The impact of a PhD on earnings and its social and private returns have been examined extensively, but there has been no hazard analysis of the determinants of withdrawals and successful completions in the UK. In Discussion Paper No. 765, Research Fellow Alison Booth and Stephen Satchell identify their determinants, which are interesting for several reasons. Completion rates are potential performance indicators for university departments; it is important to quantify the impact on completions of the substantial public funding directed to graduate students; while the speed with which new PhDs can be produced in various disciplines is relevant to labour force planning. There will be a shortfall of trained researchers as student numbers expand and the cohort of academics recruited in the 1960s and 1970s reaches retirement.

Booth and Satchell estimate a competing risks model of PhD completion and withdrawal on `duration data' for a sub-sample of 484 graduates from UK universities drawn from the 1986-7 National Survey of 1980 Graduates and Diplomates. These indicate the number of months spent on research and whether individuals had completed, withdrawn or failed, or were still engaged in research at the survey date. They find that public (research council) funding had a significant positive impact on completion rates for men but no impact for women; both male and female completion rates were higher in the sciences and lower in arts and languages than for the base of social sciences; `ability' when proxied by a first-class undergraduate degree raised completions for men but raised both completions and withdrawals (and the latter more so) for women. They observed a significant maternal role model effect for female completions, while part-time registration or full-time registration combined with paid employment had a significant negative impact on completions for men but not for women.

Booth and Satchell conclude that increased research council or equivalent institutional funding will increase only the number of trained male researchers; an increase in the number of PhDs can be achieved most rapidly in the sciences and least rapidly in the humanities; and the use of completion rates as performance indicators is fatally flawed. There is considerable variability across disciplines, gender, funding, ability proxies, employment while undertaking research, and maternal socio-economic group; no individual department can be held responsible for all of these factors.

The Hazards of Doing a PhD: An Analysis of Completion and Withdrawal Rates of British PhDs in the 1980s
Alison L Booth and Stephen E Satchell

Discussion Paper No. 765, May 1993 (HR)