On Tue, Aug 21, 2012 at 3:32 PM, Tilman Bayer tbayer@wikimedia.org wrote:
I'm not sure she actually said that. In any case, it is wrong - question D6 in the last survey asked if there was a Wikimedia chapter in the country where the respondent lived (https://meta.wikimedia.org/w/index.php?title=File:December_2011_Wikipedia_Ed... ).
My bad, glad it was there. And yes she did. Maybe not in those exact words, but she did put the results into context.
This was already quite criticized last time, yet the question is still there, unchanged, and without more context than it had last time. (under FINAL THOUGHTS).
We tried to avoid modifying questions in order to preserve consistence and be able to do some longitudinal analysis (as it was already begun for that question in http://blog.wikimedia.org/?p=14296 ).
So if a question is poorly phrased, we'll continue having it till the end of time to preserve consistency? Mind you, I do want that question in, I just want it within the same context frame that is given to the same question about the Foundation. And I'm also missing a question about other entities that might actually help Wikimedians that we're or we're not aware of.
Still, I was aware that there had been some objections to that question by chapter representatives (which I don't assume have to do with the fact that respondents rated chapters' performance lower than that of other entities in the two previous surveys), and looked into these concerns while the present questionnaire was prepared; I also reached out to one of the critics in person at Wikimania. But I still haven't seen a compelling argument why the way the question is asked should be biased against chapters. The argument that the opinion of Wikimedians who live in countries without chapter should not count seems weak to me, e.g. because the projects that the work of chapters aims to support are international, and because the question asked about chapters in general, not one particular chapter.
That is not the argument I was trying to make (ie. voices of Wikimedians in a country without chapter don't count). Rather, there is a long list of things the Foundation does, where people are asked whether they knew about it, or not. And after that, right when people have been made aware of everything the Foundation does, they are asked to rate the work of the Foundation. The same question about the chapters comes after absolutely nothing has been said about chapter work, which, I believe, does introduce a bias. In short, people are being asked to rate something they *at this point in the survey* have an idea about (for the WMF) although they might have had no idea about it before starting the survey. All I'm asking is that we review the context in which this question is being asked so results make more sense.
In any case, in parallel to preparing the upcoming survey, we have recently been compiling the public dataset with the anonymized responses from the last survey, which was uploaded yesterday here: http://dumps.wikimedia.org/other/surveys/editorsurveynov-dec2011/ Using this, anyone can do the analysis you suggest and check if ratings differed significantly in chapter/non-chapter countries.
That's great. Thanks. For the record, I'm not expecting the results to be so extremely different, but I think the fact that they might be or might not be is extremely important to know.
Finally, what I regret most, is that o little time was allocated to reviewing this survey collaboratively. There was about 20 days to review, comment, give feedback, translate etc. and this in the mist of the summer holiday for a big part of our community. When I see such processes as the ToS revision [1] conducted successfully allowing for 120 days of discussion, and such surveys rushed through the summer, when their data will probably be used to decide much of the strategic orientation of the next year, I'm thinking we're missing out on trying to collect data that is truly relevant to the work we do and that helps us to review our orientations and adjust how well we support our editing community.
Of course there are lots of interesting and important questions that had to be left out of this survey. As said earlier, the idea is to run the editor survey more frequently from now on, probably quarterly and in a more lightweight version, with a different focus in each.
That's good news and I hope the collaborative process can be reinforced and more time is allowed for comments, reviews, changes and finetuning.
Best,
Delphine