Talk:Learning patterns/Number of newly registered users

From Meta, a Wikimedia project coordination wiki

Are the two metrics in sync?Nemo 10:34, 9 November 2015 (UTC)[reply]

Yeah, they should be, Nemo. Wikimetrics uses the logging table to exclude proxy-registered users, consistent with the Research metric definition. Do you see any info in the pattern that should be updated/clarified? I just added a link to the Research metric page under 'See also'. Jmorgan (WMF) (talk) 18:21, 12 November 2015 (UTC)[reply]
I think the two pages should be merged, as they talk about the same thing. Alternatively, the first line (bolded words) should link Research:Newly registered user to clarify it's the same concept. Finally, if the two pages are not merged then it would be useful to clearly state which of the two is the master definition. Nemo 21:12, 12 November 2015 (UTC)[reply]

Proposal to change metric criteria[edit]

I have been discussing this metric with the Programs & Events Dashboard team.

This metric is currently defined as "Start date should be two weeks before the start day and time (in UTC) of your project or program. For some, you can set the start date to the actual start date of your project. End date should be the last day and time (in UTC) of your project. For some events, you may need to select a time after the event because users are unable to create their username or change their username later on."

Other people have their own opinions. I think this metric is problematic.

It singles out people who make accounts at or right before particular events. The rationale for this is that a premise that single experience is the start of most Wikipedians' entry into the Wikipedia community. While this may be true in locations and outreach areas where Wikipedia events are rare, for demographics which are targeted repeatedly like outreach to women or outreach in cities with highly active Wikipedia communities, new users tend to come to Wikipedia after bombardment over a long period of time with many events. There is some value in making an event which actually prompts people to sign up, but I am not convinced that the WMF should design a measurement problem which incentivizes outreach to the point of registration then de-emphasizes other outreach like follow up and early support for new users.

One problem with incentivizing account registration is that it puts Wikipedians who do outreach in competition with each other to be the ones to do the quick act of recruitment. If someone does more signups, then by this metric, they are getting more Internet points than someone who provides training or ongoing support for new users after the act of registration.

New users might have a longer period of being new than a single event. For example, they might be interested in joining a medicine event, but when they look at Wikipedia events in their city, they see that a wiki art event will happen before the medicine event and a library event will happen after. If that person attends all three events, then by the current standard, only the first event in the series will get credit for new user outreach, even if there was a concerted effort by many Wikipedians to provide new user support. It is not fair to the other organizers who get judged with this scoring system when everybody should be collaborating to provide the most support without regard to what others are doing.

Here are some proposed alternative metrics:

  • Track beginning editors by one year Instead of counting registrations, count the number of events which have any editors who registered their account in the last year. These are the people who benefit most from in-person support. Probably more important than day-old users are three-month old accounts, because these are people who are digesting the editing process and are more committed to learning it.
  • Track beginning editors by edit counts In the English wiki community anyone who has fewer than 300 edits is subject to some restrictions, then finally at 500 edits many more privileges become accessible. For other languages the numbers could be different. I am comfortable saying that anyone with fewer than 500 edits is a beginner, and any event serving this community is one that serves new editors.

Blue Rasberry (talk) 18:40, 1 February 2017 (UTC)[reply]