Student employment: social differentials and field-specific developments in higher education

Abstract

In this article, we examine social origin differences in employment patterns across different stages of higher education and compare these differences between vocational and academic fields of study. Using data from a large-scale German student survey, we study the development of inequality, according to social origins, in student employment from first-year to graduating students. We show that inequality in job quality exists and is partly attributable to the need for students from lower social origins to work in order to finance their studies. We hypothesise that initial inequalities decrease as students progress through higher education. While we find evidence for this hypothesis, we also show in multivariate models that the reduction of inequality in the student labour market is explained by prior differences between social origin groups.

Keywords:

• Higher education
• social origin
• Germany
• term time employment
• educational inequality

Notes

1. Additionally, we assume that job quality increases somewhat for all students.

2. See verdict of the Federal Social Security Court (Bundessozialgericht) 11.11.2003 – B 12 KR 24/03 R. There are some exceptions, eg working mainly in the evening or weekends. During the semester break students are allowed to work more than 20 h per week, however, at a maximum of 26 week per year.

3. More recent waves were unsuitable for our research question because detailed questions regarding students’ job quality were not asked. Moreover, we favour the 2000 data compared to later surveys (e.g. in 2006 or 2009) since two major structural changes in higher education occurred between 2000 and 2010. First, the introduction of the Bologna reforms in Germany began in 1999, gradually replacing old diploma programmes over a longer period with Bachelor/Masters programmes and resulting in greater heterogeneity of students in various programmes with different demands and durations. However, until summer term 2000 only 183 study programmes (123 Bachelor, 60 Master) had been introduced resp. changed to the new system, (estimated) 2% of all study programmes offered (HRK 2011, 7). On the level of individual students, the replacement took long and neared completion only in 2015, and the share of students in the new programmes in 2000 was negligibly small (cf. Statistisches Bundesamt 2016, 10). Second, several federal states (Bundesländer) introduced and abolished tuition fees in different years, see Baier and Helbig (2014, 100) for an overview, which may have affected students’ security about their financial situation. In the year 2000, there were no tuition fees yet in Germany. Third, with the introduction of the Bachelor-/Master scheme a growing number of students interrupted their educational career and gathered work experience, which leads to more heterogeneity in Master programmes regarding work experience and previous labour market activity.

4. Many students leave higher education after the fifth year as graduates, and thus we have to expect extremely selective departure from our sample. For example in 2003, the average study duration in long university programmes was 12 semesters (Statistisches Bundesamt 2003).

5. We assign code ‚0’ for working more than 20 h and ‚1’ for working less 20 h in order to improve comparability of our findings across the different indicators indicating ‚improvement’ of job quality over time.

6. We do not consider research assistants when analysing field-specific inequality because the applicability of skills developed during higher education should not differ between occupation-oriented fields and arts and sciences.

7. For the analyses by field of study, we collapse university diploma and church degree into one category.

8. Results for binary outcomes do not differ substantially when estimated using OLS regressions and are available from the authors upon request.

9. For the full models see Table A3 in the Appendix 1.

10. Figure 6 visualises the results. The full model in the Appendix 1 shows the contribution and significance for each category of each motive separately.