There's two types of Factor Analysis as we know, that is Confirmatory Factor Analysis(CFA) and Exploratory Factor Analysis(EFA). Do you know any other kind of it?
Dear Md. Israt Hasan ,
I think that's it. However, there are many variants of EFA. I think you may be confusing the purpose for the analysis with the analysis algorithms.
Md. Israt Hasan, Peter Samuels' advice is good. You might also see principal components analysis (PCA) being referred to in relation to factor analysis, but it is not really factor analysis. Essentially, it's an earlier kind of analysis that has, for many purposes, been replaced by factor analysis.
This can be confusing because SPSS has PCA as the first extraction method in what people might think is factor analysis because it appears under its Dimension Reduction / Factor / Extraction options.
There are lots of types of factor analysis, depending on what you mean. For example, Bartholomew et al. (in editions that go back many decades) described the general linear latent variable model, which differentiated models by whether the latent and manifest variables are metric (i.e., quantitative) or categorical variables, so for example IRT has metric latent variables and (usually) binary manifest variables ( Book Latent Variable Models and Factor Analysis: A Unified Approach
). And there are further variations among these, e.g., http://www.gllamm.org/ ). Within each of these types these are different estimation procedures (the random variables in multilevel models/Bayesian models can be thought of conceptually like the constructs/latent variables), and different ways of using them (e.g., to compare different models, to estimate specific hypotheses, or trying to find some good fitting model without knowing anything about the data [i.e., exploratory]). So lots of types, but it better to think what models you want to estimate rather then worrying about how they are labelled.
Robert Trevethan much obliged to hear from you and getting your amiable suggestion.
I think that they are the ones. However, it is of great importance to differentiate the two approaches and their appropriate application areas.
EFA is mainly used for scale reduction e.g. (developing a new scale) then using EFA the new (items/questions) have to be examined if they are properly loaded on the targeted (factors/constructs)).
In the other hand, CFA is mainly used when the aim is to validate a measurement model in the structural equation context. In other words, CFA examines if the (items/questions) proper measure the selected (factors/constructs).
Another major difference in the approaches is that using EFA, sometimes called dimension reduction, the researcher would allow the data to play the role of specifying which (items/questions) load on one (or/and) multiple (factor(s)/construct(s)) with a prior selection of a rotation method. Mostly used techniques are the oblique rotation with (oblimin) or orthogonal rotation with (Varimax). The main target is to ensure the unidimensionality of the (factor/construct) where each (item/question) loads solely on only one (factor/construct).
In the other hand, using the CFA the researcher has to specify which (items/questions) measure which (factor/construct). Later, the results of CFA are used to assess the validity and reliability of the measurement model.
I recommend
Article Exploratory Structural Equation Modeling, Integrating CFA an...
Article Exploratory Structural Equation Modeling: An Integration of ...
Article A Comparison of CFA, ESEM, and BSEM in Test Structure Analysis
Check the following web-link, please.
https://www.toppr.com/guides/algebra/factorisation/factor-analysis/
- Badges
- Science topic
- Similar topics
- Social Psychology
- Attitude