Page 202 - Quantitative Data Analysis

Quantitative Data Analysis
Simply Explained Using SPSS

Principal Component Analysis (PCA)

Basic Concept
If two or more items are highly correlated that means that they may
represent the same phenomenon. In other words, they are
measuring same underlying variance. Combining such highly
correlated items called component in PCA and factors in EFA. For
example,

Item 1 : I am very good in computer

Item 2 : I have skills to use computer
Item 3 : I feel confident when I operate computer

These three items are measuring same underlying phenomenon. It
is high likely the response pattern for these three items would be
same which means there would be high correlation among these
items. Using PCA or EFA procedure, the researcher can reduce these
three items into single variable which he/she can use in further
analysis.

Principal Component Analysis (PCA)
Principal Component Analysis is variable reduction technique. It
reduces multiple observed variables (i.e., the number of items in an
instrument) into fewer components that summarize their variance.
PCA procedure allows combining several correlated variables into
one component then those components can be used in further
analysis. The goal of PCA is to express all the set of variables into
few possible components. PCA is also useful to find the pattern of
association across different set of variables. PCA procedure also
helps to detect multi normality and multi co-linearity issues among
the set of predicators.

The Theory and Applications of Statistical Inferences 186