| 英文摘要 |
Spearman (1904) presented factor analysis more than 120 years ago to formulate the theory of intelligence with a mathematical model. With psychologists’interest and exploration of the nature of intelligence over time, the evolution of theories of intelligence has greatly impacted the development of factor analysis. We first review the interplay between early advancement of factor analysis and related theories of intelligence. After the progression of factor analysis for more than a century, methodologists have gradually reached a consensus on the implementation of this popular method. The recommended practices of factor analysis are then presented followed by a review of its current applications in Taiwan. Our review indicated that researchers frequently used the unrecommended eigenvalue-greater-than-one rule to determine the number of factors, the unrecommended principal component method to estimate initial factor loadings, and oblique rotation. In addition, more than one fifth of the analyses failed to report the methods adopted during the analysis. We therefore further summarize the reporting practices suggested for factor analysis along with considerations in designing factor analysis research to assist future researchers in properly applying this method. Factor analysis was originally proposed as a mathematical model to depict theories of intelligence. Spearman’s use of mathematical models to represent psychological phenomena is exactly the focus of Thurstone’s (1937) advocacy at the founding of the Psychometric Society thirty years later. As noted by Thurstone, a solid psychological theory may not be established simply because ingenious experimental apparatus or clever mathematics is employed; it must stand the test of time. Ultimately, we should reflect on whether the latent factors identified through factor analysis have withstood the test of time and have enhanced our understanding of human mind and behavior. |