Abstract
Current study focuses on modeling count data using Poisson regression in educational sphere. For demonstrative purposes, a model involving antecedents and precedents of class participation (CP) has been adopted. CP consisted of count data and for the purposes of the present study, it is used both as a dependent variable and an independent variable. Other study variables included are private tuitioning, scores in English in preparatory year program and scores in psychology. Salient points on when and where to apply Poisson regression technique has been discussed viz-e-viz ordinary least squares (OLS) regression. The results of the study provide input to administrative and pedagogical staff for an alternative to OLS regression for modeling count data using Poisson regression and some effective visual post estimation diagnostics. An easy-to-follow approach is adopted for the replication of proposed procedure in other subjects at all levels in higher educational institutions.