Wikipedia:Wikipedia Signpost/2011-04-04/Editor retention

Editor retention

Fighting the decline by restricting article creation?

The decline in new users and articles

Three issues ago, the Signpost reported on Sue Gardner's "March 2011 Update", which dedicated a significant chunk of prose to the results of the Editor Trends Study. The study showed a low retention rate for new editors; this joins a mass of other evidence as to the difficulty new users have with integration, and also the decreasing rate of account creation and new user participation. The number of new accounts has been dropping, with 7,428 created in February 2011 compared to 8,161 a year before; at the same time, more recent statistics suggest that the number of new accounts which make more than five edits is dropping, as well as the number of new accounts which make more than 100. This indicates an ongoing problem with the community's attraction and retention of new editors. Gardner's message and the results of the study were also highlighted in a recent "Message to community about community decline" by the Foundation's chair Ting Chen. A session at the recent Wikimedia and last week's IRC office hour with Sue Gardner were dedicated to the problem, too.

Article creation in general has been declining, as evidenced by a graph covering creations from 2001 to the present. One of the reasons given for both the decline in new articles and new users is the reception users receive at Special:NewPages. It is a commonly held belief within the community that the attitude there has the tendency to be WP:BITEy and to discourage new users from contributing further. Statistics gathered by Mr.Z-man from February 2010 show that almost a third of new users who edited (about 21,000 accounts at the time of the data snapshot) choose to create new pages immediately rather than edit existing ones, and only 0.6 percent of those whose articles are met with deletion stayed editing, compared to 4.4 percent of the users whose articles remained.

Restricting new article creation

In line with the Special:NewPages issues, a proposal at the Village Pump advocates preventing users from creating articles unless they have made a certain number of edits and been registered for a specific period of time. User:The Blade of the Northern Lights, who proposed the change, argues it would help by reducing the backlog for the seriously overworked new-page patrollers (saying that he, personally, had to patrol 200+ pages a day to keep the backlog under control) and, as a result, by reducing the bad experiences of new editors (less stressed editors are less likely to bite newbies). New editors would instead be directed to Articles for Creation, or would be asked to wait until they fulfil the criteria. Another rationale in support of the argument include that a large proportion of newly-created articles are inappropriate; sending the new editors through Articles for Creation would ensure that not only are they less likely to be bitten, but those contributions which are approved are of higher quality than current new articles.

Opponents of the idea make several arguments. First, they say, the backlog does not "properly" exist: of the 30-day buffer provided for new articles, the backlog currently leaves 24 days free – the wiki will not burn down if people take a few days off to reboot. Second, shifting newbies to Articles for Creation would not fix the problem, but would just move it somewhere else; AfC would quickly become overloaded, backlogged and understaffed, like Special:NewPages, resulting in the same issues with the community's reaction to new users that is already prevalent. Third, the proposal has the distinct possibility of disenchanting new users by creating an additional hoop for them to jump through, disinclining those who start by creating articles (nearly a third of new users) from contributing in the first place.

The proposal has since been made into an RfC (Request for Comment).

Article incubation and other efforts

In response to Sue's March update and the general concerns regarding the experience new editors receive at Special:NewPages, several projects have been created to tackle the problem. The article incubation trial, supported by Philippe and James at the Wikimedia Foundation's Community Department, is assessing whether user retention rates can be improved if users have their articles incubated rather than deleted, with full assistance and tutoring provided by the users who do the incubating. People interested in getting involved are invited to sign up and begin userfying potentially viable pages.

User:Snottywong and User:Kudpung have been working over the past few weeks to deal with one of the issues: users who show a less-than-pleasant attitude at Special:NewPages. With the help of a bot, Snottywong and Kudpung have produced a list that has looked at

“

the most recent patrols from mid-October 2010 to mid-March 2011 that were performed in the Main namespace (413,902 patrols total). It counted how many articles each user patrolled, as well as each user's edit count. It then generated this table, sorted by "Edits per patrol". The users at the top of this list are those who have both a relatively low edit count and a relatively high patrol count. It also lists the date of each user's first edit. The bot initially found a total of 3,310 unique users who have patrolled articles in this time frame, but skipped any users who have patrolled less than 40 articles. This resulted in 882 unique users.

