One database, many researchers
The same (large-scale) dataset can answer many questions. Often, the data for these datasets are collected for one primary purpose (e.g., school monitoring). However, these datasets have much potential for secondary analyses. Therefore, it is important to explore and to decide which questions could and should be analysed. The “how” is crucial too (i.e. what methods should be used). The richer the data, the more important and complex these decisions become. In the context of educational research, analyses of large-scale data can lead to important insights on the functioning of the educational system. In this presentation, we will introduce the datasets from the Luxembourgish School Monitoring Programme (Épreuves Standardisées, ÉpStan) we currently use. We will explain the methods we use, i.e. multiple linear regressions, and why they are a good fit for our research questions. Sophie focuses on the relation between language and mathematics using grade nine data and Jessica investigates value-added modelling with longitudinal grade one and three data. We will present and discuss the results we have found using these datasets and methods.