PHƯƠNG PHÁP NGHIÊN CỨU

Research Methodology

Part 3: Exploratory Factor Analysis

Pham Thai Binh, Ph.D

UEH, Nov 2024
DEFINITION

• Exploratory factor analysis, often referred to as EFA, is an interdependence

technique, whose primary purpose is to define the underlying structure among
the variables in the analysis.

2
Exploratory Factory Analysis

• When adding more variables, overlapping (correlation) between variables are more likely:
o Reduce the number of variables to be analyzed
o Select the appropriate variables to be included in the multivariate analyses

• Factor analysis provides the tools for analyzing the structure of the interrelationships
(correlations) among a large number of variables
• Factors or components: sets of variables that are highly interrelated

• A continuing debate concerns its appropriate role:

o In many applications, its role is an exploratory approach, useful in searching for structure
among a set of variables or as a data reduction method
o In other cases, a confirmatory approach is used when researcher wish to test hypotheses on
preconceived notions about data structure.

3
Exploratory Factory Analysis

• Factor analysis is different from the Principal Component Analysis (PCA) in terms of statistical
interpretation:
✓ PCA is a linear combination of variables;
✓ Factor Analysis is a measurement model of a latent variable.

Principal Component Analysis Factor Analysis: Measurement Model

C = w1(Y1) + w2(Y2) + w3(Y3) + w4(Y4) Y1 = b1*F + u1 ; Y2 = b2*F + u2 ; Y3 = b3*F + u3 ; Y4 = b4*F + u4

4
A Hypothetical Example: Conducting Survey

• A retail firm identified 80 different characteristics of retail stores and their service that consumers
mentioned as influencing their patronage choice among stores
• Conduct a survey for consumer evaluations on each of the 80 specific items.

Question may wish to address:

• Are all of these elements separate in their evaluative properties or do they group into some more general
areas of evaluation?
• For example, do all of the product elements group together?
• Where does price level fit, or is it separate?
• Do the in-store features (e.g., store personnel, service, and atmosphere) relate to one another?

5
A Hypothetical Example: Conducting Survey

In-store experience
of shoppers

Product assortment
and availability
Product quality and
price levels 6
Factor Analysis Decision Process

Stage 1 Objectives of Factor Analysis

Stage 2 Designing an Exploratory Factor Analysis
Stage 3 Assumptions in Exploratory Factor Analysis
Stage 4 Deriving Factors and Assessing overall Fit
Stage 5 Interpreting the Factors
Stage 6 Validation of Exploratory Factor Analysis
Stage 7 Data Reduction – Additional Uses of Exploratory Factor Analysis Results

7
Stage 1: Objectives of Factor Analysis

EFA focuses on four issues:

• Specifying the unit of analysis
• Achieving data summarization and/or data reduction
• Variable selection
• Using exploratory factor analysis results with other multivariate techniques.

8
Stage 1: Objectives of Factor Analysis

Specifying the unit of analysis

R Factor Analysis
• Research objective: summarize the characteristics (i.e., variables)
• Factor analysis would be applied to a correlation matrix of the variables
• Analyze a set of variables to identify the dimensions for the variables that are latent (not easily
observed).
Q Factor Analysis
• Factor analysis would be applied to a correlation matrix of the individual cases
• Combines or condenses large numbers of cases into distinctly different groups within a larger population
• Not utilized frequently because of computational difficulties
• Instead, most researchers utilize some type of cluster analysis to group individual cases.

9
Stage 1: Objectives of Factor Analysis

Achieve data summarization and/or data reduction

Data Summarization
• Derives underlying dimensions that describe the data in a much smaller number of concepts than the
original individual variables
• The fundamental concept involved is the definition of structure, through which researchers can view the
set of variables at various levels of generalization
• Variate: the linear composite of variables
• In EFA, the variates (factors) are formed to maximize their explanation of the entire variable set, not to
predict a dependent variable(s)
• The goal of data summarization is achieved by defining a small number of factors that adequately
represent the original set of variables.

