We are in the throes of interview season. At Yale, my fellow program directors and I have evaluated dozens, maybe hundreds of applications. We have tried to decide which applicants to interview on the basis of objective criteria — board exam scores, grades, deans’ letters. This practice is certainly defensible: how else can we screen the large number of applicants?
But I often wonder. Am I being completely objective? Have I passed over students who might contribute a unique worldview but have not met my tidy cut-offs?
Blind Spots
A few months ago, I came across the book Blind Spot – hidden biases of good people. Banaji and Greenwald, two psychologists, have studied the ways our accumulated exposure to culture, social norms, and experience influence our judgments. Even well-intentioned “good people,” who believe they are open-minded and objective, harbor hidden prejudice.
They have developed the Implicit Association Test (IAT) which has been administered to millions of people for more than 20 years and is validated through dozens of studies. The IAT uncovers hidden racial, gender, and age stereotypes — among others — not just by asking us about our attitudes toward certain types of people but by cleverly timing our associations. For example, the IAT explores black/white stereotypes by pairing black- and white-appearing faces with pleasant or unpleasant words in a timed format. You can try it yourself at implicit.harvard.edu. The results are fascinating, provocative … and humbling. Even the most sensitive and self-aware among us have implicit, hidden bias.
Implicit Bias In Deans’ Letters
As medical educators, we may think we are immune to hidden biases, and many of us openly seek to build residency programs that represent the diversity of the patients we care for. But we are not immune — we have a blind spot.
The problem with a blind spot is just that — we do not know what we do not see. David Ross, an associate program director for Yale’s psychiatry residency, evaluated the deans’ letters from 6,000 US medical students applying to 16 residency programs. Using proprietary software, he extracted keyword frequency from four thematic categories — standout traits, ability, grindstone habits, and compassion. He showed significant differences between race and gender. Deans were more likely to describe white applicants as “exceptional,” “best,” and “outstanding,” whereas they were more likely to describe black applicants as “competent.” In describing female applicants, deans frequently used words like “caring,” “compassionate,” and “empathetic.” These differences remained significant after controlling for Step 1 scores.
What are we to make of this? Is it true that women really are more compassionate or empathetic? Are black students only “competent” and not “exceptional?” I doubt it. Is the whole system riddled with bias? Almost certainly.
By the time a student applies to residency, they have made it through college and medical school — and experienced the compound effect of implicit bias through their entire academic career. This phenomenon is deeper than politics and affirmative action and cuts to the core of our educational culture. Is the whole system riddled with bias?
Whom Have I Missed?
Implicit bias is also bidirectional. As I evaluate applicants, what do applicants see when they come to my program? I am a white middle-aged man representing an Ivy League institution. But do applicants know that I am an ordained minister and was one of the first to officiate same-gendered weddings in Connecticut? That I can play the opening riff to “Roundabout” by Yes, that I have a mean backhand on the tennis court?
Does any of that really matter to students seeking a residency program? I hope so. In the same way, I realize that I need to look beyond board scores and evaluations. The solution, in part, may be self-awareness. The only way to “see” a blind spot is to acknowledge one exists. What are my biases? What are yours? Who else should seek to remedy the implicit bias of the health care system but the program directors — the last gatekeepers of the health care profession?
Our residents should reflect the diversity of our patients. I want great students in my residency program, but mostly I want great people. Who will contribute to the culture of the program? Who will bring a joyful presence to morning rounds? Who will spend the extra time with a patient — and enjoy doing so? There are plenty of smart medical students, but who will bring wisdom? These are qualities that are hard to discern in a world where blind spots are so common. We falter because we do not see. Over the years, whom have I missed?
Benjamin R. Doolittle, MD, is an expert in burnout and wellness in residents and physicians. He is an associate professor and program director of internal medicine and pediatrics at the Yale University School of Medicine and the medical director of the Yale Medicine-Pediatrics Practice.
