October 2008

Tim Lee at Princeton's Freedom to Tinker blog has been writing a lot about Wikipedia, and it's sparked a series of interesting debates. A lot of great discussion in the comments, too. I want to highlight one particular back-and-forth in the comments on Tim's post about why he thinks free-riding isn't a problem on Wikipedia.

The argument is a pretty common one when it comes to online collective action, and it comes down to this: Wikipedia shouldn't exist at all. At least if we accept traditional economic models. Here's this from commenter Mitch Golden:

The issue with Wikipedia (and Firefox, and Linux, and all the other open source software) isn't the free rider problem, it's the fact that it disproves the model of homo economicus that is the basis for classical economics and the political philosophy of Libertarianism. There is just no rational reason for people to contribute to Wikipedia – they aren't paid anything, and the value of any benefit they might be getting from the trivial publicity associated with making an edit is overwhelmed by the cost of the often significant time it takes to do the work. So the question isn't one of free riding, the issue is to explain why Wikipedia exists at all. (It is, after all, the free riders that are behaving rationally!)


Mitch goes on in an interesting back-and-forth with Frater Plotter here, here, and here. In this play, Mitch is in the part of the typical economist, and Frater is in the part of the wise foil for narrow and outdated economic theories. It's really very entertaining.

But, Mitch makes a good and economic point: looking purely at tangible or monetary gain, it would seem like Wikipedia 'breaks' the economic model in which individuals are rational, self-interested people who therefore should free-ride all the time.

Here's how I respond. First, as an aside, the homo economicus model seems a little outdated to me. Most reasonable observers of public goods and collective action these days admit that bounded rationality is more like it. People use a variety of mechanisms other than pure self-interested cost / benefit analysis to make decisions in the real world. They act on biases, preferences, values, and norms. No surprises there. It's just that idealized economic models don't usually account for these distinctly messy, human, social factors.

But anyway, we have Mitch's main question to answer: what is different about writing information versus digging coal? (for example) Well, the example isn't quite right, because coal isn't a public good. But in the spirit of the question, I say there are two main differences that make Wikipedia work:

  1. On the internet, the costs of participation can be extremely small. In fact, once you account for fixed costs like your computer and the internet connection, the marginal cost of contributing is just time. And the risk of carpal tunnel. You may choose to devote a huge amount of time and effort, but you sure don't have to. There's lots you can do at a very low cost. Not only that, but the massive communication network makes the costs of aggregation, coordination, and distribution similarly small. 1,000,000 people writing one word each can be as efficient as a few people writing it all.
  2. When the costs are tiny, even small rewards can tip the balance towards participation. This is the key point. One of the commenters linked above points out some of the social psychological rewards people get from contributing: fun, feeling smart, status, power, reputation, group belonging. These may, on average, provide only very small rewards. On Wikipedia, that's ok because a very small reward is enough to offset a very small cost. Certainly some people will get far larger rewards from these things. Some people will develop reputations that they can later convert to material gain. Some people are rational zealots who believe so powerfully in open knowledge that they'll spend all their time working on Wikipedia. But for most people, the small rewards will do the trick.

So, what happens if we take this 'small costs, small rewards' point of view and apply it to Wikipedia? We find out that the traditional model actually does a pretty good job of explaining things. If, that is, we're taking a wider view of what benefits are in the real world.

Even this isn't a complete answer to the question, though. I think the larger project should be really coming up with a good answer to the question 'What makes Wikipedia work?' It'll be lots of things, but I think we know a lot about that. I'll be testing a lot of ideas about this in my dissertation work, and I'll be posting about it here.

Ryan pointed me to a recent post called 'The Trouble with Free Riding' on the Freedom to Tinker blog at Princeton. I'm going to try and summarize it's arguments in a neat bulleted list and respond to each one, and in the process I'll miss some of the detail. But, wow. There's a lot there, and it's such a thought provoking blog post. I disagree with most of it though. It's pretty imprudent to do away with 100 years of theory on public goods, collective action, social dilemmas.

  1. The notion of free-riding is 'nonsensical' on Wikipedia. It doesn't even make sense to talk about it in that context.
  2. This seems like a rhetorical tactic to me. Wikipedia is a public good. It's produced by a process of generalized exchange which means that one's rewards aren't contingent on they give. Therefore, it introduces a social dilemma and the possibility of free riding. It's a non-starter to say that somehow online public goods are so different that we have to abandon previous notions, though this is a disturbingly popular point of view. Instead, let's talk about the ways that public goods that consist of digital information on the internet have some unique and special properties.

