Introduction
When analysing frequencies of themes within your data, sometimes you analysis just needs to count all of the respondents in your data. Other times, it also makes sense to compare the results amongst different subgroups within the data. For example, you may want to compare younger vs older respondents, customers vs non-customers, Product A vs Product B and so on.
The crosstab visualisation allows you to perform this kind of analysis by generating a table of data with your themes as rows and sub-groups as columns.
Default display
By default the "crosstabs" visualisation in themeit will display two columns:
- "Theme" - which contains the themes and nets within your codeframe
- "% Respondents (Matches)" - which contains the counts and percentages. The percentages are using the column totals by default. For more details about the percentages, see below "Crosstab Settings"
Adding columns to crosstabs
By add further columns (or "breaks") to the crosstab, click the "Add Columns" button in the header.
This will display the "Add Context Columns" dialog, which will allow you to define subgroups using the context variables of the current task.
Note: Only variables defined as context variables for the current task will be displayed for selection. If the variable you need is not displayed, use the "Project Details" page to add it to the list of context variables for your task.
Optionally. you can group multiple context values together to form a subgroup in a single column, by ticking the "Group the values into one column" checkbox. This can be useful for applying to continuous variables like Age where you want group a set of values together rather than analyse by separate values.
Crosstab Filtering
It is possible to filter the crosstab results by using the Filter toolbar. Click here for more details.
Crosstab Settings
By default, the percentages are calculated based on the number of respondents for each specific column.
This can be edited in the Crosstab Settings by clicking on the "Cog" icon:
Select "Overall totals" to calculate the percentages based on the total number of respondent across the sample.