Felix is an Associate Professor of Psychology at Cornell University. His research interests are causal inference and missing data. In addition to various journal publications, he also published an R package on CRAN to analyze regression discontinuity designs.
In November, he will teach the course “
Causal Mediation Analysis” in the GESIS Workshop program.
How did you become interested in your subject?
I first became interested in mediation analysis when I learned about it from one of my advisors during graduate school, David MacKinnon. It must have been in my third or fourth year in graduate school when I worked in the mediation lab, and I believe it was right around the time I graduated that the first papers on causal mediation analysis were published. It was an exciting time to see new developments on this topic frequently show up in journals. My general interest in causality was seeded by my Ph.D. advisor, Steve West, at Arizona State University and then further enhanced by Rolf Steyer at the University of Jena, where I spent a year at the beginning of my research career. Meetings with both Donald Rubin and Judea Pearl at various conferences were also significant early influences on me. I still recall being impressed that both were generous with their time toward younger researchers.
Unlike in my early years, I feel that research on causality, specifically causal mediation, is much more mainstream and has matured. I remain fascinated by it, as I believe that causal inference is a cornerstone for sound, theoretical research.
What lessons can participants draw from your GESIS course?
I hope workshop participants will feel confident applying the causal mediation models we cover in the lectures. Necessarily, they will learn about the underlying causal assumptions of the methods and will be able to judge their plausibility in applied contexts. They will be able to apply the models to real data using computer code in R and, importantly, be able to draw the correct conclusions from the results of their analyses. And of course, I hope that they will be able to enjoy the workshop and realize that causal mediation is a fascinating topic in the domain of quantitative methods.
What do you enjoy most about being a social scientist?
I am a quantitative psychologist, so I work a lot with data, sometimes collected by fellow researchers and sometimes simulated by myself. I thoroughly enjoy all aspects of data analysis, from reading data into R, wrangling it, if needed, analyzing it with various methods and models, and generating graphs and reports. Doing all my analyses in R / RStudio is also a source of fun because of the amazing community around R, the new packages being developed, and the plethora of tutorials online. It seems like the envelope of what R is capable of doing is constantly pushed by the community.
We thank Felix for his interesting insights and look forward to his class.