FACTOR ANALYSIS AND CONSTRUCT VALIDITY IN EDUCATIONAL RESEARCH: "A CRITICAL ANALYSIS OF ITS USE IN EDUCATIONAL AND PSYCHOLOGICAL MEASUREMENT"

Abstract

Factor analysis occupies a central position in psychological and educational research as a method for examining the internal structure of measurement instruments. However, its value for making inferences about construct validity depends on the consistency between statistical evidence and the theoretical and content foundations of the construct being measured.

Against this background, the present study analyzes actual practices in the use of factor analysis in psychological and educational research. It also examines the relationship between factor-analytic findings and theoretical assumptions in instrument development and interpretation. The study adopts a critical analytical-documentary approach. It analyzes 50 studies, including academic theses and scientific journal articles. The analysis focuses on the type of factor analysis used, factor extraction methods, the foundations of confirmatory factor models, and researchers' responses to agreement or disagreement between theoretical structures and empirical findings.

The results show that exploratory factor analysis was the most frequently used approach (56%). A further 24% of the studies combined exploratory and confirmatory factor analysis, indicating a tendency to move from structural exploration to structural testing. Principal component analysis was the most frequently used method among the exploratory studies (70%). This finding raises a methodological concern about the distinction between data reduction and the identification of latent factors. The confirmatory models relied on previously established theoretical models in 59% of the cases. The comparison between theoretical assumptions and factor-analytic findings also showed agreement in 74% of the studies, while the remaining studies showed disagreement.

The study concludes that factor-analytic indicators provide important evidence regarding internal structure, but they cannot independently establish construct validity. Sound inference requires methodological integration of theoretical foundations, appropriate factor-analytic procedures, scientific interpretation of findings, and multiple sources of validity evidence. The study therefore proposes an integrative approach that treats factor analysis as one component of a broader validity argument rather than as a decisive and independent statistical criterion.