ponents in the production function of skills. While the former is accounted for in several studies, the latter is not often considered in this strand of literature, with some exceptions such as Fryer Jr and Levitt (2004, 2006). Controlling in this way for unobserved school-level characteristics reduces selection bias. This enhances our comprehension of how individual characteristics relate to skill development. This chapter proceeds as follows: Section 3.2 provides a background on achievement gaps and their development throughout schooling. Section 3.3 discusses the data, Section 3.4 clarifies the empirical strategy, Section 3.5 gives the results, and Section 3.6 concludes. 3.2 Background 3.2.1 Achievement gaps in education Achievement gaps in education are well-documented in the literature. We briefly summarize the literature in this section related to the three distinctions we study in this chapter: i.e., those by SES, sex, and migration background. First, we focus on achievement gaps related to SES, typically measured using indicators like parental education and income. In this chapter, we define SES using parental education, as this is a more stable indicator for SES than income. Research consistently shows that students with high-educated parents outperform their peers with low-educated parents (e.g., Elliott & Bachman, 2018; Jordan & Levine, 2009; Ng et al., 2021; Passaretta & Skopek, 2018b; Passaretta et al., 2022). Explanations 45
RkJQdWJsaXNoZXIy MjY0ODMw