Learn: A Tutorial on Running and Interpreting Crosstabulations With Two Variables

Compared with other Americans, residents of the nation's largest metropolitan areas (both central cities and their suburbs) report 10-15 percent fewer group memberships, attend 10-15 percent fewer club meetings, attend church about 10-20 percent less frequently, and are 30-40 percent less likely to serve as officers or committee members of local organizations or to attend public meetings on local affairs. (Figure 50 and figure 51 illustrate these differences.)

Open the DDB dataset. Again, you will be investigating the relationship between the "Size of city residence" (citysize) and "Worked on a community project" (commproj). However, before you can run your crosstabulation, you must recode "commproj" so that it is simplified into a variable that has only the two codes below (return to Exercise 2 if you need to refresh your memory on how to recode):

• 0 = None (Change the old code from '1' into '0' to make more sense.)

• 1 = Worked on a community project in past twelve months (Change the old codes '2 thru 7' into '1'.)

Once you have completed your recode, go back to the SDA analysis window and select "Run frequency or crosstabulation (with charts)" and type "your_unique_name" for your recoded variable after "Row" and type "citysize" after "Column" [instead of typing "year" like you did in Exercise 2]. Under "CHART OPTIONS" and "Type of chart," select "bar chart" and under "Bar chart options" select the "horizontal" orientation. Finally, select "Run the Table."

As you can see below, your results look somewhat similar to figure 50 on page 206, with the exception that in SDA analysis the horizontal bar chart is inverted. The percentages appear on the top rather than the bottom of the chart. Consequently, rural towns appear on the top left, rather than the bottom left.

Now scroll up in your results window and examine the table that SDA has produced. Your results should match those depicted in the screen capture below.

The first thing to note about this table is that the very bottom row contains "COL TOTAL" or column totals. You see 100 percent totals in each column because in SDA analysis, the default for percentages is the column. This is important because it assumes that you have entered your independent variable (the "suspect" of city size in this case) in the column. When investigating relationships with crosstabulations, it is essential that you examine the percentages within the categories of the independent variable. In this case, what is the difference between rural and urban dwellers in terms of working on a community project? Setting your table up the way the table above is set up, with the independent variable in the column, is a good practice to follow. This way you can train your investigative eye to look across--from left to right. Let's try this with the table you produced.

Focus on the second row for those who worked on a community project (the first row is simply the "mirror" image, much like you saw in your line graphs for Exercise 2). Once you have focused your investigative eye on the second row, move it across from left to right, in order to examine the percentage differences across categories of the independent variable or the categories of city size. How big of a percentage difference is there between rural dwellers and the largest city dwellers? The answer: 39.9 percent compared to 31.6 percent, which is a 8.3 percent difference. If you compare this to figure 50 on page 206, you can see that the percentage differences are a bit larger in figure 50 compared to your analysis with community projects as your dependent variable. This is not surprising, given that serving as officers and attending public meetings are civic activities that require even more engagement than working on a community project, and consequently one might expect even larger differences between rural and city dwellers.

There is, however, one more important thing to note about the table you produced. It does not compare differences across time. Instead, you are looking at the totals for all the surveys conducted by the DDB from 1975 to 1998. Each survey that the DDB conducted, from 1975 to 1998, represents a "snapshot" in time, much like taking a picture. The technical term for this in survey research is a "cross-section."

• Is it possible that the differences between rural and city dwellers has become more pronounced over time?

Open the DDB dataset. Run a crosstabulation between "Size of city residence" (citysize) and your recoded variable for "Worked on a community project," but before you run the table type "year(1975)" in the "Selections Filter(s)" box. Under "CHART OPTIONS" and "Type of chart," select "bar chart" and under "Bar chart options" select the "horizontal" orientation.

• What are the percentage differences in this table between rural and city dwellers?

Run the same crosstabulation for 1998 and compare your results with 1975. Your investigation will reveal that a similar relationship existed in 1975 and 1998. City life, even in 1975, leads to less engaged citizens. This analysis reveals, so far, that urbanization may not be the prime suspect. But the story is more complicated than this. There are a lot of other potential suspects or explanatory factors for civic disengagement. Some of these factors may be tangled up with urbanization, something like "accessories" to a crime. Return to the main page for Exercise 3 to learn about controlling for other suspects.