NCSS Statistical Software NCSS.

com

Chapter 420

Factor Analysis
Introduction
Factor Analysis (FA) is an exploratory technique applied to a set of observed variables that seeks to find
underlying factors (subsets of variables) from which the observed variables were generated. For example, an
individual’s response to the questions on a college entrance test is influenced by underlying variables such as
intelligence, years in school, age, emotional state on the day of the test, amount of practice taking tests, and so on.
The answers to the questions are the observed variables. The underlying, influential variables are the factors.
Factor analysis is carried out on the correlation matrix of the observed variables. A factor is a weighted average of
the original variables. The factor analyst hopes to find a few factors from which the original correlation matrix
may be generated.
Usually the goal of factor analysis is to aid data interpretation. The factor analyst hopes to identify each factor as
representing a specific theoretical factor. Therefore, many of the reports from factor analysis are designed to aid
in the interpretation of the factors.
Another goal of factor analysis is to reduce the number of variables. The analyst hopes to reduce the interpretation
of a 200-question test to the study of 4 or 5 factors. One of the most subtle tasks in factor analysis is determining
the appropriate number of factors.
Factor analysis has an infinite number of solutions. If a solution contains two factors, these may be rotated to form
a new solution that does just as good a job at reproducing the correlation matrix. Hence, one of the biggest
complaints of factor analysis is that the solution is not unique. Two researchers can find two different sets of
factors that are interpreted quite differently yet fit the original data equally well.
NCSS provides the principal axis method of factor analysis. The results may be rotated using varimax or
quartimax rotation. The factor scores may be stored for further analysis.
Many books are devoted to factor analysis. We suggest you obtain a book on the subject from an author in your
own field. An excellent introduction to the subject is provided by Tabachnick (1989).

420-1
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
Factor Analysis

Technical Details

Mathematical Development
This section will document the basic formulas used by NCSS in performing a factor analysis. The following table
lists many of the matrices that are used in the discussion to follow.

Label Matrix Name Size Description

R Correlation pxp Matrix of correlations between each pair of variables.
X Data Nxp Observed data matrix with N rows (observations) and p
columns (variables).
Z Standardized data Nxp Matrix of standardized data. The standardization of each
variable is made by subtracting its mean and dividing by
its standard deviation.
A Factor loading pxm Matrix of correlations between the original variables and
the factors. Also represents the contribution of each
factor in estimating the original variables.
L Eigenvalue mxm Diagonal matrix of eigenvalues. Only the first m
eigenvalues are considered.
V Eigenvector pxm Matrix of eigenvectors. Only the first m columns of this
matrix are used.
B Factor-score coefficients pxm Matrix of regression weights used to construct the factor
scores from the original variables.
U Uniqueness pxp Matrix of uniqueness values.
F Factor score Nxm Matrix of factor scores. For each observation in the
original data, the values of each of the retained factors
are estimated. These are the factor scores.
The principal-axis method is used by NCSS to solve the factor analysis problem. Factor analysis assumes the
following partition of the correlation matrix, R:
R = AA ′ +U
The principal-axis method proceeds according to the following steps:
1. Estimate U from the communalities as discussed below.
2. Find L and V, the eigenvalues and eigenvectors of R-U using standard eigenvalue analysis.
3. Calculate the loading matrix as follows:
1
A =VL 2
4. Calculate the score matrix as follows:
− 12
B =VL
5. Calculate the factor scores as follows:
F = ZB
Steps 1-3 may be iterated since a new U matrix may be estimated from the current loading matrix.

Initial Communality Estimation

We close this section with a discussion of obtaining an initial value of U. NCSS uses the initial estimation of
Cureton (1983), which will be outlined here. The initial communality estimates, cii, are calculated from the
correlation and inverse correlation matrices as follows:

420-2
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
Factor Analysis
p

 1 
∑ max r jk
over j ≠ k
( )
c ii = 1 - ii 
k =1
 R  p
 1 
∑ 1 - R
k =1
kk 

where Rii is the ith diagonal element of R-1 and rjk is an element of R. The value of U is then estimated by 1-cii.

Missing Values and Robust Estimation

Missing values and robust estimation are done the same way as in principal components analysis. Refer to that
chapter for details.

