Visualisation in network psychometrics is effective

How do you map out psychological phenomena in a way that makes their relationships and interactions clear? Sacha Epskamp conducted research in this area known as 'network psychometrics', which explains various forms of psychological networks. Epskamp obtained his PhD on Wednesday 5 April at the University of Amsterdam.

Network psychometrics is concerned with estimating network structures in psychological data. The discipline emerged from the call not to perceive psychological phenomena as traits that are affected by an 'underlying' complaint or mental state, but rather as a consequence of direct interactions between the traits measured. Take problems such as insomnia, fatigue and poor concentration, for example – not traits that are influenced by an underlying depression, but traits that interact with each other: after all, poor concentration often results from fatigue, and fatigue from insomnia.

Psychology is extremely complex: analysing the intertwined moods, feelings and disorders in human beings and summarising these in model form is a fairly new discipline. How to visualise such a network is part two. Wouldn't it be great if you could use advanced software, based on digitally processed questionnaires for example, to present a network that could be understood at a glance? Epskamp studied the feasibility of various elaborated approaches, which give a clear picture of connecting structures and variables. The network called LASSO comes out on top: in a simulation study, the researcher shows that the method described works well.

Epskamp also warns against over-interpreting estimated networks based on psychological data. Estimates underlying the data may show connections that are not actually there; crucial information for the user of network models in clinical practice.