Page MenuHomePhabricator

'Retention after 7 days' figures are wrong
Closed, InvalidPublic

Description

This event— Editatona Ciudad Juárez —has a suspiciously high proportions of editors who qualify for "Retention after 7 day"s: 7 of 7 (new editors)

So I checked some of the new users. Their stats are below. None should qualify for "Retention after 7 days," since they all registered fewer than 7 days ago.

  • Mslulyana, registered 6 days ago, April 2
  • Georgina.mtz, registered 2 days ago, April 5
  • Daphselegna, registered April 3, 2019 (5 days ago)
  • Paola-perez76, registered April 5, 2019 (2 days ago)
  • Joseline Cassio, registered April 5, 2019 (3 days ago)
  • Val Tolentino, registered April 5, 2019 (3 days ago)

The definition of Retention after 7 days is: The number of New Editors who make at least one edit in any Wikimedia project in any namespace at any time beginning 7 days after the event until the report is run.

That sounds like it's counting from the end of the event. But end or beginning doesn't matter for our purposes, because the event in question took place on 2019-04-06, and today is just 2019-04-08. So no one should qualify yet for retention after 7 days, no matter how you count it. (BTW, the Retention figure was a Grant Metrics figure that we didn't really touch. This event is defined with Participants, as well, so should not have been an issue.)

Testing

For testing purposes, here is another event with a too-high proportion of Retained new editors: Australian National University - 3-4 April 201 and

Event Timeline

Restricted Application added a subscriber: Aklapper. · View Herald Transcript

Actually, Niharika just explained that everyone is counted as Retained until they are proven to be not retained. I.e., since this event is fewer than 7 days ago, it's "correct" that everyone is retained. That sounds weird to me, but if that's the way the figure has been then that's the way it will stay. I'm closing this as invalid (I'll watch what happens to this event in 5 more days).