Research On FIFA And Comparing It With CLFA In Exploring EPQ Factor Structure

Content: The full item factor analysis (FIFA) method is an application that the item responds the theories in factor analyze. This text introduced systematically the analytical mathematics foundation of the full item factor analysis method for binary item and its superior to the classical linear factor analysis (CLFA). Through simulation study, we find FIFA can overcome CLFA's disadvantage that is to overestimate the dimension and to underrate loadings. Besides that, the full item factor analysis can handle with guess parameters. Aiming at the weakness of the chi-square standard in the full item factor analyzes, this text put forward a standard of quasi-eigenvalue. That standard not only is viable theoretically, and through simulation study, we confirm the standard is practicable. As an application of FIFA, Structural Equation Modeling was firstly applied in the data based on 2311 Chinese college students' response to Eysenck Personality Questionnaire (EPQ), and no evidence was found to support the hypothesis of EPQ. On this foundation, we used FIFA and CLFA to explore the structure of EPQ, and found that the FIFA method overcomes the faults of classical linear factor analysis, such as overestimating dimension number, low estimating factorial loading and so on. Through the FIFA method, we found there were six factors in EPQ and these factors were explained well.