confirmatory factor analysis nov 2008

[...]... observe its symptoms Exploratory Factor Analysis Exploratory factor analysis is used to identify the underlying factors or latent variables for a set of variables The analysis accounts for the relationships (i.e., correlations, covariation, and variation) among the items (i.e., the observed variables or indicators) Exploratory factor analysis is based on the common factor model, where each observed... Introduction Comparison of Confirmatory Factor Analysis With Other Data Analysis Techniques Confirmatory factor analysis is strongly related to three other common data analysis techniques: EFA, PCA, and SEM Although there are some similarities among these analyses, there are also some important distinctions that will be discussed below Before we begin discussing the data analysis techniques, we need to define... common factors (i.e., the underlying latent variables) and one unique factor (i.e., error- or item-specific information) It partitions item variance into two components: (1) Common variance, which is accounted for by underlying latent factors, 9 10 Confirmatory Factor Analysis and (2) unique variance, which is a combination of indicator-specific reliable variance and random error Exploratory factor analysis. .. on the common factor model, and consequently, CFA may not work well when trying to replicate structures identified by PCA There is debate about the use of PCA versus EFA Stevens (2002) recommends PCA instead of EFA for several reasons, including the relatively simple mathematical model used in PCA and the lack of the factor indeterminacy problem found in factor analysis (i.e., factor analysis can yield... (2003) refers to the data analysis as an EFA, but then states “The responses of the 1988 sample to the entire 121 POS items were examined using principal components factor analysis [with varimax rotation] for the purpose of identifying value dimensions (factors) within the POS” (p 647) Based on this analysis, “The 10 items having the highest loadings on each of the four remaining factors were retained,... caused by the latent factor, there may also be some unique variance in an indicator that is not 23 24 Confirmatory Factor Analysis accounted for by the latent factor( s) This unique variance is also known as measurement error, error variance, or indicator unreliability (see E1 to E6 in Figure 2.1) Other parameters in a CFA model include factor variance, which is the variance for a factor in the sample... includes structural or causal paths between latent variables CFA may be a stand-alone analysis or a component or preliminary step of a SEM analysis Software for Conducting Confirmatory Factor Analysis There are several very good software packages for conducting confirmatory factor analyses, and all of them can be used to conduct CFA, SEM, and other analyses Amos 7.0 (Arbuckle, 2006a) is used in this book... EFA yielded a two -factor solution, with five items on the first factor and four items on the second factor; the two factors were significantly correlated (r = 0.47; p < 0.00001) The CFA was conducted using LISREL 8.54 and maximum likelihood (ML) estimation (estimation methods are discussed in Chapter 2) with the second half of the sample The CFA resulted in several modifications to the factor structure... negatively and positively worded items, data analysis may suggest that there are two factors when only one was expected based on theory The Rosenberg Self-Esteem Scale (SES) provides a good example of this problem The Rosenberg SES includes a combination of positively and negatively worded items Early exploratory factor analysis work consistently yielded two factors—one consisting of the positively worded... prior research are crucial to specifying a CFA model to be tested As noted in Chapter 1, the one -factor solution of the Rosenberg Self-Esteem Scale was tested based on the conceptualization of selfesteem as a global (i.e., unitary) factor, although the existing exploratory factor analysis (EFA) work found two factors Early in the process of measurement development, researchers may rely entirely on theory . Creating a Confi rmatory Factor Analysis Model 21 3 Requirements for Conducting Confi rmatory Factor Analysis: Data Considerations 36 4 Assessing Confi rmatory Factor Analysis Model Fit and. 9 Comparison of Confi rmatory Factor Analysis With Other Data Analysis Techniques Confi rmatory factor analysis is strongly related to three other common data analysis techniques: EFA, PCA,. its symptoms. Exploratory Factor Analysis Exploratory factor analysis is used to identify the underlying factors or latent variables for a set of variables. The analysis accounts for the rela- tionships