  3. People who free-ride on Wikipedia provide an audience, and the presence of that audience is a motivator for some contributors.
  4. This is a great point. I don't know of any experimental research that looks at an audience-as-motivator effect in public goods. Actually, I'd love to do that research. An interesting thing about Wikipedia is that estimating the size of the audience may be very hard, and may depend on how much you know about the system. So, we wouldn't expect the audience effect to work the same for my grandma who fixes a typo as it does for the Barnstar-waving expert contributor.

  5. Being an audience benefits free-riders because they value the content, and the presence of an audience benefits contributors because they appreciate that someone is out there reading. Therefore free-riding is not a relevant issue.
  6. Erp. Now we've gone too far. The possibility of an audience effect doesn't make free riding irrelevant. First of all, other contributors are just much an audience (maybe more so!) as anyone. So, there's no necessary relationship between free riding and an audience. Also – it's kind of a snitty point – but the author makes an analogy to softball games, community orchestras, and poetry readings. These are certainly non-rival in the sense that one's partaking doesn't prevent others from doing the same. But they're all completely excludable, so they're not really public goods, so free-riding doesn't mean the same thing.

  7. The scale and function of the Internet changes the nature of collective action.
  8. That's for sure. Part of what's so interesting about online public goods is that the internet reduces the coordination and distribution costs so much that 10,000 people contributing a little can more or less be the same as 10 people contributing a lot.

  9. "The concept of "free riding" emphasizes the fact that traditional offline institutions expect and require reciprocation from the majority of their members for their continued existence. A church in which only, say, one percent of members contributed financially wouldn't last long. Neither would an airline in which only one percent of the customers paid for their tickets."
  10. Well, this really doesn't have much to do with how free riding is defined. Moreover, I would argue that most churches get the vast majority of their money from 1% of contributors or less. These are what Oliver and Marwell called 'privileged groups' who have the resources to provide the good on everyone's behalf. They also said that the chances of the group including these people go up as the size of the group goes up.

  11. As long as there are enough contributors to sustain the public good, the number of free-riders doesn't matter because of jointness of supply.
  12. Well, this completely depends on what your goal is. On the one hand, sure, once you reach critical mass, the marginal cost of providing the good is zero (or near-zero), so who cares how many free-riders there are. On the other hand, there are lots of benefits to adding to the group of contributors. Wikipedia isn't perfect – not even close. It's wrong on a lot of topics. It's poorly written in many places. It's skewed heavily towards CS and popular culture, and away from things like history and literature. There's a lot to be gained for Wikipedia by converting free-riders to contributors. And let's not foget about the many, many systems that never get to critical mass.

  13. "Every project would like to have more of its users become contributors."
  14. I understand that this may have been an afterthought, but I doubt this is true. I'd rather the group that writes the Linux kernel stay very, very small. And, it could be that if contribution rates went from 5% to 50%, we'd have a much worse encyclopedia on our hands. This is part of the reason why I think Knol is such a bad idea. Money appeals to everyone, but you don't want everyone contributing. Maybe one reason why Wikipedia is as good as it is, is because the social psychological incentives that are at work there (status, reputation, feeling smart, unique, like your knowledge is valued) work much better for the kind of people we want to contribute. Of course, there are lots of ways to contribute, so what we really want is to turn some free-riders into content contributors, others into editors, still others into fact checkers, and lots of them into proof-readers.

  15. "We've never before had goods that could be replicated infinitely and distributed at close to zero cost."
  16. Really? No goods before that were non-rival and non-excludable? What about national defense or clean air? Classic examples of public goods. Here's the fundamental problem. Yes, online public goods are interesting and unique in lots of ways. But, they don't require us to rethink 100 years of theory. It all still applies, though maybe in new and interesting ways. So, let's get over the hypoerbole and talk about those.

Recently many folks were laughing / sneering / cringing at Wired editor Chris Anderson when he stupidly proclaimed the end of theory and the scientific method. He said: why do we need theory when we can just throw massive data into models and see what falls out? I've already reviewed why this is idiotic.

Today, ComputerWorld has a story about Allan Greenspan's testimony on Capital Hill about the financial crisis: Greenspan, Cox tell Congress that bad data hurt Wall Street's computer models

Business decisions by financial services firms were based on "the best insights of mathematicians and finance experts, supported by major advances in computer and communications technology," Greenspan told the committee. "The whole intellectual edifice, however, collapsed in the summer of last year because the data inputted into the risk management models generally covered only the past two decades — a period of euphoria."

