Wikipedia:Bots/Requests for approval/SuggestBot 5


 * The following discussion is an archived debate. Please do not modify it. To request review of this BRFA, please start a new section at WT:BRFA. The result of the discussion was Symbol keep vote.svg Approved

SuggestBot 5
Operator:

Time filed: 14:51, Wednesday August 24, 2011 (UTC)

Automatic or Manual: Automatic unsupervised

Programming language(s): Perl, Python (pywikibot rewrite framework), MySQL, Bash shell scripts

Source code available: Currently not

Function overview: Automatically post article suggestions with additional information during research study.

Links to relevant discussions (where appropriate):

Edit period(s): twice a day

Estimated number of pages affected: 300

Exclusion compliant (Y/N): Y

Already has a bot flag (Y/N): Y

Function details: As part of a research study we are looking to add information to SuggestBot's recommendation messages posted to user's talk pages (or to a sub-page in their userspace if they prefer that). The proposed changes will only affect some of the users who have signed up to receive recommendations using one of our sign-up templates. More specifically, we plan to do this study in two stages, where the first stage adds information about article popularity and/or quality to the recommendation message (you can see a draft here). Because we want to compare the results of different amounts of information, we plan to split our users into five different groups (using a random draw to achieve balanced groups), with each group receiving different types of information (one group sees only popularity, one group only quality, and so on, with one group being our control).

In the second stage we are interested in adding specific suggestions on improving the articles (you can find a draft here), e.g. "this page does not appear to have enough sources", but we also plan to do a simple design study with some of our users to understand what specific suggestions are actually useful. Because some of these article improvement suggestions duplicate the work tags currently used, we are interested in experimenting with selecting articles based on discrepancy between article popularity and quality (currently we select articles based on work tags). And again we plan to split our users into groups who receive different amounts of information to be able to compare results to measure effects.

This research study has been approved by the Institutional Review Board at the University of Minnesota. We have added information about the study to the list of research projects on meta, specifically this page. We are also in the final stages of adding the necessary functionality to SuggestBot.

Discussion

 * Question, so this affects only some new signups? How do you decided which get the "new" versions of the suggestions and which get the "old" version? Headbomb {talk / contribs / physics / books} 17:09, 24 August 2011 (UTC)


 * I see that I could have explained this better, and have added to the description to attempt to make it clearer. We do plan to add the information to SuggestBot's recommendations for most of the users who have already signed up (a control group won't see any difference, obviously), which is why the number of affected pages is around 300.  The assignment of users to groups will of course be handled by a random draw to ensure unbiased groups.  Cheers, Nettrom (talk) 18:50, 24 August 2011 (UTC)


 * So basically you are testing an improved suggestion system, which for the end user won't really make a difference, except they might get better suggestions? Seems uncontroversial and fine by me. — HELL KNOWZ  ▎TALK 07:07, 25 August 2011 (UTC)


 * Yes, your basic summary is correct. The changes we are planning to do are not substantial, and users can also freely choose to not participate if they want to.  Cheers, Nettrom (talk) 21:27, 25 August 2011 (UTC)

Uncontroversial and potentially affects only an opt-in list for user talk pages. — HELL KNOWZ  ▎TALK 07:14, 26 August 2011 (UTC)


 * The above discussion is preserved as an archive of the debate. Please do not modify it. To request review of this BRFA, please start a new section at WT:BRFA.