In the past decade, network approaches to psychological constructs have rapidly gained popularity and have seen a proliferation of psychometric techniques designed to estimate such networks from data. In this talk, I will review and critically evaluate this development. Networks are clearly a psychometric success story in terms of rate at which they have been adopted by the psychological research community, and offer powerful explorative data analysis tools and visualizations of complex multivariate dependencies. However, network analysis is also showing signs of developing some of the same problems that have plagued other psychometric methodologies in the past: unrealistic expectations of what data analysis can do in the first place and a lack of formalized theory that is powerful enough to guide model applications. This hampers the full realization of the potential of the network approach. I will outline a number of ways in which statistical network models can inform network theories of psychological phenomena in connection to recent attempts to create methodologies for theory formation.
about the speaker
Denny Borsboom is Professor of Psychological Methods and Director of the Social and Behavioural Data Science Centre at the University of Amsterdam. His work has focused on conceptual analyses of psychometric models, the measurement problem in psychology, and on substantive psychological research in a number of domains, including intelligence, personality, and psychopathology. He is the founder of the Psychosystems Project, which is dedicated to the development of network approaches to psychometric problems and the construction of statistical models; in addition, he directs the Theory Methods Lab, which focuses on the development of methodological tools designed to facilitate the development of formalized theories in psychology.