Endogeneity Violation on the Comparison of Ordinary Least Square and Maximum Likelihood Extraction Method of Factor Analysis

Abstract

One of the main objectives of factor analysis is to reduce the number of parameters. The number of parameters in the original model is equal to the number of unique elements in the covariance matrix. The study compared ordinary least square and maximum likelihood method of extraction of factor analysis under two approaches such that the variables employed were assumed to be independent of error i.e endogeneity assumption in the first approach while the endogeneity assumption is violated by omitting the important variable HLT in the second approach. The result showed that the extracted factors under the violation of endogeneity has similar factors loading pattern which accounted for a great deal of variance and the factors do a good job of representing the original data and the Bayesian information criterion also showed that the maximum likelihood method of extraction slightly outperforms ordinary least square.