The welcome survey was deployed to Arabic Wikipedia on 2019-07-15 with an experiment in which half of users receive Variation A and half receive no survey. After gathering about a month of data, we will analyze the differences in abandonment rate and differences in activation rate between the treatment and control groups.
This analysis should cover the same metrics as were pursued for other target wikis in T206380.
| Status | Subtype | Assigned | Task | ||
|---|---|---|---|---|---|
| Resolved | Trizek-WMF | T227573 Deploy Growth team experiments on Arabic Wikipedia | |||
| Resolved | MMiller_WMF | T227625 Check incoming data from Arabic deployments | |||
| Resolved | nettrom_WMF | T226221 Setup Welcome Survey for Arabic Wikipedia | |||
| Resolved | nettrom_WMF | T228117 Analyze welcome survey experiment in Arabic Wikipedia |
@nettrom_WMF has sent me the analysis on this, in which it looks like the Arabic abandonment and activation rates are not affected by the survey! That's good news. We're currently deciding where/when to publish the results.
@nettrom_WMF is going to publish the results to this page: https://www.mediawiki.org/wiki/Growth/Analytics_updates/Welcome_survey_editor_activation_rate
The results have now been published: https://www.mediawiki.org/wiki/Growth/Analytics_updates/Welcome_survey_editor_activation_rate#Addendum%3A_Arabic_Wikipedia
One part of our analysis that is not part of the report is site abandonment. In Arabic Wikipedia, we did not find a significant increase in site abandonment. Rather, we found a slightly lower (but perhaps not significant) abandonment rate in the survey group, both on desktop and mobile.