Effect of Small Sample Size on Parameter estimation in a three-level Organizational Framework: A Simulation Study using M-Plus

The use of Multilevel Modeling has become extremely popular in Social Science research owing to the natural hierarchy which often exists in the dataset. As has been pointed out by most researchers, number of units (sample size) at upper levels of hierarchy becomes extremely crucial. The current study employs a Random-Intercept model (using M-Plus) to study the effect of level-3 sample size on parameter estimation in a three level organizational framework. Number of sampling units at the third level was varied to check the impact on fixed effects, variance components, and their associated standard errors