On 3/29/06, Erik Moeller <eloquence(a)gmail.com> wrote:
For someone who values empirical data, you are very
quick to reject a
hypothesis using very little data. I suspect that this is because you
are predisposed towards a particular conclusion. Furthermore, that
there will be an increase in vandalism is not the only relevant
hypothesis.
I do not reject the idea that April fools causes some vandalism.
I reject the hypothesis that the vandalism is to such an enormous
extent that it would necessitate a dramatic change in operations.
I did not fail to consider the below points,
We need to consider many different factors:
1) What types of vandalism are there? A simple page blanking is very
different from a subtle alteration of facts. Double edit vandalism
(two vandal edits in a row) is harder to detect than single edit
vandalism. And so forth.
2) Accordingly, what is the persistency of the vandalism on April
Fools vs. the persistency of vandalism on other days of the year?
3) How much vandalism is committed by regular users? How much by
anonymous users? What is the availability of regular users and
administrators to fix vandalism?
4) What is the situation in other projects/languages than en.wp? How
about WM wikis with small communities that aren't used to dealing with
vandalism?
5) How will the rate of vandalism this year be impacted by the
Seigenthaler incident and other media coverage about vandalism in
Wikipedia?
But I can not accept their relevance. The bulk editing patterns for
April 1/2 last year both in terms of total edits and obviously
reverted edits were well within the normal deviation at the time.
I have no doubt that there will be specific articles which see a
somewhat elevated level of vandalism, but I can't support the notion
that we should change sitewide policy on the basis of events which
were not numerous enough to impact the project wide statistics.
I also remember excellent examples were discussed last
year of serious
vandalism that was particular to April Fools and that was a lot of
work to clean up after. But, unless there is a clear sign from above
that a fast and radical decision is likely given enough convincing
reasons, I am skeptical as to whether it makes sense to dig up these
threads, and to expand upon the above points. We will all find out in
3 days. You should be happy - more data for you to chew on.
Where are these examples? If there were such examples which are so
much worse than our typical vandalism I would expect someone to
produce links pretty much off the top of their head...
Without hard data human perception and memory are poor analysis
devices.. "It's april first and I saw vandalism, oh no, april first
must be a heavy vandalism day".
So again, I'm not rejecting the notion that april first causes some
harm... only that, given the basic data I provided and without a solid
smoking gun, I consider the likelihood of the harm being significant
enough to warrant a substantial behavioral change to be low enough to
be laughable.