Mr. Anderson – Game. Set. Match.

I find Sam Wang's polling analysis at PEC to be smart and informed, and his commentary to be pretty interesting. But his confidence scares me. Wang seems to be doing all the right things. He's making his data and his models freely available. He's keeping his models purposely simple in order to avoid the chance of a mistake. And he's not fiddling with his model as time goes on. All good things.

But… Wang thought his model was right in 2004 until it turned out it wasn't. And people probably didn't think much of the problems in polling methodology until the Dewey / Truman fiasco, and then they did. What don't we know now? I think the data problem is the more dangerous one. As a commenter on PEC points out, polling is a 'black art', and though we gain a lot by doing a meta-analysis, we're still at the mercy of the underlying numbers. Another commenter points out that for Obama, a win is not a win. We want to crush McCain. Doing so sets a completely different political landscape than a narrow victory would.

Anyway, I don't suggest that Wang should be doing anything differently on the statistics. But, and I'm sure he hates the constant comparisons to Nate Silver at 538, where Nate just says that McCain is in trouble, Wang advises that Presidential race is in the bag, and we should forget about it and start focusing on Senate races. A little more temperance would make the commentary much more palatable for me.

As reported on electoral-vote.com:

Politico went through the financial report the RNC just filed with the FEC and discovered that the Republican National Committee has spent $150,000 for clothes and accessories for Sarah Palin since she was tapped for the VP slot in late August. One shopping trip to Neiman Marcus cost them $75,062.63, for example. They also spent over $4700 on her hair and makeup. Remember how the Republicans howled at John Edwards $400 haircut (which included a house call by the barber)? Google for: Edwards "$400 haircut" and you'll get 27,000 hits. That was major news for a week. That aside, a far more damaging effect of this revelation is that Palin keeps saying she is just an ordinary small-town hockey mom. It is likely that if Joe-the-plumber's wife were to rack up $150,000 in clothing expenses in a single month, Joe might ask how she was planning to pay the credit card bill since the median annual salary for plumbers is $37,514. Palin is already being ridiculed all over the place, and this provides more fodder for the comics.

1st, I'm perfectly aware that I'm using blogging as a way of avoiding the stress of preparing for quals., thank you.

Now, I'm hopping back on-board the Andrew Keen is an Idiot bus. His latest narrow minded commentary is that the economic downturn will mean the end of Web 2.0's 'free labor' movement. Knol will thrive, Wikipedia will fail, etc., etc., etc.

Here are a few reasons why this will absolutely not happen, Wikipedia will be fine, and we'll all keep participating on the internet:

  1. Keen's commentary is narrow minded in the same way that Wired editor Chris Anderson's is (See this recent Wired article), but in the opposite direction. Anderson said the future of the web is FREE. Keen says the future of the web is $$. The truth is: both AND neither. People who participate in Wikipedia or Flickr or blogs may not reap their rewards in cash (some of them do, of course) but they do reap rewards. They connect with other people. They come to identify with social groups. They feel smart, they feel like their knowledge and opinions are valued and unique. They get reputational benefits and status rewards. Yes, I know – the economists are chomping at the bit to tell you that reputational benefits are just future cash rewards when your smart blog helps get you a better job. Sure, that's part of it, but not all of it. The point is, both Anderson and Keen are zeroed in on cash at the expense of all the other rewards (in econospeak: externalities) out there. Maybe that's because they're trying to cultivate a readership. But it's also a common point of view – reduce everything to numbers because that's what fits in my spreadsheet.
  2. Other rewards – let's call them social psychological rewards – are insulated from economic crisis. Sure, if you're about to starve or the police are knocking at your door to evict you, you might be spending less time on the internet. Social psychological rewards are great, though, because you can keep right on getting them when you lose your job and can't pay your bills. You can take solace in the constancy of community, and the fact that you still make a difference somewhere. Do people stop watching TV when they lose their jobs? Do they stop eating, smoking, knitting, running, hanging out with friends, or whatever it is they do to feel normal, even good? Nope. In fact, they often bury themselves in those things. Keen admits this himself, but for him it's because people who are out of a job will have nothing better to do. For me it's because of the rewards.
  3. Furthermore, most people don't have the expectation of monetary reward. This is the fundamental economic fallacy – people don't go around trying to convert their lives into economic gain. Sure, most models will assume that, but it's only part of the picture. There are so many other preferences, values, social interactions that figure in to the choices we make. We're not all marching around with dollar signs in our eyes. I just think it wouldn't occur to most people to think 'Now that I'm out of a job, I'd better get paid for blogging!' because that's not what it's about for them. Keen seems to be the only one screaming that these people should be paid, and that's because (as I said above), he's so stubbornly ignorant of the other benefits they receive.

