Keith Crank, ASA Research and Graduate Education Manager

Last month, I wrote about The Path Forward: The Future of Graduate Education in the United States, a recent report from the Educational Testing Service (ETS) and the Council of Graduate Schools (CGS). I expressed my disappointment with the report, especially with its authors’ failure to use data appropriately. This month, I want to bring up one more issue with the data in the report and relate it to what’s going on in statistics and biostatistics.

On Page 14, the authors write that there was a 47% increase in the number of master’s degrees awarded and a 26% increase in the number of PhDs awarded between 1995 and 2005. The authors go on to say (Page 17) that, between 2008 and 2018, there is a projected increase of 18% in the number of jobs requiring a master’s degree and 17% in the number of jobs requiring a PhD. So, if the number of people earning (master’s) degrees is increasing at an annual rate of more than 4% and the number of jobs for them is increasing at a rate of less than 2%, why should we be putting more resources here?

In terms of actual numbers, rather than percentages, the authors note (Page 14) that more than 650,000 degrees were awarded at the master’s and doctoral levels, combined, in 2007. They go on to write that, between 2008 and 2018, projections indicate there will be an additional 2.5 million jobs that require either a master’s or doctoral degree. If the number of degrees awarded stays constant over this time period, there will be 6 million new recipients of these graduate degrees (assuming everyone who earns a PhD was also awarded a master’s degree). This means 3.5 million people would need to be absorbed into jobs that currently exist. While this may be feasible, the authors do not comment about the number of current jobs that are unfilled or become vacant due to retirement or attrition each year.

In statistics and biostatistics, the growth in degrees awarded is even more extreme. As I wrote in my January 2008 Amstat News column, the growth in master’s degrees awarded between 2000 and 2005 was more than 60% (readers pointed out the growth was actually more than 70%). Similarly, for PhDs, the growth over that time period exceeded 40%. The Bureau of Labor Statistics (BLS) projects growth in the number of statistics jobs (which includes biostatistics) at only 13%, so why do I still think we need to increase the size of our graduate statistics and biostatistics programs?

First, production has not caught up with demand. As far as I can tell, graduate students in statistics and biostatistics, at both the master’s and doctoral levels, do not have much difficulty finding jobs. And even in the current economy, salaries appear to be at least stable.

Second, I don’t think the BLS numbers capture the right information. The job title is one area in which BLS misses a large number of statisticians and biostatisticians. As an example, less than two-thirds of the job titles for nonfaculty academic biostatisticians included the word “statistician” or “biostatistician” in the 2009 salary survey of biostatisticians. Those without the word statistician or biostatistician are jobs BLS would not classify under the statistician category. The statistician category does not include statistics and biostatistics faculty, either.

Another area in which BLS would underestimate demand for statisticians (or workers more generally) is in areas that don’t exist now, but will appear and grow rapidly between now and 2018. (An example over the past decade is the Internet and search engines.) Finally, companies with headquarters outside the United States, who add offices and jobs here, also are likely to be missed in the BLS projections.

I believe our discipline (statistics and biostatistics) is still one of growing demand. Even though the supply of statisticians has grown rapidly over the past 10 years, it has not caught up with demand yet. While the same may be true of graduate education more generally, the report from the ETS and CGS did not provide cogent arguments for why this should be the case.

To contact me, send an email to keith@amstat.org. Questions or comments about this article, as well as suggestions for future articles, are always welcome.