| David Most
University of California, Los Angeles
Funding and finishing the PhD: The role of various patterns of graduate support mechanisms
FINAL REPORT:
This study investigated how the type, amount, and timing of doctoral student funding are associated with the likelihood of completing the Ph.D. in five fields. The associations between completion and other factors, both department-level and additional student-level, were also explored. Student-level data were obtained from the longitudinal database of the AAU/AGS Project for Research on Doctoral Education. A total of 5323 students clustered in 69 departments across 16 universities were selected for inclusion. Department-level data were obtained from two NSF surveys, and the NRC Study of Research-Doctorate Programs.
Multilevel (random coefficient, fixed effects conditional logit, and GEE) discrete-time survival analysis models were used to estimate the effects of funding. In general, the substantive findings were found to be robust to the choice of analytic approach for handling clustering. Ignoring the clustering of students within departments resulted in underestimated standard errors and instances of coefficient bias. While many of the general substantive conclusions would be unaltered if clustering were ignored, there are enough exceptions to warrant taking it into account.
For models in which funding is measured as the cumulative proportion of funding received by mechanism, relative to students with teaching assistantships, students funded by research assistantships or fellowships were more likely to complete the Ph.D. over time, and students who received no funding were less likely to complete. For models in which funding is measured as the annual dollar amount of funding received by mechanism, the effects of receiving additional funding were found to vary over time. In years 1-3 of a doctoral program, higher amounts of funding are not associated with the chances of completion. In years 4-6, additional teaching funds are associated with decreased odds and additional fellowship funds are associated with increased odds of completion. Additional teaching and research funds are associated with a lower likelihood of completion during years 7-9. The effects of both measures of funding were not found to vary across field or demographic group.
Holding constant all else, having a masterÕs degree prior to entering a doctoral program and a lower departmental student-faculty ratio are associated with a higher likelihood of completion. Foreign students are more likely and underrepresented minorities are less likely to complete the doctorate than domestic White students. Gender, age at entry, and GRE scores, and all department-level variables other than field and student-faculty ratio are not associated with completion.
One primary finding, which has potential policy implications, is that the association between receiving additional dollars and completion can vary by time and mechanism. The results indicate that fellowship funds might be more effective with respect to completion if distributed during the middle years rather than in the early years. Given limited fellowship funds, and assuming that higher completion rates is a goal, these findings suggest that supporting students with assistantships early on followed by fellowship money during the dissertation years would maximize the return to such investments.
Back to Funded Dissertation Grants Page |