Re-printed from an old Straight Dope message board. Ha!

The first computer class I took in college was on punch cards. This wasn't before the earth's crust cooled, it was only 1979. Anyway, it's probably why it took me so long to sit down at a computer again; I had nightmares about those @#$&*^~! cards.

Oh yeah? The first programming I learned, we saved our programs on spools of paper tape. We didn't even have a CRT, just a Teletype machine. (This would have been in 1975 or '76). We eventually got a 9 1/2 inch floppy drive and a CRT, and were some happenin' dudes. I've still got some of the spools, and the enormous floppy disk.

You had a Teletype machine? We would have KILLED for a Teletype machine! Once I had to write an entire operating system using nothing but a wall full of toggle switches! Keyboards… HA! You had it easy!

You all had electricity? All we had was a bunch of gears that had to be roatated by hand… Took years just to get the thing to add 1 + 1.

You had gears??? All we had was rocks and sticks. Rock = 0 and stick = 1. You'd get a line of code all written and then a dog would come and carry off your 1's and you'd have to start all over. You guys had it easy.

You had rocks and sticks? Jeez, try growing up in the desert… Grains of sand made the coding damn near impossible!

New word from the Pew Internet and American Life project is that the internet isn't driving a wedge between family members after all. Quite the opposite actually. It turns out mobile devices, digital media, and the internet are giving families new ways to keep in touch, new activities to share. Pew does really nice surveys, and this report has been written by some smart people. (Here's an overview of new findings from CNet.)

I hope, then, that this will really put an end to anyone talking about the HomeNet study. Back in 1998, a group of researchers from CMU published a study (called HomeNet) that claimed, among other things, that the Internet increased social isolation. In fact, they claimed all sorts of negative social psychological influences. Of course, the media took the story up with gusto, writing headlines like 'Sad, Lonely World Discovered in Cyberspace' (from the New York Times). Oh no, the world will come crashing down.

But wait. Just a few years later, the researchers re-did the study with a new sample, and found that nearly all the negative influences had disappeared. Hallelujah, we're saved! Rather than refuting their original results, however, the authors explained the reversal by noting that the internet had changed, and network effects combined with broader diffusion had eliminated the isolating effect of internet use. Hmmm. That might be part of the story.

But if you think that explanation seems too simple to be complete, I'd agree with you. Last year some colleagues and I re-examined the original HomeNet data, which the study's authors kindly make freely available. We applied a technique called Matched Sampling in order to look at the influences of internet. Some day I'd really like to get around to publishing the results of our work here. To make a long story short, when you compare apples to apples – like people to like people – you find almost no negative influences of internet use in the original HomeNet data.

Now, I would never state this as a definitive result. Our method has flaws too. But it's a good one, and just as carefully done as the CMU folks'. When the results differ in that situation, I start to worry that the results are model-dependent, and not reflecting a true inference. So, it's not that the internet changed, though it most certainly did, it's that the original inference might have been wrong. And that study was done by some very smart, careful researchers who used excellent statistical methods.

Why do I tell this long, geeky story now? Because I think refuting the HomeNet results is important for understanding the history of the internet. We should stop telling a story (if anyone is actually telling it), in which the internet isolated people at first, and then tapped them in. The internet is a case study in the diffusion of a new communication technology, and we should get the story straight. These new Pew data show what I believe has been true all along: people who are more social offline are more social online. The internet, mobile devices, etc. give people more ways to communicate. And rather than diffusing that communication over many new people, most people are using email, SMS, MySpace, Facebook, Twitter, etc. to communicate with people they already know: their family and friends.

The Drops of God is a Japanese Manga series about wine. That's right, fancy wine. Today's the first I've heard of it, but it seems like the weekly series matches the ways I think about wine. Down to earth, un-pretentious (I hope), and with lots of visual imagery. Some of the description in just the sample images reminds me of the way my sister-in-law Christine talks about wine. What does the wine invoke? A large room with a leather wingback set in front of a roaring fire. Misty rain outside, someone was smoking a cigar in that chair a few days ago.

Anyone know how I can get my hands on the series? In English, hopefully…

(Click for larger images)

See: NY Times, Xorsyst , Daily Mail

Next Page »