HIGHLY EDUCATED BUT IN THE WRONG FIELD?: Educational specialisation and labour market risks of men and women in Spain and Germany Open Access

Two major trends in higher education have emerged during the past decades in all EU Member States: first, tertiary education has expanded enormously, and second, women's educational attainment has reached parity with men's (OECD 2004). Nevertheless, gender-specific stratification processes are still evident in the field- and institutional-specific differentiation of men and women within tertiary education, often leading to unequal labour market outcomes (Charles and Bradley 2002). In this context, most studies have focused on the role of the field of study or college major to explain gender differences in earnings (Daymont and Andrisani 1984; Kalmijn and van der Lippe 1997; Bobbitt-Zeher 2007), while only a few have analysed labour market outcomes other than income (Kim and Kim 2003; van de Werfhorst 2004; Smyth 2005). The consideration of non-monetary labour market outcomes seems central because almost all EU countries have been confronted with deregulation and flexibility processes leading to fragmented labour market integration and working careers (Esping-Anderson and Regini 2000). Furthermore, it is also important to consider cross-national differences in education and training systems, which may partly be responsible for gender-specific educational returns.

The paper adds to the literature on the relationship between gender, tertiary education, and the labour market by inspecting two previously neglected labour market outcomes: the likelihood of being unemployed and that of holding a low-status job. To examine how far educational system characteristics translate gender inequalities differently into the labour market, we chose to compare Germany and Spain because they vary considerably in the nexus between educational qualifications and the labour market. Furthermore, both countries differ with respect to tertiary level expansion, as well as to the relative share of female tertiary graduates. Finally, the distribution of men and women across fields of study and different types of tertiary degrees also varies between Germany and Spain, which might influence gender-specific labour market outcomes in both countries.

In the next section we discuss important characteristics of the respective institutional framework of Spain and Germany. This is followed by the theoretical background and a set of hypotheses concerning the role of gender and educational differentiation for labour market achievements in the selected countries. We then introduce the data and methods we use for our analyses, present the results and, finally, summarise and discuss our findings.

An extensive number of sociological studies have established that countries differ in how they match the output of the educational system to the demands of the labour market (Maurice et al.1986; Allmendinger 1989; Shavit and Müller 1998). However, the aforementioned literature does not explore how and in which way differences in the field of study or the type of tertiary degree should influence gender-specific labour market outcomes. Moreover, this link is not explicitly addressed from a cross-national perspective.

In this context, the German education system can be characterised by a high vocational orientation and a strong connection between education and labour market outcomes. This is also mirrored in the institutional differentiation of tertiary education: Germany represents a binary tertiary system consisting of universities of applied sciences (‘Fachhochschulen’) and traditional research universities. Particularly universities of applied sciences offer specialisation in a few subjects with a more practical focus and provide students with work-oriented training (Müller et al.2002). Moreover, returns for lower-level tertiary degrees, earned in a shorter time (from universities of applied sciences), are somewhat lower than those for higher-level tertiary degrees (from universities). In Spain, however, educational institutions are less vocationally oriented and educational credentials – particularly tertiary qualifications – structure the transition to the labour market less clearly (Iannelli and Soro-Bonmatí 2003). This may be due to the fact that higher education has developed into a sequential system that differentiates between three cycles and four different types of university institutions as well as a broad range of non-university establishments. It is also oriented towards rather flexible curricula and programmes, which sometimes make it hard for employers to evaluate the skills a graduate has gained. With respect to educational returns, short-cycle courses, offered at so-called ‘escuelas’, similarly lead to returns lower than those for long-cycle courses (Mora et al.2000).

A further important aspect is that in Germany the demand for more qualified labour has been partly absorbed by the vocational training system below the tertiary level, while in Spain the expansion took place mainly in tertiary education. The share of persons aged 25–34 with a tertiary degree increased from 19.6 percent in 1991 to only 22.3 percent in 2000 in Germany, whereas that share more than doubled (from 16.3 to 34.1 percent) in Spain during the same period. This development was due in part to higher rates of female participation, most visible in Spain, where the share of female tertiary graduates had surpassed that of men already in the early 1990s, as opposed to Germany, where even in the year 2000, women still lagged behind men (OECD 2004). Furthermore, higher female participation rates are accompanied in Spain by an overrepresentation of women with lower level tertiary degrees, while their German counterparts are overrepresented in the traditional research universities (also see Table 1, next section). With respect to the unequal distribution of women and men across fields of study, the Duncan Dissimilarity Index reveals that in Germany 32 percent of all females and males would have to ‘switch’ fields in order to achieve gender parity across fields, compared to only 24 percent in Spain.¹

