keywords: Absolute bias, RMSE, OLSE, R2-Coefficient of determination, panel data
This research is on the behavior of panel data estimators in a simulated study. The major objective is to investigate these estimators on individual effect and other remaining disturbance term in one way error component. The estimators considered were pooled OLSE, within effect and between effect estimators. The data was simulated using R Statistical software after considering all statistical properties with sample size of N=20, N=50 and fixed time T=10. The methods of validation of result include R-squared, Root Mean Square Error (RMSE) and Variance. The results of the analysis reveals that at sample size N=20 and T=10, the absolute bias of the estimators indicate that the pooled OLSE is better than all other estimators. It also reveals that within effect estimator better explain the fitness of the model due to the highest R-squared value. The between effect estimator perform better than other estimators because it has the least Variance values. The between estimator is also more efficient than other methods of estimation as its value of variance and RMSE of 0.9710397 and 0.9854 are respectively low. As the sample size increase to N=50, the results still remain the same. In conclusion the between estimator is the best and most efficient estimator among all other estimators considered.