spelling and math scores. Therefore, Figure 4.1 suggests a positive relationship among the three skills: students who excel in one skill are likely to also perform well in the other skills. 4.2.3 Sample We consider a balanced panel dataset of 17,358 Dutch students. These students were enrolled in three different cohorts. The first cohort of students was in first grade in 2014 and in sixth grade in 2019, whereas cohorts two and three were in first grade in 2015 and 2016, respectively.5 During their time in primary education, these students took semi-annual standardized tests on reading, spelling, and math. The data contain mid-term and end-of-term tests. The former tests are mostly administered in January-February, whereas the latter tests are mostly administered in May-June. We have data from these students from the end-first grade to mid-sixth grade.6 Therefore, we have ten semi-annual observations per student, resulting in 173,580 observations. Table 4.1 presents the descriptive statistics of the 17,358 students in our sample. Regarding sex, 51% of the students are male, while 49% are female. In terms of parental education, a significant portion of the students come from high-educated backgrounds (47%), followed by 5Our estimation sample consists of students that had a normal flow through primary education, meaning that we have excluded students from the sample that repeated or skipped a grade. We further only consider students with complete skill proficiency data, excluding those with missing observations. 6We do not include the mid-term test in first grade, as it is not available for all domains. Similarly, the end-term test in grade six is excluded, since schools often omit it due to the primary education exit test. 87
RkJQdWJsaXNoZXIy MjY0ODMw