Extension talk:QuickSurveys

From reading the code, it seems like users are assigned to a bucket only once, and the bucket is sticky thanks to a browser cookie. Does this also mean that changing the coverage partway through the survey, eg. from 10% to 100% will fail to pull in users who have already been excluded from the survey?

I don't know the answer to your question, but if this is a desirable feature, then please consider documenting it in a subtask of phab:T89970.

Thanks for the response—I've created a subtask on your suggestion, task T316920