STA780 Wk6 Factor Analysis-SPSS

STA780
Applied Multivariate Statistical Analysis
Factor Analysis in SPSS
1
Obtaining a Factor Analysis in SPSS
• Open data “personality.sav”
2
Initial investigation using Scatter plot
• Graphs > Legacy Dialogs > Scatter/Dot
• Matrix Scatter > Define
• Matrix Variable : ANX, AGORA, ARACH, ADVEN, EXTRA, SOCIA
• Click OK
3
Obtaining a Factor Analysis in SPSS (Ouput)
From this graph, we can see that items
(AGORA, ARACH and ANX) and items
(EXTRA, ADVEN AND SOCIA) have some
positive correlation with each two groups.
Due to relatively high correlations among
items, this would be a good candidate for
factor analysis. Recall that the goal of
factor analysis is to model the
interrelationships between items with
fewer (latent) variables. These
interrelationships can be broken up into
multiple components
4
• Analyze > Dimension Reduction > Factor
• Select six variables into ‘Variables’
• Click Descriptives
• Statistics : Univariate descriptives and initial solution
• Correlation matrix : coefficient, KMO and Bartlett’s test of sphericity
and reproduced
• Click Extraction
• Method : Principal components
• Analyze : correlation matrix
• Display : unrotated factor solution and scree plot
• Extract : based on eigenvalues greater than 1
5
• Click Rotation
• Method : Varimax
• Display : Rotated solution and Loading Plot(s)
• Click Factor Scores
• Save as Variables > Method : regression
• Click Option
• Coefficient display format : sorted by size and suppress small
coefficients
6
Obtaining a Factor Analysis in SPSS (Output)
Most correlations are well above 0.3 (a good indication that we

will obtain a result) and there appears to be two groups of
variables (as can be seen in scatter plots)
• Anxiety, agoraphobia and arachnophobia
• Adventure, extraversion and sociability
7
H0 : correlation matrix is an identity matrix (no intercorrelation between the
items/variables)
H1 : Otherwise
The KMO value indicates that it is pretty poor – just above miserable, however,
Bartlett’s test of sphericity with p-value of < 0.001 indicates that there is an
intercorrelations among the variables. Therefore, we can proceed with the Factor
Analysis
8
Estimates of that part of the variability in each
variable that is shared with others and which is
not due to measurement error or latent variable
influence on the observed variable). The initial
value can be ignored.
9
The first two factors have eigenvalues more than 1

and these two factors explain more than 96% of
the total variability in the data. This lead to the
conclusion that the two factor solution will be
adequate.
10
The unrotated factor shows high

positive and high negative loadings on
the first factor.
11
Which tests have high loadings on

each of the rotated components?
12
• These two rotated factors are just as good as the initial factors in explaining and reproducing the
observed correlation matrix.
• In the rotated factors, Adventure, extraversion and sociability all have high positive loadings on
the first factor (and low loading on the second), whereas, anxiety, agoraphobia and
arachnophobia all high positive loadings on the second factor (and low loadings on the first)
• The cumulative % are the same however different division in % of variance. 13
In this case, the factors have been rotated through 45
which can be calculated by treating the correlation
coefficient as a cosine. Cosine(45) = 0.707
14
Naming the two rotated factors :
Factor 1 – Outgoingness
(extraversion, adventure and
sociability have high loadings)
Factor 2 – Neuroticism (anxiety
and phobias all high loadings)
15
Use can use Factor Scores (FAC1_1 and FAC2_1) to carry out other analysis
such as multiple linear regression etc.
16
Confirmatory Factor Analysis using SEM
(AMOS)
17
(AMOS)
• From the plot, we can remove ADVEN, EXTRA and SOCIA
since the factor loadings are < 0.5
• Re-build the model, one factor only include ANX, AGORA and
ARACH and the other factor with ADVEN, EXTRA AND SOCIA
18
(AMOS)
19

STA780 Wk6 Factor Analysis-SPSS

Uploaded by

Copyright:

Available Formats

STA780 Wk6 Factor Analysis-SPSS

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

STA780 Wk6 Factor Analysis-SPSS

Uploaded by

Copyright:

Available Formats

STA780

Applied Multivariate Statistical Analysis

Factor Analysis in SPSS

• Open data “personality.sav”

Most correlations are well above 0.3 (a good indication that we

The first two factors have eigenvalues more than 1

The unrotated factor shows high

Which tests have high loadings on

You might also like