More data needed on the STEM ‘shortage’

(Courtesy: iStock/geopaul)

By Margaret Harris

“Science has always been the Cinderella amongst the subjects taught in schools…not for the first time our educational conscience has been stung by the thought that we are as a nation neglecting science.”

Sounds like something David Cameron or Barack Obama might have said last week, right? Wrong. In fact, it comes from a report by the grandly named Committee to Enquire into the Position of Natural Sciences in the Educational System of Great Britain, which presented its findings clear back in…1918.

I came across this quotation thanks to Emma Smith and Patrick White, a pair of education researchers at the University of Leicester who have spent the past few years studying the long-term career paths of people with degrees in science, technology, engineering and mathematics (STEM). Smith and White presented the preliminary findings of their study at a seminar in Leicester yesterday, and one of the themes of their presentation – reflected in the above quote – was the longevity of concerns about a shortage of STEM-trained people, especially university graduates. As Smith pointed out, worries about the number and quality of STEM graduates are not new and, historically, reports of a “STEM crisis” have been as much about politics as they have economic supply and demand.

Smith and White’s results pose some tough questions for modern-day proponents of a “STEM shortage”. In particular, they found that:

• The career destinations of STEM and non-STEM graduates are not radically different, and a degree in a STEM subject does not guarantee students a better-than-average chance of getting a job;

• Only a minority of STEM graduates go into jobs that require a high level of STEM skills, and over the course of their careers, the proportion falls as members of the cohort leave to take up non-STEM roles;

• STEM graduates whose first job is outside STEM are very unlikely to move into a STEM field later in their careers.

The full report hasn’t been published yet, but in their talk, Smith and White gave some tantalizing details. One is that STEM subjects are not monolithic. Although there was little overall difference between the proportions of STEM and non-STEM graduates entering “graduate-level” jobs, there were significant variations within STEM. For example, engineers are more likely than their social-science counterparts to obtain graduate jobs, but biologists do worse, while physical-science graduates are somewhere in the middle. The researchers also found some major institutional differences. Overall, slightly less than half of engineering graduates actually become professional engineers, but the percentage is much higher among those who attend universities in the Russell Group (a collection of 24 research-intensive UK universities, including Oxford and Cambridge).

Smith and White’s findings are based on three main sources of data. The first is the UK Higher Education Statistical Agency (HESA) “First Destination” survey, which gathers employment data from all students six months after they have left university. This survey has a very high response rate (around 80%) and covers the years between 1994 and 2011 (it has continued since then, but in a different format, which makes recent results hard to compare with earlier data). The second source of data is the 1970 British Cohort Study, which follows the lives of 17,000 people who were born in a certain week of April 1970. Smith and White looked at employment data from a series of five “sweeps” taken when members of this cohort were aged between 26 and 42. The final data source was the National Child Development Study, which gathers similar employment data as the 1970 study, but for a cohort born slightly earlier, in March 1958.

As rich as these data sets are, they have some frustrating limitations. One drawback is that the most comprehensive source, the HESA survey, doesn’t tell us about the eventual career destinations of students who go on to do Master’s degrees or PhDs. It only covers what graduates are doing six months after they finish their undergraduate courses. People with higher degrees are included in the two cohort studies, but their numbers are tiny; White told me that depending on how you “slice” the data, you could end up with only one or two higher-degree individuals in your sample. That makes it really hard to say anything about whether students with advanced STEM degrees are more likely to obtain (and then stay in) jobs that require their skills. It’s also difficult to tell what’s happening to STEM graduates who take STEM jobs and then leave the field later on; Smith and White strongly suspect that many of them are going into management, but without analysing the data on an individual level, they can’t say for certain.

Perhaps the biggest gap, though, is data about what STEM graduates actually want to do. If half of them are taking non-STEM jobs after graduation because they’ve decided their interests lie elsewhere, or because they think they can make more money doing finance or whatever, then in my book, that’s a “win”. It’s good to have scientifically literate people in all walks of life, and if their new roles make them happy, I certainly wouldn’t want to stop them. But if they’re leaving because they can’t find a good job that actually uses their skills, then that’s a problem. By the end of the seminar, the words going through my head weren’t those of the 1918 committee, but those of Arthur Conan Doyle, who had his creation Sherlock Holmes cry, “Data! Data! Data! I can’t make bricks without clay.”