Gender differences in higher education in Germany: are women under- or overrepresented at university, and why?

Figures & data

Table 1. Descriptive statistics of variables by gender.

Variables	Men (N = 5,955)		Women (N = 5,941)		Difference mean (women minus men)^a
	Mean	SD	Mean	SD
A. Independent variables
Age at study start (17–25)	20.81	1.58	20.31	1.50	−0.51**
Parental education (1–9)	4.99	2.02	5.02	1.98	0.03
Vocational school (0–1)	0.13		0.07		−0.05**
Exam grade (1–6 highest)	4.60	0.65	4.80	0.62	0.20**
Study field
Humanities and arts (0–1)	0.09		0.30		0.22**
Social sciences, business, law (0–1)	0.27		0.33		0.06**
Science, math, computing (0–1)	0.22		0.17		−0.05**
Health and welfare (0–1)	0.06		0.10		0.05**
Engineering, manufacturing (0–1)	0.36		0.09		−0.27**
Job goal: Good remuneration (1–6)	4.70	1.00	4.60	0.97	−0.10**
Job goal: Good career chances (1–6)	4.95	0.97	4.75	1.03	−0.21**
Job goal: Pleasant working hours (1–6)	4.45	1.15	4.53	1.05	0.09**
Job goal: High job security (1–6)	4.66	1.18	4.86	1.08	0.20**
B. Dependent variable
Traditional university (0–1)	0.58		0.72		0.14**

Source: NEPS, Starting Cohort First Year Students; weighted data, own computations.

^a Test on statistical difference in mean between men and women (regression of variable on gender; logistic when 0–1, linear otherwise).

*p < 0.05. **p < 0.01.

Table 2. Logistic regressions of the likelihood to enter traditional university rather than university of applied sciences.

	Model 1	Model 2	Model 3	Model 4	Model 5
Female (ref. male)	0.608** (0.049) [0.137]	0.517** (0.051) [0.107]	−0.032 (0.059) [0.006]	−0.043 (0.062) [−0.007]	−0.050 (0.063) [−0.008]
Age at study start		−0.264** (0.017)	−0.268** (0.018)	−0.159** (0.020)	−0.156** (0.020)
Parental education		0.218** (0.013)	0.213** (0.014)	0.167** (0.015)	0.162** (0.015)
Vocational school				−1.592** (0.094)	−1.584** (0.094)
Exam grade (1–6 highest)				0.554** (0.048)	0.517** (0.049)
Study field
Humanities and arts			ref.	ref.	ref.
Social sciences, business, law			−2.089** (0.103)	−2.149** (0.108)	−2.079** (0.110)
Science, math, computing			−0.970** (0.116)	−0.970** (0.120)	−0.937** (0.121)
Health and welfare			−1.399** (0.128)	−1.547** (0.133)	−1.511** (0.134)
Engineering, manufacturing			−2.678** (0.110)	−2.649** (0.115)	−2.582** (0.117)
Job goal: Good remuneration					−0.104** (0.037)
Job goal: Good career chances					−0.126** (0.034)
Job goal: Pleasant working hours					−0.070* (0.028)
Job goal: High job security					0.016 (0.029)
Constant	0.330** (0.035)	4.766** (0.364)	6.786** (0.402)	2.415** (0.526)	3.855** (0.571)
N	11,896	11,896	11,896	11,896	11,896
pseudo R^2	0.016	0.081	0.184	0.230	0.235

Note: Standard error in parentheses, average marginal effect for gender in square brackets.

Source: NEPS, Starting Cohort First Year Students; weighted data, own computations.

*p < 0.05. **p < 0.01.

Table 3. Percentage of the gender difference in the choice for traditional university versus university of applied sciences explained by background characteristics (Fairlie decomposition).

	% Gender gap explained (detailed)^a	% Gender gap explained (summed)^a
Control variables		9.7%
Age at study start	10.8%
Parental education	−1.0%
Educational factors		89.7%
Vocational school	11.3%
Exam grade	9.7%
Humanities and arts	ref.
Social sciences, business, law	−14.8%
Science, math, computing	0.4%
Health and welfare	−5.7%
Engineering, manufacturing	88.7%
Job goals		5.0%
Good remuneration	2.1%
Good career chances	3.0%
Pleasant working hours	−0.3%
High job security	0.2%
Total explained	104.4%	104.4%

Source: NEPS, Starting Cohort First Year Students; weighted data, own computations.

^a Observed gender gap in choice of traditional university (mean women minus men) is 0.137. Explained gender gap is greater than 100% since gender effect becomes slightly negative (though not significantly) when all covariants are included (cf. Model 5, Table 2).

Figure 1. Share of female and male higher education students choosing traditional university by field of study (observed percentages).

Data and code availability

The data that support the findings of this study are available on request from the Leibniz Institute for Educational Trajectories (LIfBi; https://doi.org/10.5157/NEPS:SC5:14.0.0). The data are not publicly available because they could compromise the privacy of research participants, but can be requested at LIfBi.