BUSINESS RESEARCH

CHAPTER 7

MULTIVARIATE TECHNIQUE 1: FACTOR

ANALYSIS
RESEARCH MEASUREMENT TOOLS: PART 2
Factor Analysis
● Factor analysis (FA) is also known as exploratory factor analysis (EFA). It is called
EFA to differentiate it from confirmatory factor analysis (CFA), a technique based
on covariance-based structural equation modeling (CB-SEM). CFA is not covered
in this chapter's discussion.
● FA or EFA is a major technique in multivariate statistics. In fact, EFA and multiple
regression analysis (MRA), another multivariate technique which will be discussed
after FA, are the foundations for developing CB. SEM. Thus, FA and MRA are the
first order and CB-SEM is the second order in terms of statistical development. If a
student wants to learn CB-SEM, he or she should learn and understand FA and
MRA first.
FA and CFA are based on interdependence techniques. In an
interdependence technique, the variables are not referred to as
independent, dependent, or moderating. The variables are not
labeled and their relationships are manifested with a line with
two-headed arrows.
 ANALYSIS AND INTERPRETATION OF EFA RESULTS
The EFA results will be analyzed and interpreted based on the following
output:
1. Measure of sampling adequacy (MSA) through Kaiser-Meyer-Olkin (KMO)
Test and Bartlett's test of sphericity.

2. Total variance explained

The cumulative variance explained should be 60% or higher to make it valid
and reliable. The cumulative variances for the unrotated (extraction sums of
squared loadings) and the rotated (rotation sums of squared loadings) have
the same final value. However, for this study, the latter is the one being
applied for interpretation.
3. Communalities of variables
Burns and Burns (2008) stated: "Each observed variable's communality is the proportion of
variance in that variable that can be explained by the common factors. It is calculated by summing
the squared factor loadings for each variable. Communality below .50 is interpreted as failing to
explain the other variables"

4. Scree plot
The scree plot is a graphical presentation of the factors as they are shown based on their
eigenvalue. Ideally for a factor to be considered significant, it should have an eigenvalue of
greater than 1 (Kaiser criterion). Below the eigenvalue of 1, the factors are not considered
anymore in the analysis. Normally, the scree plot would show a steep first factor and the
remaining factors would drop drastically down to eigenvalue 1 (called the elbow of scree plot as it
looks like an elbow) and gradually go further down (below eigenvalue 1) to the last factor. The
scree plot normally depicts the distribution of unrotated factors. When the factors are rotated, the
high eigenvalue of the first factor will be shared to the other factors, thus tapering its height.
5. Unrotated or rotated factor loadings
The researcher can opt for unrotated factor loading when interpreting it. However,
based on the advice of statisticians and researchers, it is better to rotate technique
orthogonally. If the researcher is using SPSS, the favorite is orthogonal as it would
make the interpretation easier. In general, the favorite rotation is Varimax.
STEPS IN PROCESSING FACTOR ANALYSIS

1. Encode in an Excel template

2. Import the Excel data from the SPSS environment or interface through:
a. Open the SPSS software; On the command menu: File > Open > Data >
Look in: (where the file is located - either drive C, drive D or E, or desktop); In
the Files of type: change it from SPSS (.sav) to Excel (.xls, xlsx, xlsm).
b. Press Open > Worksheet - select the sheet your file is stored - then press
Ok.
c. On the SPSS interface, accept or change the data name, type, width,
decimal, label, values, and measure at the variable view.
3. Click Analyze > Dimension Reduction > Factor (and shows Factor
analysis dialog box).
FA OUTPUT AND INTERPRETATION

The result showed that the KMO's Measure of Sampling Adequacy is .832
and is interpreted as meritorious, whereas the Bartlett's test of sphericity
has a chi-square value of 415.947 and is significant (P value .000). Both
tests showed that the data are fit for factor analysis.
COMMUNALITIES

The percentage of variance for a specific item that may be attributed to

the factors is shown by adding up the squared loadings for that item on
a factor matrix. refer to this as the communality. The extent to which the
extracted components account for the variation of the item increases
with the communality value

The percentage of variance explained by each component is indicated

by the mean score for the sum of the squared factor loadings.
TOTAL VARIANCE EXPLAINED

All of the different major component variances added together make up

the total variance. The ratio between the variance of a principal
component and the total variance is known as the fraction of variance
explained by that principal component. Divide the sum of the variances of
multiple primary components by the variance of the entire set.
SCREE PLOT

A scree plot is a graphical representation of the eigenvalues of factors or

principal components in a dataset. It helps in determining the number of
factors or components to retain in a factor analysis or principal component
analysis, respectively. Typically, the scree plot shows eigenvalues on the
y-axis and the number of factors or components on the x-axis. The
"elbow" of the plot, where the eigenvalues start to level off, indicates the
optimal number of factors or components to retain, as it signifies the point
of diminishing returns in explaining the variance of the data.
ROTATED COMPONENT MATRIX

