John Koval, PhD

Dr. Koval

Professor Emeritus

P: 519.661.2111 ext: 86271
F: 519.661.3766
E-mail: jkoval@biostats.uwo.ca

Appointments 

  • Department of Epidemiology and Biostatistics

Research Interests

  • I am chiefly interested in the building of epidemiological regression models, that is, of making statistical interpretations of words like confounders, covariates, effect modifiers, mediators, etc., then seeing how to determine which variables in particular datasets follow these definitions of epidemiological terms.
  • Thus I am also interested in methods of selecting variables for inclusion in regression models, particularly, logistic regression models, that is deciding which measurements best predict an outcome, whether it be occurrence of disease, recovery from illness, etc.This is related to the analysis of correlated outcomes in the presence of covariates, both continuous outcomes such as handled in regression analysis with correlated dependent variables, and discrete outcomes such as modeled by logistic regression analysis with correlated dependent binary outcomes. These models are used in the analysis of family data, for example, in studying occurrence of hypertension within families; dental research, for example, in studying diseases affecting teeth; data obtained from repeated measures designs such as drug trials with results measured monthly; cluster randomization studies such as studies in which the same treatment is applied to all members of a medical practice; and single-stage cluster samples such as those obtained when neighborhoods or city blocks are selected at random.Related to this topic of statistical models for correlated binary variables, and mixtures of binary and continuous variables. These models would serve well for many epidemiological studies, in which the measurements taken on subjects are usually binary, or mixtures of binary and continuous measurements.
  • Another interest is the development of computing algorithms for fast numerical computation, and the use of computer graphics for both statistical research and data analysis.

Education

  • BMath University of Waterloo
  • MMath University of Waterloo
  • MPhil University of London, United Kingdom
  • PhDdd University of Western Ontario

Publications (selected)