Maarten Marsman (Nov 2018-Nov 2019)
Assistant professor in Psychological Methods at the University of Amsterdam


My research
My research is focused on the development of formal psychometric theory of psychological constructs such as intelligence and depression using networks, and by extension the field of network psychometrics. I also work on Bayesian procedures for estimating complex statistical models –such as network models and latent variable models– and comparing ordinal hypotheses using the Bayes factor. My fields of application are psychological- and educational measurement.
IAS fellowship (Nov 2018-Nov 2019)
How to build formal psychometric models for psychological constructs such as intelligence that are consistent with established cross-sectional phenomena using idiographic network models?
Psychometrics has been very successful in uncovering major phenomena in psychological- and educational measurement. A primary example is the pattern of positive correlations, the positive manifold, between scores on cognitive tests. Another important example is the Matthew effect in education, which concerns the accumulation of individual differences initiated by minor differences in early success in, for instance, reading. Unfortunately, these successful revelations have not culminated into a formal theory of intellectual development that explains the phenomena at hand. Without an understanding of the generative mechanisms involved, undesirable phenomena such as the Matthew effect can be diagnosed, but cannot be treated or prevented.
The main challenge in the development of psychometric theory is to unite well-established phenomena from psychometric research with models for the individual in a formal framework. To address this challenge, I will build upon the random-cluster model from statistical physics that can be used to model individuals by their personal network of relations (edges) that are formed between cognitive skills or knowledge (nodes), say.
During my fellowship I hope to connect with researchers from different disciplines —such as physics and informatics— as my research builds upon ideas from disciplines that are not my own.