Tuesday, July 28, 2009

Book Review: “The Wisdom of Crowds” by James Surowiecki

“The Wisdom of Crowds”
By James Surowiecki
Anchor; Reprint edition (Aug 16 2005)

James Surowiecki’s book is an entertaining summary of many, recent and not so recent, studies on crowd behavior. The book comes with many references, and provides quite a balanced assessment of current knowledge. Unlike what the title might suggest, Surowiecki’s book is not a praise of “The Wisdom of Crowds,” but rather an examination under which circumstances crowds are wise, and for which purposes this wisdom might be useful.

Surowiecki distinguishes between three different problems posed at a crowd: problems of cognition, coordination and cooperation. The book is divided into two parts. The first part offers a lot of examples for these problems and the crowds’ attempts to solve them. The second part looks into the question which conditions are necessary for a successful solution. The author identifies three such conditions: diversity, independence (of individuals from each other), and decentralization – though with qualifiers:
“[D]ecentralization works well under some conditions and not very well under others. In the past decade, it’s been easy to believe that if a system is decentralized, then it must work well. But all you need to do is look at a traffic jam – or, for that matter, at the U.S. intelligence community – to recognize that getting rid of a central authority is not a panacea. Similarly, people have become enamored of the idea that decentralization is somehow natural or automatic. [However,] it’s hard to make real decentralization work, and hard to keep it going, and easy for decentralization to become disorganization.”

This paragraph makes clear that understanding the ways crowds make decisions is necessary to set up a system such that decision making is smart. Intelligent organization requires thinking – and scientific research.

Surowiecki warns of factors that dumb down the decisions of groups, most notably skewing information, groupthink, and herding, all of which lead to suboptimal decisions, and potentially disastrous failures.

The cases discussed in the book draw on many examples, the recurring ones are betting markets, and the financial and economic system. The author also dedicates a chapter to the academic system, and ends with discussing politics. While the elaborations on the financial markets are extensive and insightful though somewhat repetitive, those on politics are well meant but vague, and those on academia are hopelessly naïve:

“The coin of the realm, for most scientists, is not cash but rather recognition. Even so, scientists are undoubtedly as self-seeking and as self-interested as the rest of us. The genius of the way science is organized, though, makes their self-interested behavior redound to the benefits of all of us. In the process of winning notoriety for themselves, they make the group – that is, the scientific community and then, indirectly, the rest of us – smarter.”

The chapter comes with very few, basically irrelevant references, and leaves me with the impression the author has next to no experience with academic research. Unfortunately, “the genius of the way science is organized,” is today severely affected by various pressures researchers are subject to in their decision making process, most notably financial and time pressure. That, together with, herding, lacking independence, a meanwhile completely thermalized information basis, and specialization and fragmentation which promotes groupthink, several of the factors Suroviecki previously identified as necessary for smart decision making are not fulfilled. And that doesn’t even touch on the question how well the “smartness” of the scientific community reaches “the rest of us.” For more details, see my post We have only ourselves to judge each other – which suggests the steps to be taking to allow the scientific community to be indeed a decentralized, smart crowd.

A central theme of “The Wisdom of Crowds” is that leadership by a single or few persons is unlikely to be superior to using the full available knowledge (of the group, the company, the community, ie the crowd). Humans tend to assign success and failure to single persons where it might instead have been simply a result of lucky or unlucky circumstances. A consistently better performance, so the author argues, cannot be achieved by picking “the” right person, but by accessing and aggregating the wisdom of the crowd.

It remains somewhat unclear throughout the book what role Surowiecki assigns to experts, though he writes in the afterword that the expert is of course necessary to provide information, for without information neither a crowd nor anybody else can make qualified decisions. He cautions however that many studies have shown that specialists - as all other people - tend to be overconfident about their knowledge and fall for the “Illusion of Knowledge,” ie they fail to see the limits of their own knowledge. If the role of the experts is to provide the crowd with information, then one should keep in mind that the way how information is communicated can spoil the ability to make good decisions, eg by making information seem more (or less) important than it actually is, or by attaching irrelevant details like what specific persons thought about it. (And let’s not even talk about the problem of simply inventing information.)

In any case, the book makes a very compelling case that there is a large unused potential in the wisdom of crowds and that, if we know how to tap onto this potential, we could use it to improve decision making processes in certain situations. Especially scientists as members of large communities, and as members of academic institutions, can learn a lot from better understanding under which circumstances which decision making processes have been successful. The book conveys an optimistic, but also a cautious message, since there are many examples of stupid crowds as well. While the examples in the book fit very well into the argument the author is leading, I am (as often) left to wonder whether there are examples that did not fit into the theme and thus are not to be found in the book.

All together, “The Wisdom of Crowds” is a very recommendable book, informative and well written. If this was an Amazon review, I’d give five stars.

