Oh, dear.
From the study's report (not the press release about the study):
QUOTE
We recognize that retention rate does not take into account all important features
of content quality in Wikipedia, including, for example, “edit warsâ€, in which two or
more contributors continually change the content of a topic-entry, sometimes merely
using the wiki to return the article to a previous version of the text.
So, when they say that edits have a 72% retention rate, that may be after the content has been warred over countless times? More importantly, the study then equates the retention rate of an edit with its quality. They then conclude:
QUOTE
It is important to note that the control variables are also significant in explaining
the quality of contributions. The shorter a contribution is the higher its quality, for both
registered and anonymous users. Quality is also higher when the topic article being edited
is larger, regardless of the type of contributor.
That is, if a tiny edit can be "hidden" in a big, fat article, the less likely anybody is going to find that it's of low-quality or vandalism. That's my interpretation, anyway.
There is some good news, however:
QUOTE
The findings of lower quality for anonymous contributors with high participation
indicate a strong but negative interest in the collective good which, if left unchecked,
could destroy the open source good much as other commons can be destroyed by similar
collective action problems.
But then they continue with this already-famous misunderstanding:
QUOTE
To deal with the negative impact of this group of contributors
Wikipedia has instituted a policy that requires contributors to register after some number
of anonymous contributions.
Well, what do you expect from a study that actually labels "anonymous editors who contribute only once" as "Good Samaritans"? They must be disciples of
George Lakoff, the way they framed that premise!
Greg