|
|
|
|
People of the Methodology Center |
|
Home  People John J. Dziak, Ph.D.
John J. Dziak, Ph.D.
Research Associate, The Methodology Center
 |
The Methodology Center
The Pennsylvania State University
204 E. Calder Way, Suite 400
State College, PA 16801
|
 |
This e-mail address is being protected from spambots. You need JavaScript enabled to view it
|
|
814-863-9806 |
 |
814-863-0000
|
|
 |
| Education: |
|
Ph.D., The Pennsylvania State University, 2006 (Statistics)
M.A., The Catholic University of America, 2001 (Psychology)
B.S., The University of Scranton, 1999 (Psychology)
|
| Research Interests: |
|
Currently most of my work seems to have to do with studying statistical power, especially in challenging situations such as factor-screening experiments and/or clustered datasets. I am also interested in model selection, latent class and variable models, and the modeling of longitudinal data. My thesis had to do with on the application of the Smoothly Clipped Absolute Deviation variable selection complexity penalty to objective functions other than likelihoods, such as generalized least squares and quadratic inference functions.
Methodology Center Research Areas: latent class analysis, latent transition analysis, variable selection
|
| Current Projects and Collaborators: |
|
One of my main roles here at the Center is working on software development, in particular for our SAS procedures. In 2008 I helped create PROC SCADLS, our regression variable selection and estimation procedure, based on the work of Runze Li and his advisor Jianqing Fan on the SCAD complexity penalty. I’m currently working on expanded functionality for our latent class analysis procedure PROC LCA, particularly standard error estimation.
I’ve been also working with Linda Collins, Runze Li, and Inbal Nahum-Shani on some comparisons of the power of different kinds of experimental designs for studying multiple-factor interventions, especially under challenging circumstances such as cluster-correlated data.
I’ve been helping with a project led by Young Kyoung Min and Stephanie Lanza which involves using PROC LCA in an extensive series of Monte Carlo simulation experiments to investigate the performance of LCA tests and estimates at different sample sizes and given different true population structures.
Last, I’ve been working on a small paper or two in which I hope to review some approaches for power calculation with clustered binary data and to review some problems in discussing the idea of statistical power for secondary reanalyses of datasets.
My thesis advisor was Runze Li, also of the Methodology Center. We investigated some applications of penalty-based variable selection methods with longitudinally correlated data.
|
| Publications: |
|
Peer-reviewed Papers |
|
Collins, L.M., Dziak, J.J., and Li, R. (2009). Design of experiments with multiple independent variables: a resource management perspective on complete and reduced factorial designs. Psychological Methods.
|
|
Book Chapters |
|
Dziak, J.J. and Li, R. (2007). An overview on variable selection for longitudinal data (chapter). In D. Hong (Ed.), Quantitative Medical Data Analysis. Singapore: World Sciences.
|
|
Other Publications |
|
Dziak, J.J., Lemmon, D.R., and Li, R. (2008). PROC SCADLS, Version 1.0.5_, and PROC SCADLS Users Guide. University Park, PA: The Methodology Center, The Pennsylvania State University.
|
|
|
|
|
