The research unit Psychometrics and Statisticsis is concerned with quantitative techniques for the analysis of empirical data. The focus is on techniques for the analysis of various types of data, such as hierarchically and non-hierarchically organized data, and test - and questionnaire data. The techniques of interest are: component - and factor analysis techniques and their generalizations, item response theory, multilevel analysis, Bayesian statistics, and time series analyses.
The research on these techniques is problem oriented, proceeding in the following three successive steps:
- evaluation of the usefulness of existing techniques for answering the data analysis question at hand;
- improvement of existing techniques or development of new techniques;
- evaluation of the usefulness of the resulting new or improved techniques.
The usefulness of techniques is evaluated in terms of technical optimality and usability.
As far as technical optimality is concerned, the main question is:
“Does the technique give the best possible parsimonious representation of the data, at least under certain specified circumstances?”
As far as usability is concerned, the main question is:
“Can (social science) researchers use the techniques without insurmountable problems of applicability or interpretability?”
Goals and objectives
The goal of the group is, first of all, to provide (social science) researchers with concrete advice as to which techniques to use for particular data analysis problems. Findings are reported in specialized scientific journals and at international conferences for experts in psychometrics and statistics in general, and in scientific journals and/or books aimed at the wider audience of users of statistical and psychometric techniques. A second goal is to improve, where necessary, access to particular data analysis techniques. Specifically, developing and evaluating publicly accessible computer software is an important objective. A third goal is methodological consultation. On the one hand, this leads to the further dissemination of results. On the other hand, it provides insight into which data analytic problems (social science) researchers are faced with, thus providing input for new research on data analytic techniques. Coordination of research plans is done via the program leader and has resulted in recent projects that make use of the unique expertise of the different group members with a keen eye towards obtaining competitive funds.
To train potential new researchers, we supervise PhD student projects, consisting of individually carried out supervised research as well as advanced courses in collaboration with IOPS . Also, we offer a two-year Research Master program “Psychometrics and Statistics" as part of the Research Master “Behavioural and Social Sciences” , which can be seen as an ideal preparation for a PhD student in psychometrics and statistics.
Scope of Research
The research unit is concerned with quantitative techniques for the analysis of empirical data. The focus is on techniques for the analysis of complex data types, such as hierarchically organized data (e.g., three-way data, multilevel data) and relational data (e.g., sociometric data). A list of topics of interest is:
- item response theory
- use of psychological test course in selection and education
- multiway, multilevel and multiset component analysis
- resampling and cross-validation for component analysis techniques
- confirmatory factor analysis and structural equation models
- Bayesian modelling
- multilevel regression
- use and usability of statistics
- handling missing data in various techniques
- optimization procedures for data analysis techniques
|Last modified:||29 July 2021 11.26 a.m.|