Factor recovery by principle component analysis and harris component analysis Dholakia Stuti G.*, Dr. Bhavsar Chetna D.** *Research Scholar and Associate Professor, New L. J. Commerce College, Ahmedabad, Gujarat, India, stutigdholakia@gmail.com **Professor, Department of Statistics, School of Sciences, Gujarat University, Ahmedabad, Gujarat, India, chetna_bhavsar@yahoo.co.in Online published on 17 July, 2017. Abstract Exploratory factor analysis is highly used multivariate technique in the field of marketing, psychology, education and social sciences. The researchers have to make certain crucial decisions regarding number of variables to include, number of factors to retain, adequate sample size, selection of factor extraction method and rotation method. This study is intended to identify the effects of these decisions on unveiling the hidden structure between the variables. Basic aim of this study is to compare the two extraction methods namely, highly used Principle component analysis (PCA) with squared multiple correlations (SMC) as initial estimates and Harris Component analysis. This study reveals that Harris method produced better factor recovery than PCA for majority of conditions. This paper also helps the researcher to understand the importance of proper prior decisions so that the stable factor estimates can be obtained. Top Keywords EFA, SAS, Principle component analysis, Harris Component analysis, comparison indices. Top |