10
Stage 1: Objectives of Factor Analysis

Achieve data summarization and/or data reduction

Data Summarization
• Without interpretation:
o Based solely on the intercorrelations between variables
o Principal components regression employ this procedure strictly to reduce the number of variables in the
analysis with no specific regard to interpretability
o Create a parsimonious set of variables that aid with model estimation and development
• With interpretation:
o For managerial purposes or utilize the procedure to assist in scale development
o Scale development: A specific process focused on the identification of a set of items that represent a
construct (e.g., store image) in a quantifiable and objective manner
o Common factor analysis is most applicable to interpretability since it analyzes only the shared variation
among the variables
11
Stage 1: Objectives of Factor Analysis

Achieve data summarization and/or data reduction

Data Reduction
Purpose:
• To retain the nature and character of the original variables, but reduce the number of actual values
included in the analysis
Techniques:
(1) Identify representative variables from the much larger set of variables represented by each factor for use
in subsequent multivariate analyses, or
(2) Create an entirely new set of variables, representing composites of the variables represented by each
factor, to replace the original set of variables.
When combine using the result of data summarization, loadings and factor scores help to identify the final
set of variables.

12
Stage 1: Objectives of Factor Analysis

Variable Selection

• Variable Specification
• Factors are always produced
• Factors require multiple variables

13
Stage 1: Objectives of Factor Analysis

Using Factor Analysis with other multivariate techniques

• Highly correlated variables affect the incremental predictive power of sequential variables
entered in multiple regression and discriminant analysis
• EFA addressed the intercorrelations by adding new subset of variables, reducing the
dimensions and increasing interpretability

14
Stage 2: Designing an Exploratory Factor Analysis

Involve 3 basic decisions:

1. Design of the study in terms of the number of variables, measurement properties of variables, and the
types of allowable variables
2. The sample size necessary, both in absolute terms and as a function of the number of variables in the
analysis
3. Calculation of the input data (a correlation matrix) to meet the specified objectives of grouping variables
or respondents.

15
Stage 2: Designing an Exploratory Factor Analysis

Variable Selection and Measurement Issues

What types of variables can be used in factor analysis?

How many variables should be included?

• A correlation value must be computable among all variables

• Dummy variables (coded 0-1) are created for non-metric variables

o If all variables are dummy, Boolean factor analysis is more appropriate

• Include several and reasonable number of variables that represent each proposed factor

• Identify key variables that closely reflect the hypothesized underlying factors.

16
Stage 2: Designing an Exploratory Factor Analysis

Sample Size

• Absolute size: ≥ 50 observations, preferably ≥ 100; or ≥ 200 observations

• Ratio (observations to variables): 5:1, preferably 10:1 or 20:1

• Communality - the amount of a variable’s variance explained by its loadings on the factors, calculated as the
sum of the squared loadings across the factors.

• Fabrigar and Wegener guideline:

For 100 observations:
a. Communalities of 0.70 or above and 3 variables with high loadings on each factor: sufficient
b. Communalities in the range of 0.40 - 0.70: sample size should be at least 200
c. Communalities below 0.40 and few high loadings per factor: sample sizes of up to 400 are appropriate

• Increase the sample as the complexity of the factor analysis increases

17
Stage 2: Designing an Exploratory Factor Analysis

Correlations among Variables or Respondents

For Q-type factor analysis:

• Identify groups or clusters of individuals that exhibit a similar pattern on the variables included in
the analysis

• Difference to Clustering: based on intercorrelation between respondents rather than distance-

based similarity.

18
Stage 2: Designing an Exploratory Factor Analysis

Correlations among Variables or Respondents

19
Stage 3: Assumptions in Exploratory Factor Analysis

Assumption: Some underlying structure does exist in the set of selected variables

• Just because the variables are statistically correlated doesn't mean they are conceptually relevant
• Ensure that the observed patterns are conceptually valid
• Ensure that the sample is homogeneous with respect to the underlying factor structure
• Example:
o Mixing independent and dependent variables in a single factor analysis, then using that factor to
explain the dependence relationship is not appropriate
o If males and females respond differently to a set of items, analyzing them together may create
misleading results. Instead, separate analyses should be done for each group to understand their
unique patterns, then compare the results.

