Dynamic network models for dyadic data
Many disciplines in the behavioral sciences involve the study of dyadic relations. For example, one can think of the interactions that take place between mother and child or within romantic couples. Given dyadic data, interesting research questions pertain to who causes what. For instance, are the parents steering the behavior of the children, or is it exactly the opposite? Or is the relation in fact bidirectional? To fully grasp such interpersonal processes, the dyad is best recognized as a dynamic system. Modeling dyadic dynamics is quite challenging, however, because many variables may be involved and because interaction patterns may be different in specific subgroups.
In the envisaged project, we deal with these challenges by developing a dynamic network modeling framework for dyadic time series data. This approach produces an easy-to-read visualization of the results of the analysis, unraveling the structure of the interaction pattern. In a next step, the proposed methodology will be extended to handle a large number of variables. Furthermore, a clusterwise version of the network approach will be developed to reveal subgroups of dyads with similar interaction patterns. Finally, we will promote the use of the new network tools by developing software and by applying them to empirical data sets in close collaboration with substantive researchers.
Prof. dr. Eva Ceulemans
Prof. dr. Francis Tuerlinckx
1 October 2013 – 1 October 2017