Multidimensionaal Leren Volgen

Psychological and educational assessments historically are an important part of human life, and dimensionality assessment is one of the key steps for developing assessment tools in modern psychological and educational assessments. In the past decades, there has been an increased attention for multidimensional psychometric techniques, aimed to clarify relations between variables and underlying human behavior, therefore laying the foundations for building general behavioral theories. However, the fast advancement in assessment technologies has resulted in several challenges affecting the effectiveness of existing methods. Thus, this thesis aims to facilitate the methodological development of multidimensionality assessment to tackle relevant issues. It compares widely applied methods to clarify their properties in evolved assessments, proposes computational psychometric methods based on statistical psychometric methods and machine learning; and provides practical guidelines on linking methodological developments to real data applications with implementation illustrations by R codes. Specifically, this thesis is divided into five sections. Section 1 with two chapters provides a general introduction to the history and trends of psychology and educational assessments and dimensionality assessment. After that, the theoretical background of subsequent studies is introduced, including a conceptual introduction to multidimensional structures, statistical psychometric methods, and machine learning methods. Section 2 with two chapters concentrates on exploratory methods that can be used to explore what the data structure could be underlying behavioral data. This part includes one comparison study of factor retention methods and another comparison study of a statistical exploratory method and a computational exploratory method proposed by this thesis. Section 3 with three chapters focuses on confirmatory methods, such as a comparison study of statistical confirmatory methods, a comparison study of a statistical confirmatory method and a computational confirmatory method proposed by this thesis, and a comparison study of dimensionality validation methods in cognitive diagnostic models. Section 4 with a single chapter presents practical applications, offering a tutorial on multidimensionality assessment with R codes and real data examples. Section 5 with one chapter ends this thesis with a general discussion.