20
Stage 3: Assumptions in Exploratory Factor Analysis

Assumption: Normality, homoscedasticity, linearity

• Departures only apply when they diminish the observed correlations

• Some degree of multicollinearity is desirable
Measures of intercorrelations:
• Visual inspection
o Partial correlation - correlation that is unexplained when the effects of other variables are taken into account,
high correlation when > 0.70
o High partial correlations indicates no underlying factors
o Anti image correlation matrix – negative value of partial correlations
• Barlett test
o Bartlett test of sphericity - a statistical test for the presence of correlations among the variables.
o Provide the statistical significance of the correlation matrix on the significant correlations between variables

21
Stage 3: Assumptions in Exploratory Factor Analysis

Measure of Sampling Adequacy (MSA) – Kaiser-Meyer-Olkin (KMO) test

• Ranges from 0 to 1, reaching 1 when each variable is perfectly predicted without error by the other
variables.
• Guidelines:
o .80 or above: meritorious
o .70 or above: middling
o .60 or above: mediocre
o .50 or above: miserable
o Below .50: unacceptable
• The MSA increases as: the sample size increases, the average correlations increase, the number of
variables increases, or the number of factors decreases

22
Stage 3: Assumptions in Exploratory Factor Analysis

Variable-specific Measure of Intercorrelation

• The MSA extended to measure individual variables

• The process start with deleting the variable that has MSA that is below .50, then recalculate the
factor analysis. Repeat the same process until all unit have acceptable MSA value.

• This is done because low-MSA-value variables are often end up as single variable factor

23
Stage 4: Deriving Factors and Assessing overall Fit

Selecting the Factor Extraction Method

• Common variance - Variance in a variable that is shared with all other variables in the analysis
• Unique variance - Variance associated with only a specific variable and is not represented in the
correlations among variables
• Variables with high common variance are more amenable to exploratory factor analysis
• Specific variance - Reflects the unique characteristics of that variable apart from the other variables in
the analysis
• Error variance - Variance that is due to unreliability in the data-gathering process, measurement error,
or a random component in the measured phenomenon.
• Total variance – has two basic sources, common and unique variance (with 2 subparts - specific and
error variance)

24
Stage 4: Deriving Factors and Assessing overall Fit

Selecting the Factor Extraction Method

Principal Component Analysis - Variance in a variable that is shared with all other variables in the
analysis
Common Factor Analysis - Variance associated with only a specific variable and is not represented in the
correlations among variables
Selection criteria:
1. The objectives of the factor analysis
2. The amount of prior knowledge about the variance in the variables.

25
Stage 4: Deriving Factors and Assessing overall Fit

Selecting the Factor Extraction Method

26
Stage 4: Deriving Factors and Assessing overall Fit

Selecting the Factor Extraction Method

When is Principal Component Analysis most appropriate?

• Data reduction is a primary concern, focusing on the minimum number of factors needed to
account for the maximum portion of the total variance represented in the original set of variables\
• Prior knowledge suggests that specific and error variance represent a relatively small proportion of
the total variance
• The principal component results are used as a preliminary step in the scale development process

27
Stage 4: Deriving Factors and Assessing overall Fit

Selecting the Factor Extraction Method

When is Common Factor Analysis most appropriate?

• The primary objective is to identify the latent dimensions or constructs represented in the common
variance of the original variables, as typified in the scale development process
• The researcher has little knowledge about the amount of specific and error variance and therefore
wishes to eliminate this variance.

Common Factor Analysis suffers from factor indeterminacy and invalid communalities (less than 0 or
above 1) that requires deletion of the variable.
Empirical research showed identical results for most applications using the 2 methods.

