Where Do All the STEM Graduates Go? Higher Education, the Labour Market and Career Trajectories in the UK

Abstract

Problems with the supply of highly skilled science, technology, engineering and mathematics (STEM) workers have been reported by employers and governments for many decades, in the UK, the USA, and elsewhere. This paper presents some key findings from a project funded by the Nuffield Foundation that examined patterns of education and employment among STEM graduates in the UK. Five large-scale secondary datasets—comprising administrative, survey, cross-sectional and longitudinal data—were analysed in order to provide the most comprehensive account possible. The findings suggest that there is no overall shortage of STEM graduates but there is considerable variation in the career outcomes and trajectories of different groups. Recruitment to STEM degrees has stalled over the past 20 years but most STEM graduates never work in highly skilled STEM jobs—in any case, the majority of professional STEM workers do not have (or presumably need) degrees. Some groups of STEM graduates are currently under-represented in the highly skilled STEM workforce and increased recruitment from these groups could grow the numbers entering STEM occupations. However, employers may have to modify their views on exactly what constitutes a valuable or desirable employee and to what extent it is their responsibility to train their workers.

Appendix

The five datasets that we draw upon in this paper are listed below. While we have done our best to summarise the key features of our data, we encourage the interested reader to seek out a more developed account in our other publications that are cited throughout the paper.

• Universities and Colleges Admissions Service (UCAS) data

• Destinations of Leavers from Higher Education data (DLHE) data

• The Annual Population Survey (APS)

• The 1958 National Child Development Study (NCDS58)

• The 1970 British Birth Cohort Study (BCS70)

Any student who wishes to study an undergraduate (bachelors) programme at a UK university has to make their application through UCAS. UCAS releases annual data on applications and acceptances to university. These data provide information on a number of applicant characteristics including sex, institution and subject studied. Data from 1988 to 2012 were included in this analysis. Further information on these data, and their potential limitations, can be found in Smith (2010) and Gorard (2008).

DLHE is an annual graduate destination survey administered by the Higher Education Statistical Agency. The DLHE survey gathers information on the activities of graduates six months after they graduate. Response rates tend to be relatively high; for the physical sciences they are over 80%. This dataset only considers destinations at six months after the student has left university, and it is recognised that career trajectories may be very different in the subsequent period. However, it tells us a great deal about the sorts of jobs that are immediately available to STEM graduates. DLHE data from 1992 to 2012 were included in our analyses.

The APS provided us with a detailed cross-sectional snapshot of the occupational status of STEM graduates in the labour market. These data contain information on the occupational status of respondents between the ages of 25 to 64 years. APS data from 2004, 2006, 2008 and 2010 were included in our analyses. As there were no substantial differences between these four years, the findings presented here are for a combined dataset of 803,634 cases.

The BCS70 follows the lives of around 17,000 individuals who were born in Great Britain in one week in April 1970, and who were aged in their mid-40s at the time of this research. Since the birth survey in 1970, there have been eight ‘sweeps’ that have gathered data on cohort members’ health and physical, social and educational development, as well as their economic, personal and occupational circumstances (CLS 2016). Data from five sweeps of the study were used in this analysis and provide a detailed account of the cohort members’ career trajectories at ages 26, 30, 34, 38 and 42. The BCS70 provides the best available source of data on the career trajectories of a representative sample of the British population. As with most longitudinal studies, there are some issues with data quality, particularly with dropout and non-response. However, response rates for all sweeps of the BCS70 are around 75%.

The NCDS58 follows the lives of a cohort of individuals who were born in March 1958. To date, there have been nine ‘sweeps’ of the NCDS, with the latest sweep taking place in 2012/13 when participants were 54 years old. Data were analysed at six points, starting in 1981 and captured the first employment destinations of the NCDS graduates as well as career patterns in the intervening decades. For reasons of space, and because most of the findings for the BCS70 were similar to those for the NCDS, this paper emphasises the results from the former study.