A rotated component matrix is a table that displays the loadings of

variables on each factor after a rotation method, such as Varimax or
Promax, has been applied in factor analysis. It helps in interpreting the
relationships between variables and factors by showing which variables
are most strongly associated with each factor. The rotation aims to
simplify the structure of the factors and make them easier to interpret.
Each cell in the matrix represents the strength of the relationship between
a variable and a factor, with higher values indicating stronger
associations.
INTERPRETATION OF FACTOR ANALYSIS
The researcher applied the KMO's measure of sampling adequacy (MSA) and the Bartlett
test of sphericity to determine whether the data are appropriate for exploratory factor
analysis.
The 9 factors derived from factor analysis using principal component analysis

Factor 1. ( Labeled as Mall Reasonable Prices )

Factor 2. ( Labeled as Mall Promotion and Entertainment )
Factor 3.( Labeled as Mall Atmospherics and Comfort )
Factor 4. ( Labeled as Mall Product Variety)
Factor 5. ( Labeled as Mall Service )
Factor 6. ( Labeled as Mall Density )
Factor 7. ( Labeled as Mall Accessibility to Public Vehicles )
Factor 8. ( Labeled as Mall Accessibility to Private Vehicles )
Factor 9. ( Labeled as Mall Smell )
FACTOR ANALYSIS VALIDATION

The Validation of factor Analysis was done internally using the split
sample method. The sample was split into two equal group using
SPSS

Validating FA using split samples through SPSS.

BUSINESS RESEARCH
CHAPTER 8

MULTIVARIATE TECHNIQUE 2: MULTIPLE

REGRESSION ANALYSIS
RESEARCH MEASUREMENT TOOLS: PART 3
Multiple regression analysis

Multiple regression analysis (MRA) is a multivariate tool based on the dependence

technique, which will explain or predict the relationships of the independent
variables and the dependent variable. The independent variables are called
predictors or explanatory variables, whereas the dependent variable is called the
criterion variable or explained variable.
According to Janssens et al. (2008) "Regression analysis is a technique which is
used to determine the causality between one dependent interval or ratio scaled
variable (the explained variable) and one or more independent interval or ratio
scaled variables (the explanatory variables). In other words, one tries to explain
the variation in one dependent as much as possible on the basis of the variation in
a number of relevant independent variables. If there is only one independent
variable, then this is a 'simple regression', while 'multiple regression' is the term
used when multiple independent variables are involved"
APPLICATION OF MRA USING THE STEPWISE METHOD

The writers use the study of Te (2011) about Positioning the HEIs in Davao
City as an example of how to use multiple regression analysis.

Main Problem: What factors influence the quality education of HEIs in Davao
City?

Subproblems:

1. Do qualified faculty, better school facilities, campus culture, and

classroom climate influence quality education?

2. Does sex moderate the influence of the independent variables (qualified

faculty better school facilities, campus culture, and classroom climate) to
the dependent variable (quality education)?
STEPS IN PROCESSING MRA
1. Import the Excel data from SPSS environment or interface.
2. Open the SPSS software; On the command menu: File > OPen >
Data > Look in: (where the file is located - either Drive C, Drive
D or E, or desktop); In the Files of type: change it from SPSS
(.sav) to Excel (.xls, xlsx, xlsm).
3. Press Open > Worksheet - select the sheet where your file is
stored - then press Ok.
4. On the SPSS interface, accept or change the data name, type,
width, decimal, label, values, and measure at the valuable view
(see Figure 8.3).
THE MRA EQUATION

Note: To apply the mean for projection purposes, subtract its

standard deviation to get the minimum value and add its
standard deviation to get the maximum value
HOW TO APPLY THE MODERATING (DUMMY) VARIABLE

Null hypothesis: Does sex moderate the influence of the

independent variables qualified faculty, better school facilities,
campus culture, and classroom climate) to the dependent
variable (quality education)?

Based on the null hypothesis, the categorical variable Sex is

transformed first as a dummy variable before it can be processed
using the MRA Enter method. The following steps will guide the
student in transforming the categorical variable Sex:
1. Go to the SPSS Menu.
TIPS FOR MODERATING VARIABLES

As much as possible, minimize your moderating variables

(use it with moderation). Include only those that have
practical or theoretical relevance to your research.
Remember that each category of your moderating
variable/s will have to be measured. Note that for two
categories, measure only one; for three categories,
measure only two, and so on.
MRA VALIDATION

Based on split-samples, the data were validated for

generalization. The original data were split into two samples
using SPSS.
The data for split sample 1 and split sample 2, when
compared with the data on the main sample, are not very
different. Thus, the data can be generalized across the
population.
Just like factor analysis, use SPSS to split the samples for
validation purposes.
THANK
YOU!