Department of Methodology and Statistics
The Department of Methodology and Statistics of the Faculty of Psychology and Neuroscience (FPN) works in close collaboration with the Department of Methodology and Statistics of the Faculty of Health Medicine and Life Sciences (FHML) under common leadership. Together we are a group of around 20 staff members from PhD student to Professor. We provide the statistics education in five different Bachelor programmes and various master and PhD courses in the FPN and FHML. We also give statistical advice and support in data analyses to researchers from both faculties. Initially our own statistical research focused on study design and methods of analysis for nested and longitudinal data, in particular on sample size calculations and mixed (multilevel) regression analysis. In the last years our scope has broadened to include methods for multilevel interrater agreement and reliability, multiple imputation methods for missing data, Bayesian statistics for high-dimensional data, and innovative methods for growth curve analysis and for detecting person-situation interactions.
The department coordinates and carries out all three Statistics courses in the Bachelor’s programme in Psychology, plus two advanced statistics courses in the Research Master in Cognitive and Clinical Neuroscience, plus an advanced statistics course in the Master in Forensic Psychology. Furthermore, the department provides a PhD course on Structural equations modelling in the Faculty of Health, Medicine and Life Sciences. Finally, we participate in the first bachelor year course Methods and Techniques of Research, the 2nd year Research practical and we give statistical advice to students writing their thesis.
- Bayesian analysis of longitudinal data
- New methods for detecting person by situation interactions
- Critical assessment of conventional methodological practice in psychological research
- Optimal design and sample sizes for nested designs (cluster randomised and multicentre trials)
- Correcting for baseline group differences in nonrandomised studies (Lord’s ancova paradox)
- Optimal treatment regimen
Our colleagues from the Methodology and Statistics dept. of the Faculty of Health, Medicine and Life sciences, with whom we cooperate, furthermore conduct research on:
- Optimal design and sample sizes for nested designs, longitudinal studies and surveys
- Multilevel reliability and agreement analyses with covariates
- Longitudinal data analysis (mixed/multilevel regression, group-based trajectory modelling)
- Multiple imputation methods for missing data in nested study designs and in meta analyses
We publish our research in various prestigious journals such as Statistical Methods in Medical Research, Statistics in Medicine, Journal of Clinical Epidemiology, Biostatistics, Biometrics, Psychological Methods, NeuroImage, Multivariate Behavioral Research, Computational Statistics and Data Analysis
Also, various research results have already found their way in free software developed by our staff, and are frequently applied in research in health, medicine and psychology, in which we participate as statistical advisors and coauthors.
Whom to contact for first statistical advice?
For staff in the dept. Clinical Psychological Science:
dr. Nick Broers
For staff in the dept. Work and Social Psychology and dept. Neuropsychology and Psychopharmacology:
dr. Jan Schepers
Depending on the expertise needed, as well as on the timing of the request and the prior existence of collaboration, the researcher may subsequently be referred to another statistician in our dept. for further support.
The department offers statistical advice on study design (a.o. sample size) and data analysis to all researchers in the FPN, as well as support in data analysis and reporting. A single consult is offered for free. Repeated consults or support in data analysis and reporting are offered in exchange for co-authorship, sometimes a co-promotorship.
Members of the department have expert knowledge not just on standard statistical methods, but also on advanced methods such as mixed/multilevel regression, structural equations modeling (SEM), group based trajectory modeling, missing data handling, and Bayesian analyses. Software used includes SPSS, R, STATA, SAS, MLWin, MPLus and Lisrel for SEM, Matlab, JAGS and WinBUGS (Bayesian stats), and PASS (sample sizes).
Researchers who consider statistical help in their data analyses, are invited and advised to consult us already in the phase of study design. In statistics, just like in health care, prevention is the better cure!