”

The idea of the list is for it to be a starting point for finding editors who have a tendency to bite newbies while patrolling and nudge them towards more friendly, non-BITEy behaviour.

The list can be found here; editors are invited to use it to gently correct those who may be making mistakes or who are showing signs of burn-out. This is intended not to attack new-page patrollers, but to allow problems to be tackled quickly and for the long-term benefit of the community, the new contributors, and the editor in question.

Other Requests for Comments are ongoing as part of the Wiki Guides initiative: Allow socializing, Change CSD to userspace drafts, and Minimize talk page templates.

In this issue

4 April 2011 (all comments)

Editor retention

Discuss this story

These comments are automatically transcluded from this article's talk page. To follow comments, add the page to your watchlist. If your comment has not appeared here, you can try purging the cache.

We are not a playground for software writers anymore. Many people have internet access nowadays, and so many childish and immature people have access to it as well, now. We are not a collection of stubs anymore, too; our writing needs to be quite skillful. It needs education to write about science, to understand the article's rationale. Sports & Entertainment might need it too, but I'm not an expert on that. Wikiproject Geology (Rocks and minerals, Earthquakes, and Volcanoes) has only around a dozen active editors. Wikipedia is a small world and the vandalism is violent. Its backlog is a time sink, it is boring, and it bites on your patience. It is not fair that the quality work of a professor gets spoiled by a kid. We have many trends here; one, through the economic crisis we have less viewers (potential editors); two, the internet got popular, and the viewers are less educated on average; three, Wikipedia has a higher quality now, editors need to be more skillful; and four, vandalism get us BITEy. --Chris.urs-o (talk) 13:11, 5 April 2011 (UTC)[reply]

I don't think the economic crisis has made a particular impact on viewing figures, which continue to grow (unless you know something I don't?) and it's a bit more complicated than the-popularity-means-viewers-are-less-educated; I'd go for "the popularity means the demographics of the community have significantly altered". I wrote a couple of blog posts on this which seek to identify the ideological shift (and, if you come to Wikimania, you get to see me perform live! :P). Ironholds (talk) 15:57, 5 April 2011 (UTC)[reply]
- Ok, the economic crisis was important and will be important again, later on (but traffic and number of active visitors, can be different things). Before the Eternal September (Sept. 1993) internet was really, really elitist. I won't go to Wikimania 2011, Israel, I'm afraid ;) --Chris.urs-o (talk) 16:22, 5 April 2011 (UTC)[reply]
  - Well, yes, but your argument was "the economic crisis has caused a reduction in traffic...and a resulting reduction in the number of potential users". Is there any evidence that traffic has been reduced? The blog posts I did on ideological changes in the community may interest you; that's partly where my interest in this area comes from. Ironholds (talk) 16:48, 5 April 2011 (UTC)[reply]
    - I had the period 2009/2010 in mind, if people need two jobs to make the ends meet than you lose potential users. People don't pay hourly their internet connection anymore, this compensates it on the traffic side. --Chris.urs-o (talk) 18:10, 5 April 2011 (UTC)[reply]
      - But again, can you demonstrate any actual evidence? Ironholds (talk) 18:12, 5 April 2011 (UTC)[reply]
        Must search for it, sorry. Nothing ready at hand, Sir... ;p --Chris.urs-o (talk) 18:15, 5 April 2011 (UTC)[reply]
        As far as I'm aware, Wikipedia's alexa rankings keep going up and up and up. We're reaching more potential contributors; the problem is turning them into actual contributors. Ironholds (talk) 18:35, 5 April 2011 (UTC)[reply]
        I remember that we get more visits, but are they different visitors (with english as mother language, of course)? --Chris.urs-o (talk) 18:49, 5 April 2011 (UTC)[reply]
        More individual visitors, yes, not just more views from the same group. Ironholds (talk) 19:00, 5 April 2011 (UTC)[reply]
        We must not forget that we get more vandalism since the 2005-2007 period. We ourselves generate traffic fighting this vandalism. --Chris.urs-o (talk) 19:33, 5 April 2011 (UTC)[reply]
        ...sorry, you think any increase in website traffic can be attributed to us turning up more to kill vandals? Exactly how many vandal-fighters do you think we have? Ironholds (talk) 21:02, 5 April 2011 (UTC)[reply]
        It must be somehow significant, science articles get 40+ more visits on the day after a change. We got vandals and their reverted edits, we got visitors from non english speaking countries and we got ourselves checking recent changes. Your "any increase in website traffic", there is a difference between different visitors and visits, my interest is different visitors and potential new editors. --Chris.urs-o (talk) 00:13, 6 April 2011 (UTC)[reply]
        As said, the statistics we've got are "more individual editors, yes, not just more views from the same group". Ironholds (talk) 01:05, 6 April 2011 (UTC)[reply]
        Ok, our own visits can be significant, though. c. 10% of the edits per minute are being reverted, and articles with changes get c. 40 visits more the next day. --Chris.urs-o (talk) 02:31, 6 April 2011 (UTC)[reply]

