NEPS technical report for mathematics: Scaling results for the additional study Baden-Wuerttemberg

The National Educational Panel Study (NEPS) aims to investigate the development of competencies across the whole life span. It also develops tests to assess different competence domains. In order to evaluate the quality of these competence tests, a wide range of analyses are carried out by using item response theory (IRT). This paper describes the data and results of analyzing the mathematics competence test that was used in the additional study Baden-Wuerttemberg. The test was designed to test first-year students in higher education; here, three consecutive waves (2011–2013) of secondary-school students were tested in their final year of Gymnasium (type of school leading to upper secondary education and Abitur). In sum, 4,915 students participated in these three waves. The mathematics test consisted of 21 items representing different content areas as well as different cognitive components. A partial credit model was used for scaling the data. Item fit statistics and differential item functioning were investigated. The results show that the items exhibit good item fit and measurement invariance across various groups. However, the reliability is somewhat mediocre, which might be due to the fact that test targeting is not perfect. The paper also provides information about the data available in the Scientific Use File, ConQuest- and TAM-syntaxes for scaling the data, and appendices that describe the scaling of each wave separately.