28
Stage 4: Deriving Factors and Assessing overall Fit

Stopping criteria: Criteria for number of factors to extract

• Priori criterion – Set the number of factors based on prior research/objectives

• Latent Root criterion
o Also known as Kaiser rules
o Disregard any factors that have latent roots or eigenvalues less than 1; only factors that have
eigenvalues more than 1 are considered significant
o Eigenvalues – sum of square loadings of variables on factor
o Criticism: too simplistic
o Most reliable when there are 20 – 50 variables and communalities above .40

29
Stage 4: Deriving Factors and Assessing overall Fit

Stopping criteria: Criteria for number of factors to extract

• Percentage of Variance criterion – Retains factors until a specific cumulative percentage of total
variance is explained.
o In natural sciences: Should not be stopped until extracted factors account for at least 95% of the
variance
o In social sciences: considers solutions that account for 60% of the total variance
o Variant approach: selecting enough factors to achieve a prespecified communality for each of the
variables
• Scree Test criterion
o Identify the optimum number of factors that can be extracted before the amount of unique variance
begins to dominate the common variance structure
o “Elbow”: the point where the unique variance begins to dominate the common variance structure.

30
Stage 4: Deriving Factors and Assessing overall Fit

Stopping criteria: Criteria for number of factors to extract

31
Stage 4: Deriving Factors and Assessing overall Fit

Stopping criteria: Criteria for number of factors to extract

• Parallel Analysis
o Form a stopping rule based on the specific characteristics
o Generate a large number of datasets from random values and then factor analyze each of the
dataset
o Result: average eigenvalue for the first factor, the second factor and so on across the simulated
datasets.
o Compare the eigenvalues of the original data with the simulated datasets → retain factors that have
eigenvalues above the eigenvalues of simulated datasets.
o Variant: taking the upper bound (95th percentile) of the simulated eigenvalues as a more
conservative threshold
o With PCA, Parallel Analysis tend to be “stricter” than Latent Root criterion as the number of factors
increases; with CFA, the opposite has been observed.

32
Stage 4: Deriving Factors and Assessing overall Fit

Stopping criteria: Criteria for number of factors to extract

Heterogeneity of the Respondents

• If the sample is heterogeneous with regard to at least one subset of the variables, then the first factors
will represent those variables that are more homogeneous across the sample
• Variables that are better discriminators load on later factors, often not selected by the criteria discussed
• Objective: identify factors that discriminate among the subgroups of a sample → additional factors
should be extracted and the additional factors’ ability to discriminate is examined

33
Stage 5: Interpreting the Factors

Three processes of Factor Interpretation

1. Estimate the Factor Matrix

• Initial unrotated factor matrix is computed, containing the factor loadings for each variable on each factor
• Factor loadings: correlation between a variable and a factor
• When EFA is used strictly for data reduction, there is no need for interpreting the factors.

2. Factor Rotation
• Rotational methods are employed to achieve simpler and theoretically more meaningful factor solutions
• Rotation of the factors improves the interpretation by reducing some of the ambiguities that often
accompany initial unrotated factor solutions

34
Stage 5: Interpreting the Factors

Three processes of Factor Interpretation

3. Factor Interpretation and Respecification

• Researcher evaluates the (rotated) factor loadings for each variable to determine that variable’s role and
contribution
• The need may arise to respecify the factor model owing to:
o The deletion of a variable(s) from the analysis
o The desire to employ a different rotational method for interpretation
o The need to extract a different number of factors
o The desire to change from one extraction method to another.
• Respecification: involves returning to the extraction stage (stage 4), extracting factors, and then
beginning the process of interpretation once again.

35
Stage 5: Interpreting the Factors

Factor Rotation
Redistribute the variance from earlier factors to later ones to achieve a simpler, theoretically more meaningful
factor pattern

Axes are maintained at 90 degrees 36

Stage 5: Interpreting the Factors

Factor Rotation

The clustering between two groups is more obvious: V1, V2 and V3, V4, V5.