1) edits, reverted or no, are statistically minute in terms of page views and 2) yes. Articles with changes, likely to be articles people are interested in, get more hits. Result; we know that people look at things they're interested in. Ironholds (talk) 02:37, 6 April 2011 (UTC)[reply]
- Something is wrong here, you can't have the worst economic crisis, a paper money devalorization and a FAO Grain Price Index going up and more different persons visiting Wikipedia. Maybe Wikipedia is just under tigher observation. Maybe I'll find some reference, but I wanted to just look for minerals, their localities and the Pacific Fire Rim :[ --Chris.urs-o (talk) 06:58, 6 April 2011 (UTC)[reply]
  - Or maybe this is not the worst economic crisis, devaluation is a nation-specific thing and internet access is a small thing for those in western nations. Consider the possibility that if the facts don't agree with you, it isn't the facts that are incorrect. Ironholds (talk) 18:42, 7 April 2011 (UTC)[reply]

Ironholds, I just finished reading your blogs explaining your theory of the history of Wikipedia. I was here for most of that time, & your history doesn't really explain what happened. Although it is better than some accounts I've read. -- llywrch (talk) 23:08, 5 April 2011 (UTC)[reply]

It's meant to be a broad-strokes approach, rather than pointing out any specific incidents; a full history of Wikipedia would take months to accurately produce. Note that I've been here for rather a while too - since 06-ish, anyhoo. Are there any particular bits you felt I got wrong, or was it more things I should've included but didn't? Ironholds (talk) 00:08, 6 April 2011 (UTC)[reply]
- I read it too, thx Ironholds. --Chris.urs-o (talk) 00:37, 6 April 2011 (UTC)[reply]
  - Well, Ironholds, how detailed do you want me to be? The short version is that it appeared you were telling the history as you thought it should have been, rather than how it was; for example, there were surprisingly few "tech-heads" at the beginning. (FWIW, I've been here since October 2002.) Perhaps I should write out my response elsewhere, such as on your Talk page or here. -- llywrch (talk) 16:11, 6 April 2011 (UTC)[reply]
    - As mentioned, the "tech-heads" were only one segment of gen pop, as it were; feel free to write it up somewhere, though. Ironholds (talk) 17:09, 6 April 2011 (UTC)[reply]

This article illustrates exactly why Wikipedia fails to attract new blood: Most people nowadays are busy - they don't have the time to read long, elaborate articles with tons of links to other articles. How about providing a summary, for those of us who can't spend all day engaging in Wikipedia politics? Ottawahitech (talk) 13:38, 5 April 2011 (UTC)[reply]

Well, ideally, it wouldn't be a matter of politics: it's a matter of practicality. —Tom Morris (talk) 13:46, 5 April 2011 (UTC)[reply]

Without an accompanying analysis of how many of these accounts are helpful users versus pure vandals, we get the wrong impression. If 99% of new accounts are "MY beSt friend is..." writers, than losing only 94% of them is bad. If the question is how many serious writers we are losing we need a different metric than total edits. Rmhermen (talk) 13:58, 5 April 2011 (UTC)[reply]

