Increasing the enrollment of underrepresented minorities (URMs) in the health professions is a challenging and urgent issue. Six percent of nurses, 9% of physicians, and 5% of dentists are of black, Hispanic, or American Indian background, yet these ethnic groups collectively represent one-quarter of the U.S. population.1 The proportion of URM students in matriculating classes in U.S. medical and dental schools showed no net gain between 1995 and 2005, failing to keep up with the growth in minority populations.1
Many factors contribute to the lack of greater diversity in U.S. medical and dental schools, including disparities in educational attainment and academic preparedness before college.2–4 Nonetheless, despite the many adversities in the educational pipeline, the number of URMs receiving baccalaureate degrees in the United States has steadily increased during the past decade. In fact, URM college student interest in biological science majors is at an all-time high and has doubled in fields such as general biological sciences, biochemistry, and biophysics.4 Blacks and Latinos enter higher education with the same level of interest in science, technology, engineering, and math fields as their non-URM peers. However, URM college students fail to persist in these majors at the same rate as their white and Asian-American classmates.5 They also tend to struggle more in their final years to complete a bachelor’s degree. These difficulties are in part explained by URM students being less academically prepared than their white and Asian American counterparts. However, even after controlling for academic preparation and achievement before college, URM college students who begin work toward a science degree are more likely to switch into another field or lose interest than are Asian American and white students.6 Studies have also found that URM students are often discouraged from pursuing degrees in science because of preconceived notions about scientists and lack of career planning.7–10
Our study focuses specifically on the college stage of the health professions educational pipeline, examining the “leakage” of URM students at this stage. Although it is known that URM college students in general have lower grade point averages (GPAs) than their non-URM counterparts, prior research has not investigated in detail URM academic achievement in specific prehealth “gateway” courses required for application to medical or dental school. Although gateway courses, such as chemistry and calculus, carry the reputation of being courses that “weed out” noncompetitive students in prehealth pathways, previous studies have not investigated whether particular courses in the gateway course series pose unique academic challenges to URM students. However, it has been known that a negative experience in a gateway course can lead to premed students losing interest in pursuing a health professions education, particularly among URM and female students.7
We examined the academic achievement of undergraduate students enrolled in prehealth gateway courses at several University of California (UC) and California State University (CSU) campuses. The study objectives were to determine (1) whether URM students receive lower grades in gateway courses than do non-URM students, (2) whether URM students perform disproportionately worse in particular gateway course(s), (3) the extent to which lower grade performance, if found, is explained by the differences in precollege academic achievement, and (4) whether URM students are less likely than non-URM students to persist in completing the full requisite number of gateway courses for eligibility for application to medical or dental school.
Computerized administrative records on undergraduate students were obtained from the central registrar or institutional resource center offices at six California college campuses: University of California, Berkeley; University of California, Los Angeles; University of California, Riverside; San Francisco State University; California State University, Los Angeles; and California State University, Dominguez Hills. We selected these campuses for the study because they had relatively high numbers of URM students relative to other UC or CSU campuses, covered a wide geographic scope within the state, and were willing to provide administrative data for the study. Data on students were requested and collected from January 2006 until March 2007 for the academic years 1999–2000 through 2005–2006. Although the terms “black” and “African American” were used interchangeably in the data reported by the schools, in this study we use only “black” for consistency.
To be included in the study, students at these campuses had to have been new matriculants (either a freshman or transfer student) in the 1999–2000 or 2000–2001 academic years and enrolled at any time after matriculation in at least one course from the sequence of prerequisite or “gateway” courses required for application to medical or dental school (general chemistry, organic chemistry, general biology, introductory physics for nonengineering majors, and calculus). Data files were collected for each eligible student and contained the following information: self-reported ethnicity, gender, whether the student matriculated as a transfer student, high school of graduation (a specific school identifier if a California public school, with other schools classified into two broad categories of either California nonpublic school or out-of-state schools), intended and declared major, and college entrance standardized test scores (SAT or ACT). The data file also included information on all courses taken at the campus and the corresponding grades awarded. Students with missing ethnicity data were excluded from the analysis of academic achievement. Records that had a missing outcome, grade in a gateway course, were also dropped.
