Hok Chio (Mark) Lai
Research & Practice Areas
Quantitative methods, multilevel modeling, latent variable modeling, psychometrics, Bayesian statistics
I have a broad interest in quantitative methodology in behavioral and social sciences. My research focuses on the applications and methodological innovations in analyzing clustered and longitudinal data that are common in the behavioral and social sciences, as well as the integration of multilevel modeling with latent variable models. Specific topics including measurement invariance, measurement error adjustment, effect size estimation, and data harmonization. I am committed to communicating the state-of-the-art methodological practices to substantive researchers.
- Ph.D. Educational Psychology, Texas A&M University, 2015
- B.Sc. Psychology, University of Macau, 2011
Summary Statement of Research Interests
Advances in quantitative methods are inseparable from advances in scientific research. As data are collected and stored in more and more complicated forms, researchers need new statistical methods to most effectively analyze such data and answer new research questions. On the other hand, without rigorous measurement of constructs, interesting research findings may just be a product of measurement bias, no matter how advanced the statistical techniques are. Therefore, development of new statistical models should go hand in hand with development of psychometric methods to evaluate measurement; only together will they enhance the usefulness of research.
My research focuses on two of the most distinctive features of quantitative data in behavioral and social science research: their complex, multilevel structure due to clustering by social context and time, and the need to account for their imperfect measurement. Recent methodological efforts have raised researchers’ awareness of data dependency and measurement issues. However, much more research effort is needed to translate conventional procedures and statistics for multilevel data and to make evaluations of measurement properties an integrative part of any data analysis. Motivated by such a mission, I use statistical theories and Monte Carlo simulations to develop and test statistical tools for multilevel data and data based on imperfect measurement. Because by nature research methods cut across multiple disciplines, a lot of my research involves collaborations in projects that study methods with a wide range of applications and projects that apply advanced quantitative methods to various areas of science. Below I discuss the different aspects of my work and describe my plans for continuing my research program over the next five years.
In my research, I develop methods to compute effect size statistics to quantify multilevel treatment and intervention effects, methods that account for survey designs and design properties of national and international survey data, and methods that quantify the impact of measurement bias in psychological testing.
Contracts and Grants Awarded
- CAREER: Advancing Latent Variable Methods for Integrative Data Analysis, (National Science Foundation), Hok Chio Lai $400,001, 2021-2027
- The Impact of Ignoring Parameter Uncertainty on Sample-Size Planning for Cluster-Randomized and Mult, (Spencer Foundation), Hok Chio Lai $49,796, 2020-2022
- Developing a Multidimensional Psychometric Framework on the Impact of Item Bias on Classification, (Department of Defense, Army Research Institute), Hok Chio Lai $104,382, 2019-2022