This shows the necessity to make a good first impression. Yes, most n00b creations are not useful and do not belong to Wikipedia. However, there are nice and ugly ways to deal with this problem. Template messages are part of the problem, not only part of the solution. -- Luk ^talk 14:57, 5 April 2011 (UTC)[reply]

Indeed; I wrote a report for the WMF on this a while back (which I think James may have shoved somewhere - not sure where) and identified templating as one of the issues with perceived bitiness. I think more of a hands-on effort is probably the answer, not less. Ironholds (talk) 15:57, 5 April 2011 (UTC)[reply]

Quick question: Is there any breakdown on the number of new articles by new users that are garbage/deleted soon after? Forex, is Wikipedia seeing fewer new articles by new users, but more of those are "real" than a year ago? JKBrooks85 (talk) 05:29, 6 April 2011 (UTC)[reply]

Re to Ottawahitech and Chris urs. People don't disappear in a recession, but many of them have less money to spend, and while some people are busier than ever millions are unemployed or underemployed in part time jobs. Recessions are a great time for products and services that people trade down to, where I come from vegetable seed sales have recently overtaken the sales of flower seeds for the first time in a generation. A free online encyclopaedia is a fairly cheap hobby for those of us who are currently underemployed, it certainly comes cheaper than an evening in the pub. I for one am not surprised that our viewing figures continue to rise in line wit the growth of the Internet, though I do worry that we are not recruiting editors as rapidly as we lose them. For people studying another language editing in it must be great way to get experience and feedback on that skill; They are just one group of potential editors who I would be hoping to see more of in a recession. Ϣere SpielChequers 08:28, 9 April 2011 (UTC)[reply]

Not necessarily causal

With regard to: "...almost a third of new users who edited (about 21,000 accounts at the time of the data snapshot) choose to create new pages immediately rather than edit existing ones, and only 0.6 percent of those whose articles are met with deletion stayed editing, compared to 4.4 percent of the users whose articles remained." Note that the deletion is not necessarily the causal factor in the non-continuation of the affected editors. It stands to reason that frivolous jokesters creating intentionally dumb material which is deleted would not be expected to stick around even if that material were retained. Moreover, those that are serious about editing at Wikipedia over the long haul, having an interest coming in, would be less apt to have their articles deleted in the first place. So the BITEyness and fast trigger to delete at New Pages may OR MAY NOT have an impact upon retention. The only way one could tell for sure is by "exit polling" those with an article deleted about whether the deletion changed their intentions. Carrite (talk) 16:27, 7 April 2011 (UTC)[reply]

Agree ;) --Chris.urs-o (talk) 18:18, 7 April 2011 (UTC)[reply]

The other way to do it is to compare the retention rates by looking at the reasons for deletion: if your hypothesis is correct, those people would probably be reported under G3, G5 and G10: vandalism, hoaxes, banned/blocked users and attack pages. That's basically your vandalism CSD pages. Do those people just do a drive-by article creation? We probably don't really care that the "MY FRIEND IS GAY!" people don't come back. But what about good-faith editors who put up stuff that gets deleted under A7 or G2? What about people who make a page, go away, come back to find it's been PRODded or AfDed in their absence and give up. The vandals and "frivolous jokesters" can be analysed apart from the ordinary editors by looking at retention for deleted articles excluding G3/G5/G10, and by excluding blocked users. Impossible exit polls are not the only way to find this out. —Tom Morris (talk) 18:24, 7 April 2011 (UTC)[reply]

The difficulty is that errors are made both at newpage patrol and at speedy deletion. If something is correctly tagged and deleted per the speedy deletion criteria then we haven't lost a worthwhile article and while the new editor may not like having their contribution rejected, at some point they may be back determined to do better. The problem is when an article is incorrectly tagged and even deleted for "poor formatting", "should probaby fail AFd" or any of the other misuses of the speedy deletion process. I think the debate over new articles by new editors needs some stats on the proportion of incorrect speedy deletions. Ϣere SpielChequers 08:01, 9 April 2011 (UTC)[reply]

I don't get it

Just how is one supposed to use that list? Questioningly, GeorgeLouis (talk) 06:29, 12 April 2011 (UTC)[reply]

The Signpost needs your help putting together the next issue.

Home

About