# Advantages Of Second Order Factor Analysis

727 Words3 Pages
The utility of second-order factor analysis models and bi-factor models Second order analysis and bi-factor ( discuss al little later) analysis are substitute methods of representing primary constructs that containing a number of highly related domians (Chen, West, & Sousa, 2006). A second order factor analysis much alike first order factor analysis is statisticalmethod used to ratify a theorized construct into a certain number of subcomponents. The metodology is widley used in the study of practical concepts like personality (DeYoung, Peterson, & Higgins, 2002; Judge, Erez, Bono, & Thoresen, 2002), self-concept (Marsh, Ellis, & Craven, 2002) as cited in (Chen, West, & Sousa, 2006). Using this technique the investigator may look to approximate…show more content…
When a researcher is conducting a second order CFA, for a single factor it is imparative that the parameters are restricted to 1; this is however not necessary if the CFA is being conducted for pooled constructs. Where constructs are pooled on a single reference point is necessary regardless of the amount of components in the…show more content…
Bi-factor analysis has rose to popularity in recent years as it has been used in research about intelligence. This techniques is apllied firstly in instances where, there are multiple domain specific factor, which are speculated to have a unique effect on the domian. Secondly, in research where the researcher is curious about domain specific factors as well as commn factors of focal interested. Thirdly, a general factor which is assumed to account for commanalities exsists (Chen, West, & Sousa, 2006). Chen, West and Sousa (2006) make use of a study by Sterwart and Ware (1992) to illustrate how the bi-factor analysis works. The study by Stewart and Ware relaties to quality of life. In the factor model the a single quality of life factor underlies 17 factors. While the distinct domain factors are, Cofnitive, vitality, mental health and worry of diease. The give model is consider to be conical, bi-factor and is assumed to be orthogonal (statistically independent) as the afore mentioned domain factors, each contribute to the overall general factor. In bi-factor CFA the equation for observed variable is as follows: y=Λ_y η+ϵ; much like the equation ealier discussed in the seaction about second order factor analysis; however here, Λ_y represents the factor loadings for general and domain specific factors. While the vector η represent the general and domian specific factors and ϵ