Information about the compensation of postdocs is annoyingly elusive. The reasons for this are widely known and commonly repeated. The vast array of postdoc funding sources, the lack of central institutional control of the employment status, and creative packaging of awards all contribute to the complexity of any effort to do comparative studies of postdoc funding.
One graduate dean has quietly picked up the challenge, adding queries about postdoctoral fellowships to a national survey of graduate-assistant stipends and fellowships. Since 1981, the University of Nebraska-Lincoln, (UNL) has conducted a biennial survey of graduate teaching and research assistantship stipends. Compiled by the Office of Graduate Studies  at UNL, the survey has changed and grown through the years. At first, only the 63 members of the Association of American Universities  were surveyed, and in 1989, the survey team incorporated graduate and postdoctoral fellowships.
The AAAS 2001 Salary Survey of Life Scientists offers additional insights into the compensation of young scientists and a rather dismal view of postdoc salaries . The full story is available on the AAAS Science Careers  Web site.
Last month the university released the 1999-2000 National Survey of Graduate Assistant Stipends, Graduate Fellowships, and Postdoctoral Fellowships  to great fanfare in the graduate education community. According to survey coordinator Carla Coorts, however, several major universities choose not to participate because they believe the findings will have limited utility at their institution or because of their concerns about the effort required to collect the requested data. Furthermore, only five academic institutions (the Universities of Colorado-Boulder, Michigan, and Wisconsin-Madison, Baylor University, and the Massachusetts Institute of Technology) of the 25 estimated to have the largest number of postdocs participated in the survey.
Nonetheless, Peter Syverson, vice president for research and information services at the Council of Graduate Schools , says the survey is "as good as we've got" and that it provides useful benchmarks.
The Committee on Science, Engineering, and Public Policy  (COSEPUP) estimates that there are 52,000 postdoctoral scholars in the United States. The UNL study includes data on 2987 fellowships at 106 universities. We know from the COSEPUP report that fellowships support only a small portion of all postdocs; most postdocs are supported by research grants. For example, of the nearly 4500 postdocs supported by the National Science Foundation in 1999, about 200 were supported by fellowships; the remainder are paid from PI grants.
The 1999-2000 UNL survey does not contain information about the sources of fellowship funds. And the methodology has changed: the 1997-1998 survey reports the number of fellowships awarded by departments; earlier surveys reported the number of departments awarding fellowships. Thus the survey does not offer insights about the growth or decline in the number of fellowships being awarded.
The distribution of fellowships by area of interest in the UNL study is comparable to estimates reported by COSEPUP. The largest percentage of postdoctoral fellowships is awarded in the biological sciences, followed closely by the physical sciences (see Table 1). Together, the biological and physical sciences represent more than 60% of the total.
The survey corroborates what many already suspect about fellowships in the aggregate--postdocs in the physical sciences have the largest awards, and those in the arts and humanities the smallest (see Table 2).
Some of these variations are likely the result of local or specialized fellowship programs, not necessarily variations in larger national fellowship programs.
Fellowship amounts vary considerably by area of interest and within areas of interest (see Table 3). For example, one fellowship, in business management, paid $57,000; another, in finance, paid only $3500. The survey does not differentiate the source of fellowship funds; thus awards may include a mix of local as well as public and private funds.
The UNL survey is the only one of its kind. Merlin Lawson, dean of graduate studies and international affairs at UNL, has supported the survey with institutional funds and a small grant from the Council of Graduate Schools. Lawson believes he has an obligation to sustain the effort given how important the information is to the higher education community.
Additional information about the UNL survey is available at www.unl.edu/gradstud/stipend/survey.html .