How Many Factors

Several methods have been proposed for determining the number of factors that should be kept for further
analysis. Several of these methods will now be discussed. However, remember that important information about
possible outliers and linear dependencies may be determined from the factors associated with the relatively small
eigenvalues, so these should be investigated as well.
Kaiser (1960) proposed dropping factors whose eigenvalues are less than one since these provide less information
than is provided by a single variable. Jolliffe (1972) feels that Kaiser’s criterion is too large. He suggests using a
cutoff on the eigenvalues of 0.7 when correlation matrices are analyzed. Other authors note that if the largest
eigenvalue is close to one, then holding to a cutoff of one may cause useful factors to be dropped. However, if the
largest factors are several times larger than one, then those near one may be reasonably dropped.
Cattell (1966) documented the scree graph, which will be described later in this chapter. Studying this chart is
probably the most popular method for determining the number of factors, but it is subjective, causing different
people to analyze the same data with different results.
Another criterion is to preset a certain percentage of the variation that must be accounted for and then keep
enough factors so that this variation is achieved. Usually, however, this cutoff percentage is used as a lower limit.
That is, if the designated number of factors do not account for at least 50% of the variance, then the whole
analysis is aborted.

Varimax and Quartimax Rotation

Factor analysis finds a set of dimensions (or coordinates) in a subspace of the space defined by the set of
variables. These coordinates are represented as axes. They are orthogonal (perpendicular) to one another. For
example, suppose you analyze three variables that are represented in three-dimensional space. Each variable
becomes one axis. Now suppose that the data lie near a two-dimensional plane within the three dimensions. A
factor analysis of this data should uncover two factors that would account for the two dimensions. You may rotate
the axes of this two-dimensional plane while keeping the 90-degree angle between them, just as the blades of a
helicopter propeller rotate yet maintain the same angles among themselves. The hope is that rotating the axes will
improve your ability to interpret the “meaning” of each factor.
Many different types of rotation have been suggested. Most of them were developed for use in factor analysis.
NCSS provides two orthogonal rotation options: varimax and quartimax.

Varimax Rotation
Varimax rotation is the most popular orthogonal rotation technique. In this technique, the axes are rotated to
maximize the sum of the variances of the squared loadings within each column of the loadings matrix.
Maximizing according to this criterion forces the loadings to be either large or small. The hope is that by rotating
the factors, you will obtain new factors that are each highly correlated with only a few of the original variables.

420-3
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
Factor Analysis

This simplifies the interpretation of the factor to a consideration of these two or three variables. Another way of
stating the goal of varimax rotation is that it clusters the variables into groups; each “group” is actually a new
factor.
Since varimax seeks to maximize a specific criterion, it produces a unique solution (except for differences in
sign). This has added to its popularity. Let the matrix G = {gij} represent the rotated factors. The goal of varimax
rotation is to maximize the quantity:

 p ∑p g 4 - ∑p g 2 
k  ij ij 
Q1 = ∑ i =1 i =1
j =1 p 
 
This equation gives the raw varimax rotation. This rotation has the disadvantage of not spreading the variance
very evenly among the new factors. Instead, it tends to form one large factor followed by many small ones. To
correct this, NCSS uses the normalized-varimax rotation. The quantity maximized in this case is:
 p  g 4 2

- ∑  g ij 
p

k ∑
 p  ij 
 i =1  c i 
 
i =1  c i  
QN = ∑ 
j =1 

p2 
 
 
where ci is the square root of the communality of variable i.

Quartimax Rotation
Quartimax rotation is similar to varimax rotation except that the rows of G are maximized rather than the columns
of G. This rotation is more likely to produce a “general” factor than will varimax. Often, the results are quite
similar. The quantity maximized for the quartimax is:

 p  g ij  4 
∑
k  i = 1  c i  
QN = ∑
j =1
 p 
 
 

Miscellaneous Topics

Using Correlation Matrices Directly

Occasionally, you will be provided with only the correlation matrix from a previous analysis. This happens
frequently when you want to analyze data that is presented in a book or a report. You can perform a factor
analysis on a correlation matrix using NCSS.
NCSS can store the correlation matrix on the current database. If it takes a great deal of computer time to build
the correlation matrix, you might want to save it so you can use it while you determine the number of factors. You
could then return to the original data to analyze the factor scores.

