I am trying to compare various indicators as they change before/after an intervention that takes place in a process for a specific period of time, specifically a protection spell at Wikipedia. This is when edits are temporarily restricted to only admins and other verified users. I.e., anonymous contributions are not allowed. Through text mining I have identified various variables, for example the level of "conflict" that occurs within an article over time.

Now I want to compare the level of "conflict" before a protection spell (a certain time interval before the protection spell), during the protection spell (a limited time interval), and after the protection spell (a limited time interval).

What would an appropriate technique for doing this be? Would a repeated measures ANOVA be helpful here?


1 Answer 1


I think this is a good question but needs some fleshing out in terms of whether order matters.

If you don't care about the amounts of time or order then, yes, an RM ANOVA will work, but only if your data meets the assumptions of the ANOVA and garners relatively normally distributed residuals. The benefit is that the RM ANOVA only assumes that the repeated measures are linked to a subject, or in your case a wiki-page. It makes no assumptions about the ordering of the items.

If order, time, or amount of time is something of theoretical importance for any of these stages, then I would use a multilevel model aka mixed-effects model. But again, be careful of assumptions.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.