When setting up a statistical test a base size and a respondent count limit can be specified to not run the test on a too low sample size.


Previously the base size/respondent count had to be greater than the specified limits to run the test which made it impossible to perform a test against a 0 % result (as that respondent count is 0).


 The logic of these limits has been updated, the test will run if the base size/respondent count is equal to or greater than the specified limits.


Note, the updated logic is applied to newly created projects only and not to existing projects. In the project settings page you can choose which logic to be used meaning that you can apply the new logic also in older projects.


Here we see the new option in the Project settings page that can be used to apply the new logic to older projects. The setting is applied to Storyteller, StoryCreator and Cross table tool.


 

Here we see the setting in the Project settings page used to change logic, this option is ticked by default in all newly created projects but not in all projects created before this logic was introduced.



 

Note in case you use the old logic:

The base limit for hiding data is hiding the result if the base size is less than the number inserted. 

If the base limit for hiding statistical tests is hiding it if the base is less than or equal than the number inserted, the limit for hiding series should be 1 higher than the base size limit used in significance testing to make this work logically together. If you set this up differently, it will work, but the result will be confusing for the users so it is not recommended.