. R. Brown, C. A. Moss-Racusin, & J. L. Smith, “Quality of evidence revealing subtle gender biases in science is in the eye of the beholder,” PNAS 112:43 (2015); doi: 10.1073/pnas.1510649112
14. C. A. Moss-Racusin et al, “A “Scientific Diversity” Intervention to Reduce Gender Bias in a Sample of Life Scientists,” CBE-Life Sci Educ 15:ar29 (2016); doi: 10.1187/cbe.15-09-0187; Bias Interrupters Tools for Organizations, http://biasinterrupters.org/toolkits/orgtools/
15. F. Dobbin & A. Kalev, “Why Diversity Programs Fail,” Harvard Business Review, July-August, 2016, pp. 52–60.