Are you interested in a postdoctoral position where you will develop a state-of-the-art methodology for the analysis of psychological networks? We are looking for a postdoc to join the Bayesian Graphical Modeling (BGM) lab (bayesiangraphicalmodeling.com) headed by Dr. Maarten Marsman.
The psychological literature on network modeling is growing rapidly, and Bayesian statistics offers concrete solutions to outstanding methodological problems (e.g., doi.org/10.1007/s11336-022-09861-x and psyarxiv.com/ub5tc). The network structure will be modeled with a graphical model or a Markov random field. The postdoc will develop a Bayesian methodology to assess structural changes in the network after an intervention and to model the influence of grouping or predictor variables. These variables may be observed, as in ANOVA and regression, or unobserved, as in mixture modeling or latent class analysis. After successfully deriving the Bayesian models and methodology, the candidate will implement the new methods in the open source statistical software R and JASP. In addition, the candidate will collaborate with the BGM lab on joint projects and help supervise Ph.D. students. The BGM lab will consist of two postdocs and three Ph.D students.
The BGM lab is part of the Psychological Methods Program Group at the University of Amsterdam, one of the largest and most successful research centers in psychological methods. Its main research areas are network analysis, Bayesian statistics, mathematical psychology, and psychometrics. In addition to developing novel research methods, we are actively developing psychological theory on several topics, including intelligence, psychopathology, and decision-making. The group is committed to open science and has made important methodological contributions over the past decade to facilitate transparent research practices.
What are you going to doYou will:
- derive Bayes factors or Bayesian variable selection procedures to assess the impact of interventions and predictors on graphical models;
- analyze the performance of proposed Bayesian solutions using formal methods and simulations;
- implement the new Bayesian procedures in open source software such as R and JASP;
- regularly present research results at (international) workshops and conferences and publish them in internationally renowned journals;
- help supervise Ph.D. projects in the BGM lab.
Teaching is limited to 5% of the time, assisting in courses and supervising M.Sc. students. More teaching is possible if desired by the candidate.