Finally, the selection process for a specific field of study or type of degree should be placed in a broader institutional context. Particularly for women, considerations related to future perceived labour market risks and support in terms of opportunities to reconcile work and family are crucial for educational decisions. In this respect, both countries represent different ‘welfare regimes’ that give women different opportunities to combine work and family tasks. Germany is often classified as a conservative, family-oriented welfare state where women's role is primarily seen as that of caregiver and part-time income earner. Whereas in Spain, the ‘familialistic’ welfare regime provides women with weak public support for reconciliation of work and family, limited opportunities for part-time employment, and strong cultural barriers to mothers’ working (Esping-Andersen 1999).

Several explanations of gender-specific educational and occupational choices, such as gender-specific socialisation, future family responsibilities, and anticipated discrimination practices on the labour market have been discussed extensively. However, the main focus of this article is not to explain why women choose certain fields of study and/or occupations, but how the gender differentiation in the achievement of tertiary degrees translates into different labour market risks for men and for women in two distinct institutional contexts.

According to assumptions of human capital theory (Becker 1991), gender inequality in the labour market should be reduced when gender-specific differences in human capital achievement are reduced. However, increasing women's educational attainment may have little impact on gender differences in employment if women are more likely than men to choose fields and institutions associated with higher labour market risks. Human capital investment consequently should reflect both a vertical (degree levels) and a horizontal dimension (the chosen field of study) of tertiary qualifications. Based on the fact that previous research (Isengard 2003; Wolbers 2003; Azmat et al.2006) has shown that women are generally disadvantaged as to the selected labour market outcomes, we hypothesise that the expected female disadvantage in terms of unemployment or a low-status job will be at least partly mediated by gender differences in the distribution across different fields of study as well as different types of tertiary degrees(Hypothesis 1).

A further strand of research deals with consequences of gender-atypical occupational choices. In this context, we particularly focus on the question of whether or not women with a degree in a male-dominated field of study are disadvantaged compared to men.² While some scholars (e.g., Hayes 1986) assume that women who choose male-dominated occupations thus increase their opportunities for higher pay and career advancement, others (Blalock 1967; Kanter 1977) indicate that minorities, such as women in predominantly male organisational settings, are not perceived as individuals but rather as ‘tokens’ of their category. Traditionally privileged (male) majorities may feel threatened by the (female) minority, which is therefore subject to various kinds of hostile behaviour. Moreover, research suggests many jobs continue to be ‘gendered’ and beliefs persist that women are not ‘suitable’ for certain male occupations (Reskin and Roos 1990; Whittock 2002; Hultin 2003). This argument can also be applied to the field of study: women in typically male fields of study constitute a minority and might therefore be perceived as being inadequate for typically male jobs and thus face discrimination. Especially at the stage where jobs are allocated to men and women, beliefs or prejudices regarding the performance of women might be prevalent and discrimination relatively easy to implement (e.g., Petersen and Saporta 2004). Therefore, we predict that the likelihood of unemployment or employment in a low-status job will be higher for women in typical male fields than for men in the same fields (Hypothesis 2).

While we expect to find evidence for our hypothesis in both countries, the described characteristics of the German and Spanish education system are likely to mediate gender-specific returns to tertiary education. With respect to our first hypothesis, we assume that gender differences according to the field of study should be more consequential in Germany, given the high vocational (skill-specific) orientation and the strong association between the educational system and the labour market. Yet in Spain, gender differentiation in tertiary education may also be consequential for our labour market outcomes due to the high share of tertiary-degree holders (the lesser exclusiveness of tertiary education), which might increase the relevance of the field of study and the degree earned as an ‘educational signal’ (e.g., Reimer et al.2008). Furthermore, in Spain female disadvantages might be particularly related to women's overrepresentation among holders of lower status tertiary degrees.

