Richard Gonzalez
Center Director, Research Center for Group Dynamics, Institute for Social Research
Co-Director, BioSocial Methods Collaborative
Amos N Tversky Collegiate Professor, Psychology and Statistics, LSA
Professor of Integrative Systems and Design, College of Engineering
| E-mail: | Email Richard Gonzalez |
| Address: | Research Center for Group Dynamics Institute for Social Research University of Michigan 426 Thompson Street Ann Arbor, Michigan 48106 |
| Phone: | 734-647-6785 |
Articles in Statistics/Methods
Tutorial chapter on dyadic data analysis
This was a tutorial piece on issues surrounding dyadic data analysis.
Every one (1) matters when developing structural equation models: The effects of scale
This was a fun paper to write. Most people who use structural equation modeling don’t realize that a model’s parameterization can influence the standard errors and significance tests. The effect can be quite large and can make significant effects nonsignificant or nonsignificant effects significant. People assumed that parameterization can be ignored because the overall goodness of fit is unaffected. We point out the problem and present a simple way to test parameters in structural equation modeling that is unaffected by parameterization.
Relating dyadic models of interpersonal behavior to personality
In this paper we review our work on statistical models for dyadic data. We relate the models to traditional psychometric models such as multitrait multimethod models.
The value of brain imaging in psychological research
We discuss the role of brain imaging techniques in psychological research, reviewing both the promise and the hype.
When multidimensional scaling goes random, random effects that is….
This paper expands individual-difference multidimensional scaling to include random effect terms.
Psychological theory wagging the statistical tail
The transactional model has had a major influence in the theories of development over a couple of decades. In this chapter I explore some areas of opportunity for developing statistical methodology to facilitate novel tests of the model’s predictions. The transactional model focuses on dynamic multivariate processes with individual differences (heterogeneity) across multiple interacting individuals; these are areas of active methodological research.
Is advice treated the same way as evidence in a learning task?
This paper we investigate several mathematical models of learning and extend them to include advice from others as part of the learning mechanism. We find that a type of reinforcement learning model does well at accounting for the explore-exploit behavior present in the experimental task, and accounts for the data better than Bayesian models. We designed a second study to tease apart model predictions.
How does the brain facilitate rumination and what is the relation to depression?
This paper compares how depressed and healthy controls differ in their ability to forget information, particularly negative information, which may be a precursor for psychological processes such as rumination. We present a new approach to fMRI data analysis that focuses on spatial variability in activation. We find that participants with major depressive disorder show more spatially variable activation in the inferior frontal gyrus compared to healthy control participants using a directed-forgetting task.
Dyadic data analysis
I presented a short, one hour overview on dyadic data analysis at the methods preconference at the 2011 Society for Personality and Social Psychology Meeting.
Visualizing data
Alex Chavez and I wrote a chapter highlighting some modern graphical tools for getting a better handle on both data and model. This was an invited piece.
