Assessing Demand for PhD Statisticians and Biostatisticians

Back to Blog List

Assessing Demand for PhD Statisticians and Biostatisticians

By Steve Pierson posted 04-14-2016 08:43

Recommend

[7/18/16 update: The 2015 new doctorate unemployment data were added to the first table from the Report on the 2014–2015 New Doctoral Recipients.]

We recently received a couple inquiries on the demand for PhD statisticians and biostatisticians. A new piece in Inside Higher Education (IHE), The Shrinking Ph.D. Job Market, also addresses the broader market for PhDs and so I thought it would be helpful to share what we know on the statistics and biostatistics PhD market. I think it is true that the IHE piece's summarizing statement—“As number of new Ph.D.s rises, the percentage of people earning a doctorate without a job waiting for them is up”—doesn't hold for statistics and biostatistics.

Thanks to the Annual Survey of the Mathematical Sciences by the American Mathematical Society, we know that at least for new doctoral recipients, the unemployment rate for statistics and biostatistics doctoral recipients is less than that of mathematical sciences overall, as shown in this table:

	Mathematical Sciences	Statistics & Biostatistics
2008	3.8%	2.0%
2009	4.9%	2.9%
2010	6.9%	2.7%
2011	4.3%	4.2%
2012	6.9%	4.0%
2013	5.7%	2.8%
2014	6.2%	2.8%
2015	7.3%	3.1%

The employment data are from the fall following the previous academic year's graduates (e.g., the graduates receiving a PhD in mathematical sciences during the period July 1, 2012, through June 30 were asked about their employment plans in fall 2013.) The AMS states the following about its unemployment calculations: “the individuals employed outside the U.S. have been removed from the denominator used in the calculation of the rate, in addition to the routine removal of all individuals whose employment status is unknown.” For 2014, the biostatistics data was presented separately and showed an unemployment rate of 1.8% (2/112). I should also note that in the AMS surveys, they have a statistics and biostatistics section where they look at statistics and biostatistics degrees granted by statistics and biostatistics departments. This group covers a large fraction of the overall number of such degrees granted but generally has a lower unemployment rate.

The number of PhD recipients going into postdoctoral positions is sometimes considered as an indicator for PhD demand though there are many other factors affecting the number of PhD recipients going onto postdoc for a field. The NSF data on the number of postdocs in mathematical sciences is in the following table [follow the prior link for an update on these data]:

	2009	2010	2011	2012	2013
Mathematical Sciences	737	756	805	902	932
Mathematics/Applied Math	675	680	706	782	787
Statistics	62	76	99	120	145

For comparison, the number of PhD’s awarded in Math/Applied Math was 1122 in 2009 and 1270 in 2013; for statistics, 483 in 2009 and 573 in 2013 (according to the AMS survey.) This gives ratios of postdocs to PhDs for Math/Applied Math of 60% in 2009 and 62% in 2013; for statistics, 13% in 2009 and 25% in 2013.

The following statement from the 2013 AMS annual survey is also informative (and similar for other years): "46% of Statistics/Biostatistics Ph.D.’s are employed in Business/Industry, compared to 23% in all other groups."

For rough comparison, the March 2016 Physics Doctorates Initial Employment from the American Institute of Physics (AIP) reports a 4% unemployment rate one year after degree for the classes of 2013 and 2014. The document also reports, “Almost half of PhD recipients from the classes of 2013 and 2014 combined were in postdoctoral fellowships after receiving their degrees .” According to another AIP report, “The proportion of astronomy PhDs accepting postdocs has risen over the last three decades. Three-quarters of new astronomy PhD recipients in the classes of 2010, 2011 and 2012 combined accepted a temporary postdoctoral appointment after receiving their degrees.” More information on physics PhD employment is available at this excellent AIP website.

For PhD Computer Scientists, the 2014 Taulbee Survey from the Computing Research Association states, “The unemployment rate for new Ph.D.s again this year was below one percent.” [The 2016 Taulbee Survey states: "The unemployment rate for new Ph.D.s again this year was below 1 percent."] On postdocs, the same CRA report, reports, “When academic and industry postdocs are combined, the result is that 15.6 percent of 2013- 14 doctoral graduates took some type of postdoctoral position, down from 18.1 percent last year. Approximately 14 percent of these were industry postdocs, a slightly higher fraction than was reported last year.”

For PhD chemists, I found a 2012 report, Advancing Graduate Education in the Chemical Sciences, that included this summary from an American Chemical Society report: “New PhDs in chemistry were unemployed, but seeking employment, at about a 4% rate before 2008, but that rate had reached 9% in 2011. Of those employed, about 40% were in postdoctoral appointments before 2008, but in 2011 the postdoctoral fraction accounted for 47%. About half of new PhDs reported that they were in full-time permanent employment in 2008, but in 2011 only a third so reported.”

More generally, according to an April 2014 NSF InfoBrief, the unemployment rate for the estimated 805,500 individuals in the United States who held research doctoral degrees in science, engineering, and health (SEH) fields was “2.4% in October 2010, up from 1.7% in October 2008 and similar to the rate in October 2003… Moreover, the 2010 unemployment rate of the SEH doctoral labor force was about one-third of the October 2010 unemployment rate for the general population aged 25 years or older (8.2%)." See their table:

Karl Broman recently considered the issue for biomedical data science: Job opportunities in biomedical data science. Among his resources are the many reports rating degrees in statistics highly, including some of the following that This is Statistics also covers:

Blog Viewer