predictive value of admissions criteria as they contemplate increasing diversity in the medical student population.

For example, a recent analysis of predictors of medical school performance (White et al., unpublished) showed that for underrepresented minority students at one medical school, academic performance (i.e., the undergraduate grade point average) is a better predictor of academic performance in medical school (measured at three time points, progressively) than standardized test scores (i.e., the Medical College Admissions Test, or MCAT), which reliably predicted only scores on future standardized tests (i.e., USMLE Step 1 examination).

We ask medical educators to consider how studies reported in the higher education and psychology literatures (and to a lesser degree in the medical education literature), many of which are presented here, can be used to develop approaches that ensure medical student achievement of cultural competence. The methods and pedagogies discussed in this paper can potentially help medical students transform their perspectives about diversity and at the same time progress toward achievement of self-author-ship, which provides a foundation for understanding the perspectives of those from other cultures and backgrounds.

However, in considering a comprehensive educational approach within the medical education context, the “negative reinforcement” issues also must be addressed. When students begin clinical training, they leave behind the sheltered environment of the classrooms and clinical skills laboratories, and the day-to-day influence of the faculty members who have taught them in the first 2 years. As they enter the clinics, hospitals, and operating rooms, the “hidden curriculum” (Hafferty, 1998) and “hidden culture” (Taylor, 2003) largely dictate the cultural norms students will need to survive or even thrive. As noted by Hafferty (1998), the values underlying these informal curricula might be in direct conflict with the values underlying the goals of the formal curriculum. The challenge for medical educators is how to extend learning into a very influential milieu over which they have limited control and which in fact may serve to undermine or undo learning that has occurred.

In this paper we have encouraged medical educators to contemplate connections between attributes associated with specific developmental levels and those associated with educational standards and goals for cultural competence. As noted, language in the standards describing desired attributes of physicians closely mirrors language that Perry (1970), Kegan (1994) and others use to describe a high (or in some cases the highest) level of cognitive and interpersonal development—one in which individuals be-