This paper compares the performance of multilevel regression model and traditional regression model of hierarchical data. Multilevel regression has been used to describe an analytical approach that allows the simultaneous examination of the effects of group-level and individual-level variables on individual level outcomes. The aim of this paper is to introduce the multilevel regression model that allow to explicitly incorporate the hierarchical nature of the data into the analysis, to incorporate variables measured at different levels of hierarchy, and to examine how regression relationships vary across clusters, and compare their performance with traditional regression model on a hierarchically dataset. The comparison is based on two main criteria: the bias of the estimated regression coefficients and the size of testing significance of each regression coefficient. The results suggest that the underestimation of standard error in standard regression artificially increases the significance of hypothesis tests, and the multilevel regression models had a better model fit than standard regression.   A multilevel model of this type is provided by many computer packages, including MLwin 2.36.