Work log[edit]

Wednesday, February 5th -- Initial comparison with common definitions for "survival".
Monday, March 10th -- Definition and test spec of trail and survival periods
Thursday, March 13th -- Results of trial and survival periods specs

Thursday, March 13th[edit]

Here's the survival periods I described in the last log over time for enwiki and dewiki.

Survival by parameters. The proportion of surviving newly registered editors is plotted by registration date for a set of different trial and survival periods.

It looks like the periods just push a common trend around. Increasing the duration of the trial period and decreasing the duration of the survival period decreases the survival proportion. Let's do some factor analysis. --Halfak (WMF) (talk) 22:57, 13 March 2014 (UTC)[reply]

Here's the factor comparison. For survival factor, I locked the trial period at 3 months and for the survival factor, I locked the trial period at 3 months.

Trial period factor. The factor of difference between proportions of productive new editors for different trial periods is plotted (based on trial months = 3).

It looks like both enwiki and dewiki have a bit of trend going on where the number of users surviving for 1 or 2 trial months in relation to 3 or more is changing. This not extreme and therefor might not matter. But it does suggest that even users who survive 1-2 months are getting less likely to survive 3.

Survival period factor. The factor of difference between proportions of productive new editors for different survival periods is plotted (based on survival months = 3).

For the survival period duration, we don't see any meaningful change over time. It's clear that we should probably just pick the shortest survival period that we can reasonably make use of. --Halfak (WMF) (talk) 23:19, 13 March 2014 (UTC)[reply]

Discussion[edit]

Summing activation edits and surviving edits[edit]

This seems wrong: SUM(activation_edits) + SUM(surviving_edits) > @m AS surviving. Because what if surviving_edits is 0 but activation_edits > @m?

Good catch! I've updated the SQL. --Halfak (WMF) (talk) 17:05, 24 April 2014 (UTC)[reply]

The $n,m$ edit thresholds[edit]

TODO: Discuss why 1 edit is good enough
TODO: Discuss which namespace edits count

@Halfak (WMF):, I moved this section out of the main page for now, as discussed. DarTar (talk) 14:49, 31 March 2014 (UTC)[reply]

Other wikis[edit]

I've not checked yet in the usual places: were these graphs ever produced for any other wiki, and especially Wikisource or Wiktionary or Wikiquote? --Nemo 07:33, 11 November 2016 (UTC)[reply]