Boys Lag Behind: How Teachers’ Gender Biases Affect Student Achievement

I use a combination of blind and non-blind test scores to show that middle school teachers favor girls when they grade. This favoritism, estimated in the form of individual teacher effects, has long-term consequences: as measured by their national evaluations three years later, male students make less progress than their female counterparts. Gender-biased grading accounts for 21 percent of boys falling behind girls in math during middle school. On the other hand, girls who benefit from gender bias in math are more likely to select a science track in high school. * MIT, IZA and CEP. Address: MIT Department of Economics, 50 Memorial Dr., Cambridge, MA 02142, USA. Telephone number: +1 (857) 331-0924. Electronic address: [email protected]. I would especially like to thank my advisor, Marc Gurgand. This paper also benefited from discussions with and helpful comments from Joshua Angrist, David Autor, Esteban Aucejo, Elizabeth Beasley, Thomas Breda, Ricardo Estrada, Alan Manning, Eric Maurin, Stephen Machin, Sandra McNally, Steve Pischke, Corinne Prost, and anonymous referees and participants at various seminars and conferences. I am especially grateful to Francesco Avvisati, Marc Gurgand, Nina Guyon, and Eric Maurin for sharing their dataset, as well as to the Direction de l’Evaluation, de la Prospective et de la Performance (DEPP) of the French Ministry of Education for giving me access to complementary data used in this paper. A previous version of this paper circulated as a CEP Discussion Paper no. 1341 March 2015. Camille Terrier acknowledges support from the Walton Family Foundation under grant 2015-1641. Boys are increasingly lagging behind girls at school.1 This disadvantage has important consequences: boys who fall behind are at risk of dropping out of school, not attending college or university, and/or being unemployed. In OECD countries, 66 percent of women entered a university program in 2009, versus 52 percent of men, and this gap is increasing (OECD (2012)). In Europe, 43 percent of women aged 30–34 completed tertiary education in 2015, compared to 34 percent of men in the same age range. Because this gap has increased by 4.4 percentage points in the last ten years, there is a growing interest in identifying its roots.2 Some recent studies have highlighted the role of school-related inputs, such as school quality (Autor et al. (2016)), peer socio-economic status (Legewie and DiPrete (2012)), teacher gender (Dee (2005)), or teaching focus on literacy or numeracy (Machin and McNally (2005)). This article complements this literature by demonstrating how teachers’ gender biases affect their pupils’ progress and schooling decisions. A number of papers have shown that stereotyping can bias teachers’ assessment and grades, but the impact of such biases has yet to be addressed.3 Prior research on that topic is limited, and it has focused on specific mechanisms through which a gender bias could affect progress. Research shows that teachers’ biases generate self-fulfilling prophecies (Jussim and Eccles (1992)), produce stereotype threats4 (Steele and Aronson (1995), Spencer et al. (1999), Hoff and Pandey (2006)), affect students’ interest in a subject (Marsh and Craven (1997), Trautwein et al. (2006), Bonesrønning (2008)), and affect students’ levels of effort5 (Mechtenberg (2009)). To my knowledge, this is one of the first pa1In OECD countries, "15-year-old boys are more likely than girls, on average, to fail to attain a baseline level of proficiency in reading, mathematics and science” (OECD (2015)). 2In France, 49.6 percent of women aged 30–34 have completed tertiary education in 2015, compared to only 40.3 percent of their male counterparts. 3See for instance Bar and Zussman (2012), Burgess and Greaves (2013), Hanna and Linden (2012) on teachers’ gender bias, and Tiedemann (2000) and Fennema et al. (1990) for the existence of a gender bias in mathematics. Several papers have exploited blind and non-blind scores (teachers’ grades) to test the existence of such biases in teachers’ grades, a methodology introduced in a seminal paper by Lavy (2008). Some papers find that girls benefit from grade discrimination (Lindahl (2007), Lavy (2008), Robinson and Lubienski (2011), Falch and Naper (2013), Cornwell et al. (2013)), while others find no gender bias (Hinnerich et al. (2011)). Ouazad and Page (2013) and Dee (2007) observed that gender biases depend on teachers’ genders. Breda and Ly (2015) found that discrimination depends on the degree to which the subject is “male-connoted”. 4The latter arises when girls or minority groups perform poorly for the sole reason that they fear confirming the stereotype that their group performs poorly. The apprehension it causes might disrupt women’s math performance. Therefore, over-grading girls can reduce their anxiety to be judged as poor performers when they undergo a math exam. 5Mechtenberg (2009) provided a theoretical model of how biased grading at school can explain gender differences in achievements. School results are defined as a combination of talent and effort, the latter being the channel