Home > Teaching > SYG 3300 > Assignments > Lab 4

Lab 4: Measuring Association, Crosstabulation

1. You'll recall that the first lab assignment asked you to produce frequency distributions for the variables RACE and AFFRMACT, in the dataset RACE96.  At that point, we were interested only in univariate analysis -- that is, how responses to each variable were distributed across the sample.  But you might have wondered whether the two variables were associated with one another.  You can use crosstabulations to find out.

1. State a hypothesis about the relationship between RACE and AFFRMACT.  You may want to consult your work from Lab 1 to remember just what these variables measure.
2. Which of these variables would you regard as the independent variable, and which as the dependent variable?
3. Now, use SPSS to crosstabulate the variables and hand in the resulting table with your work.  Follow the convention of assigning the independent variable to the columns and the dependent variable to the rows of the table.  Is there an association between RACE and AFFRMACT?  That is, do the percentage distributions of the dependent variable vary across categories of the independent variable?  Explain.
4. It's possible that the association between RACE and AFFRMACT might vary across categories of a third variable.  For example, does level of education alter the relationship between RACE and AFFRMACT?  To find out, add EDUCR3 as a control variable and run the Crosstabs procedure again. Does the association between RACE and AFFRMACT vary across levels of education?

2. Do men and women differ in the number of children they would like to have?  We can find out using the variables CHLDIDEL and SEX in the dataset FAMILY96.

1. First, produce a frequency distribution for CHLDIDEL and hand it in with your work.  Notice that more than half of all respondents answered two or fewer children was best.  Before running Crosstabs, we might want to recode this variable to have just two response categories: two or fewer, and three or more children.  So, create a new variable, CHILD2, with the following values and labels:
   

   CHLDIDEL	CHILD2
   Old Values	New Values	Value Labels
   0 through 2	1	0-2
   3 through highest	2	3 or more

2. Run the Crosstabs procedure for the independent variable SEX and the dependent variable CHILD2.  Is there an association between these two variables?  If so, how strong is the association?
3. Does the relationship between SEX and CHILD2 depend on values of other variables like education? Add EDUCR3 and run the Crosstabs procedure again.  Explain how the relationship between SEX and CHILD2 varies across levels of education.

3. Are there generational differences in attitudes about both men and women having careers?  To begin looking at this question, open the dataset WOMEN96 and run Crosstabs for TWOINC and AGER2.

1. Report and interpret the measure of association, Gamma.  How strong is the association between TWOINC and AGER2?
2. Report and interpret the Chi-square statistic.  Would you be confident in generalizing your findings to the population from which the sample was drawn?