I am really not sure how many of them are clean starts and socks. Probably
not a lot, but I also doubt that the number is insignificant. Given privacy
policies and people deliberately covering their tracks when using a new
identity, we probably can only guess at real numbers.
Hazarding a guess I would therorize the "returning" editor population to be
around 5-10% at any given time, at most.
Editors have a certain attachment to their identity so starting over isn't
exactly a choice taken lightly - its something done because events connected
with an old name make it more difficult to continue editing under it than it
is to break the attachment to ones identity.
Sent from my mobile device.
On Mar 27, 2011 11:06 PM, "Sarah" <slimvirgin(a)gmail.com> wrote:
On Sun, Mar
27, 2011 at 17:27, Sarah <slimvirgin(a)gmail.com> wrote:
> It says: "Between 2005 and 2007, newbies started having real trouble
> successfully joining the Wikimedia community. Before 2005 in the
> English Wikipedia, nearly 40% of new editors would still be active a
> year after their first edit. After 2007, only about 12-15% of new
> editors were still active a year after their first edit."
> A simple explanation is that a significant percentage of new accounts
> after 2007 were not new people, but people returning with new
> identities, sometimes multiple ones.
On Sun, Mar 27, 2011 at 19:17, Jon Davis <wiki(a)konsoletek.com> wrote:
> Wouldn't someone leaving & returning as a new username be a loss of 1 and
gain of 1?
Thereby being a net change of zero?
The conclusion of the study is that losses after one year were more
likely to happen after 2007. That could be (and almost certainly is)
because a higher proportion of accounts created after 2007 were second
accounts, which were then abandoned for third accounts, or to return
to the first one.
> I'm sure there is some username churn in the stats, but I'd be surprised
> it was a significant portion (more than 1%) of
tens of thousands of
I would dispute that, Jon, based on experience. That's why it would be
helpful to make some effort to identify how many people we're talking
about, as opposed to user names.
foundation-l mailing list