(4) Filtering: Look back at (2) step 3 in Analysis of Data Using SDA, where we typed "year (96)" in the Selection Filter(s): box. This selected the GSS 1996 data from the complete data set (1972-96) for analysis. We could use the filter to select for other variables to focus our analysis. We could perform an analysis on Catholics, for example, by filtering in the following way:

Our output tables would contain only that group we filtered for. In this case, Catholics.

(5) Recoding and Control Variables: There is virtually no difference between Protestants and Catholics in terms of how they feel about abortion for any reason. Let's explore this further by introducing another variable into the analysis -- frequency of church attendance. The first thing you will do is to run a frequency distribution for ATTEND.

1. Type the name of the variable, ATTEND, in the Row box
2. Type year(96) in the Selection Filter(s): box
3. Click the Percentage Total box
4. Click the Run the Table button

The problem is that there are too many categories of attendance, so we'll reduce the number of categories by recoding the variable. Let's combine the first four categories (0 to 3) into one category and call that seldom, combine the next two categories (4-5) into a second category and call that sometimes, and combine the next three categories (6-8) into a third category and call that frequent. The last category is the missing values category and anything that is not recoded is automatically defined as missing. Now we'll have three categories:

1. seldom (categories 0-3)
2. sometimes (categories 4-5)
3. frequent (categories (6-8)

(6) Now we're ready to recode ATTEND and introduce this variable into the analysis as a control variable.

1. Type the name of your dependent variable, ABANY, in the Row box
2. Type the name of your independent varible, RELIG, in the Column box
3. Type the name of your control variable, ATTEND, in the Control box. In order to recode ATTEND, the full command should look like this:


attend(r:0-3 "seldom"; 4-5 "sometimes"; 6-8 "frequent")

1. r: indicates that you want to recode this variable
2. 0-3 indicates that you want to combine values 0 to 3
3. text inside the quotation marks is the value label and will appear on the output
4. ; separates one recode from another recode
5. everything is included within parentheses
4. Type year(96) in the Selection Filter(s): box
5. Click the Percentage Column box
6. Click the Statistics and Color coding boxes
7. Click the Run the Table button

Back | Other Home