The rise and decline of Wikipedia?

Halfaker, Geiger, Morgan, and Riedl have a new paper on this topic (pdf), here is the abstract:

Open collaboration systems like Wikipedia need to maintain a pool of volunteer contributors in order to remain relevant. Wikipedia was created through a tremendous number of contributions by millions of contributors. However, recent research has shown that the number of active contributors in Wikipedia has been declining steadily for years, and suggests that a sharp decline in the retention of newcomers is the cause. This paper presents data that show that several changes the Wikipedia community made to manage quality and consistency in the face of a massive growth in participation have ironically crippled the very growth they were designed to manage. Specifically, the restrictiveness of the encyclopedia’s primary quality control mechanism and the algorithmic tools used to reject contributions are implicated as key causes of decreased newcomer retention. Further, the community’s formal mechanisms for norm articulation are shown to have calcified against changes – especially changes proposed by newer editors.

This is an interesting paper, but I think it undervalues the hypothesis that potential contributors simply prefer to be in on things which are both new and cool.  Wikipedia, which is no longer new, cannot be so cool.  That is why Beethoven’s 5th does not top the pop charts, though if it were new it might.

For the pointer I thank David Siegel.


Do the authors consider that over time there's simply fewer contributions to be made?

Or what is the optimal number of contributors?

The Wikipedia home page states that there are 5,000,000 articles in English. The figure in the paper plots a decline in active editors in English language Wikipedia from a high of 5,000,000 to about 3,500,000. This hardly seems catastrophic.

Wikipedia's own analysis of its size/growth is interesting:

Self-reporting of size is notoriously fraught.

It's not plausible for Wikipedia to falsify basic statistics like this, given the transparency and decentralized nature of the system. Plus, their stats show a decline in activity that fits with the article. There is just more context and interesting analysis.

+10 Thats obviously why there are less contributors !

I once heard that back in the Middle Ages (or maybe Rennaisance) that in order to earn the equivalent of a Ph. D one had to basically know the sum of all human knowledge. Which I suppose is a finite number. Or is it?

In the time of "Renaissance Men" it was certainly finite.

But now there is (possibly) so much more knowledge that it got infinite?

The libraries weren't very big back then, and did not in fact include a very high share of human knowledge (excluding stuff by Arabs and Chinese, who at that time were pretty far ahead in most areas).

Good Point! That explains all of the points made in the abstract. Maintenance clearly takes less effort. Even those who bring in new content, have less to contribute because of the vast areas they can now reference.

Beethoven's 5th has, of course, lasted this long because of the excellence of musical and spiritual experience it manifests; it will far outlast the current childish parodies of music being produced by that garbage dump we call "popular culture". It will most certainly far outlast Wikipedia (or Google or Facebook for that matter).

The reason it is not "at the top of the charts" has nothing to do with how "new" or "cool" it is--it has everything to do with the mindless "pseudo-culture"--one verging on sort barbarism--that we have created. In fact, it is a reasonable assumption that much of our young today could not grasp a work such as the 5th, warmly and generously accessible as it is, if their lives depended on it. Our collectivist political masters have seen to that. We cannot have the young come close to any understanding or appreciation of the great legacy of Western Civilization.

What guides the high expressions of our civilization has little to do with "coolness". While you may be right about Wikipedia, tossing Beethoven into the mix bespeaks of a most trivial and superficial understanding of the human spirit as expressed in our civilization. You, reflecting no doubt your own wallowing unpopular culture, verge towards the edge of barbarism yourself.

As for Wikipedia, that viper pit of "techno Maoists" that run it have with their PC nonsense chased away most of the sane, It has gotten to the point that outside of mundane and uncontroversial entries such as agreed upon information (e.g., the Periodic tables, Physical Laws, historical dates, etc.) it is really quite useless as a actual research encyclopedia, and is held as such by serious scholars and teachers. Even their entries on Music are often quite bizarre. When it comes to history, politics or biography their entries are quite often little more that Marxist agitprop.

It is really the product of the same sort of BoBo crowd that once gave us "The Whole Earth Catalog", and will in the end have a similar fate.

