Confirmatory factor analysis decision sequence.

Abstract

Although the basic concepts of factor analysis are roughly a century old (see Spearman, 1904), more recent extensions of the exploratory factor analysis (EFA) concepts have created the basic methods for confirmatory factor analysis (CFA; see Joreskog, 1969). Both EFA and CFA are part of the general linear model (GLM). Thus, many of the concepts introduced previously with regard to EFA apply in CFA as well. Indeed, an important emphasis in this book is elaborating the similarities between EFA and CFA as part of a single underlying GLM, while also highlighting the differences in the two sets of methods.Of course, the basic difference in CFA is that one or more underlying models (i.e., how many factors, which measured variables are thought to reflect which latent variables, whether the factors are correlated) must be specified even to run the analysis. There are also statistical analyses that are possible in CFA but impossible in EFA (e.g., allowing error variances to be correlated).And some procedures that are routine in EFA, such as factor rotation, are irrelevant in CFA. This is because the a priori models themselves typically specify simple structure, by constraining certain factor pattern coefficients to be zero while freeing other pattern coefficients to be estimated.Confirmatory factor analysis is a very important component within a broader class of methods called structural equation modeling (SEM), or some 109

Perceived context of reception among recent Hispanic immigrants: Conceptualization, instrument development, and preliminary validation.

“…Following Conway and Huffcutt (2003), we used a multi-stage process to decide on the number of factors to extract. First, we examined the scree plot (Thompson, 2004), where the “leveling-off-point” on the scree line represents the last factor that should be extracted. Second, we considered only factors with eigenvalues above 1.00, where the eigenvalue represents the product of the number of items entered into the analysis and the percentage of variability accounted for by the factor (Kaiser, 1958).…”

“…Because three or more items are required to create a reliable factor (Thompson, 2004), it is important to generate more items for positive context of reception in future work. Third, we sampled from two U.S. metro areas with large Hispanic populations and with fairly long histories of Hispanic immigration.…”

“…Both phases used a sample of recent Hispanic immigrant families from Miami and Los Angeles. Phase 1 of the study involved (a) developing items assessing perceived context of reception based on sociological writings on this construct and (b) using exploratory and confirmatory factor analyses on randomly selected half-samples to examine the factorial validity of scores generated by the measure that these items comprised (Thompson, 2004). In Phase 2, discriminant and incremental validity were evaluated between perceived context of reception and perceived discrimination in three ways: (a) examining bivariate associations between these two constructs; (b) regressing both of these constructs on Hispanic and American practices, values, and identifications to ascertain whether perceived discrimination and context of reception would relate differently to indices of acculturation; and (c) mapping the associations of perceived discrimination and context of reception with depressive symptoms (a correlate of perceived discrimination; Torres & Ong, 2010) to examine whether perceived context of reception would contribute additional variance in depression beyond that accounted for by perceived discrimination.…”

Factor Structure, Reliability, and Convergent Validity of the Engagement in Meaningful Activities Survey for Older Adults

“…Evaluating factor structure by means of EFA provides a key piece of evidence for establishing the validity of a measuring instrument (Thompson, 2004). Typically, scale developers will provide some information about the manner with which the instrument is intended to measure the construct of interest.…”

Adipocytokines as features of the metabolic syndrome determined using confirmatory factor analysis

“…In this approach, a hypothesized model is visually constructed by linking different observable variables (indicators) with hypothesized latent (unobservable, underlying) factors [4]. The analysis provides estimates of the strength of the relationships between the different components and goodness of fit indices that indicate the adequacy of the model [40]. …”