Basics of Factor Analysis imagine that we have test scores in
What is factor analysis? algebra, geometry, trigonometry,
A statistical technique or a way for and calculus, if someone excels in
researchers to identify relationships one, they tend to excel in the others
between many variables to simplify data as well) factor analysis can group
by finding underlying patterns. them into one category, which
First step might be called “Factor M” that
It begins by observing numerous represents mathematical ability.
individuals and quantifying those Second step
observations. It might be about people’s Determine how much individual score is
height, weight, test scores, or job related to each of the identified factors
performance, and these measurements and this is measured using:
are turned into numbers (e.g., If Factor loadings
researchers want to know how each o correlation of scores with
measurement relates to the others, for factors
instance, if they have 1,000 o it shows how strongly a
measurements for 5,000 people, they specific score is tied to a
do it by calculating the correlation particular factor
coefficient, a mathematical expression o helps us understand how
that expresses the degree of “pure” each factor is—
correspondence between two sets of essentially
scores.)
How does factor analysis help in the
research data? The axes on which the scores are plotted are usually
Factor analysis helps us find these turned or rotated into a specific mathematical
patterns (relation of two variables), relationship with each other, in order to the
making it easier to see the bigger mathematically derived factors to have psychological
picture in complex data. meaning. The rotation can be either orthogonal or
Some things might be high and oblique, But orthogonal rotation are favored by the
positive or closely related (leg advocates of Five-Factor Theory
length and height), some might be
near zero while others might be
negative or not related at all. (e.g.,
leadership ability might be
positively correlated to social
confidence, suggesting that both
traits are influenced by self-
confidence.
It groups related variables into a
smaller set of “factors” or “traits” that
represent a cluster of closely related
variables. So instead of analyzing all
1,000 variables, we can just consider
one broader trait that explains the
data precisely.
Having a huge set of data can be
overwhelming in analyzing all the
relationships between them, like
1,000 separate measures (e.g.,