The world has advanced so thoroughly and fantastically that you have no idea what you're looking at.

Come mothers and fathers/
Throughout the land/
And don’t criticize/
What you can’t understand/
Your sons and your daughters/
Are beyond your command/
Your old road is rapidly agin’/
Please get out of the new one if you can’t lend your hand/
For the times they are a-changing'/

In the case of some Wikipedia articles, literally no idea.

Barbarism! Tehcho Maoists! BoBo crowd! Collectivist political masters! Marxist agitprop!

I can't tell if this guy is under- or over-medicated.

He's attempting a parody of Cicero's parody.'s_law

Well played.

There's always Conservapedia

Here's my favorite entry:

Thanks for introducing me to Conservapedia. Here is what it says for "Global Warming:"

"The global warming theory is the liberal hoax[2] that the world is becoming dangerously warmer due to the emission of greenhouse gases"

MIchael Stipe once said that if the latest REM album were released by another band, that band would be the next hot thing. But that's not true. It would be the next derivative REM cover band.

Two alternative explanations:

1) Wikipedia faces more and more competition from other things you can do on the Internet with each passing year.

2) As Wikipedia grows, the marginal utility of writing more articles, or expanding existing articles, diminishes. The low-hanging fruits have all been picked.

Somewhat implicit in the quote above, but I would guess by now many contributions by new participants (even those in good faith) are reverted or deleted by existing editors, which after some attempts probably means departure. It might be that Wikipedia has reached the limit of its model.

It's not just Wikipedia. I believe there has been a decline in both blogging and forums. People have switched from those platforms to social media.

The issue is that there was a lot of semi-expert information in blogs and forums. For example, if I have a problem with my car, I generally go to a forum for my make and model and search, and generally find an explanation for what's wrong and detailed information on how to fix it.

I've noticed in the social media age people asking questions of a technical nature on Facebook. It's just nowhere near as effective.

With the decline of Wikipedia and forums, and the rise of Facebook and Twitter, we have become more stupider.

Phones and tablets are for consuming not creating. Try typing anything longer than a tweet on your iphone.

> It’s not just Wikipedia. I believe there has been a decline in both blogging and forums. People have switched from those platforms to social media.

Like the coach said about the traded basketball player: I think that move caused the average quality of both communities to increase.

You can't trust wiki on any divisive issue. William Connolley destroyed wiki's credibility on climate, for example. They suspended him for a short period but that they allowed him back spoke volumes about their tolerance crazy activists hijacking pages.

I mean .. seriously. Have a little self-recognition that when you are a hold-out, a fringe believer, your normal mode of play will be to reject consensus and authority right and left.

Leftists have infested the site and you run into 1984 type stuff on some pages. There was a dustup over Neil Tyson's quote fabrication, for instance. It was officially scrubbed last I checked.

Someone should write an article on the paranoid style in internet comments.

William? I thought he was Weasel Connelley: we live and learn.

Wiki doesn't need more writers, it needs editors and fact-checkers.

"Opinion is not truth." Plato

The little bit I've read tended to project 21st century, amoral biases on historic movements which didn't apply to how people were motivated/thinking hundreds of years ago.

Editors would ensure context/background and add value. They might provide that articles don't contain distortions, exaggerations, fabrications, misrepresentations, omissions. It probably won't help. The NYT has editors and the paper is unadulterated bullshit.

In conclusion, if you don't read Wiki you are uninformed. If you read it, you will likely be misinformed.

Does the decline in new contributors really suggest a decline in value, though? I have not noticed any appreciable decline in the quality of Wikipedia articles despite this decline in users, of which I previously did not know.

I think that decline in participation is a result of the generally good nature of wikipedia articles. the low hanging fruit is gone.

I find the learning curve for the editing and backchannel discussion interface too high. Once I learn, by the time I'm ready to make new edits I've forgotten, so I pass.

I think you're seeing the same thing happening on Facebook - the algorithms shove out the many in favor of the active few.

There have always been Wikipedia naysayers. Like many successful things, as success after success was proven, naysaying had to shift.

Sure Wikipedia may be slowing. The last refuge of naysaying?

The bigger story though is still overwhelming success.

