Crowd Wisdom, Fake News, Information Disparity, and Antarctic Ice Shelves. What’s the connection?

Are Crowds Looking Better These Days?

Facebook is reported to be using crowdsourcing to keep Fake News in check. See https://headleaks.com/2016/12/facebook-tries-crowdsourcing-fact-checkers-to-fight-fake-news/

Trust in numbers. That’s what democracy is all about, right? In a representative democracy, crowds pick their representatives by majority rule. (I’m talking about the principle, not a debacle like the 2016 U.S. Presidential election.) Wisdom flows from the crowd. . . all of us persuaded of crowd wisdom are prepared to hand over most decisions to the crowd. Thus the popularity of ballot initiatives, such as the ones to legalize marijuana in several states in the 2016 election—let the voters decide, directly. Real democracy. Obtains the wisest results. If two heads are better than one, a million heads are better than. . . yours.

Or are they? There are a couple of things that call that into question crowd wisdom when applied to our real, complex, modern world.

Who’s Calling the Shots?

First: It depends on which crowd—and I’m not just talking about a Trump rally. Is it representative of the population that will be affected by the decision—whether it’s a national election, or where to put the town library? Example: does the vote tally in a U.S. national election convey the will of the population at large? Even without voter suppression, a large portion of the electorate does not vote. Historically, the turnout in U.S. Presidential elections, as a proportion of the voting age population, has hovered between 50% and 60% for decades (and far less in “off year” elections, where the wisdom of the crowd is not to bother with electing legislators).

If we could peer into the minds of the 40-50% who do not vote, what would they tell us (besides f-word-you)? Is it important? Does the simple fact of not voting deliver as much information as we could expect to get, namely, I don’t care? Not caring could be a form of wisdom—comparable to that embodied by holy men in caves in the Himalayas, meditating on the Four Noble Truths?

If that last bit sounds silly, if indeed the non-voters have something more substantive to communicate than they don’t care about our lowly world, then the crowd wisdom of our nation is not represented by votes. The voters are an incomplete “crowd.” Add to the I-don’t-care’s the victims of voter suppression, then the vote is even less representative. . . and therefore less wise?

Hillary Clinton won the national popular vote in 2016. Set aside the undemocratic machinery of the electoral college, and liberals like myself would be satisfied that the national crowd had made the wisest presidential pick. But what about those shadowy non-voters? Some of them say it’s wisest not to vote at all, as a protest against the powers-that-be. If they had been forced to vote (22 countries, to include Australia and Greece, have mandatory voting) would they have gone with Trump, who declared “the system is rigged?” Would they have written in Bernie Sanders? Or just checked “none of the above.”

What do We Know and How Do We Know It?  Ox Weight and What Google Knows

Crowd wisdom is skewed by what I’ll call information disparity. Let’s look at the famous example of the event observed by Sir Francis Galton, who in 1906 found, in a contest to guess the weight of an ox at a fair, the average of what 787 fairgoers guessed was almost exactly the right number. In that case, everyone presumably had the same information: a tangible, physical, visible ox. I’d guess anyone could, if not walk all around it, view it from several relevant angles. What’s more, the audience—of villagers in a rural area, not the general public—already had a good idea of what an ox would weigh.

But in today’s increasingly complex world, there are few things of significance about which you even can have all the relevant information. Some people know a lot about certain subjects, some people don’t know much about any meaningful subject, some people know a lot about a lot of meaningful subjects, etc. But few have complete information about consequential subjects, such as the factors that go into, say, the comparative value of the dollar against foreign currencies. Donald Trump, for example, has incomplete information of the compound reasons for flux of the dollar against the yen, not to mention against the congeries of multiple other foreign currencies (you can bet the farm on that).

The number of people who understand most of the factors are few and far between, and even they have trouble putting that knowledge together into a correct interpretation. Should the wisdom of the few be discounted because they are a distinct minority?

So there’s a lack of information parity. You really don’t know what goes into the pricing of an iPhone; you trust (implicitly) that the market dictates the fairest (if steep) price. But the market operates half-blind when the supplier has an ad campaign going that beats out the competition in marketing but not in value of the product. Those in the industry, and especially the advertiser, know a lot more about what’s going on with value and pricing than you do. The iPhone, whatever you may think about it, is not the ox in the village. Information disparity.

There’s a case to be made that the Internet has enabled us to get more complete information than we ever could in the past. Sure, but there’s such a glut of data pouring down on us through the Internet, it’s hard to sort out what’s important and meaningful (information as opposed to raw data). Unless, that is, you have superbly clever computers to do the sorting for you, such as Google’s. Then Google transfers the information they select back out into the world, often targeting individuals. Such as you. So information disparity is about as big as ever. The quip that Google knows more about you than you know about yourself gets truer every day.

This information disparity is akin to the problem with Adam Smith’s famous “Invisible Hand,” the idea that the accumulated decisions of individuals determine, through the market, the price of goods and even public benefits. It’s why the market is alleged to work for the good of all. It has long been argued that the Invisible Hand is a poor decision-making tool when all actors do not have the same information—it’s missing some fingers and has a tremor.

Slippery Slopes

Back to the matter of Facebook crowdsourcing to identify Fake News. Just how informed and vigilant can the Facebook crowd be? Fake News that is preposterous, such as Ted Cruz’s father was in on the Kennedy assassination, is easy to spot. But there’s plenty of more nuanced Fake News that can fly under Facebook’s radar. Statistics can be slightly tweaked in favor of some agents that would have significant although barely noticeable effects.

You’re probably alarmed, if you read in today’s (Jan 6th) Washington Post that the split in the Antarctic Larsen C Ice Shelf had grown by eleven miles in a month, and a complete break “could be imminent.” (!!) If you then heard that the Larsen B [Bee] shelf calved back in 1996, you might think, whew! it’s bad but not happening as fast as some people think—twenty years between breaks. However, the Larsen B actually went kaput in 2002, just fourteen years ago, in this century. Fourteen years, not twenty, between breaks. (I got these numbers from the Post’s story, so find them highly credible, since the Post has professional reporters and editors who fact-check these things. They don’t have crowds in the newsroom.)

It’s the kind of little tweak that could slip past most Facebook members—it’s a plausible number for an event that happened in a retreating past (year 19-something), and it has less impact than the real number. The person in the street may think something like, oh, ice shelves break off every once in a while, what’s the big deal? A slight shading of the truth, and by the time some better informed member of The Crowd spots it, the misperception may already have sunk in. A subtle difference that creeps into the subconscious and skews one’s perception of Antarctic ice loss henceforward. It takes knowledgeable individuals in the crowd to point out such small discrepancies, and point them out in time to change minds.

To sum it up, I am troubled by an embrace of the Wisdom of the Crowd, such as ballot initiatives to answer questions of great consequence. Direct democracy is a blunt instrument. The Colombian referendum on the peace accord between the government and FARC is instructive. Eventually their Congress passed the accord despite the plan’s small loss in the referendum. Some big points scored for representative democracy. As for Brexit—I rest my case.

 

Relevant reading: James Surowiecki, The Wisdom of Crowds [and a long subtitle]