420-4
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
Factor Analysis

Principal Component Analysis versus Factor Analysis

Both principal component analysis (PCA) and factor analysis (FA) seek to reduce the dimensionality of a data set.
The most obvious difference is that while PCA is concerned with the total variation as expressed in the correlation
matrix, R, FA is concerned with a correlation in a partition of the total variation called the common portion. That
is, FA separates R into two matrices Rc (common factor portion) and Ru (unique factor portion). FA models the Rc
portion of the correlation matrix. Hence, FA requires the discovery of Rc as well as a model for it. The goals of FA
are more concerned with finding and interpreting the underlying, common factors. The goals of PCA are
concerned with a direct reduction in the dimensionality.
Put another way, PCA is directed towards reducing the diagonal elements of R. Factor analysis is directed more
towards reducing the off-diagonal elements of R. Since reducing the diagonal elements reduces the off-diagonal
elements and vice versa, both methods achieve much the same thing.

Data Structure
The data for a factor analysis consists of two or more columns. We have created an artificial data set in which
each of the six variables (X1 - X6) were created using weighted averages of two original variables (V1 and V2)
plus a small random error. For example, X1 = .33 V1 + .65 V2 + error. Each variable had a different set of
weights (.33 and .65 are the weights) in the weighted average.
Rows two and three of the data set were modified to be outliers so that their influence on the analysis could be
observed. Note that even though these two rows are outliers, their values on each of the individual variables are
not outliers. This shows one of the challenges of multivariate analysis: multivariate outliers are not necessarily
univariate outliers. In other words, a point may be an outlier in a multivariate space and yet you cannot detect it
by scanning the data one variable at a time.
This data is contained in the dataset PCA2. The data given below are the first few rows of this dataset.

PCA2 dataset (subset)

X1 X2 X3 X4 X5 X6
50 102 103 70 75 102
4 2 5 11 11 5
81 98 94 5 85 97
31 81 86 46 50 74
65 50 51 60 57 53
22 30 39 17 15 17
36 33 39 29 27 25
31 91 96 50 56 85

420-5
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
Factor Analysis

Procedure Options
This section describes the options available in this procedure.

Variables Tab
This panel specifies the variables used in the analysis.

Input Variables
Variables
Designates the variables to be analyzed. If matrix input is selected, indicate the variables containing the matrix.
Note that for matrix input, the number of rows used is equal to the number of variables specified. Other rows will
be ignored.
Data Input Format
Indicates whether raw data is to be analyzed or if a previously summarized correlation or covariance matrix is to
be used.

• Regular Data
The data is to be input in its raw format.

• Lower-Triangular
The data is in a correlation or covariance matrix in lower-triangular format. This matrix could have been
created by a previous run of an NCSS program or from direct keyboard input.

• Upper-Triangular
The data is in a correlation or covariance matrix in upper triangular format. The number of rows used is equal
to the number of variables specified. This matrix could have been created by a previous run of an NCSS
program or from direct keyboard input.

Covariance Estimation Options

Robust Covariance Matrix Estimation
This option indicates whether robust estimation is to be used. A full discussion of robust estimation is provided in
the PCA chapter. If checked, robust estimates of the means, variances, and covariances are formed.
Robust Weight
This option specifies the value of ν1 for robust estimation. This parameter controls the weighting function in
robust estimation of the covariance matrix. Jackson (1991) recommends the value 4.
Missing Value Estimation
This option indicates the type of missing value imputation method that you want to use. (Note that if the number
of iterations is zero, this option is ignored.)

• None
No missing value imputation. Rows with missing values in any of the selected variables are ignored.

• Average
The average-value imputation method is used. Each missing value is estimated by the average value of that
variable. The process is iterated as many times as is indicated in the second box.

