Comparing physician implicit and explicit attitudes about race by gender, race and ethnicity

Background: Recent reports suggest that health care provider implicit attitudes about race may contribute to racial and ethnic health care disparities. A social psychological perspective may help to understand provider contribution to health care disparities.

Objective: We used a measure of social cognition, the Race Attitude Implicit Association Test (IAT), to measure implicit and explicit (self-report) attitudes about race among MDs. We hypothesized that MDs, similar to others in society, would show a strong implicit preference for white Americans relative to blacks Americans and that MD implicit and explicit attitudes about race would differ.

Research Design: This study measured implicit and explicit attitudes about race using the Race Attitude IAT for a large sample of test takers (N=404,277), which included a sub sample who self-identified as MDs (n=2535). Test takers voluntarily visited the Project Implicit? demonstration web site (https//implicit.harvard.edu/implicit/) from January 12, 2004 through May 12, 2006. We compared implicit and explicit attitudes about race by MD gender, race and ethnicity (African American, Hispanic, Asian and White).

Measurements: The IAT measure produces a score which is calculated as a standardized difference in timed responses to two key sections of the IAT that pair the concepts of White American with “good” and African American with “good.” A positive IAT score that differs from zero indicates some level of implicit preference for white Americans relative to black Americans. Explicit attitudes about race were measured by self-report. Standardized effect size was measured using Cohen's d.

Results: We found that non-African American MDs showed a statistically strong implicit preference for white Americans relative to black Americans (M= 0.39, Cohen's d= 0.89). This result was similar to that of the total sample of test takers (M= 0.35, Cohen's d= 0.81). African American MDs did not show an implicit preference for either blacks or whites (n= 206, M= 0.05, Cohen's d= 0.11). Among all racial and ethnic groups, females showed less implicit preference for white Americans than did males. Implicit and explicit attitudes about race were dissimilar.

Conclusions: MDs, similar to others in society, with the exception of African American MDs, hold strong implicit pro-White attitudes. This finding lends support to speculation that provider implicit attitudes about race may represent one pathway to unequal treatment in health care. Future research is needed to determine whether MD implicit attitudes about race are related to quality of medical care.