Table I. Mean Evaluation of Most People, Reagan, and Mondale
 on Basic Dimensions of Human Nature

                  Evaluation of:              Difference:
           :Most People Reagan Mondale  [Reagan-     [Mondale-
Dimension  :                            Most People] Most People]
===========:=====================================================
Trustworthy:   5.01      4.75    4.58      -.258        -.433
 (S.D.)    :  (1.05)    (1.84)  (1.66)     (2.09)       (1.94)
           :
(Un)selfish:   3.76      3.92    4.17       .170         .420
 (S.D.)    :  (1.58)    (1.71)  (1.52)     (1.96)       (1.80)
           :
In Control :   4.50      4.63    4.00       .121        -.502
 (S.D.)    :  (1.63)    (1.67)  (1.62)     (1.85)       (1.91)

* All differences between Reagan and Most People, Mondale and Most People, and Reagan and Mondale are statistically significant at the .05 level or better, two-tailed.

Replicate the results reported in the table by analyzing the 1984 Gallup Survey Data with SPSS.  Present a table of your own that is similar to the table above.  In addition, add one column of statistics about the differences between Reagan and Mondale (use Reagan-Mondale) on the three dimensions. 
You will need to use the following variables to replicate the table:
  Trust Unselfish  Control
People v158  v161   v164
Reagan v159  v162   v165
Mondale v160  v163   v166

Make sure you appropriately define missing values and reverse scales, if necessary.  For example, for the "Unselfishness" dimension, do this:

MISSING VALUES V161 V162 V163 (8,9).
COMPUTE URP=V162-V161.
COMPUTE UMP=V163-V161.
COMPUTE URM=V162-V163.
DESCRIPTIVES V161 V162 V163 URP UMP URM.

which will give you, among other things, the number of valid cases, the mean, and the standard deviation of the six variables.  In the syntax above, "URP" is simply a variable name for the difference in respondent evaluations of unselfishness (U) between Reagan (R) and Most People (P); UMP and URM are similar, with M standing for Mondale.

Remember that the syntax above is only for the unselfishness dimension; you will need to use similar syntax for "trustworthy" and "in control."
 

B.  Based on the 1984 Gallup Survey, test whether there is a statistically significant difference between men and women with regard to the Reagan-Mondale differentials on the dimensions of Trustworthiness, Unselfishness, and in-Control:

(1)  Compute the Reagan-Mondale differentials on the three dimensions: 

COMPUTE TRM=V159-V160.
COMPUTE URM=V162-V163.
COMPUTE CRM=V165-V166.

(2) Compute the variable SEX using '1' for men and '2' for women:

COMPUTE SEX=2
IF ((V039 = 1) OR (V039 = 3) OR (V039 = 5)) SEX=1

(3) Compute the descriptive statistics for the two samples:

MEANS TABLES=TRM URM CRM BY SEX

(4) Construct the 95% confidence interval for the difference of the two means on each dimension according to W&W (8-20). Note that the "pooled variance" sp2, given in W&W (8-21), can be more conveniently calculated as
 

        (n1-1)s12 + (n2-1)s22
sp2 = -------------------------
             n1 + n2 - 2
 

where s12 and s22 are the variances of the two samples. Although (8-20) assumes that the two populations have the same underlying variance, in general, s1 is not equal to s2.

(5) Based on the confidence intervals, what conclusions can you draw about the differences between men and women regarding the three dimensions?
 

C. Read Kawato, "Nationalization and Partisan Realignment in Congressional Elections." Pay some attention to the Appendix: Stokes' Variance-Components Model. This article is technically difficult, but try to answer the following questions:

(1) What is the problem the author tries to attack?

(2) On what statistical method is the "model for the assessment of electoral forces" based?

(3) What is the dependent variable in the model? What are the independent variables? What kind of values do these variables take? In other words, what kind of scales are these variables?

(4) Why is the statistical method justified in this research?