420-6
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
Factor Analysis

• Multivariate Normal
The multivariate-normal method. Each missing value is estimated using a multiple regression of the missing
variable(s) on the variables that contain data in that row. This process is iterated as many times as indicated.
See the discussion of missing value imputation methods elsewhere in this chapter.
Maximum Iterations
This option specifies the number of iterations used by either Missing Value Imputation or Robust Covariance
Estimation. Robust estimation usually requires only four or five iterations to converge. Missing value imputation
may require as many as twenty iterations if there are a lot of missing values.
When using this option, it is better to specify too many iterations than too few. After considering the Percent
Change values in the Iteration Report, you can decide upon an appropriate number of iterations and re-run the
problem.

Factor Options
Factor Rotation
Specifies the type of rotation, if any, that should be used on the solution. If rotation is desired, either varimax or
quartimax rotation is available.
Number of Factors
This option specifies the number of factors to be used. On the first run, you would set this rather large (say eight
or so). After viewing the eigenvalues you would reset this appropriately and make a second run.

Communality Options
Communality Iterations
This option specifies how many iterations to use in estimating the communalities. Some authors suggest a value
of one here. Others suggest as many as four or five.

Reports Tab
The following options control the format of the reports.

Select Reports
Descriptive Statistics - Scores Report
These options let you specify which reports are displayed.

Report Options
Minimum Loading
Specifies the minimum absolute value that a loading can have and still remain in the Variable List report.
Precision
Specify the precision of numbers in the report. A single-precision number will show seven-place accuracy, while
a double-precision number will show thirteen-place accuracy. Note that the reports are formatted for single
precision. If you select double precision, some numbers may run into others. Also note that all calculations are
performed in double precision regardless of which option you select here. This is for reporting purposes only.

420-7
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
Factor Analysis

Variable Names
This option lets you select whether to display only variable names, variable labels, or both.

Plots Tab
These sections specify the pair-wise plots of the scores and loadings.

Select Plots
Scores Plot - Loadings Plot
These options let you specify which reports and plots are displayed. Click the plot format button to change the
plot settings.

Plot Options
Number Factors Plotted
You can limit the number of plots generated using this parameter. Usually, you will only have interest in the first
three or four factors.

Storage Tab
The factor scores and/or the correlation matrix may be stored on the current dataset for further analysis. This
group of options let you designate which statistics (if any) should be stored and which columns should receive
these statistics. The selected statistics are automatically stored to the current dataset. Note that existing data are
replaced.

Data Storage Columns

Factor Scores
You can automatically store the factor scores for each row into the columns specified here. These scores are
generated for each row of data in which all independent variable values are nonmissing.
Correlation Matrix
You can automatically store the correlation matrix to the columns specified here.

420-8
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
Factor Analysis

Example 1 – Factor Analysis

This section presents an example of how to run a factor analysis. The data used are shown in the table above and
found in the PCA2 dataset.
You may follow along here by making the appropriate entries or load the completed template Example 1 by
clicking on Open Example Template from the File menu of the Factor Analysis window.

1 Open the PCA2 dataset.

• From the File menu of the NCSS Data window, select Open Example Data.
• Click on the file PCA2.NCSS.
• Click Open.

2 Open the Factor Analysis window.

• Using the Analysis menu or the Procedure Navigator, find and select the Factor Analysis procedure.
• On the menus, select File, then New Template. This will fill the procedure with the default template.

3 Specify the variables.

• On the Factor Analysis window, select the Variables tab.
• Double-click in the Variables box. This will bring up the variable selection window.
• Select X1 through X6 from the list of variables and then click Ok. “X1-X6” will appear in the Variables
box.
• Check the Robust Covariance Matrix Estimation box.
• Enter 6 in the Maximum Iterations box.
• Select Varimax in the Factor Rotation list box.
• Enter 2 in the Number of Factors box.
• Enter 6 in the Communality Iterations box.

4 Specify which reports.

• Select the Reports tab.
• Check all reports and plots. Normally you would only view a few of these reports, but we are selecting
them all so that we can document them.

5 Run the procedure.

• From the Run menu, select Run Procedure. Alternatively, just click the green Run button.

Robust and Missing-Value Iteration Section

This report is only produced when robust or missing value estimation is used.