37
Stage 5: Interpreting the Factors

Orthogonal Rotation Methods

• Preferred method when the research goal is data reduction
• VARIMAX
o Simplify the columns of the factor matrix
o Maximizes the sum of variances of required loadings of the factor matrix
o Logic: Interpretation is easiest when correlations are close to either +1 or -1; or close to 0
• QUARTIMAX
o Simplify the rows of the factor matrix
o Ensure that each variables have high loadings on one factor and as low as possible on the others
o Create a “general factor”, which does not align with the goal of factor rotation
• EQUIMAX
o A compromise between the QUARTIMAX and VARIMAX

38
Stage 5: Interpreting the Factors

Oblique Rotation Methods

• Oblique rotations are similar to orthogonal rotations, except that oblique rotations allow correlated
factors instead of maintaining independence between the rotated factors
• Best suited to the goal of obtaining several theoretically meaningful factors or constructs, as
realistically, few constructs are uncorrelated.
• Statistical packages:
o IBM SPSS: OBLIMIN
o SAS: PROMAX and ORTHOBLIQUE

39
Stage 5: Interpreting the Factors

Judging the Significance of Factor Loadings

Ensuring Practical Significance

• Factor loadings < ±.10: can be considered equivalent to zero for purposes of assessing simple
structure.
• Factor loadings ±.30 - ±.40: are considered to meet the minimal level for interpretation of
structure.
• Loadings ≥ ±.50: are considered practically significant.
• Loadings > ±.70: are considered indicative of well-defined structure and are the goal of any
factor analysis.

40
Stage 5: Interpreting the Factors

Judging the Significance of Factor Loadings

Assessing Statistical Significance

• Factors have substantially large standard errors:
➔ should be evaluated at considerably stricter levels
• Employ the concept of statistical power

41
Stage 5: Interpreting the Factors

Interpreting a Factor Matrix

Simple structure general guidelines (Thurstone, 1947):

• Each variable: should have at least one very low loading (under ± .10)
• Each factor: should have at least as many very low loadings as there are factors
• Each pair of factors:
a. Some variables have a significant loading (greater than .3 or .4) on one factor and very low on
the other.
b. A substantial percentage of very low loadings on each factor when there are four or more
factors.
c. Relatively few cross-loadings (i.e., variables with significant loadings on each factor).

42
Stage 5: Interpreting the Factors

Interpreting a Factor Matrix

5-step procedure to achieve simple structure:

1. Examine the Factor Matrix of Loadings
2. Identify the Significant Loading(s) for Each Variable
3. Assess the Communalities of the Variables
4. Respecify the Factor Model if Needed
5. Label the Factors

43
Stage 5: Interpreting the Factors

Interpreting a Factor Matrix

1. Examine the Factor Matrix of Loadings

• Factors are arranged as columns
• Each column of numbers represents the loadings of a single factor.
• Two matrices of factor loadings in Oblique Rotation:
o Factor pattern matrix: loadings that represent the unique contribution of each variable to the
factor
o Factor structure matrix: loadings contain both the unique variance between variables and
factors and correlation among factors
o It is more difficult to distinguish which variables load uniquely on each factor in the factor
structure matrix with greater factor correlations.

44
Stage 5: Interpreting the Factors

Interpreting a Factor Matrix

2. Identify the Significant Loading(s) for Each Variable

• Start with the first variable on the first factor and move horizontally from left to right
• Cross-loading: a variable has more than one significant loading
• Rotation methods can eliminate cross-loadings
• Detect cross-loadings detracting from simple structure objective:
o Compare variances, not loadings
o Compare as a ratio of variances:
- Larger variance/Smaller variance
- Between 1.0 - 1.5: problematic cross-loading
- Between 1.5 - 2.0: potential cross-loading
- Greater than 2.0: ignorable cross-loading

45
Stage 5: Interpreting the Factors

Interpreting a Factor Matrix

2. Identify the Significant Loading(s) for Each Variable

46
Stage 5: Interpreting the Factors

Interpreting a Factor Matrix

3. Assess the Communalities of the Variables

• Find any variables that are not adequately accounted for by the factor solution.
• Approach:
o Identify any variable(s) lacking at least one significant loading
o Examine each variable’s communality
• e.g. – A researcher may specify that at least one-half of the variance of each variable must be
taken into account → identify all variables with communalities less than .50 as not having
sufficient explanation.