With regard to the second hypothesis, we expect that in countries with a weaker tie between the skills obtained in the education system and the labour market, women in typically male fields will by be less constrained by their choice of field. They should have more ‘switching opportunities’ to other jobs, which might lower their risks of becoming unemployed or obtaining a low-status job. Consequently, possible disadvantages for women with a degree in typically male fields might be more marked in Germany than in Spain. Furthermore, in both countries typically male occupations offer limited opportunities for combining work and family, which might also influence gender-specific outcomes in different fields. For Spanish women, work-family considerations are typically related to the decision to stay employed or to leave (full-time) employment during the care period (Cousins 1999). Thus, a degree in a male-dominated field might not necessarily lead to higher labour market risks because women with children either withdraw from the labour market entirely or continue to work full-time. In Germany however, female graduates in a male field who wish to work part-time might not be able to find employment of this kind or might have to accept lower-status positions.

Finally, it should be emphasised that even though we will interpret our results in light of these plausible underlying mechanisms, our research design does not allow for a direct test.

For our analyses we use data from the Spanish and German Labour Force Survey for the year 2000. These surveys provide detailed information on the social and economic situation as well as on the educational achievements of the population in each country. We restrict our sample to the economically active population and exclude foreigners and students.³ In order to reduce problems due to unobserved differences between men and women in their respective labour market careers, only tertiary graduates aged 20–35 are included in the analyses. We chose to include respondents as young as 20 because several graduates in the Spanish sample had completed the first cycle of tertiary education by that age.

In order to test our assumptions, we first examine the effect of gender and fields of study on the risk of being unemployed, using the International Labour Organization's (ILO) standard definition of unemployment.⁴ The second labour market hazard is the probability of a person with a tertiary degree holding a ‘low-status job’. The low-status variable is based on the International Socio-Economic Index of Occupational Status (ISEI) developed by Ganzeboom et al. (1992). The ISEI consists of weighted averages of standardised measures for the income and education of incumbents of each occupation comprised by the ISCO88 classification. We dichotomised the resulting index: tertiary graduates with fewer than 50 points were classified as holding a low-status job, whereas graduates with 50 points and more were classified as being adequately employed.⁵ We use the same set of independent variables in all multivariate analyses. To avoid problems with a small N in certain fields, we coded our field variable in 12 categories. Our measure for the level of tertiary degree was coded differently in the two countries. For Germany, graduates of traditional universities are assigned to the higher-level tertiary degree group, whereas graduates of universities of applied sciences comprise the lower tertiary degree group. Due to the different structure of tertiary education in Spain, only graduates with higher-level university degrees (long-term courses of study lasting five to six years) and doctoral degrees are coded into the higher tertiary degree group. Finally, age was included as a control variable and as a proxy for potential work experience.⁶

Table 1 reports the means of the dependent and independent variables for both men and women in Germany and Spain. Clear gender-specific variations can be observed in both countries with respect to the examined labour market outcomes. In general, there is a relatively high level of unemployment among tertiary graduates in Spain, particularly among women. In Germany, female graduates are also more often unemployed than are male graduates, but the overall level of unemployment is much lower than in Spain. Spanish graduates also hold low-status jobs more often than German graduates. This may be partly related to the different Spanish occupational structure as well as to the fact that there are considerably more tertiary graduates in Spain than in Germany. Women in both countries are more likely than men to work in low-status jobs. This could be partly linked to the aforementioned difficulty of work-family compatibility in both countries that often leads women into part-time or precarious employment. Moreover, particularly in Spain, the generally high rate of youth unemployment might increase the willingness of female tertiary graduates to accept lower-status jobs.

The different attainment of tertiary degrees in Germany and Spain is also reflected in the composition of our analytical sample, where only 43 percent of all German respondents with tertiary degrees are female, as opposed to 59 percent in Spain. The distribution also shows that German women more often attended a university (higher-level degree) than a university of applied sciences (lower-level), which is most likely a result of the different curricula offered at higher- and lower-level tertiary institutions. In Spain, however, about half of all women in the sample have a higher-level tertiary degree, compared to 56 percent of all men. The table also reports country differences in the concentration of graduates across different fields. Taken together, 27 percent of all German graduates in the sample acquired a degree in either engineering, the natural sciences, or the life sciences, compared to only 7 percent in Spain. In Spain, on the other hand, more students graduated in law, economics, and the social sciences (40 percent). When considering the gender composition of each field of study (e.g., ‘row percentages’; not reported), in both countries the social sciences, education, liberal arts, and the fine arts are typically female fields because more than half of all graduates are female. In contrast, engineering and architecture are typically male fields, with a low number of female graduates. The life sciences, law, economics, and ‘other’ fields are considered to be integrated fields of study because the shares of men and women are nearly similar. Nonetheless, the share of women in these integrated fields is generally higher in Spain. With respect to country differences, medical science emerges as a typically female field of study in Spain, while it is an integrated field in Germany. Furthermore, the natural sciences can be considered a typically male field in Germany, but an integrated field in Spain.

