[Foundation-l] Personal research grant collaboration request

Chitu Okoli cokoli at jmsb.concordia.ca
Fri Jul 30 14:32:11 UTC 2004


Hi Angela,

Thanks for your responses. Here are my comments:

>> I am currently working on a couple of research projects involving
>Wikipedia.
>> Two specific projects involve:
>> 1. Obtaining a scholarly evaluation of the quality of its articles by
>> comparing Wikipedia articles with those of other encyclopedias; and
>
ANGELA> This part sounds great.
CHITU> I think this study is very important both for Wikipedia's publicity,
and also for my research program, so that I can "prove" to other scholars
the legitimacy of researching Wikipedia. You know the usual reactions: "You
mean ANYONE can write whatever they like? How do they control this?" ... and
so on. "Scholarly types" are even more uncomfortable than average with these
ideas. I believe such a study would help provide verification in terms they
understand of what I already believe (but am trying to prove with scholastic
rigour): that Wikipedia is CURRENTLY a top-notch encyclopedia, and its only
uphill from here.

>> 2. Mapping the sociological networks of Wikipedians among each other in
>> their wiki activities, and the effects of these networks on their
>individual
>> and group performance in Wikipedia.
>
ANGELA> This part concerns me slightly. Could you explain what you mean by
ANGELA> this please? What would you be doing other than surveying people?
Will
ANGELA> it cause any disruption to Wikipedia? Will the users be aware they
are
ANGELA> being studied? Will they be able to opt out?

CHITU> Actually, this part is much more benign than it sounds. This project
(which I am currently working on as we speak) doesn't involve any
involvement at all by Wikipedians or the Wikimedia Foundation. I've
replicated the English Wikipedia using a database dump, and I'm analyzing
the historical entries directly via SQL queries and special social
networking software to observe these relationships. No problemo. As much as
possible of my research will work directly from my replicated Wikipedia
installations and will involve no direct interaction with Wikipedia or
Wikipedians.


>> In return, they would
>> have to send you a letter of support indicating that they will be
>helping
>> you to collect the data you need via surveys, etc.
>
ANGELA> We might be able to help, but we can't guarantee any particular
ANGELA> response rate since completion of the surveys would obviously have
to
ANGELA> remain completely optional to the users.

CHITU> Of course; response rates are a basic risk with any survey research.
Which is why I LOVE the database dumps--the data is all right there! But for
this kind of investigation, we would have to beg the users to respond as
usual. For the SourceForge hacker survey
(http://www.bcg.com/opensource/BCGHACKERSURVEY.pdf), they got 526 responses
out of 1,648 developers contacted--an awesome response rate of 34.2%. On the
other hand, when they surveyed the Linux Kernel group, they got 134 out of
around 4000--a measly 2.4%.

>> However, some of the further research I might need to do could need
>> substantial help from the Wikimedia Foundation. One particular idea I
>have
>> in mind would be to conduct a survey of Wikipedians to figure out who
>they
>> are, and why they do what they do. 
>
ANGELA> What sort of support were you expecting from us here? Do you want us
ANGELA> to host the surveys? Do you want us to create the database where the
ANGELA> responses will be stored? Or would simply advertising them and
linking
ANGELA> to them on an external site be enough? Is there any reason the
surveys
ANGELA> would need to be on Wikipedia itself?

CHITU> I was thinking of something like a temporary link (perhaps for a week
or two) on all pages (kind of like the fundraising banner) asking for
participation in the survey. This link could link to an external Web survey
host that I would manage directly, so that it would not need to expend any
further Wikipedia/Wikimedia resources other than the links. Of course, the
survey questionnaire itself would have to be fully approved by the Wikimedia
Foundation. I was thinking of at least two different surveys--one for
Wikipedia readers, available for every article page, and a second one for
Wikipedia contributors, available only when they click "Edit this page".
Just thoughts out load. Either way, I would only do what was approved by the
Wikimedia Foundation.





More information about the wikimedia-l mailing list