Exercise 1:

On the basis of these results, are people who took the more conservative view on helping the poor more likely to classify themselves as conservatives, and vice versa? Yes or No. ______________

Obtain the significance level, and Kendall’s tau-b, and write them in the spaces provided.

Prob. = ___________

Kendall’s tau-b = ________________

Is this relationship statistically significant? Yes or No. _______________

Do you reject the null hypothesis? Yes or No. __________________

Exercise 2:

Relationship Between Conservative Self-Classification and Conservatism on Economic Issues by Education Category

EDUCATION CATEGORY              KENDALL’S TAU-B            PROB

NO HS GRAD                                                0.092                    0.072

HS GRAD                                          _______________        ___________

COLL EDUC                                       _______________        ___________

Examine these results and the original results you obtained before controlling for education. Does the control variable, education, affect the relationship between conservative self-identification and conservatism on economic issues? Yes or No. __________________________

If you answered yes above, describe how the relationship between conservative self-identification and conservatism on economic issues varies by education level. If you answered no, explain why you think education had no impact on the relationship between conservative self-identification and conservatism on economic issues.

__________________________________________________________________________________________________________________________________________________________________________

Exercise 3:

Write the column percentages in the following table.

Fundamentalist                Moderate             Liberal

Kerry                    ____%                                ______%               ____%

Bush                    ____%                                ______%               ____%

Which religious tradition shows the lowest support for the Democratic candidate, Perry? Circle one.

Fundamentalists              Moderates              Liberals

Which religious tradition shows the highest support for the Republican candidate, Bush? Circle on.

Fundamentalists              Moderates              Liberals

Write in the significance level for this relationship.           Prob. = __________________

Write in Cramer’s V for this relationship.                             Cramer’s V = ____________________

Is this relationship statistically significant? Yes or No. ________________

Do you reject the null hypothesis that there is no relationship between 2004 presidential voting choice and the liberalism of the respondent’s religious tradition? Yes or No. ___________________

Exercise 4:

The first table shows the relationship between 541) WHO IN 04? and 116) R.FUND/LIB for white Americans. Write column percentages in the following table, which is labeled Control Category: White. Also, check the summary statistics and write in the significance level and Cramer’s V.

Control Category: White

Fundamentalist                Moderate             Liberal

Kerry                    ____%                                ______%               ____%

Bush                    ____%                                ______%               ____%

Prob. =_____________

Cramer’s V = _______

Among white Americans, is the relationship statistically significant? Yes or No. _____________

Now move to the next table and write the column percentages in the table that is labeled Control Category: African American. Again, write in the significance level and Cramer’s V.

Control Category: African Americans

Fundamentalist                Moderate             Liberal

Kerry                    ____%                                ______%               ____%

Bush                    ____%                                ______%               ____%

Prob. =_____________

Cramer’s V = _______

Among African Americans, is the relationship statistically significant? Yes or No. _____________

Exercise 5:

To compare the percentages in the tables, it is often helpful to compare a particular row from each table. Here, let’s compare the Kerry row from all tables, including the original table before we controlled for anything. In the following table, write in the percentages from the Kerry row from each of the preceding tables. To provide a starting point, the percentages from the original table have been provided.

Fundamentalist                Moderate             Liberal

Original Table                  __40.7__%                        __45.0____%       __56.3__%

White Americans             ____%                                ______%               ____%

African Americans           ____%                                ______%               ____%

Compare the results that you just wrote in this table for whites and African Americans. Are the patterns in those percentages basically the same as or basically different from each other? Circle on.

Basically the same                         Basically different

Among white Americans, which group was least likely to vote for Kerry? Circle one.

Fundamentalists              Moderates              Liberals

Among African Americans, which group was most likely to vote for Kerry? Circle one.

Fundamentalists              Moderates              Liberals

Given these results, which of the following statements is most accurate?

a.     There is no important difference between white Americans and African Americans in terms of the relationship between 2004 presidential voting choice and liberalism of the respondent’s religious tradition.

b.     The relationship follows the same basic pattern among white Americans and African Americans, but it is a stronger relationship among white Americans.

c.     Among white Americans, the more liberal the religious tradition form which one comes, the more likely one is to vote for Kerry; however, African Americans voted overwhelmingly for Kerry regardless of religious tradition. Further, among African Americans, to the extent that there was any difference among the three religious traditions, it was the opposite of the pattern among white Americans.

On the basis of that comparison of percentages, does race affect the relationship between 2004 presidential voting choice and liberalism of the respondent’s religious tradition? Yes or No. _____________

Exercise 6:

Go back to the table for whites and the table for African Americans and, in the following table, write the Cramer’s V and the significance level for each.

Cramer’s V                        Significance Level

Original Table                  .128                                    .000

White Americans             _______                            ______

African Americans           _______                            ______

Which of the following conclusions is most justified by the results?

a.     The relationships are basically the same for all the categories of the control variable and for the original table. Thus, the control variable does not affect the relationship between 2004 presidential voting choice and liberalism of the respondent’s religious tradition.

b.     The relationship between 2004 presidential voting choice and liberalism of the respondent’s religious tradition is different for white Americans and African Americans. Specifically, among white Americans, liberalism of one’s religious tradition is related to 2004 presidential voting choice; that relationship does not exist among African Americans.

Exercise 7:

Compare the column percentages horizontally. Are these results in accord with what we hypothesized? Yes or No. ___________________

Exercise 8:

Prob. = ______________

Kendall’s tau-b = _______________

Is the relationship between degree of conservative self-identification and support for increased governmental services statistically significant? Yes or No. _________________

Do you reject the null hypothesis? Yes or No. __________________

Exercise 9:

Percentages Classifying Themselves as Conservatives by Level of Support for Increased Government Services by Political Information level

                           More Services   Same       Fewer Services    Kendall’s Tau-B              Prob.

High Pol. Inf.      18.4%                 ______%    ________%                    _______                ______

Av. Pol. Inf.      ______%             ______%    ________%                     _______              ______

Low Pol. Inf.    ______%             ______%    ________%                     _______              ______

Exercise 10:

Examine the preceding results. Which of the following conclusions is most appropriate? (Circle the letter of your choice)

a)     The level of political information of respondents has no effect on the relationship between degree of conservative self-identification and support for increased governmental services.

b)    The level of political information of respondents affects the relationship between degree of conservative self-identification and support for increased governmental services: The higher the level of political information of respondents, the stronger the relationship between degree of conservative self-identification and support for increased governmental services.

c)     The level of political information of respondents affects the relationship between degree of conservative self-identification and support for increased governmental services: The higher the level of political information of respondents, the weaker the relationship between degree of conservative self-identification and support for increased governmental services.

Exercise 11:

We need ordinal statistics here. Write the significance level for this relationship.

Prob. = _______

Is this relationship statistically significant?          Yes        No

Exercise 12:

How many of these four tables still have a warning about a potential significance problem? (Circle one.)

None                   One                      Two                     Three                   All Four

Exercise 13:

How many of these four tables still have a warning about a potential significance problem? (Circle onne).

None                  One                      Two                     Three                   All Four

In that situation, did collapsing the variables reduce the number-of-cases problem that can occur when we examine a relationship between two variables while controlling for one or more other variables?

Yes                       No

 
<