Robust and Missing-Value Estimation Iteration Section

Trace of Percent
No. Count Covar Matrix Change
0 30 4907.795 0.00
1 30 4907.795 0.00
2 30 4423.718 -9.86
3 30 4423.718 0.00
4 30 4353.748 -1.58
5 30 4353.748 0.00
6 30 4335.77 -0.41

This report presents the progress of the robust iterations. The trace of the covariance matrix gives a measure of
what is happening at each iteration. When this value stabilizes, the program has converged. The percent change is
reported to let you determine how much the trace has changed. In this particular example, we see very little

420-9
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
Factor Analysis

change between iterations five and six. We would feel comfortable stopping at this point. A look at the
Descriptive Statistics section will let you see how much the means and standard deviations have changed.
A look at the Residual Section will let you see the robust weights that are assigned to each row. Those weights
that are near zero indicate observations whose influence have been removed by the robust procedure.

Descriptive Statistics Section

Descriptive Statistics Section
Standard
Variables Count Mean Deviation Communality
X1 30 42.83667 23.18579 0.997983
X2 30 53.25062 27.93123 0.999791
X3 30 57.13034 26.3737 0.999585
X4 30 43.5617 24.56474 0.992023
X5 30 43.20835 25.75021 1.00007
X6 30 48.61827 32.49559 0.999944

Count, Mean, and Standard Deviation

These are the familiar summary statistics of each variable. They are displayed to allow you to make sure that you
have specified the correct variables. Note that using missing value imputation or robust estimation will change these
values.
Communality
The communality shows how well this variable is predicted by the retained factors. It is similar to the R-Squared
that would be obtained if this variable were regressed on the factors that were kept. However, remember that this
is not based directly on the correlation matrix. Instead, calculations are based on an adjusted correlation matrix.

Correlation Section
Correlation Section
Variables
Variables X1 X2 X3 X4 X5
X1 1.000000 0.271780 0.127016 0.881604 0.814686
X2 0.271780 1.000000 0.988909 0.683206 0.778649
X3 0.127016 0.988909 1.000000 0.568933 0.677480
X4 0.881604 0.683206 0.568933 1.000000 0.986945
X5 0.814686 0.778649 0.677480 0.986945 1.000000
X6 0.484907 0.973093 0.928454 0.831949 0.901975
Phi=0.769781 Log(Det|R|)=-29.547320 Bartlett Test=773.15 DF=15 Prob=0.000000

Variables
Variables X6
X1 0.484907
X2 0.973093
X3 0.928454
X4 0.831949
X5 0.901975
X6 1.000000
Phi=0.769781 Log(Det|R|)=-29.547320 Bartlett Test=773.15 DF=15 Prob=0.000000

Bar Chart of Absolute Correlation Section

Variables
Variables X1 X2 X3 X4 X5
X1 |||||| ||| |||||||||||||||||| |||||||||||||||||
X2 |||||| |||||||||||||||||||| |||||||||||||| ||||||||||||||||
X3 ||| |||||||||||||||||||| |||||||||||| ||||||||||||||
X4 |||||||||||||||||| |||||||||||||| |||||||||||| ||||||||||||||||||||
X5 ||||||||||||||||| |||||||||||||||| |||||||||||||| ||||||||||||||||||||
X6 |||||||||| |||||||||||||||||||| ||||||||||||||||||| ||||||||||||||||| |||||||||||||||||||
Phi=0.769781 Log(Det|R|)=-29.547320 Bartlett Test=773.15 DF=15 Prob=0.000000

420-10
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
Factor Analysis

Variables
Variables X6
X1 ||||||||||
X2 ||||||||||||||||||||
X3 |||||||||||||||||||
X4 |||||||||||||||||
X5 |||||||||||||||||||
X6
Phi=0.769781 Log(Det|R|)=-29.547320 Bartlett Test=773.15 DF=15 Prob=0.000000

This report gives the correlations alone for a test of the overall correlation structure in the data. In this example,
we notice several high correlation values. The Gleason-Staelin redundancy measure, phi, is 0.736, which is quite
large. There is apparently some correlation structure in this data set that can be modeled. If all the correlations are
small (say less then .3), there would be no need for a factor analysis.
Correlations
The simple correlations between each pair of variables. Note that using the missing value imputation or robust
estimation options will affect the correlations in this report. When the above options are not used, the correlations
are constructed from those observations having no missing values in any of the specified variables.
Phi
This is the Gleason-Staelin redundancy measure of how interrelated the variables are. A zero value of ϕ2 means
that there is no correlation among the variables, while a value of one indicates perfect correlation among the
variables. This coefficient may have a value less than 0.5 even when there is obvious structure in the data, so care
should to be taken when using it. This statistic is especially useful for comparing two or more sets of data. The
formula for computing ϕ3 is:
p p
∑ ∑ r ij 2 -p
i =1 j =1
ϕ=
p(p - 1)

