In 1999, the Bill and Melinda Gates foundation started the Gates Millennium Scholars Program (GMS), a 20-year initiative which intends to expand access to higher education for high achieving, low-income minority students. In addition to its academic objectives, GMS also has the goal of creating future leaders in minority groups. The program is administered by the United Negro College Fund (UNCF). In 2000, the first year of the program, over 4,000 grants were awarded to minority students who were entering college or continuing their undergraduate or graduate studies during the 2000-2001 academic year. Since then, an additional 1,000 scholarships have been awarded to outstanding freshmen every year. Awardees can receive the scholarship for up to 5 years as an undergraduate and 4 years as a graduate student. The scholarship is renewable through graduate school in math, science, engineering, library science, and education.
To be eligible, students had to meet several qualifications. They must (1) be of African American, American Indian/Alaska Native, Asian American, Hispanic/Latino, or Pacific Islander background; (2) be full-time students entering college or university; (3) have a GPA of at least 3.3 on a 4.0 scale; (4) be eligible for Pell Grants; (5) be leaders in community service, extracurricular, or other activities. The purpose of the GMS Tracking and Longitudinal study is to gather data on the lives of scholars and selected non-recipients in order to analyze the effects on the educational, civic, and personal lives of selected sample members. The Gates Foundation hopes to generate research that will help improve education attainment and achievement of minority students.
Cross-sectional data collection contains the responses of all 4,053 scholarship recipients in the first year of Gates Millenium Scholars and additional sampling of 2,744 non-recipients. A stratified sample design was used for non-recipients in order to enable powerful comparisons at the level of racial/ethnic group between freshman and continuing undergraduate students. All Pell-Grant-in-place non-recipients were included in the samples for each racial/ethnic group. The remainder of cases was drawn from the pool of non-recipients without Pell Grants in place. The goal was to obtain 300 completes from each racial/ethnic group except American Indians. Due to their small number, all non-recipients were included in the samples of both populations. The designs are likely to produce a design effect for non-recipients because the samples are not proportional to their representation in the population as a whole. This effect is likely to be offset by the stratification.