... which focuses on the development and use of new multivariate statistical tools needed for analyses of data collected in the four other programs.
- Mark de Rooij
The aims and objectives of the unit are twofold:
1) The development of new multivariate statistical methodology and the investigation of properties of existing statistical methods;
2) Empirical application of advanced multivariate statistical methods in psychological research in cooperation with staff members of the other units within the Psychological Institute.
Although these are two different goals, they are highly interconnected, since issues from goal 2 will lead to new research questions in 1, and new methodology developed in 1 can be applied in 2.
Within the first goal different areas of investigation can be distinguished. The first is Statistical Learning and prediction (De Rooij, Dusseldorp, Fokkema, Busing, Wilderjans) where new methodology is developed for both supervised as well as unsupervised statistical learning. The second is applied psychometrics including latent variable modeling (De Rooij, Kelderman, Bakk, Stevenson, Van Putten) where statistical models are being developed for both manifest and latent variables. Special interest in models for longitudinal data analysis fall within this second area. A third area is Methodology for FMRI research (Rombouts, Weeda, De Rooij), where investigations focus on the use of MRI scanners in psychological research, experimental fMRI design, and data analysis of fMRI data.
Within the second goal we have many collaborations with staff members from developmental psychology and clinical psychology. Moreover, one important area of research within this second goal is Research into Mathematics Education (Van Putten). Using advanced psychometric models solution strategies used by children in division problems are investigated. Changes in solution strategies are linked to a declined general mathematics level in Dutch children and changes in the curriculum of Dutch elementary education.
Connection with other research
- Multivariate analysis
- The δ-machine: A new competitive and interpretable classifier based on dissimilarities
- Development of Generalized Squared Distance Logistic Model (GSDLM).
- Stepwise latent class analysis
- Brain networks and the initial stages of dementia
- New Methods for (f)MRI Analysis
- Resampling Methodology for Longitudinal Data Analysis