Page 212 - Quantitative Data Analysis

Quantitative Data Analysis
Simply Explained Using SPSS

While in exploratory factor analysis, each variable’s loading is
defined as,

Factor analysis assumes that the covariation on in the observed
variables is due to the effect of one or more construct (latent
variable) that utilize casual influence on these observed variables.
Principal Component Analysis Exploratory Factor Analysis
PCA uses total variance of all EFA divides total variance into
1 variables and treats it as a true measurement error and true
variance. factor variance.
EFA can give multiple solutions
PCA gives one and only one
2 depends on the estimation
unique solutions
methods of communality.
the component analysis is
simply a transformation of the
3 original variables into the new Whereas, factor analysis, the
factors can only be estimated.
set of linear combination
(Stevens, 2012)
The goal of FA (EFA/CFA) is to
The goal of PCA is to extract as explain maximum covariance
4 much as variance with few with a few possible factors that
possible components are tied specifically to assumed
constructs.
EFA states the observed
PCA states the component(s) as
5 variables as a function of
a functions of variables
factor(s).

The Theory and Applications of Statistical Inferences 196