Research talk:The effects of member turnovers in WikiProjects

From Meta, a Wikimedia project coordination wiki

Some notes[edit]

Hey! Just looking over the project page. Here are some notes about stuff that I think is problematic.

  1. You reference "hypothesis" in Methods, but you don't mention what those hypotheses are.
  2. You haven't posted your recruitment message. The consent form is great, but it would be nice to review the message you plan to post on use talk pages. I recommend creating Research:The effects of member turnovers in WikiProjects/Recruitment and posting your proposed message there. See Research:Peer mentorship and snuggle/Recruitment message for an example.
  3. Your user page (User:Bobo.03) is empty. It would be great if you could add a little bit about who you are and how to get in contact with you.
  4. Please take your consent form and interview questions and copy them to a wiki page. That will make it easier for Wikipedians to reference this study in the future (e.g. 10 years from now). Wiki pages are a lot more stable than google docs. I recommend creating Research:The effects of member turnovers in WikiProjects/Consent form and Research:The effects of member turnovers in WikiProjects/Interview questions for these artifacts.

I have a couple of questions too.

  1. How will you be targeting Wikipedians to interview?
  2. How many interview recruitment messages do you plan to send out?

Otherwise, this looks good to me. --EpochFail (talk) 15:55, 2 November 2016 (UTC)[reply]

Hi EpochFail, thanks for your help and input so much! I have made couple changes according your comments - adding recruitment message, switching to wiki page instead of Google doc, etc.
Regarding your two questions. Here are my responses.
1. How will you be targeting Wikipedians to interview?
I actually had a writeup for the sampling methods. Please see below.
We also conducted an interview study to understand turnover in WikiProjects from the perspective of editors. In order to recruit Wikipedia editors who are experienced in the operations of WikiProjects, i.e., editors who do administrative work, we first set a threshold to filter out WikiProjects whose project and project talk pages were edited by fewer than 100 editors since the project creation date. 100 turned out to be a reasonable number after we hand-checked the scope of our sampled projects. This resulted in 689 WikiProjects. We then randomly sampled 10 projects. Finally, we needed to identify one editor from each project to interview. We did this by sorting editors engaged in the project by the number of edits they had made to the project page and project talk page. We then selected three candidate editors from each project who were still active in Wikipedia (by manually checking their user page and user talk page). We then invited these editors to be interviewed by positing a message on their user talk pages, and we interviewed the first editor from each project who responded. In total, we interviewed 10 Wikipedia editors.


2. How many interview recruitment messages do you plan to send out?
This should be answered in the writeup above. Not sure about the spamming limitation on Wikipedia. I think I will send out 5 messages everyday.
Bobo.03 (talk) 16:49, 2 November 2016‎ (UTC) signature and some formatting added by EpochFail (talk)[reply]
OK. This looks good except for one remaining issue.
  1. You reference "hypothesis" in Methods, but you don't mention what those hypotheses are.
--EpochFail (talk) 16:59, 2 November 2016 (UTC)[reply]
@EpochFail Yes.. honestly, hypotheses are not finalized yet.. as we are still pending on our independent variables. Our initial plan was to identify different types of leavers/newcomers based on their productivity(article page edits)/communication(talk page edits) level, which we had hypotheses, but that turned out to be similar to the existing work. Recently I found the work you and Diyi did for role identification (best paper in ICSWM 2016). We thought to look at the turnover effects of editors of different roles. But after chatting with Diyi (we know each other during our internship this summer at MSR), she said you have another ongoing project which is to use editor's editing intentions to identify roles of the editors. She said it a better approach. So I am sort of waiting for her progress for her algorithms. Therefore, we don't have finalized hypotheses at this point.. To save time, we decided to carry out interviews at the same time for our qualitative analysis. If you have any suggestions, please let me know. And also hope the project you have with Diyi is going well.
OK, I've made an edit to the concerning section. Does that look OK to you? See Special:Diff/16032933. Otherwise, it looks like you are ready to continue with sending out ~5 invites per day. Please make sure you check your talk page on English Wikipedia regularly as that will likely be where anyone raises an alarm should your recruitment be disruptive. Godspeed. --EpochFail (talk) 19:42, 3 November 2016 (UTC)[reply]