"When a thing is new, people say: 'It is not true'.
Later, when its TRUTH becomes obvious, they say: 'It's not important.'
Finally, when its importance cannot be denied, they say 'Anyway, it's not new.'" -- William James

Great saying :)

Though the growth RATE may be in decline, the absolute growth is anything but.
In 2015, the english Wikipedia grew by 362.000 articles. (compared to 270.00 in 2014).
That growth is still very impressive.
At that rate, Wikipedia adds an entire Encyclopaedia Brittanica every 2 months.

"...ironically crippled the very growth they were designed to manage. Specifically, the restrictiveness of.. the encyclopedia’s primary quality control mechanism... as key causes of decreased newcomer retention... {and} are shown to have calcified against changes – especially changes proposed by newer editors."


Iron Law of Oligarchy.

Over time, insider managers always bend an organization to their own personal goals & views, in contradiction to the original objectives of that organization.
Insiders always end up broadly ruling the larger organization of people, despite any formal commitment to open participation and fair representation.

Also has parallels to the general web blog reader-comments issue of Open commenting versus Moderated/Restricted/No commenting. Restrictions and "Quality Control" efforts drive away the good commenters as well as the bad ones.

Also ends up explaining political corruption.

Historically, dynasties came to power via violent uprising/conquest with an associated large army/power base. Over time, the supporters of the new regime purged themselves to maximize individual returns, until one day they were too few to stop the next invading/rebelling army, and the cycle continued.

Nowadays, politics is becoming increasingly the domain of moneyed elites, until the hitherto-disengaged masses realize they can vote themselves into power. This happened already about 100 years ago, and is pretty much what is happening within the Republican Party right now (the Democrats have a bit more organized of a party apparatus and a less radicalized voter base, so they haven't been hit as hard).

Actually, this is a common problem in social media and a current area of research.

Here is a current example: Getting customers to write reviews of products purchased online, or their travel experience and recommendations. The bulk of the recommendations actually become a "good" which is marketed by the web seller (How often have you gone to Amazon to see reviews but not purchase).

Re reviews and member contributions: You either get none, a few from those who experienced something bad, or a confirmatory large number of reviews which falls off over time.

Consider what happens to an online merchant when the first few reviews are actually sets the tone for further reviews (you can conduct an experiment by writing a positive review and a negative review for two products and see how it cascades).

In the Wiki realm, part of the problem is that the last additional contribution is likely to be small or a picky detail, making it not worth the effort to add to the anthill. Same with the nth+1 review where n is very large.

In a parallel to paid reviews, we learned in the last year or two that Congressional staffers are tasked with updating Wikipedia. That's mostly good, I think.

Proof of concept: The disco version, fifth of beethtoven peaked at 1 on the US hot 100.

"Who will judge?" - That is the eternal problem in formal debate. Wikipedia is designed to move topics toward expert consensus. But many of the topics that people consider important and interesting are controversial. Consensus as a tool is useful for some topics but has a declining marginal usefulness as Wikipedia increasingly tries to apply it to topics where it is not so useful. Stanford's philosophy encyclopedia ("Plato"), is an alternative design more like traditional encyclopedias. Experts are selected and paid to produce articles on specific topics. It is (also) outstanding, perhaps because there are many extremely intelligent people involved in selecting the authors - people who are also very high in terms of critical thinking skills. But when it comes to extremely complex tasks different designs nearly always involve different trade-offs. The number of topics at Plato are much less, and the topics sometimes seem to feel (to me at least) as presenting a more unified point of view than is warranted. Each designs, Wikipedia and Plato, are amazing in their own way. Thinking now more expansively, as the historical Plato was in favor of the centralization of political power in the hands of philosopher-kings, and mass participation or democracy might be associated with Wikipedia -- Perhaps a way can be found to have different branches within the governance of an internet encyclopedia such that there are checks and balances between the different modes allowing each design mode to do what it does best. Perhaps, in effect, that is already happening.

Great analysis Richard. You hit the nail on the head. I think there is a huge opportunity for a wikipedia for controversial topics. In fact, I'm building one at the moment - see:

I'm not alone in this area. A friend of mine is building with a very similar goal, but slightly different approach.

