Chapter 4. The Interplay of Reading, Spelling, and Math in Primary Education: A Panel VAR Analysis 4.2 Data 4.2.1 Data This chapter uses data from the Student Administration System (LVS) in the Netherlands, spanning from 2014 to 2021. The LVS dataset contains semi-annual standardized skills scores in reading comprehension, spelling, and mathematics. Standardized tests are administered, taken, and scored consistently across all students. For this reason, test scores among different students and grades are consistent, and so comparable.2 These skill scores are obtained by converting students’ raw test scores to skill scores using Item Response Theory (IRT).3 In this way, the LVS data contain valuable information on the skill development of primary school students. In addition, we enrich the LVS data with additional information from the Netherlands Cohort Study on Education (NCO), which contains data on students’ demographic characteristics from Statistics Netherlands. 4.2.2 Skill scores We standardize the obtained skill scores across test period and domain.4 Consequently, all skill variables have a mean of zero and a 2We assume that the time periods between the mid- and end-of-term tests for every student in every grade, as well as between tests, are similar for all observations. 3IRT is used to transform raw test scores into skill scores by estimating each student’s ability on a unidimensional scale. This process involves calibrating the difficulty of test items and calculating the likelihood of a student correctly answering each question based on their ability level. 4A test period refers to a specific point in time within a given grade. For instance, middle grade two is one point in time, end grade two is the second point in time, middle grade three is the third point in time, and so forth. Domain refers to the 84
RkJQdWJsaXNoZXIy MjY0ODMw