47
Stage 5: Interpreting the Factors

Interpreting a Factor Matrix

4. Respecify the Factor Model if Needed

• Problems may arise:
o A variable has no significant loadings
o A variable’s communality is deemed too low even with significant loadings
o A variable has a cross-loading:
- Ignore those problematic variables
- Possible deletion
- Alternative rotation method
- Decrease/increase the number of factors retained
- Modify the type of factor model used

48
Stage 5: Interpreting the Factors

Interpreting a Factor Matrix

5. Label the Factors

• Variables with higher loadings → more important and have greater influence on the name or
label selected to represent a factor.
• Intuitively developed by the researcher based on its appropriateness for representing the
underlying dimensions of a particular factor
• Result: meaning names or labels that accurately represents each of the derived factors

49
Stage 5: Interpreting the Factors

Example

• Nine measures were obtained in a pilot test based on a sample of 202 respondents
• Through prior research, 3 factor solution was deemed appropriate
• Task: interpreting the factors

50
Stage 5: Interpreting the Factors

Example – Step 1 & 2

• Sample size: 202 → factor loadings ≥ .40

• V1: cross-loading
.5052
• Ratio: .4622 = 1.19 → problematic cross-
loading

51
Stage 5: Interpreting the Factors

Example – Step 3

• V3: low communality (.299)

• Retained in this context

52
Stage 5: Interpreting the Factors

Example – Step 4 & 5

Step 4:
• Factor 1: V7, V8, V9
• Factor 2: V2, V3, V5
• Factor 3: V4, V6.
• Due to cross-loading, V1 is eliminated
and the loadings are recalculated
Step 5: Well-defined three distinct group of
variables

53
Stage 6: Validation of exploratory Factor Analysis

Use of Replication or a Confirmatory Perspective

• Direct methods of validating the results
(a) To assess the replicability/generalizability of the results, either with a split sample in the original dataset
or with a separate sample
(b) Pursue a confirmatory analysis.
Assessing Factor Structure Stability
• Factor stability is primarily dependent on the sample size and on the number of cases per variable
• Split the sample into two subsets (if sample size permits), estimate factor models for each, and compare the
resulting factor matrices to assess solution robustness.
Detecting Influential Observations – If justified, omitting the outliers will lead to better generalization

54
Stage 7: Data Reduction—Additional Uses of Exploratory
Factor Analysis Results

Selecting Surrogate Variable for Subsequent Analysis

• Surrogate variable: variable with the highest factor loading on each factor
• Disadvantages of choosing a surrogate representative for a factor or component:
o Does not address the issue of measurement error encountered when using single measures
o Runs the risk of potentially misleading results by selecting only a single variable to represent a perhaps
more complex result

55
Stage 7: Data Reduction—Additional Uses of Exploratory
Factor Analysis Results

Creating Summated Scales

• Benefits:
o Reducing measurement error – the extent to which observed values deviate from actual values
due to errors (e.g., data entry) or individuals' inability to provide accurate information
- Multiply indicators, reduce the reliance of a single response
o Represent Multiple Aspects of a Concept in a Single Measure - Combine the multiple
indicators into a single measure representing what is held in common across the set of measures

56
Stage 7: Data Reduction—Additional Uses of Exploratory
Factor Analysis Results

Creating Summated Scales

• Conceptual definition - defining the concept being represented in terms applicable to the
research context
o Content validity – assessment of the correspondence of the variables to be included in a
summated scale and its conceptual definition
• Dimensionality – items are unidimensional, strongly associated with each other and represent a
single concept