In a first step, we estimate separately for Germany and Spain a set of nested logistic regression models for the risk of being unemployed. In model A, we include gender and age. In model B, we add the dummy variable for the higher tertiary degree in order to examine whether gender differences in unemployment can be explained by the different distribution of men and women over the level of tertiary degree. In model C, dummy variables for the field of study are introduced to test whether the gender-specific distribution over fields explains the expected female disadvantage.

The results in Table 2a reveal that in both countries women are more likely than men to be unemployed. With the introduction of the level of tertiary degree (model B), the gender coefficient does not change substantially, even though in Germany graduates of universities are more likely to be unemployed than are graduates of the universities of applied sciences. In contrast to our expectations, the overrepresentation of women among holders of lower tertiary degrees in Spain is not related to a higher risk of unemployment among females. However, the introduction of the field of study (model C) leads to a marked reduction of the gender coefficient (from 0.542 to 0.495 in Germany, and from 0.426 to 0.306 in Spain). Compared to respondents in the reference category (education), graduates in the fine arts and ‘other’ fields are significantly more likely to be unemployed in Germany, while in Spain this only holds true for graduates in the life sciences and in the liberal arts.

In order to provide a parsimonious test for our second hypothesis and to find out whether there are gender differences in the probability of being unemployed in typically male vs. typically female or integrated fields, we specify an additional model with interactions between gender and the level of tertiary degree, as well as between gender and the field of study. Therefore, we recode the field of study variable into integrated, male-dominated and female-dominated fields.⁷

Table 2b reveals that, as expected, in both countries women in male-dominated fields seem more likely than men to become unemployed. Furthermore, the effect for the type of tertiary degree does not differ between men and women in either Germany or Spain. Because the interpretation of interaction coefficients in nonlinear logit models can be problematic (Ai and Norton 2003), we plot differences in the predicted probabilities of becoming unemployed for men and women with a higher tertiary degree and mean age based on the model specified in Table 2b.

Figure 1 shows that in both countries regardless of the gender composition of fields women have a higher probability of becoming unemployed. However, this female disadvantage is most marked in the male-dominated fields, while differences between men and women in integrated and typically female fields are considerably lower.⁸ Furthermore, there is no indication that the female disadvantage in male fields is greater in Germany than in Spain. Taking into account the overall higher level of unemployment in Spain compared to Germany, the observed cross-national similarity in the pattern of female disadvantages across the three field groupings is remarkable.

In a next step, we turn to the multivariate analysis of low-status jobs (Table 3a), estimating the same set of models as in the previous analyses. Model A shows that while women in both countries are significantly more likely than men to hold a low-status job, the female disadvantage is more pronounced in Germany. Contrary to the analysis of unemployment, model B reveals that the addition of the type of tertiary degree significantly affects the status outcome. In both countries, graduates with a higher-level tertiary degree are less likely to enter low-status jobs. This is also reflected in the considerable increase in the model fit in Germany and Spain. In Germany, surprisingly, the main effect for gender increases slightly when the type of tertiary degree is added to the model. This implies that women, where they not overrepresented among holders of higher-level tertiary degrees, would be even more disadvantaged with respect to low-status jobs. The opposite effect can be observed in Spain, where the gender coefficient is marginally reduced upon the introduction of the level of tertiary degree.

The introduction of the field of study in model C leads to a similar reduction of the gender main effect in both countries. However, the effect for females loses its statistical significance in the case of Spain. In Germany, controlling for other variables, graduates in the typically male fields (natural sciences, engineering, and architecture), as well as those in the more gender-integrated fields (medical sciences and law), are less likely to hold a low-status job than are education graduates. Liberal arts and ‘other’ graduates, however, are more likely to be in a low-status job. The disadvantage of economics graduates is somewhat surprising given their supposedly good labour market prospects. The Spanish case shows that graduates in law, economics, the social sciences and liberal arts are disadvantaged compared to education graduates, whereas only medical science and architecture graduates are at an advantage with respect to low-status jobs.

As in the previous analysis, an additional model with interactions between gender and the level of tertiary degree and the collapsed field of study variable is specified.