The high school identifier was used to determine the Academic Performance Index (API) for each student who graduated from a California public high school. The California Department of Education assigns an API to public schools in the state on the basis of a variety of academic measures.11 For each student, we constructed a final high school variable that included categories for each quartile of API scores and additional indicator categories for students who did not attend California public high schools and for transfer students whose academic files did not include high school data. To construct a variable on college admission test scores, we divided scores from the SAT and ACT into quartiles and assigned students to a quartile ranking based on their score (student files had either an SAT or ACT score, but not both). SAT or ACT scores were used as a predictor to capture the student’s academic history before entering college, and API index was used to capture the academic quality of the high school. College major was categorized into three groups (biological science, physical science, and other).
We analyzed grades for each course as a dichotomous outcome comparing “A” or “B” grades versus “C,” “D,” or “F” grades. We dichotomized grades because a B grade is usually considered the minimum acceptable grade for a student to be considered competitive for medical or dental school admission. In the case of duplicate entries for a given course indicating repeat enrollment, we kept the record with the highest grade for analysis. We calculated overall GPA for the gateway courses completed as the total grade points earned divided by the total courses attempted. Nongateway courses were excluded from the GPA calculation. We counted total gateway course sequences completed by summing the number of terminal courses taken (e.g., the second semester of a two-semester general chemistry course) in each gateway science discipline. Completion of four or more gateway courses is considered a reasonable measure of persistence in the gateway series. Although five gateway courses are required for eligibility for application to medical or dental school, consideration was given to those students who may have completed one of their gateway courses at another college or placed out of this requirement.
Comparisons were made between ethnic groups on the outcome variables of percentages achieving grades of A or B in each course, mean overall gateway GPA, and percentages completing four or more gateway courses. Single variable regression models were estimated to compute unadjusted odds ratios for the association between ethnic group and outcome variables and test for the statistical significance of these odds ratios. We then estimated multivariate regression models to account for the potential effects of other variables that might confound an association between ethnicity and outcomes. These models were constructed based on a priori expectations about potential confounding variables. For the models predicting course grades and overall gateway GPA, regression models included covariates for gender, transfer status, high school category, and college admission test score. For the models predicting whether students completed four or more gateway courses, regression models included covariates for gender, GPA in the first two gateway classes completed (a class being defined as a single term of a subject, as opposed to a course which is defined as a linked sequence of two semesters or three quarters of classes in the subject), college major, college admission test score, and transfer status. All regression models adjusted the standard errors to account for clustering of students at the campus level. Statistical analyses were performed using Stata (College Station, Texas).
The study included 15,000 students, with about 83% of the students attending UC campuses (Table 1). Five percent (729) were black, 12% (1,743) were Latino, and 7% (1,055) were Filipino.
Grade performance stratified by course and ethnicity is shown in Table 2. The table displays the percentages of students achieving a grade of A or B in the course, the unadjusted odds ratio of achieving a grade of A or B with white students as the referent group, and the adjusted odds ratio after controlling for admissions test scores, high school API, and other variables.
In all courses, black and Latino students were significantly less likely to achieve a grade of A or B than white students. For example, in biology, only about 29% (2,108/729) of black and 36% (312/1,743) of Latino students received a grade of A or B, compared with 65% (3,106/4,280) of white students. The unadjusted odds ratios essentially show the same results, only displaying odds relative to white students rather than crude percentages. Comparing the results of adjusted and unadjusted odds ratios demonstrates the degree to which the lower grade performance among blacks and Latinos is explained by factors such as lower college admission test scores and the type of high school attended. For each of the courses, adjusting for these factors diminishes, but does not eliminate, the gap in grade achievement between black and Latino students as compared with white students. For example, for biology, the unadjusted odds ratio for black students is 0.23, and the adjusted odds ratio is 0.46.
Filipino students also achieve consistently lower grades than white students in the gateway courses. As for black and Latino students, adjusting for other variables diminishes but does not fully account for the grade differences between Filipinos and whites.
The course-specific grade patterns are reflected in the overall GPA for all gateway courses combined (Figure 1). The mean gateway GPAs for black and Latino students were 1.70 and 1.94, respectively, compared with the mean GPA of 2.57 for white students. The mean GPA for Filipino students fell between those of Latino and white students. The results of the regression model predicting overall gateway course GPA is shown in Table 3. The coefficients represent the difference between the mean GPA of students in each ethnic group relative to white students, after adjusting for the other variables in the model. As was the case for course-specific grades, the adjusted model narrows but does not eliminate the grade gap between white and URM students. For example, the difference between the black and white students’ GPAs in Figure 1 is −0.63, whereas the coefficient in Table 3 is −0.53. The results in Table 3 also demonstrate that the variable for admission test scores is a strong and highly significant predictor of GPA.