57
Stage 7: Data Reduction—Additional Uses of Exploratory
Factor Analysis Results

Creating Summated Scales

• Reliability – assessment of the degree of consistency between multiple measurements of a

variable
o Test – Retest: consistency is measured between the responses for an individual at two points
in time, ensure that responses are not too varied across time periods
o Single items: item-to-total correlations exceed .50; inter item correlations exceed .30
o Cronbach’s alpha: should exceed a threshold of .70, although a .60 level can be used in
exploratory research
o CFA measures

58
Stage 7: Data Reduction—Additional Uses of Exploratory
Factor Analysis Results

Creating Summated Scales

• Construct validity – extent to which a scale or set of measures accurately represents the concept
of interest
o Convergent validity – degree to which two measures of the same concept are correlated
o Discriminant validity – degree to which two conceptually similar concepts are distinct
o Nomological validity – degree that the summated scale makes accurate predictions of other
concepts in a theoretically-based model

59
Stage 7: Data Reduction—Additional Uses of Exploratory
Factor Analysis Results

Creating Summated Scales

Calculation
• The items with high loadings are summed or averaged, with averaging being the most common
approach
• Negative loadings are reverse scored
Example: V1 - positive loading and V2 - negative loading
• First case: V1 – 10, V2 – 0
• Second case: V1 – 0, V2 – 10
• Scale score (non-reversed): V1 – 10, V2 – 10
• Scale score (reversed): V1 – 20, V2 – 0

60
Stage 7: Data Reduction—Additional Uses of Exploratory
Factor Analysis Results

Computing Factor Score

• Factor score: the degree to which each individual scores high on the group of items with high
loadings on a factor
• Computed-based on the factor loadings of all variables on the factor, whereas the summated scale
is calculated by combining only selected variables
• Disadvantage: not easily replicated across studies
• Scoring procedure saves the scoring coefficients from the factor matrix and then allows them to
be applied to new datasets.
• SAS: PROC SCORE

61
Stage 7: Data Reduction—Additional Uses of Exploratory
Factor Analysis Results

Method Selection

• If data are used only in the original sample, interpretation is less important or orthogonality must be
maintained, factor scores are suitable
• If generalizability or transferability is desired, then summated scales or surrogate variables are more
appropriate.
• If the summated scale is untested and exploratory, with little or no evidence of reliability or validity,
surrogate variables should be considered if additional analysis is not possible to improve the summated
scale

62
Stage 7: Data Reduction—Additional Uses of Exploratory
Factor Analysis Results

63
 Practice with Stata

Factor Analysis using Stata’s factor command

• Data were collected on 1428 college students

(complete data on 1365 observations) and are
responses to items on a survey.
use https://stats.idre.ucla.edu/stat/stata/output/m255, clear

ssc install factortest

factortest item13-item24

factor item13-item24, ipf factor(3)

64
Practice with Stata

Factor Analysis using Stata’s factor command

• Data were collected on 1428 college students

(complete data on 1365 observations) and are
responses to items on a survey.
use https://stats.idre.ucla.edu/stat/stata/output/m255, clear

factor item13-item24, ipf factor(3)

65
Practice with Stata

Factor Analysis using Stata’s factor command

• Data were collected on 1428 college students

(complete data on 1365 observations) and are
responses to items on a survey.
use https://stats.idre.ucla.edu/stat/stata/output/m255, clear

factor item13-item24, ipf factor(3)

rotate, varimax horst

66
Practice with Stata

Factor Analysis using Stata’s factor command

• Data were collected on 1428 college students

(complete data on 1365 observations) and are
responses to items on a survey.
use https://stats.idre.ucla.edu/stat/stata/output/m255, clear

factor item13-item24, ipf factor(3)

rotate, varimax horst

rotate, promax horst blanks(3)

estat common

67
Practice with Stata

Factor Analysis using Stata’s factor command

• Data were collected on 1428 college students

(complete data on 1365 observations) and are
responses to items on a survey.
use https://stats.idre.ucla.edu/stat/stata/output/m255, clear

factor item13-item24, ipf factor(3)

rotate, varimax horst

rotate, promax horst blanks(3)

