The traditional approach to market research analysis is to create banners, combining multiple questions across the top, as shown in the example below. While it is possible to create such tables in Q, it is often more efficient to instead get Q to create lots of tables and either sort them by statistical significance or to delete insignificant tables.
Creating lots of crosstabs and deleting insignificant tables
- Create lots of table. For example, crosstab every question in the study by the questions you would ordinary put in a banner. The fastest way to do this in Q use Tables - Crosstabs.
- Automatically delete tables that are not statistically significant. For example, using either of:
Using Smart Tables to sort tables according to statistical significance
Smart Tables can be used to automatically crosstab one question with other questions, sorting them according to level of significance. Thus, the user can run a large number of tables and just focus on the ones that are most significant.
- Categories are automatically combined.
- You can compare multiple questions at the same time. For example, if there were, say, three key questions in a study, you could use Trees to rank order the other questions in the study according to how well each question jointly explain these three questions. This is particularly useful if comparing segmentations, as it allows you to statistically compare segmentations in terms of their explanatory power across multiple dependent variables.
- Trees are much slower than Smart Tables and it is usually impractical to use Trees with all the questions in a study at the same time.
- Information Criteria are used instead of statistical tests.