Contrary to our expectations, Table 3b indicates that neither in Germany nor in Spain do women in male-dominated fields seem to be more disadvantaged than men in these fields. Nor does the gender effect differ for the type of tertiary degree. However, it is interesting that in both countries the interaction between gender and female fields is significant, indicating that for women rather than men, obtaining a degree in a typically female field considerably reduces the risk of entering low-status employment. Particularly in Spain, women seem to fare considerably better than men in female-dominated fields. Figure 2 reports gender differences in the predicted probabilities to obtain a low-status job for persons with a higher tertiary degree and mean age based on the model specified in Table 2b.

The results reveal that the predicted probabilities to obtain a low-status job are noticeably higher for women than for men in integrated as well as in typically male fields in Germany and Spain. Moreover, in both countries the female disadvantage is highest in integrated fields. Some country differences can be observed with respect to typically female fields: While there are almost no gender differences in Germany, in Spain the predicted probabilities to enter low-status jobs are even considerably lower for women than for their male counterparts in these fields.⁹

For the application we use coefficient estimates from a pooled model (⁠$β^*$⁠) (e.g., Oaxaca and Ransom 1998) to calculate average predicted probabilities for outcome Y for both groups under consideration. Next, a random sub-sample of the group with a smaller N is drawn, equal to the size of the larger group.¹¹ Because the results of the decomposition depend on the sub-sample that is selected, it is advisable to draw a large number of random sub-samples in order to achieve reliable results.

Presented in Table 4 are results of the non-linear decomposition of unemployment for Germany and Spain, using the pooled sample of men and women.¹² The results are based on the mean values of decompositions with 1,000 different sub-samples.¹³

We are mainly interested in the respective contribution of the field of study and the higher tertiary degree to the gap. Overall, the differences in the average values of the independent variables account in Germany for about one-third of the raw gender gap in unemployment (31.2 percent), at a raw gender gap of only 2.2 percent, while in Spain the same differences account for about half of the gender gap in unemployment (52.8 percent), at a raw gender gap of 8.5 percent. Interestingly, in both countries only a very small and statistically not significant part of the gender gap can be attributed to gender differences with regard to a higher tertiary degree. The results also show that the field of study accounts for 22.7 percent of the raw gender gap in Germany, as opposed to 32.6 percent in Spain. In other words: the female disadvantage with respect to unemployment would be reduced by about one-third in Spain and by roughly one-fourth in Germany if the female distribution across fields of study were equivalent to the male distribution. Virtually no change in the female disadvantage in unemployment would occur if men and women were equally distributed with respect to the type of higher tertiary degree. It should also be noted that part of the gender gap, at least in Spain, is due to the fact that in the sample men are slightly older than women.

In Table 5, we present results of the decomposition of the gender gap in low-status positions. Differences in means in the independent variables explain 30.9 percent of the gender disparity in low-status jobs in Germany, and 84.4 percent of the disparity in Spain. In Germany, a higher tertiary degree explains only a very small proportion of the gender gap, whereas in Spain the level of tertiary degree explains 45.3 percent of the low-status gap. The field of study accounts for 23 percent of the gap in Germany, and for 25.1 percent of that in Spain.

Again, the gender-specific distribution across the fields of study does not explain more of the gender disparity in low-status jobs in Germany than in Spain. In both countries the difference in the distribution across fields of study is accountable for at least one-fourth of the gender gap. Contrary to the analysis of unemployment, however, the level of tertiary degree does play a significant role in explaining the gender gap in low-status employment in Spain – though not in Germany.¹⁴ In line with our expectations, the overrepresentation of women among holders of lower-level tertiary degrees in Spain translates into a disadvantage with respect to the status outcome, while German women do not seem to benefit from their overrepresentation in higher tertiary degrees.

The main purpose of this paper was to analyse how gender differences among tertiary graduates in the chosen field of study as well as those with regard to the level of degree earned affect the risk of being unemployed or holding a low-status job. Moreover, we wanted to examine whether women with a degree in a male-dominated field of study are disadvantaged compared to men. In order to assess whether characteristics of the educational system mediate the impact of gendered educational choices, we chose to compare Germany and Spain, two countries that differ with respect to the nexus between education and work, as well as tertiary level expansion and gender segregation across fields and types of tertiary degrees.

