EXPLORATORY VERSUS CONFIRMATORY FACTOR ANALYSIS

There are two basic types of factor analysis: exploratory and confirmatory. Exploratory factor analysis (EFA) is used when the researcher does not know how many factors are necessary to explain the interrelationships among a set of characteristics, indicators, or items (Gorsuch, 1983; Pedhazur & Schmelkin, 1991; Tabachnick & Fidell, 2001). Therefore, the researcher uses the techniques of factor analysis to explore the underlying dimensions of the construct of interest. This was the approach that Leske (1991) used in her conceptualization of the dimensions of needs of families of the critically ill. EFA is the most commonly used form of factor analysis in health care research. It is what we will use to examine the dimensions of Concerns About Genetic Testing.

In contrast, confirmatory factor analysis (CFA) is used to assess the extent to which the hypothesized organization of a set of identified factors fits the data (Nunnally & Bernstein, 1994; Pedhazur & Schmelkin, 1991). It is used when the researcher has some knowledge about the underlying structure of the construct under investigation. CFA could also be used to test the utility of the underlying dimensions of a construct identified through EFA, to compare factor structures across studies, and to test hypotheses concerning the linear structural relationships among a set of factors associated with a specific theory or model. Pett, Wampold, Turner, and Vaughan-Cole (1999), for example, used CFA to test a hypothesized model predicting the paths of influence of divorce on young children’s psychosocial adjustment.

When undertaking a factor analysis using EFA, it is common practice to use more traditional statistical computer packages (e.g., SPSS, SAS, and BMDP) for the statistical analyses. CFA, on the other hand, requires a comprehensive analysis of covariance structures (Byrne, 1989). This form of measurement model is available in structural equation modeling (SEM). LISREL (Jöreskog & Sörbom, 1989) and EQS (Bentler, 1985) are two statistical computer packages that are used to undertake SEM analyses.

Pet et al. (2003)