| How To Control For A Variable |
 |
| Opening Files | Find Info | Cross Tabs | Control Variables | Plot Charts | Chart Transfers | Combine Variables | Omit Variables |
When a relationship between two variables is found the next step is to check it for spuriousness. To do this we must control for another variable. First we must "Open" a data set. For this tutorial we will use "Born9.dat" found in the Cen1990 folder. Then choose two variables to cross tab into a perfect table. Go to "Command" then "Cross Tab" and choose your variables. For this example we will use "RaceLat" and "Educ". Then go to "Table" and "Percent Across". A table similar to the one below should appear.
In these results we find that Latinos lowest chance of becoming a college graduate, but this may be caused by some other variable like the amount of children a person has. So, we must control for the variable "Child". To do this go to "Table" then "Control" like shown below.
Here the available control variables will appear on your screen. We highlight "Child" and click "Select".
You will return back to your normal screen. From here you must go to "Table" and then "Percent Across". This will give you the results of cross tabbing "RaceLat" and "Educ" while controlling for "Child". Now, you can see how the percentage differences have changed in each category of "Child" and determine if the relationship is really spurious.
Return To Top