An additional important measure in evaluating the health professions pipeline is to assess the number of candidates who are eligible to apply to medical or dental school on the basis of completing the sequence of gateway prerequisite courses. Overall, about 20% (3,049/15,000) of students who took at least one gateway class in college went on to complete four or more courses (bottom rows of Table 2). This relatively low percentage reflects not only attrition of students from a prehealth pathway but also the fact that many students who are not interested in health careers may take one or two classes in the gateway series as part of their general education coursework or requirements for non-health-oriented science and engineering majors.
In contrast to the findings for course grades and GPA, black, Latino, and Filipino students were not significantly less likely than white students to complete four or more gateway courses (Table 2). In fact, after adjusting for admission test scores, college major, and grades achieved in the initial gateway classes, black and Filipino students were significantly more likely than white students to persist in completing four or more gateway courses, as indicated by the adjusted odds ratios of greater than one. Latinos also tended to have greater odds (1.23) than whites of completing four or more gateway courses in the adjusted model, though this difference did not reach statistical significance.
The same analyses were repeated including only students who were declared or intended biology majors. The results were similar when limited to biology majors. Because UC and CSU students may differ in characteristics that we were not able to measure, such as family socioeconomic status, and because UC and CSU campuses also are different academic environments, all the analyses were repeated with the inclusion of a variable indicating whether each student attended a UC or CSU campus. Inclusion of this variable had almost no effect on the findings across ethnic groups for any of the study outcomes.
Black and Latino students at a diverse group of UC and CSU colleges receive significantly lower grades on average in premedical and predental gateway courses than white students. For all five of the gateway courses examined, there was an achievement “gap” of between 25% and 30% when comparing the percentages of black and Latino students receiving grades of A or B and the percentages of white students receiving these grades. This gap was relatively consistent across all five courses. No single course stood out as being disproportionately more challenging for URM student academic achievement relative to white student achievement. Filipino student grade achievement in these courses was somewhat higher than that of blacks and Latinos but lower than that of white students.
The disparity in gateway grade achievement did not seem to be fully explained by URM students entering college with less academic preparation. The gap in grades persisted even after adjusting for measures of prior academic performance and preparation such as standardized college admission test scores and high school API score. This finding suggests that there are factors operating within the college environment itself that may contribute to the lower grade achievement of URM students in prehealth courses and that these differences in academic achievement in college are not fully attributable to disadvantages experienced in the precollege stages of the educational pipeline. These college-level factors may be potentially modifiable by interventions introduced at this stage, such as academic and social supports, to enhance URM student achievement.
Despite the greater challenges experienced by URM students in gateway courses, URM students were nearly as likely as white students to persist in the course series and complete at least four gateway courses. In fact, URM students seemed to be less deterred than white students by receiving low grades in their initial gateway classes. After accounting for the lower grades of URM students in their initial courses, URM students were more likely than white students to complete four or more gateway courses. This finding suggests a degree of resiliency and determination among URM college students that may be assets for ultimately succeeding in the health professions educational pipeline. Many URM students at the campuses studied did not seem to simply give up on their career aspirations and abandon a prehealth curriculum in the face of academic adversity.
This study has several limitations. The data available did not include a measure of student socioeconomic status, a factor which might have also explained some of the differences in grade achievement between ethnic groups. The database did not contain consistent information on high school grades, which might have provided some predictive value for the regression models beyond that provided by SAT or ACT scores. The database also did not permit us to directly identify premedical and predental students. Not all students who take gateway courses are considering a prehealth college pathway. Many students in the study may therefore have had no intention of completing more than one or two courses in the gateway series. Although this contributes to the finding that a minority of students in all ethnic groups completed four or more gateway courses, there is no reason to believe that this introduces a bias in the pattern of findings across ethnic groups. Although five courses are required to gain admission into most health professions schools, we used completion of four courses as the outcome variable. Students in the study sample who attended CSU campuses were less likely than UC students to complete five gateway courses (data not shown), and many CSU students attend community college before matriculation in a CSU. We do not have data on college courses completed before matriculation at the study schools, and therefore we considered completion of four gateway courses a reasonable outcome measure. In addition, although an attempt was made to identify the key courses fulfilling prehealth requirements, there may have been courses not included in the study that fulfilled prehealth requirements. For example, math or science courses offered in engineering schools were not included. Another limitation of this study is grouping students by college major measured at a single point in time, when students may change major during their college years. However, we did not have a running history of major changes for each student in the database. Finally, although a wide range of UC and CSU campuses was included in the study, the results may not be generalized beyond these campuses.
