After obtaining a Master Degree in Psychology (specialisation quantitative research, KU Leuven) in 2005, I started working on a PhD in quantitative psychology (funded by a four-year PhD Fellowship from the Research Foundation – FWO – Flanders). In 2009, I defended my Doctoral Dissertation entitled “Component and hierarchical classes analysis of coupled N-way N-mode data blocks”. The next three years I was employed as a post-doctoral researcher at KU Leuven (partly funded by a grant from the Research Council of KU Leuven). In 2012, I received from the FWO a three-year Postdoctoral Fellowship to continue my research on quantitative methods.
A common theme in my research pertains to the development of novel data analytical methods to tackle complex substantive research questions that arise in the social and behavioral sciences. These questions mainly focus on (1) the understanding of the mechanisms underlying human behavior (e.g., the relation between how people react to a particular situation and the appraisal of this situation) and (2) the identification of inter-individual differences in behavior (e.g., types of people that have a different response profile across situations). Regarding the first question, I use explorative techniques like component and factor analysis to extract a set of underlying dimensions that accounts for the correlations (or more general, the associations) between the variables (e.g., behavioral dimensions pertaining to a verbal and a physical response system or to approach and avoidance reactions; a characterization of the situations in terms of underlying perceptual dimensions like the level of frustration or the degree to which others are involved in the situation). To discover inter-individual differences, clusters of subjects that have different variable profiles (e.g., varying response patterns across situations) are derived from the data (i.e., unsupervised learning techniques). By combining cluster and component/factor analytical methods, quantitative and qualitative differences in underlying mechanisms can be disclosed. Quantitative differences are encountered when the same dimensions/mechanisms (e.g., response systems or perceptual dimensions) apply to all subject and subjects only vary in the degree to which these mechanisms apply to them (e.g., all subjects show verbal behavior to highly frustrating situations, with this pattern being stronger for some subjects than for others). When (groups of) subjects differ regarding the dimensions that apply to them (e.g., different people may appraise the same situation in a different way because some people only focus on the amount of frustration involved whereas others only take into account the perception other people have from them; some persons may only distinguish between verbal and physical channels, whereas others only between approach and avoidance reactions), qualitative differences in the underlying mechanisms are encountered. Uncovering these qualitative and quantitative differences in underlying mechanisms is especially challenging when these mechanisms (and inter-individual differences therein) only can be identified by combining multiple (coupled) data sets that contain heterogeneous information regarding the same persons (e.g., data fusion by combining observational and questionnaire data, or by combining EEG and fMRI data).
The goal of my research is to develop and evaluate novel data analytical strategies to search for such qualitative and quantitative differences in underlying mechanisms. To evaluate these techniques, I make use of extensive simulation studies in which various data characteristics are manipulated. As these simulations (and novel techniques) can become computational intensive, I often rely on high-performance computing infrastructure (i.e., parallel computing to deal with high-dimensional – big – data). Other research interests of me pertain to model selection problems (e.g., how many clusters and/or dimensions/components are present in the data) and the development of software packages (e.g., in R) to facilitate applied researchers to use these novel – more complex – data analytical techniques in their research.
No relevant ancillary activities