Laboratory for latent variable modeling
Faculty: Minjeong Jeon
I am an assistant professor of Quantitative Psychology at the Ohio State University (OSU). I am also an affiliated faculty member with the Translational Data Analytics group at OSU. Prior to joining OSU, I was a postdoctoral researcher at the University of California, Berkeley where I also obtained my Ph.D (in Quantitative Methods and Evaluation) and MA (in Statistics). My research interests include developing, applying, and estimating a variety of statistical models, such as latent variable models, multilevel models, and item response theory models. I am also interested in developing computational algorithms and software.
I have been collaborating with researchers from various areas, including developmental psychology, neuro science, social psychology, clinical psychology, educational measurement, biostatistics, and statistics.
Lab members/research associates
I am a second-year graduate student in the Quantitative Psychology program. I graduated with bachelor's degrees in psychology and statistics from Virginia Tech in 2014. In general, I am interested in using advanced quantitative techniques to help the social sciences, either by applying techniques to actual data and answering applied research questions, or by improving or creating new methods for other researchers to use. More specifically, my research is based on latent variable modeling, particularly in the "GLLAMM" framework. I enjoy using item response theory, factor analysis, and SEM and seeing how they are all somewhat related to one another.
I have several projects that I am currently working on. One uses "IRTrees" to investigate the influence of response time on intelligence test research. It turns out that the speed at which a respondent answers an item may change the properties (difficulty, etc.) of that item. Another project uses item-level multilevel SEM to explore the effects of self-reported math anxiety on math performance in children. This project uses data from a huge database of twin pairs, collected by Dr. Steve Petrill's lab in the Development Psychology program here at OSU. Something else I am also looking into is whether it is possible to perform simple IRT analyses in SPSS. Many psychologists use this program of course, so it may boost the usage and the popularity of IRT if it can be readily used within SPSS.
I am a postdoctoral researcher working with Dr. Jeon, with a background in both quantitative and social psychology. Broadly, my interest concerns the cross-talk of seemingly unrelated areas of research, as well as investigating the utility of advanced statistical and mathematical methodologies in solving common problems encountered by psychological researchers. This abstract interest is reflected in my areas of research, including using item response theory to draw more information from self-reported measures, examining alternative models and parameter sensitivity within a mediation framework, as well as research into self-esteem, identity formation, and personal security.
I have two lines of research with Dr. Jeon. The first makes use of IRTrees (i.e., multinomial response trees) to examine participant response patterns when self-reporting self-esteem. At present, my primary goal for this line of research is to demonstrate the utility of such an approach to answering theoretical questions regarding the nature of self-esteem as is measured in psychology. The second applies the logic and literature of mediation to differential item functioning (DIF) to provide additional insights into what a test of DIF does (and does not) imply regarding the underlying processes.