From the given word “factor”, the strategic concept analysis
in factor analysis notes that the observation of the statistical analysis technique is
focused on the variables being analyzed.
The 9 fundamental steps in conducting Factor Analysis are academically
essential in achieving the right output data from the variables. As the general notion that factor analysis is
similar to regression analysis in terms of variable observation, we can
mentally visualize that factor analysis involves more than 1 variable for
observation. In the process of the
observed variables are compared to the component, and the nearest variables
to the component are likely to be selected as the variable for testing. In another case, if two variables overlap and
acquire the same degree of significance the combination of the two variables will
likely to occur.
STEPS IN CONDUCTING FACTOR ANALYSIS
1 Formulate the Problem- by formulating the problem
as the normal research process will happen. Problem will give the inferable
subject that can be considered as the component of the study. Likewise after the component is created,
possible variables will emerge as the pre-observed factors for testing.
2 Construct the Correlation Matrix- the correlation
matrix process will allow the testing of relatedness between the
variables. So that the factor analysis
will be appropriate variables must be correlated. If in any cases the correlations between
variables are small, therefore factor analysis may not be appropriate.
3 Determine the Method of Factor Analysis- Several
methods are also given as the variety of methods to be used in analyzing the
variables
a.
Principal Components Analysis- an approach to
factor analysis that consider the total variance in the data
b.
Common Factor Analysis- an approach to factor
analysis that estimates the factors based only on the common variance.
4 Determine the Number of Factors- in determining the
number of factors that will be used in the observation several methods can also
be used in order to compute the principal components to be included.
a.
A Priori Determination
b.
Determination based on Eigenvalues
c.
Determination based on Scree Plot
d.
Determination based on Percentage of Variance
e.
Determination based on Split-Half Reliability
f.
Determination based on significance tests
5 Rotate Factors- in rotate factor which is also
called the factor pattern matrix is observing the coefficient with a large
absolute value indicates that the factor and the variable are closely
related. The 3 major types of rotate
factors are:
a.
Orthogonal Rotation- rotation of factors in
which the axes are maintained at right angles.
b.
Varimax- An orthogonal method of factor rotation
that minimizes the number of variables with high loadings on a factor, thereby
enhancing the interpretability of the factors.
c.
Oblique Rotation- Rotation of factors when the
axes are not maintained at the right angles.
6 Interpret Factors- in this stage given factors can
be presented in various manners to show which factors deserved to be considered
in the final model. The most common
interpretation of variables is by plotting the variables in a mathematical table.
7 Calculate Factor Scores- calculating the factor
scores estimated for each respondent on the derived factors.
8 Surrogate Variables- is relating to the subset of
original variables selected for use in subsequent analysis.
9 Determine the model of Fit- determining the model
of fit will finalize the actual model to be used in the research resulting from
the factor analysis.
(Reference: Malhotra 1999 Marketing Research an Applied Orientation)
(Reference: Malhotra 1999 Marketing Research an Applied Orientation)
SPSS instructions for Solving Factor Analysis
Comments