This study has several implications for educational and workforce policy. First, the findings highlight the need for interventions at the college level that can enhance the academic achievement of URM students in prehealth science and math courses. The academic achievement gap is wide. Closing this gap will require concerted efforts to fund and implement effective interventions at the college level to promote the academic success of URM students. Good-quality evidence has demonstrated the effectiveness of many interventions at the college stage to boost URM students’ performance in science and math courses.12–16
Second, the results suggest that interventions to support URM students in gateway courses must be sustained throughout a student’s gateway course series rather than being focused on a particular step or course in this sequence. URM students are more likely than white students to struggle academically in all of the gateway courses. It is unlikely that a single dose of an academic support program administered at the first one or two courses in the gateway series will confer lasting success throughout the gateway series. Students almost certainly will benefit from booster doses throughout the prehealth curriculum.
Finally, the study raises questions about why URM students experience major challenges in gateway courses, even when they have similar prior academic preparation and achievement to non-URM students who experience greater success in these gateway courses. Further research is needed to identify the critical elements in the college environment that impede URM students’ academic success and that may be targeted by interventions to enhance URM student achievement.
URM students at the UC and CSU campuses studied have lower grade achievement in prehealth gateway courses than their non-URM counterparts. However, despite experiencing greater academic adversity in these courses, URM students are not more likely than non-URM students to be deterred from completing the full gateway course series to become eligible for application to medical and dental school. Interventions at the college level to support URM student performance in gateway courses are particularly important for increasing the diversity of medical and dental schools.
This work was supported by a grant from The California Endowment. The authors wish to thank the following individuals and their institutions for their assistance in providing the data used in the study: Pamela Brown and Denis Hengstler, University of California, Berkeley; Kelly Wahl and Caroline West, University of California, Los Angeles; Bob Daly, University of California, Riverside; Anne Shen and Susan Dmytrenko, San Francisco State University; Jen Chan and Alan Muchlinski, California State University, Los Angeles; Peter Van Hamersveld, California State University, Dominguez Hills
2Maple S, Stage F. Influences on the choice of math/science by gender and ethnicity. Am Educ Res J. 199l;28:37–60.
3Ware N, Lee V. Sex differences in choice of college science majors. Am Educ Res J. 1988;25:593–614.
5Anderson E, Kim D. Increasing the Success of Minority Students in Science and Technology. Washington, DC: American Council on Education; 2006.
6Barr D, Gonzalez ME, Wanat SF. The leaky pipeline: Factors associated with early decline in interest in premedical studies among underrepresented minority undergraduate students. Acad Med. 2008;83:503–311.
7Gardner AL, Mason CL, Matyas ML. Equity, excellence, and just plain good teaching. Am Biol Teach. 1989;51:72–77.
8Porta AR. A positive effect of role modeling on minority students using diversity among biomedical scientists as a teaching tool. Am Biol Teach. 2002;64:176–182.
9Fouad NA. Career linking: An intervention to promote math and science career awareness. J Couns. 1995;27:163–171.
10Lindner JR, Wingenbach GJ, Harlin J, et al. Students’ beliefs about science and sources of influences affecting science career choice. NACTA J. 2004;48:2–7.
13Maton K, Hrabowski F, Schmitt C. African American college students excelling in the sciences: College and postcollege outcomes in the Meyerhoff Scholars program. J Res Sci Teach. 2000;37:629–654.
14Fullilove R, Treisman PU. Mathematics achievement among African Americans undergraduates at the University of California, Berkeley: An evaluation of the mathematics workshop program. J Negro Educ. 1990;59:463–478.
15Barlow EL, Villarejo M. Making a difference for minorities: Evaluation of an educational enrichment program. J Res Sci Teach. 2004;41:861–881.
© 2009 Association of American Medical Colleges
16Matsui J, Liu R, Kane CM. Evaluating a science diversity program at UC Berkeley: More questions than answers. Cell Biol Educ. 2003;2:11.