Happy to share more information. firstlast gmail

MR comment sections present an alternative to tightly monitored quality control. So, there are tradeoffs.

Comment threads in general are the last refuge of bad ideas.

And why not, all you need is an asserted belief, repeat as necessary.

'repeat as necessary' {and your 4th post on this brief thread alone; we'll have your December MR grand total tabulated by midnight tonite}

My problem is that I too often engage bad ideas, rather than make empty ad hominems.

There goes that arrogant leftist attitude. You just can't help it.

Yeah, when Sailor says we should educate black kids differently, because they have different abilities, only a "leftist" would challenge that.

"formal mechanisms for norm articulation": what is that a euphemism for?

Good riddance. It long ago ceased to be an open collaboration platform. Now that we spend much more of our time in front of phones and tablets instead of laptops and desktops means that there's less utility in being active on Wiki.

I have actually tried to edit Wikipedia articles in areas where I am interested or areas where I have recently read books. It is not a pleasant experience: it is likely that your edits will be reverted by a more senior editor or by a robot, discussions between editors on Wikipedia are filled with bizarre jargon, and there's definitely a mistrust of new people. I am totally unsurprised that retention of new Wikipedians is low and getting lower.

On the other hand, I still use Wikipedia every single day and I find that it gets more useful and reliable each year. I don't think number of new editors is the metric by which to judge if Wikipedia is in decline or not: Wikipedia has simply entered a new phase of its existence, more about maintenance than growth.

Exactly. Same here.

What articles did you try and edit?

The established volunteers who effectively "run" wikipedia really are the living worst, though.

In Enemies of Promise (1938) Cyril Connolly advanced the ambition (with his observation that many books of the 1930s lacked staying power) to be able to write a book that could still be being read ten years after initial publication. Flannery O'Connor famously subscribed to Connolly's standard in 1962, when she wrote a note celebrating the second edition of her first novel, first published in 1952. (I have since finessed Connolly's and O'Connor's views by insisting that to qualify as Literature a work must be being read at least ten years after the author's death, a standard which testifies to the accomplishment of both Connolly and O'Connor--a standard which, incidentally, university undergraduate schools and MFA factories are unable to pronounce upon in their considered judgments of contemporary literary merit.)

Does novelty alone merit respect or elevated regard as a key to discernment or assessment? If tradition and history both have (been permitted to) become so thoroughly devalued, we do not find ourselves merely in some new or novel situation: we are, in the terms of Giambattista Vico, right back where many a declining civilization has found itself just before slipping beneath the waves of "modernity" (or read: "mindless contemporaneity") for good and forever. Both Juvenal and Lucian are ancient authors, having flourished in the second century CE: but neither Juvenal nor Lucian lack novelty today for their both being vastly under-read, even by the plentiful idiots among us who insist that the only reality worth considering is the present reality illustrated only in mirrors of contemporary manufacture. Horace's "carpe diem" has succeeded over the years not merely as an exhortation to daily fulfillment but as a dire warning: otherwise, not even Horace would have survived to be translated.

Death to "cool". Immediately and soon.

Peak computer science student population happened on 2001. Meanwhile tech companies are doing well. So, what's the optimal number of computer science graduates in the economy? What's the optimal number of Wikipedia editors? If this number is ignored, the discourse is just the simple idea that "less is less".

You can now hire pretty good programmers to code whatever software you want at good prices, often working from places like India and Russia. The new skill that's needed is recognizing what would be a good app and writing a specification so that the programmer gives you what you want.

Beethoven 9 > Beethoven 7 > Beethoven 5

Though the status of Beethoven 6 should be higher.

Agree, The Original D, with Beethoven 3 also > 5, which is really a mid-tier Beethoven symphony. And if geewiz is really such a civilized fellow he misses the boat that the symphonies are not Beethoven's greatest works of music, at least to really snobby people. As someone who blathers on about others careening to the edge of barbarism after supposedly wallowing in popular culture, he is really quite a joke, quite aside from being an ideological fanatic, although that is not a big deal around this place.

