NATIONAL SCIENCE FOUNDATION
Directorate for Social, Behavioral
and Economic Sciences
NSF 99-310 December 2, 1998
by Mark C. Regets
Has the Use of Postdocs Changed?
Both the percentage of S&E Ph.D.s ever in a postdoc and the median length of time spent in postdocs have risen over time.
Only 17 percent of 1995 postdocs reported that they were in postdocs because other employment was not available.
A postdoctoral appointment, or "postdoc," has traditionally been defined as a temporary position, after completion of a doctorate, taken primarily for additional training-a period of advanced professional apprenticeship. More recently, however, there are reports that new Ph.D.s are turning to postdocs not primarily for training, but as a form of low-pay employment while waiting for a more permanent professional position. This issue brief examines the self-reported postdoc histories of holders of science and engineering Ph.D.s from U.S. schools to address the question whether the use of postdocs has changed.
Postdocs Are an Increasingly Common Experience
Also shown in figure 1 are the postdoc rates for Ph.D. cohorts in the two fields with the highest rate of postdoc use, biological sciences and physics, as well as the aggregate for all engineering fields. Both the biological sciences and physics showed robust increases of postdoc use since the 1965-66 Ph.D. cohort: biological sciences from 51-72 percent, physics from 50-69 percent. Postdocs are notably gaining in importance for Ph.D.s in engineering fields. Postdoc use for engineering Ph.D.s went from 12 percent of 1965-66 Ph.D.s to 31 percent of the 1993-94 cohort. Seven fields, which together accounted for over 90 percent of all S&E postdoctorates in 1995, all showed a similar pattern of increasing postdoc use (table 1).
Is Increased Use of Postdocs Due to Labor Market Conditions?
More Time is Spent in Postdoctoral Appointments
The increase in median length of postdoc experience was greatest in the biological sciences-rising from 24 months for the pre-1965 cohort to 46 months for the 1989-91 cohort. However, for some fields-agricultural sciences, chemistry, and psychology-median postdoc length stayed roughly the same or even declined somewhat for Ph.D.s graduating after 1965. Postdocs in engineering fields remain relatively short, reaching only 17 months for the 1985-88 cohort.
Postdoc Positions For Want of More Desirable Employment?
Trends in the use of postdocs as employment-of-last-resort differ greatly by field. In physics, it has varied greatly over time with approximate peaks for the 1971-73 (25 percent) and 1989-91 (29 percent) graduation cohorts. But in the biological sciences, "other employment not available" was reported by only around 10 percent of each graduation cohort over the last 20 years. In the other major postdoc fields (agricultural science; earth, atmospheric and ocean sciences; and in psychology) the highest reported rates occurred in the late 1970's or early 1980's.
Most respondents report reasons for accepting postdoctoral appointment that are consistent with the more explicit purposes of the position, such as advanced training in their field, training outside their field, or working with a specific person. On the other hand, reporting that no other job was available may be a difficult step for some respondents. The data reported here suggest, however, that the rise in the number of postdoctoral appointments is not mainly due to insufficient other employment opportunities.
This Issue Brief was prepared by:
SRS data are available through the World Wide Web (http://www.nsf.gov/statistics/). For more information about obtaining reports, contact email@example.com. or call (301) 947-2722. For NSF's Telephonic Device for the Deaf, dial (703) 306-0090. In your request, include the NSF publication number and title, your name, and a complete mailing address.
 Across all S&E fields, the median postdoc salary for recent Ph.D.s was $28,000, exactly half the median salary of recent S&E Ph.D.s in industry and almost one-third less than for those in tenure track positions.
 Estimated standard errors on the percentage of total postdocs reporting "other employment not available" ranged from 0.6 to 1.3 percentage points for the various graduation cohorts. For biological sciences the standard error ranged from 0.7 to 1.4 percentage points, and for physics, from 3.0 to 4.4 percentage points.