Log(Det|R|)
This is the log (base e) of the determinant of the correlation matrix. If you used the covariance matrix, this is the
log (base e) of the determinant of the covariance matrix.
Bartlett Test, df, Prob
This is Bartlett’s sphericity test (Bartlett, 1950) for testing the null hypothesis that the correlation matrix is an
identity matrix (all correlations are zero). If you get a probability (Prob) value greater than 0.05, you should not
perform a factor analysis on the data. The test is valid for large samples (N>150). It uses a Chi-square distribution
with p(p-1)/2 degrees of freedom. Note that this test is only available when you analyze a correlation matrix. The
formula for computing this test is:

χ2 =
(11 + 2p - 6N ) Log R
e
6
Bar Chart of Absolute Correlation Section
This chart graphically displays the absolute values of the correlations. It lets you quickly find high and low
correlations.

420-11
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
Factor Analysis

Eigenvalues Section
Eigenvalues after Varimax Rotation
Individual Cumulative
No. Eigenvalue Percent Percent Scree Plot
1 3.288191 54.89 54.89 |||||||||||
2 2.701207 45.09 99.99 ||||||||||
3 0.001207 0.02 100.01 |
4 -0.000099 0.00 100.01 |
5 -0.000121 0.00 100.00 |
6 -0.000295 0.00 100.00 |

Eigenvalues
The eigenvalues of the R-U matrix. Often, these are used to determine how many factors to retain. (In this
example, we would retain the first two eigenvalues.)
One rule-of-thumb is to retain those factors whose eigenvalues are greater than one. The sum of the eigenvalues is
equal to the number of variables. Hence, in this example, the first factor retains the information contained in 3.3
of the original variables.
Note that, unlike in PCA where all eigenvalues are positive, the eigenvalues may be negative in factor analysis.
Usually, these factors would be discarded and the analysis would be re-run.
Individual and Cumulative Percents
The first column gives the percentage of the total variation in the variables accounted for by this factor. The
second column is the cumulative total of the percentage. Some authors suggest that the user pick a cumulative
percentage, such as 80% or 90%, and keep enough factors to attain this percentage.
Scree Plot
This is a rough bar plot of the eigenvalues. It enables you to quickly note the relative size of each eigenvalue.
Many authors recommend it as a method of determining how many factors to retain.
The word scree, first used by Cattell (1966), is usually defined as the rubble at the bottom of a cliff. When using
the scree plot, you must determine which eigenvalues form the “cliff” and which form the “rubble.” You keep the
factors that make up the cliff. Cattell and Jaspers (1967) suggest keeping those that make up the cliff plus the first
factor of the rubble.
Interpretation of the Example
The first question that we would ask is how many factors should be kept. The scree plot shows that the first two
factors are indeed the largest. The cumulative percentages show that the first two factors account for over 99.99%
of the variation.

Eigenvectors Section
Eigenvectors after Varimax Rotation
Factors
Variables Factor1 Factor2
X1 -0.303444 -0.662220
X2 -0.416551 0.378018
X3 -0.382768 0.491154
X4 -0.428167 -0.317929
X5 -0.448606 -0.204840
X6 -0.450912 0.185189

420-12
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
Factor Analysis

Bar Chart of Absolute Eigenvectors after Varimax Rotation

Factors
Variables Factor1 Factor2
X1 ||||||| ||||||||||||||
X2 ||||||||| ||||||||
X3 |||||||| ||||||||||
X4 ||||||||| |||||||
X5 ||||||||| |||||
X6 |||||||||| ||||

Eigenvector
The eigenvectors of the R-U matrix.
Bar Chart of Absolute Eigenvectors
This chart graphically displays the absolute values of the eigenvectors. It lets you quickly interpret the eigenvector
structure. By looking at which variables correlate highly with a factor, you can determine what underlying
structure it might represent.