As for Wiki, it is gradually improving, but I continue to find basic errors of fact in entries. As a professor I do not accept it as a source for papers from students, although I think it can be a useful research tool to suggest real sources.

Also forgets that popular music also existed in Beethoven's time, and that Beethoven was not at all reluctant to use it. (I've written about this, in the context of intellectual property and copyright theft). And the real challenge of listening to Beethoven isn't thinking about culturally warm and cuddly he is, but about realizing how radical he was, and why his contemporaries thought his music was "noise." There are ideas in Beethoven that didn't reappear until Pink Floyd. (Don't know if Waters et. al. listened to Beethoven, but that's not relevant.) Highbrow warm fuzzies? Only proves geewiz doesn't get it.

As far as Wikipedia: Yeah, I get the point about the decline in contributors. Not really convinced it's an issue. With 5M articles, it's hard to think of a topic for a new article. And with the current (probably overly academic, frankly) quality of the contributions, it's hard to think of an article to which one could contribute. I think there is certainly a need to fight off biased contributions. A better metric than the number of contributions would be some way of comparing additions to "new knowledge": for example, Wikipedia seems to have tracked reasonably well the latest developments on the ABC Conjecture (though I admit, that's a sample of 1. And I'm not a number theorist, and barely understand what the ABC conjecture means). Wikipedia also seems to have done a good job of tracking physics beyond the standard model (I'm not a physicist either, but the sample size is now up to 2).

It's not the number of contributors that's important, it's the overall body of work.

The reason the 3rd Symphony of Beethoven is greater than the 5th is that it was the intellectual-aesthetic breakthrough into Romanticism in music, which was lurking in late Mozart, but came full force in the Eroica, even if some musicologists argue that there was this or that piece (none symphonies) that preceded the Eroica to make the break from the older "Classical" tradition, which Beethoven's earlier work reflected, folllowing at least to some degree the great master of classical "Classical" School tradition as defined by Hayden.

"The Ninth. The glorious Ninth." - Alex

I remember playing the 5th in orchestra and realizing what a drab piece it is once you pull the parts together. It has an easily remembered punch line, repeated endlessly, and not much more.

"several changes the Wikipedia community made to manage quality and consistency in the face of a massive growth in participation have ironically crippled the very growth they were designed to manage."

Isn't this exactly what one would expect? Incentives matter.

culture has inertia and a snow avalanche it builds and grows and then cannot be stopped.

Wikipedia sought conformity to mainstream talking points...this started long ago. The culture of wikipedia was designed to cleave to and maintain establishment culture.
By enforcing mainstream culture and values, wikipedia thus garnered attention from the media.

That establishment-oriented wikipedia culture has now snowballed and stifled the growth it once had. That culture inhibits and prevents the use of free labor of educated youth.

Hoist by on their petard.

I tried becoming a wikipedia editor; I fixed poor grammar and spelling a couple of times. All of these edits were quickly reverted by bots.

Apparently established users with authority set up automated bots to revert new edits, then they go in and "approve" them later and get credit for making the edits, to make them look like an active user, leading to them getting admin status/ability to ban, etc. It seems a pretty toxic mechanism. I immediately stopped editing because of this.

Another factor might be that when wikipedia first began, the average editor age was around 14 years old. Those kids grew up and became wikipedia admins. There must be a very strong "I built it, it's mine!" culture there, and interlopers will be turned away.

The question is, how do you fix an institution that has so clearly become sclerotic like Wikipedia?

Somehow, Wikipedians, no matter how many contribute, cannot seem to bring the European History article up to the level not even remotely close to that attained by Britannica. I suppose if your a specialists in some sort of pharmaceutical or pop culture tidbit, Wikipedia is great, but nothing compares to the long form articles in Britannica concerning foundational topics, cultural or otherwise.

Very good for students looking for everyday research topics.

A lot of the work is already done. It's easy to understand why knowledgeable and intelligent people dedicated a lot of time to producing encylopedic knowledge for public benefit. But a lot of the hard work is done already, which easily explains why there are fewer contributors.

In the controversial stuff, I see more usage of the indication of "disputed", where you can read a debate, which seems to be a good indication for controversial issues.

Comments for this post are closed