Factor Analysis Exercise: Rating Characteristics of Criminal

advertisement
Factor Analysis Exercise:
Rating Characteristics of Defendants Accused of Burglary or Swindle
For this exercise we are going to use the data which led to this publication:

Wuensch, K. L., Castellow, W. A., & Moore, C. H. (1991). Effects of defendant
attractiveness and type of crime on juridic judgment. Journal of Social
Behavior and Personality, 6, 713-724.
The research participants were mock jurors hearing a simulated case in which a
male or female defendant was accused of either a burglary or a swindle. We
experimentally manipulated the physical attractiveness of the defendant. The evidence
made it quite clear that the defendant was guilty of the crime. The jurors were asked to
recommend a sentence (from 1 to 15 years) and were asked the following questions
about characteristics of the defendant:
Think back about the defendant, Miss/Mr. Barbara/Robert Helm, and rate her/him
on the following characteristics by circling the number that best reflects your impression
of her/him for each item.
A. dull
1
exciting
2
3
4
5
6
7
8
B. nervous
1
calm
2
3
4
5
6
7
8
C. dependent
1
2
3
4
5
6
7
8
2
3
4
5
6
7
8
2
3
4
5
6
7
8
2
3
4
5
6
7
8
2
3
4
5
6
7
8
2
3
4
5
6
7
8
9
strong
2
3
4
5
6
7
8
J. naive
1
9
intelligent
I. weak
1
9
kind
H. unintelligent
1
9
physically unattractive
G. cruel
1
9
warm
F. physically attractive
1
9
sincere
E. cold
1
9
independent
D. insincere
1
9
9
sophisticated
2
3
4
5
6
7
8
9
FA-Jury91.doc
K. sad
1
happy
2
3
4
5
6
7
8
K. unsociable
1
2
9
sociable
3
4
5
6
7
8
9
We shall use these ratings as our variables for this exercise. Download the file
JURY91.SAV and bring it into SPSS. Conduct a principle axis factor analysis on the
variables exciting, calm, independent, sincere, warm, phy_attr, kind, intellig, strong,
sophist, happy, and sociable.
1. Conduct a principal axis factor analysis with varimax rotation. Obtain a scree
plot, Kaiser’s MSA, and the anti-image correlation matrix. Retain all factors with an
initial eigenvalue of 1 or more (3 factors). Sort the coefficients by size. Look at the
rotated loadings. Note that several of the variables load .3 or more on two of the
factors. I feel better which each variable loads well on only one factor.
1. Conduct a parallel analysis. It will indicate a two factor solution.
2. Conduct a MAP test. The original test will indicate a two factor solution, the
revised test a one factor solution.
2. Run the factor analysis again, but reduce the number of factors to 2 and
suppress the output of the initial solution, the MSA, the anti-image matrices, t, and the
scree plot. Output the factor scores. This solution captures a bit less of the common
variance, but it looks cleaner to me. Look at the rotated loadings and decide what you
would call the two factors.
3. Make one more change. Apply an (oblique) oblimin rotation.
4. I don’t find this solution much different from the varimax solution, so I’ll go
back to the varimax solution. Use an independent samples t test to make the following
comparisons:
a. pa_manip group 1 versus 2 – this tests the effect of our manipulation of
the defendant’s physical attractiveness.
b. crime group 1 versus 2 – this tests the effect of the type of crime.
Karl L. Wuensch, September, 2012
Factor Analysis Exercise: Animal Rights, Ethical Ideology, and Misanthropy
Return to Factor Analysis with SPSS
Return to Multivariate Analysis with SPSS
Download