[WikiEN-l] Totally unscientific investigation...

[Geoff Burling]

On Tue, 8 Nov 2005 kosebamse at gmx.net wrote:

> Thanks for the feedback. One interesting point with my little test seems to
> be that the average quality of our content has not much improved since March
> (or since 2003, as far as I can remember).
>
One detail that attracted my attention in your results was that both times
that you id your test, you found exactly *5* articles were of "good"
quality or better. Was this the result of an unconscious bias, or would a
painstaking study show that about 25% of Wikipedia articles are satisfactory
or better?

So I did my own survey of 20 articles, picked at random.

First result I found was that Kosebamse had better luck than me: of the
20 articles I found, no matter how generous I felt, I couldn't rate
any of the articles I found as anything better than "satisfactory". I did
check Kosebamse's ratings, & the 4 articles K rated as better than
"satisfactory" were arguably good; I added 2 of them to [[WP:GA]]. But the
ones I found weren't that good; I rated about 4-6 out of 20 as "satisfactory".

Two were on the cusp: [[J. Paul Getty]] was in many ways an adequate article,
but most of the article was devoted to discussing a pair of events in Getty's
life: the fact that he had a pay phone in his mansion at one time, & the
events surrounding the kidnapping of his grandson. The narrative of his life
was very thin.

The other was [[Santa Clara Drum and Bugle Corps]], to which I had a
conflicted reaction: this was an article that I had a hard time imagining
why anyone would bother to read it. Now before anyone thinks I would
want to remove this article from Wikipedia, I will say that I would never
put it on AfD, & if I found it there I would most likely defend it from
deletion. However, after I read it, I was still as perplexed as I was
at the beginning over why anyone would write an article about it. And
that is my criticism of the article, & why I wonder if it will ever
reach FAC status: the people who care about this subject have no sense for
how to make the rest of us care about the topic.

But back to my original point: Kosebamse found 25% of the articles at
least satisfactory. I found at worst 20%. I suspect that the number lies
somewhere between the 2 numbers.

But if Kosebamse's two surveys are a fair sampling, it shows that only
25% of our articles at any time -- no matter how many there are -- would
be acceptible as they stand. I find this an interesting suggestion where
we need to focus our energies.

Geoff