الفهرس | Only 14 pages are availabe for public view |
Abstract In many applications, data could have some kind of correlation. Hierarchical data and longitudinal data are famous examples for such kinds of correlated data. For these kinds, the traditional regression modeling techniques are not the ideal methodology to analyze such data, as some sources of variance will be dropped and consequently the estimation accuracy will be affected. Multilevel Modeling is a proper technique to overcome this problem and best fit these kinds of data. To determine whether to use the Multilevel Model or the traditional regression model, hypothesis tests are used to test the significance of the random effects. The main purpose of the present study is to review and compare the previous studies presented to test the misspecification of the variance components. An extensive comparison is performed using simulation techniques considering many simulation settings. Also, an application of the tests is performed on real data obtained from the formal site of the World Bank. These data are about Labor force and Inflation rate for the largest economies in 2018 for the years from 2000 to 2017 |