estat common

screeplot

68
Practice with Stata

Factor Analysis using Stata’s factor command

• Data were collected on 1428 college students

(complete data on 1365 observations) and are
responses to items on a survey.
use https://stats.idre.ucla.edu/stat/stata/output/m255, clear

factor item13-item24, ipf factor(3)

rotate, varimax horst

rotate, promax horst blanks(3)

estat common

screeplot

alpha item13-item17, std item detail

69
Practice with JASP

DATASET DESCRIPTION

• HBAT sells paper products to two market segments: the newsprint industry and the magazine industry.
Also, paper products are sold to these market segments either directly to the customer or indirectly
through a broker. Two types of information were collected in the surveys. The first type of information was
perceptions of HBAT’s performance on 13 attributes. These attributes, developed through focus groups, a
pretest, and use in previous studies, are considered to be the most influential in the selection of suppliers
in the paper industry.
• Respondents included purchasing managers of firms buying from HBAT, and they rated HBAT on each of
the 13 attributes using a 0–10 scale, with 10 being “Excellent” and 0 being “Poor.” The second type of
information relates to purchase outcomes and business relationships (e.g., satisfaction with HBAT and
whether the firm would consider a strategic alliance/partnership with HBAT). A third type of information is
available from HBAT’s data warehouse and includes information such as size of customer and length of
purchase relationship.
• By analyzing the data, HBAT can develop a better understanding of both the characteristics of its
customers and the relationships between their perceptions of HBAT, and their actions toward HBAT (e.g.,
satisfaction and likelihood to recommend). From this understanding of its customers, HBAT will be in a
70
good position to develop its marketing plan for next year.
Practice with JASP

Variable Description
X1 Customer Type Length of time a particular customer has been buying from HBAT: 1 = less than 1 year, 2 = between 1 and 5 years, 3 = longer than 5 years
X2 Industry Type Type of industry that purchases HBAT’s paper products: 0 = magazine industry, 1 = newsprint industry
X3 Firm Size Employee size: 0 = small firm (fewer than 500 employees), 1 = large firm (500 or more employees)
X4 Region Customer location: 0 = USA/North America, 1 = outside North America
X5 Distribution System How paper products are sold to customers: 0 = sold indirectly through a broker, 1 = sold directly
X6 Product Quality Perceived level of quality of HBAT's paper products
X7 E-Commerce Activities/Website Overall image of HBAT’s website, especially user-friendliness
X8 Technical Support Extent to which technical support is offered to help solve product/service issues
X9 Complaint Resolution Extent to which complaints are resolved in a timely and complete manner
X10 Advertising Perceptions of HBAT's advertising campaigns in all types of media
X11 Product Line Depth and breadth of HBAT's product line to meet customer needs
X12 Salesforce Image Overall image of HBAT's salesforce
X13 Competitive Pricing Extent to which HBAT offers competitive prices
X14 Warranty and Claims Extent to which HBAT stands behind its product/service warranties and claims
X15 New Products Extent to which HBAT develops and sells new products
X16 Ordering and Billing Perception that ordering and billing are handled efficiently and correctly
X17 Price Flexibility Perceived willingness of HBAT sales reps to negotiate prices on paper product purchases
X18 Delivery Speed Amount of time it takes to deliver the paper products once an order has been confirmed
X19 Customer Satisfaction Customer satisfaction with past purchases from HBAT, measured on a 10-point graphic rating scale
X20 Likelihood of Recommending HBAT Likelihood of recommending HBAT as a supplier of paper products, measured on a 10-point graphic rating scale
X21 Likelihood of Future Purchases from HBAT Likelihood of purchasing paper products from HBAT in the future, measured on a 10-point graphic rating scale
X22 Percentage of Purchases from HBAT Percentage of the responding firm’s paper needs purchased from HBAT, measured on a 100-point percentage scale
X23 Perception of Future Relationship with HBAT Extent to which the customer/respondent perceives a future strategic alliance/partnership with HBAT (0 = Would not consider, 1 = Would consider)
71
 Vietnam
https://future.ueh.edu.

Thank You vn/