Analysis of EFA, CFA, SEM

In this study, Exploratory Factor Analysis (EFA) and regression analysis are applied instead of using the combination of Confirmatory Factor Analysis (CFA) and Structural Equation Modelling (SEM). In spite of the fact that CFA and SEM are often selected by researchers in order to examine relationships between variables and test hypotheses, for this study, EFA and regression analysis is the most appropriate choice because CFA and SEM cannot be conducted separately, and they are estimated simultaneously (Jeon 2015). Using CFA and SEM, researchers refer to assumptions that the discussed variables are in strict linear relationships that are set according to a certain model.

In this study, the conceptual model proposed to test the formulated hypotheses demonstrates that there can be relationships between Performance Management, Rewards, and Promotion, and Training and Development as independent variables (IVs) and Retention as a dependent variable (DV), and there can be relationships between Performance Management, Rewards and Promotion, and Training and Development (IVs), Job Satisfaction as a mediating variable (MV), and Retention (DV). In this case, different models should be developed for these relationships. While following SEM, this task can be discussed as a challenge (Jeon 2015). As a result, the analysis can become too complex and lack the possibility to explore these relationships as series that do not depend on each other (Burns & Burns 2008; Suhr 2006). The first model should demonstrate the direct relationship between Performance Management, Rewards and Promotion, and Training and Development (IVs) and Retention (DV), and the second model should demonstrate the mediating relationship, involving Performance Management, Rewards and Promotion, and Training and Development (IVs), Job Satisfaction (MV), and Retention (DV). The use of multiple regression analysis (in contrast to SEM) allows for testing both types of relationships.

If the use of SEM is not appropriate for this study, it is also important to note that the use of CFA is also unsuitable because CFA will not be effective to reduce the number of factors in this study that is a primary purpose of factor analysis in this research. CFA is used to analyze the concrete number of factors to demonstrate the relationships between them (Suhr 2006). However, in this study, the researcher allows for varying factors to support this or that construct, and the main focus is on selecting the factors that are representative in relation to the variables. Therefore, EFA can directly address this goal of factor analysis in this study in contrast to CFA that cannot complete this goal.

Thus, researchers choose to use CFA as integrated into SEM when they focus on examining structural or causal relationships, where the linear model can illustrate these relationships (Jeon 2015). However, in this study, it is important to refer to separate correlational and regression analyses to conclude about the relationships between variables not from the perspective of simple linear relationships, but from the perspective of two different types of relationships, including direct and mediating relationships. In this case, CFA cannot be used just as the basis for conducting correlational and regression analyses when EFA is appropriate to provide the researcher with the number of significant factors that can be used for further analysis. Thus, in their study on retention and factors that cause it, Kyndt et al. (2009) use EFA to reduce the number of variables and conduct the regression analysis, and they also support the results by the separate correlational analysis. This approach allows for analyzing the relationship between factors and retention in much detail. The same approach is used by Govaerts et al. (2011) to study the relationship between the working climate and retention in organizations.

From this point, researchers prove the effectiveness of using EFA, correlational analysis, and regression analysis in order to explore the relationship between certain variables and retention in organizations. In this study, the use of CFA and SEM is not appropriate to address the stated research aims and hypotheses. The use of regression analysis is the more effective choice, and it is necessary to apply EFA to conduct the regression analysis referring only to meaningful factors.


Burns, RP & Burns, R 2008, Business research methods and statistics using SPSS, Sage, New York.

Govaerts, N, Kyndt, E, Dochy, F & Baert, H 2011, ‘Influence of learning and working climate on the retention of talented employees’, Journal of Workplace Learning, vol. 23, no. 1, pp. 35-55.

Jeon, J 2015, ‘The strengths and limitations of the statistical modelling of complex social phenomenon: focusing on SEM, path analysis, or multiple regression models’, International Journal of Social, Behavioural, Educational, Economic, Business and Industrial Engineering, vol. 9, no. 5, pp. 1597-1605.

Kyndt, E, Dochy, F, Michielsen, M & Moeyaert, B 2009, ‘Employee retention: organisational and personal perspectives’, Vocations and Learning, vol. 2, no. 3, pp. 195-215.

Suhr, D 2006, Exploratory or confirmatory factor analysis?, SAS Institute, Cary.