[WikiEN-l] [WikiEn-l] Article surveys, revisited

[Mark Wagner]

I've finished my review of the 100 randomly-selected articles I surveyed
back in November: http://en.wikipedia.org/wiki/User:Carnildo/The_100  Of the
original 100, seven have been deleted, and another three have been turned
into redirects.

Over the past three months, there were 1,121 edits to the articles on the
list, giving a mean of 11.56 edits per non-deleted article.

Not much happens on most articles: of the 93 articles remaining in the
survey, the median number of edits was 3. Nine articles were completely
unchanged since November, and all but 18 of the articles had fewer than ten
changes.  About 99% of the edits were minor things: adding interwiki links,
fiddling with categories and stub tags, adjusting wikilinks, and
spelling/grammar fixes.  Only a few edits added a paragraph or more of
information.

At the other extreme of editing are the four articles with 100 or more
edits.  Unfortunately, this does not neccessarily translate into an increase
in article content.  Of the four articles, only [[Midfielder]] was expanded
significantly.  [[Aleksandr Pushkin]] and [[Lawrenceville School]] were
cleaned up, with some addition of information.  [[List of Barney & Friends
stage shows]] merely suffered prolonged vandalism.

Overall, quality has improved, but not by much. Out of the original 20
substubs, five have been deleted, and four have improved to "stub" status.
Three articles originally classified as "low" have improved to "good".  None
of the stubs has improved beyond stub status, and there are still no
articles considered "high" quality.  No article declined significantly in
quality.

The sourcing situation hasn't changed much: two articles gained sources,
while one article is now unsourced.  [[General Semantics]], the messiest,
most over-referenced article in the previous round of the survey, gained
another three sources, for a total of 17.  Fortunately, it also gained a
great deal of improvement.

The image situation has changed significantly.  Originally, free images
outnumbered non-free ones by 2:1, with only a few images of unknown
copyright situation.  The ratio of free to non-free images hasn't changed,
and the total number of images has gone up.  Of the six unsourced images in
the original survey, all of them have been sourced or removed.  However,
there are also 17 images with apparently-incorrect free-license tags: 16
images from Commons with disputed PD-self tags, and a GFDL tag on an image
that is probably not eligable for copyright.

--
Mark
[[User:Carnildo]]