[Foundation-l] Analysis of statistics

[Milos Rancic]

Bad news is that I was right almost a year ago about trends of new
Wikimedians. Relatively good news is that the statistics may be
interpreted as not so bad ones. Good news is that WMF started to act
in relation to those problems around half a year ago.

I went to en.wp stats [1] and I've seen that:
* Number of new Wikipedians is lowering since March 2007. May 2009 is
the worst month since March 2006.
* Fortunately, numbers of active and very active Wikipedians are
stable since the second half of 2007.
* The problem is that curves for active and very active Wikipedians
look like just prolonged curve of the number of new Wikipedians.

But, I wanted to be sure that this is the trend on other large projects.
* German Wikipedia [2]: worse than English in the sense of new
Wikipedians, however, very stable in the sense of active and very
active ones.
* French Wikipedia [3]: Somewhat better than German, but it just shows
the earlier phase of German Wikipedia.
* Chinese Wikipedia [4]: Almost the same as French.
* Russian Wikipedia [5]: Shows even earlier phase. Lowering number of
new Wikipedians just began.

Then, I wanted to see if there are some problems in general
demographics. So, I've found demographics pyramids of USA [6], Germany
[7] and France [8] (from 2005). If we assume that our target groups
are between 15 and 24, just number of German contributors may be ~10%
less (note that the population groups are now ~5 years older). In the
case of French contributors we should expect ~5% less contributors,
while in the case of USA we should expect ~2% more contributors.

But, this is not all. We should add another variable. A significant
number of the initial "new" Wikipedians (by "initial" I assume the
raising period, in the case of en.wp, it is up to March 2007) were
older. So, younger than them were also inside of the initial group.
But, is the number of older Wikipedians so big that we may expect just
16% (de.wp), 46% (fr.wp), 60% (en.wp),  of the peak number of new
Wikipedians (statistics from de.wp: January 2006=1960 new, May
2009=320 new; see others from the charts)?

* If our dominant groups are 15-24 years old and if we say that they
consist 80% of Wikipedians, we should expect that the number of new
Wikipedians compared to the peak should be: de.wp ~30%, ~35% fr.wp,
~40% en.wp.
* If our dominant groups are 15-29 years old, then the numbers are
~25%, ~30%, ~35%.
* If our dominant groups are 15-35 years old, then the numbers are
~15%, ~20%, ~25%.

(Note that you may move up lower age level and you'll get
approximately the same results.)

In the best scenario, just de.wp is in the dangerous zone. In the
worst scenario de.wp is far inside of the unsustainable development,
while fr.wp and en.wp are still staying relatively well. (However,
again, note that fr.wp and en.wp look a lot like the earlier phases of
de.wp.)

In all cases we need to think seriously how to educate younger
generations about Wikipedia and other Wikimedia projects.

[1] - http://stats.wikimedia.org/EN/ChartsWikipediaEN.htm
[2] - http://stats.wikimedia.org/EN/ChartsWikipediaDE.htm
[3] - http://stats.wikimedia.org/EN/ChartsWikipediaFR.htm
[4] - http://stats.wikimedia.org/EN/ChartsWikipediaZH.htm
[5] - http://stats.wikimedia.org/EN/ChartsWikipediaRU.htm
[6] - http://commons.wikimedia.org/wiki/File:Pyramide_Etats-Unis.PNG
[7] - http://commons.wikimedia.org/wiki/File:Pyramide_Allemagne.PNG
[8] - http://commons.wikimedia.org/wiki/File:Pyramide_France.PNG