Factor Loadings Section

Factor Loadings after Varimax Rotation
Factors
Variables Factor1 Factor2
X1 -0.019936 -0.998792
X2 -0.967470 -0.252572
X3 -0.994037 -0.107126
X4 -0.478418 -0.873578
X5 -0.594943 -0.803812
X6 -0.883654 -0.468080

Bar Chart of Absolute Factor Loadings after Varimax Rotation

Factors
Variables Factor1 Factor2
X1 | ||||||||||||||||||||
X2 |||||||||||||||||||| ||||||
X3 |||||||||||||||||||| |||
X4 |||||||||| ||||||||||||||||||
X5 |||||||||||| |||||||||||||||||
X6 |||||||||||||||||| ||||||||||

Factor Loadings
These are the correlations between the variables and factors.
Bar Chart of Absolute Factor Loadings
This chart graphically displays the absolute values of the factor loadings. It lets you quickly interpret the
correlation structure. By looking at which variables correlate highly with a factor, you can determine what
underlying structure it might represent.

420-13
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
Factor Analysis

Communality Section
Communality after Varimax Rotation
Factors
Variables Factor1 Factor2 Communality
X1 0.000397 0.997586 0.997983
X2 0.935999 0.063793 0.999791
X3 0.988109 0.011476 0.999585
X4 0.228883 0.763139 0.992023
X5 0.353958 0.646114 1.000072
X6 0.780845 0.219099 0.999944

Bar Chart of Communalities after Varimax Rotation

Factors
Variables Factor1 Factor2 Communality
X1 | |||||||||||||||||||| ||||||||||||||||||||
X2 |||||||||||||||||| || ||||||||||||||||||||
X3 |||||||||||||||||||| | ||||||||||||||||||||
X4 ||||| |||||||||||||||| ||||||||||||||||||||
X5 |||||||| ||||||||||||| ||||||||||||||||||||
X6 |||||||||||||||| ||||| ||||||||||||||||||||

Communality
The communality is the proportion of the variation of a variable that is accounted for by the factors that are
retained. It is similar to the R-Squared value that would be achieved if this variable were regressed on the retained
factors. This table value gives the amount added to the communality by each factor.
Bar Chart of Communalities
This chart graphically displays the values of the communalities.

Factor Structure Summary Section

Factor Structure Summary after Varimax Rotation
Factors
Factor1 Factor2
X3 X1
X2 X4
X6 X5
X5 X6
X4

Interpretation
This report is provided to summarize the factor structure. Variables with an absolute loading greater than the
amount set in the Minimum Loading option are listed under each factor. Using this report, you can quickly see
which variables are related to each factor. Note that it is possible for a variable to have high loadings on several
factors, although varimax rotation makes this very unlikely.

Score Coefficients Section

Score Coefficients after Varimax Rotation
Factors
Variables Factor1 Factor2
X1 0.268188 -0.5366089
X2 -0.3613275 0.1312937
X3 -0.4135219 0.2176106
X4 2.901571E-02 -0.3414549
X5 -0.0422971 -0.2712604
X6 -0.2641693 -8.934696E-03

420-14
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
Factor Analysis

Score Coefficients
These are the coefficients that are used to form the factor scores. The factor scores are the values of the factors for
a particular row of data. These score coefficients are similar to the eigenvectors. They have been scaled so that the
scores produced have a variance of one rather than a variance equal to the eigenvalue. This causes each of the
factors to have the same variance.
You would use these scores if you wanted to calculate the factor scores for new rows not included in your original
analysis.

Factor Scores Section

Factor Scores after Varimax Rotation
Factors
Row Factor1 Factor2
1 -1.7219 -0.2752
2 1.4002 1.0317
3 -1.2231 -0.2862
4 -1.1632 0.5302
(report continues through all thirty rows)

Factor1 - Factor2
The factor scores are the values of the factors for a particular row of data. They have been scaled so they have a
variance of one.

Factor Score Plots

This set of plots shows each factor plotted against every other factor.

420-15
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
Factor Analysis

Factor Loading Plots

This set of plots shows each of the factor loading columns plotted against each other.

420-16
© NCSS, LLC. All Rights Reserved.