Explaining Gender Grade Differentials: Evidence

Explaining Gender Grade Differentials: Evidence From
Germany
Bernhard Enzi∗
Extended Abstract. Do not cite or print without permission
Are there systematic differences in the way teachers grade their male and female students?
Experimental studies by Hanna and Linden (2009), Hinnerich, Hoeglin and Johannesson (2011a,
2011b) and Sprietsma (2013) have shown that boys and children with a migration background
tend to be graded worse conditional on same performance in the respective subject. Observational
studies like Burgess and Greaves (2013), Cornwell, Mustard and Van Parys (2013) and Lavy (2008)
suggest the presence of different grading patterns by a student’s observable characteristics as well.
Investigating grade differentials is important, because high school grades are according to Altonji
and Pierret (2001) highly correlated with wages at labor market entry. Hence, systematic differences
in grading schemes that are not caused by actual performance can induce wages that are not
reflecting productivity differences, but solely discrimination. Goldin, Katz and Kuziemko (2006)
have shown that there exist female advantages in the US school environment. By investigating
gender grade differentials in Germany, I want to research the question whether those findings have
external validity that may explain the gender role reversal in educational outcomes over the past
decade. To my knowledge no observational study exists so far for the German setting.
Using the rich data set of the German National Educational Panel Study (NEPS), I investigate
the correlation of 5th grade students’ genders and their grades in math and German. NEPS consists
of extensive questionnaires for students, parents, teachers and school principals that allow me to
control for many determining factors of grades. Besides basic characteristics (e.g. gender, age,
migration status and SES), it includes information about life satisfaction, leisure time activities and
self- and parent-reported behavioral specifics. Most importantly, it includes objective measures of
intelligence and performances in math and German.
OLS regression estimates of grades on a student’s gender and the above mentioned extensive set
of variables including the performance measures may not be biased by omitted student variables,
but are likely to be biased by omitted school and teacher factors if there’s non-random sorting of
students to their teachers and schools. Therefore I employ school and classroom fixed-effects (FE)
estimation techniques to account for these confounding factors.
I find, in contrast to previous results, indications of the presence of subject specific grading by
gender stereotypes. While girls are, conditional on all controls and a classroom FE, advantaged by
∗ Ifo Institute – Leibniz Institute for Economic Research at the University of Munich. Center for the Economics
of Education and Innovation. Poschingerstrasse 5, 81679 Muenchen. Email: [email protected]
1
12,8% of a SD in German, they are disadvantaged by 19,7% of a SD in math relative to boys. These
findings are robust to many different specifications. Investigating whether this can be explained
by heterogeneous teacher effects, I can also account for an unobserved student FE by taking the
first difference over math and German grades. However, none of the investigated teacher characteristics (e.g. migrant status, gender) can explain these gender differentials, implying that teachers
irrespective of their own background have similar gender stereotypes that affect their grading.
Accounting for unobserved student heterogeneities by taking the first difference of grades over
math and German, makes the identification of subject specific coefficients of a student’s gender
impossible. However, the difference of the effect of gender on math and German can still be
identified. Comparing our FE headline results that are probably not biased by the omission of any
major grade determinant with the estimate from a FD regression suggests that my estimates indeed
do not suffer from omitted variable bias.
References
[1] Joseph G. Altonji and Charles R. Pierret. Employer learning and statistical discrimination. The
Quarterly Journal of Economics, 116(1):313–350, 2001.
[2] Simon Burgess and Ellen Greaves. Test Scores, Subjective Assessment, and Stereotyping of
Ethni Minorities. Journal of Labor Economics, 31(3):535–576, 2013.
[3] Christopher Cornwell, David B. Mustard, and Jessica Van Parys. Noncognitive Skills and the
Gender Disparities in Test Scores and Teacher Assessments: Evidence from Primary School.
Journal of Human Resources, 48(1):236–264, 2013.
[4] Claudia Goldin, Lawrence F. Katz, and Ilyana Kuziemko. The Homecoming of American College
Women: The Reversal of the College Gender Gap. NBER Working Paper Series, 12139, 2006.
[5] Rema Hanna and Leigh Linden. Measuring discrimination in education. NBER Working Paper
Series, 15057, 2009.
[6] Björn Tyrefors Hinnerich, Erik Höglin, and Magnus Johannesson. Are boys discriminated in
Swedish high schools? Economics of Education Review, 2011.
[7] Björn Tyrefors Hinnerich, Erik Höglin, and Magnus Johannesson. Ethnic Discrimination in
High School Grading: Evidence from a Field Experiment. SSE/EFI Working Paper Series in
Economics and Finance, 733, 2011.
[8] Victor Lavy. Do gender stereotypes reduce girls’ or boys’ human capital outcomes? Evidence
from a natural experiment. Journal of Public Economics, 92(10-11):2083–2105, October 2008.
[9] Maresa Sprietsma. Discrimination in grading: experimental evidence from primary school teachers. Empirical Economics, 45(1):523–538, June 2012.
2