Effect of influential cases on factor analysis results

Abstract

When performing regression analysis, one way to examine the normality of data is to screen outliers. Outliers, on the other hand, do not always have an effect on regression results. In reality, cases with a large amount of residuals that affect regression analysis results are referred to as influential cases. It is important to detect them in the dataset because they can lead to erroneous conclusions. The influence of influential cases has already gotten a lot of attention in the regression literature, while it has gotten a lot less attention in factor analysis. The aim of this paper is to show how influential cases affect factor analysis results when they are detected using the Forward Search algorithm. The data was collected from 686 university students ranging in age from 17 to 30. The data was gathered using the Self-Regulation Scale (SRS). The results revealed that the removal of influential cases had an effect on the observed correlation matrice for the SRS items, the factorability results, the number of dimensions extracted, CFA fit indices, and the amount of factor loadings and associated errors. Later, in light of related literature, these results were discussed and the researchers were recommended to consider the effect of influential when applying factor analysis.

Motivation & work efficacy of MBA course in online teaching mode during and after pandemic

“…Factor loading indicates the strength of a component's influence on a variable, with a value above 0.5 signifying significant impact. [1][2][3][4][5]…”