Crosstab tables are more difficult to create and interpret than quick reports, but they give you a lot more control over the data you want to compare. Simmons opens to the window that allows you to create a crosstab table.
1. Choose your study. Generally, you'll want to click on Study (blue button), and change it from the most recent 6 month study to the most recent 12 month study. This doubles your sample size.
2. Find your questions and answers. You can click on the arrows next to the categories (lifestyle, print media, computers/internet, etc.) to see the questions and answers within those categories and sub-categories. If you click Search, then you can type in a keyword and limit to the categories that only mention that word. If you choose to search, change the option that says Answers Only to Questions & Answers. Below is a sample search for categories that contain the word yogurt.
Once you have clicked through all the categories and subcategories, you will see actual questions. When you click on a question, you will see the responses appear in the neighboring pane. So, in the screenshot above, the question is do you eat yogurt or smoothies? The responses are yes, no, and don't know/no answer.
3. Move responses to columns/rows.
You can now drag the answers over to the empty panes on the right. Think about whether you want these responses in the columns or the rows. You can also combine answers before putting them in the rows and columns. See the sidebar Combining Responses for more information about this advanced feature.
4. Continue to find your variables and move them to rows or columns.
Now that the yogurt-eating responses are in the rows, I can add responses to the columns. In this case, let's add gender to the columns.
5. Now, run your analysis! Once you have responses in both the rows and the columns, you're ready to click Run Crosstab and get to your data. Click the Interpret Crosstab Tables tab to see how to read and edit your results.
The pane at the bottom of the screen allows you to combine responses before you add them to rows or columns.
Using "AND." One way you can use this pane is to create a more nuanced view of your consumer. So, instead of looking just at female consumers' yogurt consumption, you can look at the yogurt consumption habits of females who are Asian and have at least a college degree. To do this, find each of these responses in the list, and then pull them down to the bottom pane. Use the And button to add AND between each response. Because the responses become an unreadable code when you move them to the bottom pane, you may also want to give this variable a name, in the Name field.
Once you've added what you want, you can click move to rows or move to columns.
Notice that the sample for this population is now down to 119 people, because I added so many factors. A potential problem with using too many ANDs is that your sample may become too small to be generalizable. In the results, Simmons will add ** or * to data points that may be too small to be usable.
Using "OR". While the AND makes your sample smaller, OR wil actually make it larger. One common use of the OR is to make a group larger. In the example below, there was not a question about whether or not there were any hamsters in respondents' households. Instead, the responses were broken down to 1 hamster, 2-3 hamsters, or 4+ hamsters in household. If you pull all these responses down to the bottom pane and combine them with the OR button, you will get the total number of people with hamsters in their household. Again, you can see that I assigned this group a name before pulling it over to the rows or the columns.