For additional reading on this topic:
“Diversity in Residency” podcast from NEJM Resident 360
The Perils and Rewards of Critical Consciousness Raising in Medical Education
Dealing with Racist Patients
Structural Racism and Supporting Black Lives — The Role of Health Professionals
#BlackLivesMatter — A Challenge to the Medical and Public Health Communities
Graduate Medical Education in the Freddie Gray Era
Bias, Black Lives, and Academic Medicine
As the father of a daughter in medical school, this post struck a chord. Certainly being aware of one’s biases (by using the IAT, for instance) is a good idea, but the research is equivocal on whether that awareness actually reduces biased decision making. In her book “What Works: Gender Equality by Design” Iris Bohnet, a behavioral economist and professor at Harvard’s Kennedy School, argues that the more effective way to reduce biased decision making in organizations is to design organizational processes such as hiring in ways that make it harder for bias to influence decisions. (Blinded orchestra auditions, where candidates now sit behind screens so their gender can’t influence the evaluation, is a familiar example of such design.) Among Bohnet’s targets are the unstructured interviews favored by hiring managers in which a candidate’s age, gender, race, alma mater, and fondness for the manager’s favorite sport team may all play on the manager’s biases. In an article in HBR, Bohnet writes “While unstructured interviews consistently receive the highest ratings for perceived effectiveness from hiring managers, dozens of studies have found them to be among the worst predictors of actual on-the-job performance.” In filling residency programs, designing the evaluation process to make biased choices harder could help achieve the goals Dr. Doolittle describes here. (Here, for those interested, is professor Bohnet’s HBR piece: https://hbr.org/2016/04/how-to-take-the-bias-out-of-interviews )
Dear Mr. Morse,
Thank you for your thoughtful comment. The post from Iris Bohnet is very helpful. There is a certain tension to this process. If we default to objective measures (ie. the structured interview, grades, MCATS), then do we miss the untraditional applicant who may offer a unique perspective? And yet we can’t have a completely subjective process either – lest we have too many candidates who reflect the make-up of the faculty. I think Bohnet’s suggestion is helpful – multiple independent interviewers. I would add that the interviewers should be of different backgrounds and interests. Also, I think an interviewer’s self-awareness is key. After each interview, we might ask ourselves the question, “Could I have a blind spot here?”
Many thanks for your comments. I wish your daughter all the best.
~ Ben
Thank you for taking on this topic. One way that we can all improve the system is to put members of minority communities and those who are underrepresented in our own programs in positions of power. Residencies will look different when selection committees look different. Deans letters will look different when deans look different. We are all biased We can’t eliminate bias. Maybe we can find ways to have bias work to bring us what we want: a health care system that works for all people.
I do appreciate the sensitivity of this discussion and the need to better develop tools to identify the best attributes of candidates and compatibility with programs. But is using race, gender and other similar attributes truly fair? Does it truly benefit societies health care needs?
Let me give an example of a real life medical school candidate I had the pleasure to meet a number of years ago. When I met him he was 27yo. He grew up in a middle class, Caucasian family in suburbia. His father was a blue collar worker. His mother was unable to work because of a slow, chronic and ultimately fatal illness. His larger family had little respect for higher education. The family could neither afford nor had the desire to send him to college. So he started in a community college and did well. However, he had a low lottery number and ended up in the military as a medic and eventually in conflicts overseas. During his tour he attended some college courses as best he could. He came back to a country that hated its Veterans so ended up with no advantages. He put himself through college and became the first in his large family to graduate. He did this mostly with student loans and working odd jobs. He had a passion for medicine acquired by his job in the military and the physicians and surgeons whom he worked with. His GPA was acceptable but not tops. His MCATs were very good. He had good letters of recommendation (which is when I met him). He tried and failed to gain acceptance to medical school three times. He did not look for “angles” to get accepted. He wanted to be accepted for who he was and what he could contribute. He finally did get accepted. Today he leads a large department in a large organization, contributes research and publications, teaches, and runs a residency program.
The case above, is but one of many somewhat similar cases I have encountered in my long career. People that do not care about the “big picture” of societal engineering. Rather, they have a strong passion for medicine and want it and will not be satisfied with anything else. No counselor can sway them to something easier to obtain.
So, my solution is simple. If readers are not doing this already, send the applications first to a third party for coding. Remove names, photos, demographic information as best as one can (not always realistic because letters can contain some of this). Use as objective a screening tool as possible. Interview based upon fit for the program and then … let nature take its course. Stop trying to engineer America’s health care provider demographics.
what is calling when doing evaluation at different stage of age
is that interns and residents change quickly over a short period of time
(a sort a recall bias)
some seniors have inconscient influence on youger colegues , but when telling thier sotry back
some residents says ” yes this man have really fascinated me , i dint tell
but he was the ma that change my view for the rest of my way