Student Level Achievement in Different Public University of Bangladesh: A Comparison Between Multilevel Models and Classical Regression Models

In this study, the main purpose is to consider generalized linear modeling where regression coefficients are also modeled. The models used here slopes and intercepts may vary by group or university. study time of selected students a dependent variable socio demographic variables independent variables. statistical method data analysis for two-level multilevel modeling. analysis, we find associated with spends in reading academic books (TSRAB), nonacademic (TSRNAB), central library (TSRCL) at 95% level significance. Using R, get three names as first model null empty when intercept fixed, second random fixed final third slopes. Comparing models, based on AIC BIC statistics, preferable two models. results coefficient determination (R-square) revealed that better than classical model.