Dear all,
some of you might remember our survey of contributors to Wikipedia last year. First of all, thanks to everyone who participated in the survey! Taking both the German and the international surveys together, we have a total sample of N=525 contributors, which would have been impossible without support from so many people. Thank you!
Although I'm still knee-deep in analyzing the data, we already have a good idea what keeps people motivated once they are involved in the project. We also see a characteristic pattern in the reasons to join the project, but we could make an even stronger point if we had a suitable reference group to compare contributors and non-contributors.
The best reference group I can think of would be Wikipedia users. Ideally, the best and cleanest way to get that sample is to publish the link on Wikipedia itself (e.g. on every 10'000th page or so). Do you see any chance to make that happen? Of course, we can also include additional questions in the survey that might be useful for the project.
Thanks a lot in advance for your ideas, and best wishes from Würzburg,
Joachim Schroer
On 1/25/06, Joachim Schroer joachim.schroer@googlemail.com wrote:
The best reference group I can think of would be Wikipedia users. Ideally, the best and cleanest way to get that sample is to publish the link on Wikipedia itself (e.g. on every 10'000th page or so). Do you see any chance to make that happen?
I think that's a question for wikitech-l.
On 1/25/06, Joachim Schroer joachim.schroer@googlemail.com wrote:
Dear all,
total sample of N=525 contributors, which would have been impossible without support from so many people. Thank you!
Cool, thanks for the update!
The best reference group I can think of would be Wikipedia users. Ideally, the best and cleanest way to get that sample is to publish the link on Wikipedia itself (e.g. on every 10'000th page or so). Do you see any chance to make that happen?
Please explain what you mean. Every 10,000th user page? A link to the entire survey? How would the link being on a page distinguish contributors from non-contributors?
If you're looking for a way to split visitors into two identical groups, you could separate out anon IPs (mainly readers, not contributors) that end in odd numbers from those that end in even numbers...
Of course, we can also include additional questions in the survey that might be useful for the project.
SJ, remembering that we still need to work on a general / universal survey to point new users to when they sign on...
Hi,
The best reference group I can think of would be Wikipedia users. Ideally,
the best and cleanest way to get that sample is to publish the link on Wikipedia itself (e.g. on every 10'000th page or so). Do you see any
chance
to make that happen?
Please explain what you mean. Every 10,000th user page? A link to the entire survey? How would the link being on a page distinguish contributors from non-contributors?
my idea was to display the link on every 10'000th article, not user pages, because ordinary users probably don't look much on user pages or discussion pages. The inviting text should make it clear that we are primarily conducting a survey of Wikipedia users (as opposed to contributors). This was just an idea, though...
If you're looking for a way to split visitors into two identical
groups, you could separate out anon IPs (mainly readers, not contributors) that end in odd numbers from those that end in even numbers...
Yes, displaying the link only to anon IPs seems a very good idea.
Best wishes,
Joachim
-- Joachim Schroer, Dipl.-Psych. University of Wuerzburg Department of Psychology II, Industrial and Organizational Psychology Roentgenring 10 97070 Wuerzburg Germany
Phone: +49 931 31 6062 Fax: +49 931 31 6063 http://www.psychologie.uni-wuerzburg.de/ao/staff/schroer.php schroer@psychologie.uni-wuerzurg.de
Hi,
Joachim Schroer wrote:
some of you might remember our survey of contributors to Wikipedia last year. First of all, thanks to everyone who participated in the survey! Taking both the German and the international surveys together, we have a total sample of N=525 contributors, which would have been impossible without support from so many people. Thank you!
Although I'm still knee-deep in analyzing the data, we already have a good idea what keeps people motivated once they are involved in the project.
Will you publish the raw data after finishing the paper?
We also see a characteristic pattern in the reasons to join the project, but we could make an even stronger point if we had a suitable reference group to compare contributors and non-contributors. The best reference group I can think of would be Wikipedia users. Ideally, the best and cleanest way to get that sample is to publish the link on Wikipedia itself (e.g. on every 10'000th page or so). Do you see any chance to make that happen?
It's not possible on server side because of caching but you could write some JavaScript code that randomly shows a message (to anonymous users only?). There are some geeks in the German Wikipedia that should know how to write such JavaScript magic:
http://de.wikipedia.org/wiki/Benutzer:BLueFiSH.as/Javascripts_%26_Stylesheet...
The code will be activated if you convice the Community of Admins to add it to
http://de.wikipedia.org/w/index.php?title=MediaWiki:Monobook.js&redirect...
Of course, we can also include additional questions in the survey that might be useful for the project.
In which languages do you want to do the survey? I'm very interested but a little but disappointed because your results are not published yet. How about an open survey where the raw data is published and everybody can analyse it?
Thanks a lot in advance for your ideas, and best wishes from Würzburg,
Greetings, Jakob
Hello,
Will you publish the raw data after finishing the paper?
yes, the dataset will be available on request as soon as the paper is published/in press. Guido Hertel (my supervisor/colleague) did just that with the data from his earlier Linux study.
It's not possible on server side because of caching but you could write
some JavaScript code that randomly shows a message (to anonymous users only?). There are some geeks in the German Wikipedia that should know how to write such JavaScript magic:
http://de.wikipedia.org/wiki/Benutzer:BLueFiSH.as/Javascripts_%26_Stylesheet...
The code will be activated if you convice the Community of Admins to add it to
http://de.wikipedia.org/w/index.php?title=MediaWiki:Monobook.js&redirect...
Thanks for the links, they are most helpful. I'm confident I'd manage to come up with a JS to display a link on every 10000th instance, but to integrate it into the existing framework seems to be more complicated... ;-)
In which languages do you want to do the survey?
I think it should be in English, because more contributors participated in the international survey.
I'm very interested but
a little but disappointed because your results are not published yet.
The American Psychological Association (APA) is extremely strict when it comes to publishing empirical results in one of their journals -- which we intend to do. One has to avoid every impression that those or similar results were published somewhere else before. Basically, I think that is a good policy, but you'd better check twice or three times before "going public".
That said, I will have an English presentation of the results ready by the end of next month, which I am going to present at the "General Online Research" conference (http://www.gor.de/) in March. I believe that this presentation file would be a suitable format to inform participants about the results.
How about an open survey where the raw data is published and everybody
can analyse it?
Once the paper is published, the data from this next planned survey would of course be available on request as well (see above).
All the best,
Joachim
-- Joachim Schroer, Dipl.-Psych. University of Wuerzburg Department of Psychology II, Industrial and Organizational Psychology Roentgenring 10 97070 Wuerzburg Germany
Phone: +49 931 31 6062 Fax: +49 931 31 6063 http://www.psychologie.uni-wuerzburg.de/ao/staff/schroer.php schroer@psychologie.uni-wuerzurg.de
wiki-research-l@lists.wikimedia.org