At most colleges or universities, a snapshot of graduates who recently entered PhD programs in the sciences won't reflect the diversity seen in the undergraduate population. Even undergraduate science majors, bound for graduate school or not, likely do not reflect the student distribution within each discipline's introductory courses.

Recognizing how common this situation is presents an opportunity to explore whether our shared challenges might be met with strategies based on best practices that have been effective at minority-serving institutions as well as at a handful of traditionally white institutions.

The idea of addressing shared challenges with shared successful strategies coalesced during an October 2004 conference on undergraduate science education at the Howard Hughes Medical Institute (HHMI), where a group of scientists resolved to disseminate information about current strategies that have successfully educated a diverse array of undergraduates who later pursued PhDs. Representing 18 institutions—from large research universities to small liberal arts colleges—the consortium of scientists, or the "Diversity in the Sciences Collaborative," organized three national symposia on "Diversity in the Sciences" for 2005-06, held at Harvard University, the University of Louisiana at Monroe, and the University of Washington. Support from HHMI, the National Institutes of Health, and the host schools enabled more than 60 institutions to participate.

A first step toward our consortium's goal of national engagement is to provide symposium participants with a framework for building an educational community that welcomes students of all backgrounds into the ranks of potential future scientists.

The consortium aims to build on the successes of existing programs with proven track records. Xavier University in Louisiana, for example, served as one model for the symposia. Current national data reveal that decades of well-funded, well-intended diversity programs at majority institutions have generally not contributed significantly to educating scientists from underrepresented groups. Some exceptions stand out.

The successes of the Meyerhoff Scholars Program at the University of Maryland-Baltimore County (UMBC) and the Biology Scholars Program at the University of California-Berkeley provided a source of rich dialog among symposia participants. The programs differ significantly (the Meyerhoff Scholars Program admits only the most qualified entering undergraduates, while all students are eligible for the Biology Scholars Program), but they also employ strategies in common-immediate and continued mentoring, study group participation, and opportunities for research. Both can claim an increased retention rate of underrepresented students with science majors and increased GPAs for these students. Additionally, Meyerhoff Scholars enter graduate study at five times the rate of students who were accepted into the program but decided to attend college elsewhere. To develop new approaches that may be better suited to other institutional settings, rigorous assessment and publication of data on effectiveness from additional successful programs must ensue.

Symposia discussions revealed important components that contribute to broad-based inclusion of underrepresented students in undergraduate science programs. Though visionary soloists can make a difference, a combination of administrative and faculty commitment is clearly the ideal starting point. Presidential or high-level administrative commitment-in both visionary and financial terms-eases the road to program assessment and change, and faculty commitment leads to inclusive curricular and research initiatives.

This starting point requires an informed, institution-specific context based on the data analysis of recent enrollments, withdrawals, grades, progression through the major, and measures of excellence such as an undergraduate thesis, honors, and peer teaching. A year-by-year analysis will reveal imbalances, allow for a plan of action to be developed and implemented, and measure changes.

Successfully designed programs include significant components of early faculty and peer mentoring, which can profoundly affect student retention and success. This early attention may begin on campus in the summer preceding the start of college, where students learn about the high expectations of university academics through science course work, exposure to scientific research, time-management training, and even professional "manners." Faculty and peer mentoring are perhaps most critical in the first six weeks of the first science course (through the first exam), when students may decide whether they "belong" in the sciences. Inclusion is further fostered by expectations of excellence, not the traditional sink-or-swim approach, combined with the assurance of students' ability to meet these high expectations. Mentoring can happen within the context of a course being taught; it does not necessarily require extensive one-on-one time.

Institutions with proven track records of sending higher percentages of underrepresented students into PhD programs fund and encourage early student research experiences, where mentoring becomes more focused. Students' immersion in science with scientists can impact everything that follows. Thus research opportunities ideally begin the summer after a student's first year.

Successful programs also encourage or mandate student participation in peer study groups for science and math courses. Study groups foster identification with the sciences and take advantage of the known benefits of mixed-group learning.

While analyzing data from a range of institutions, our consortium encountered obstructive stereotypes that must be dispelled for progress to be made. The idea that college students of color taking introductory biology or chemistry courses are interested in medicine, not basic science, is often used to justify their underrepresentation in science. This does not stand up to scrutiny since approximately 90% of all college students in introductory biology or chemistry courses for majors look toward medical rather than scientific careers. ¹

Another stereotype claims that inadequate K-12 education in the United States is to blame for students' lack of interest in pursuing college-level science study. Data from our own and other institutions indicate that students of color and from all educational and income levels enroll in introductory courses in biology and chemistry (the disciplines on which we have focused) at representative levels. Yet they disengage during their earliest experiences at the university level. The most successful programs have found ways to stem this early exodus from the sciences.

Louisiana State University's relatively new La-STEM and La-Scholars "mentoring ladder" Programs have graduate students mentor undergraduate science majors and undergraduate science majors mentor entering undergraduates. Programs such as these, along with the Meyerhoff Scholars Program and Biology Scholars Program, keep underrepresented minority students thriving in the sciences, and they can serve as models for the effective engagement of underrepresented science students nationwide.

Faculty and administrators need to confront stereotypes and analyze their own recent institutional records of undergraduate science education if the tide is going to turn away from science's longstanding track record of non-inclusion. By joining a national effort to learn about successful practices and to avoid reinventing the wheel, we are among many who hope our own institutions will soon be included in this list of proven change-makers.

Wendy Raymond is Associate Professor of Biology at Williams College and Robert A. Lue is Director of Life Sciences Education at Harvard University.