Regarding the risk of unemployment, our analyses confirm the first hypothesis that the female disadvantage is reduced substantially in both countries if the field of study is taken into account. However, it is not significantly influenced in either country by the level of the tertiary degree. These findings are also supported by the decomposition analyses, which revealed that gender differentiation across fields accounts for a substantial part of the male-female differential in unemployment in both countries. However, some cross-national differences can be observed: field of study explains about one-third of the gender gap in Spain, compared to about one-fourth of the same in Germany. This indicates that even though in Spain the education-labour-market-linkage is less close, fields seem to be more relevant for mediating male-female differentials than they are in Germany. As stated before, this could be related to the higher extent of tertiary level expansion in Spain, which might increase the relevance of field of study as an educational signal.

The analyses of unemployment also largely confirm our second hypothesis. In male-dominated fields it is women rather than men who seem to be considerably more exposed to the risk of unemployment, while gender differences in integrated and female-dominated fields are less pronounced. In contrast to our expectations regarding cross-national differences, in male-dominated fields the female disadvantage with respect to unemployment is not stronger in Germany. Thus, irrespective of the differences in the setup of the educational system and work-family compatibility, women with a degree in a male-dominated field face similar relative disadvantages in finding employment. In both countries, these disadvantages are possibly connected to still existing prejudices and institutionalised or informal barriers women encounter when entering typically male occupations and which are partly visible in established personnel practices, job descriptions, mobility ladders and the exclusion from informal mostly ‘male’ networks (Morrison and Von Glinow 1990; Ibarra 1993).

With regard to the second labour market outcome, our analyses revealed that the effect of the field of study seems to be the same in both Germany and Spain. It accounts for at least one-fourth of the gender gap in low-status jobs. Furthermore, graduates with higher-level tertiary degrees less often hold low-status jobs, even though the gender differential is influenced by this variable only in Spain. Thus, as anticipated, the overrepresentation of women among holders of lower-level tertiary degrees in Spain is partly responsible for the female disadvantages behind entry into low-status employment.

The expectations concerning our second hypothesis are only partly confirmed because differences between men and women in male-dominated fields are relatively small, particularly in Germany. In this respect, our assumption that labour market risks are higher for women than for men if they chose a male-dominated field of study is not supported. In combination with the findings from our first outcome, this might indicate that in both countries women in typically male fields of study face difficulties particularly with regard to entry into the labour market rather than entry into low-status employment. However, the more interesting finding is that opposed to men, for women a degree in a typically female field of study substantially reduces the risk of entering a low-status job. Judging from the predicted probabilities in Spain, men with a degree in a female-dominated field face even higher risks of holding a low-status job than do women, while in Germany, gender differences in female fields basically disappear. This indicates that female-dominated fields like education, which typically channel graduates into public-service occupations, enable women to enter occupational niches that offer adequate status positions and might also allow for a better work-family compatibility.

Overall, the analyses have demonstrated that differences in the chosen field of specialisation in tertiary education account for a sizable portion of the gender gap in unemployment and in low-status jobs in both countries considered. In addition, the level of tertiary degree also mediates gender differences in low-status employment in Spain. Thus, studies concerned with gender differences in non-monetary labour market outcomes that do not control for the horizontal differentiation of educational degrees might misattribute a considerable portion of possible female disadvantages to unobserved heterogeneity or to discrimination. However, depending on the labour market outcome, choosing a typically male field does not necessarily reduce female disadvantages. Another interesting finding is that German tertiary graduates, on average, fare much better on the labour market than do their Spanish counterparts. Average levels of unemployment and the proportion of graduates in low-status jobs are much lower in Germany than in Spain. This partly reflects the problem Southern European labour markets display in bringing school-leavers into employment, regardless of their educational level (e.g., Gangl 2003) and hints at additional labour market characteristics such as employment protection legislation and labour market segmentation (insider-outsider markets) that might be relevant when considering gendered educational choices. Finally, future research in this area will no doubt benefit from incorporating more country cases with a greater variance in the relevant institutional characteristics. Furthermore, recognising that unemployment and low-status employment might be only transitory states for tertiary graduates, longitudinal data with information on the field of study are needed in order to deepen our insights into the role of educational differentiation and labour market outcomes.

David Reimer is a researcher at the Mannheim Centre for European Social Research (MZES), University of Mannheim. His research interestes lie in the area of social stratification research with focus on inequalities in educational attainment, gender and ethnic discrimination and comparative methodology.

Stephanie Steinmetz works as a Post-doc researcher in the project ‘Improving web survey methodology for social and cultural research’ at the Erasmus University Rotterdam. Her main research interests are gender and ethnic inequalities, social stratification, research methods and comparative labour market research.