Read my other book reviews.

16 comments:

Uncle Al said...

Vote with the stupid! How can so mnay people bew wrong? Administer centrally, solve locally.

Thomas Themel said...


While the elaborations on the financial markets are extensive and insightful though somewhat repetitive, those on politics are well meant but vague, and those on academia are hopelessly naïve:


Do you have reason to believe that investment bankers/politicians wouldn't find the chapters about finance/politics hopelessly naïve as well?

Andrew Thomas said...

I think the "Ask the Audience" part of the "Who wants to be a millionaire?" TV show is a great example of the wisdom of crowds. If it's a very difficult question then the people who don't know the correct answer (the vast majority in this case) vote randomly, so their votes tend to cancel out as noise (i.e., they vote for answers A, B, C, and D equally and so cancel each other out). In this case, the five people in the audience who actually know the correct answer make all the difference as their few votes are enough to push the correct answer above the background noise and just emerge slightly as the correct answer. I don't think I've ever seen the "Ask the Audience" get an answer wrong, as a result.

Bee said...

Hi Thomas,

They might find it basic and not telling them anything new, but I think what Surowiecki wrote about economics is by and large well backed up by current knowledge, it is well referenced and draws on a lot of studies he explains. I guess if you know the field, you know the studies, but now take academia and all he does is mentioning the Matthew effect, and pointing out its rational aspects. Not only are there are decades of studies about citation behavior, and a huge literature about issues with academic publishing (information sharing), and the side-effects of specialization that aren't even mentioned. But what's worse is that Surowiecki isn't self-consistent in his argumentation, ie he doesn't even bother to explain why in scientific research the factors that he previously identified as necessary for good decision making would be fulfilled. He'd have found that today they are almost all not fulfilled. Best,

B.

Bee said...

Hi Andrew,

Yes, Surowiecki uses this example as well. In fact, it's the first example in the first chapter. He says the audience got it right 91% of the time. Best,

B.

Phil Warnell said...

Hi Bee,

The book sounds interesting, yet I have to say I don’t trust much the wisdom of the mob. That is to say on a personal level I find myself more oft times a victim of it rather then a benefactor. I find the majority in general don’t look deeply into anything and thus form opinion as it being the result of some sort of authority they trust and simply fool themselves into believing they know something, when in truth they don’t. I know this since to a small degree I do look at things a little closer and do try to understand things more fundamentally, instead of it being an opinion formed of someone else’s or some mistaken correlation of often unrelated facts.

I suppose this is why I like blogs such as yours, for I know the authors and some of the commenters share this in common. I would also contend this be the reason why a blog like this, despite its high content value proportionately has far fewer readers then say a less well written, researched and considered sports or celebrity gossip blog. The prime reason for this of course is they require little true understanding, rather mainly only belief as they play to the mob and not serve to enlighten them.

I’m also reminded of a little video offered over at Christine's blog where it clearly demonstrates how the pressure of the majority often alters people’s answers, even if not what they know to be true. This had me wonder how many actually only echo the mob, as to tolerate it, while not actually believing in what it has to say. Strangley enough I feel this may form to be a much larger group then we think as intimidation has always been the greatest strength of the mob with wisdom being the least.

Best.

Phil

Bee said...

Hi Phil,

Surowiecki discusses these issues as well. If people in a crowd share too much information or base their opinions on other people's opinions then the aggregated opinion is likely to be dumb rather than smart. That's the factor of "independence" that he identifies as necessary for intelligent crowds.

The issue that ones' opinion can be affected by authorities is the point that it is relevant how information is communicated to how people rate it:

"Economies and societies depend on, and thrive on, the disclosure of public information. [...] the best way to disclose public information is without hype or even commentary from people in positions of power."

As to the problem that many people don't look too closely into details, that touches on a different issue which is what decisions should be left to whom. Ie in which 'crowd' can you expect relevant information to be present at all.

In either case, many of these factors necessary for intelligent opinion making in a crowd are obviously not fulfilled in today's media world. This is one of the reasons that I've written many times we are not taking sufficient care of the factors that allow or democracies to work.

Best,

B.

Phil Warnell said...

Hi Bee,

The dilemma being, we still need leaders and authorities most people trust, rather than ever hoping they be able to understand much for themselves; while having a media which knows only too well the ones they serve have little interest in ever being able to. I would agree however the media should not to be seen as being the fifth estate, yet rather a service which shows the world unbiased for what it is. The problem of course it the media does see itself as a power and often one that serves the interests of particular people or a group , rather than simply being the mirror to the world.

Best,

Phil

Plato said...

It's a question then of understanding origins?

I thought it significant in terms of the one idea becoming the many, until "such a movement" is pervading.

Best,

Giotis said...

Overall I could say that a well informed and well educated large assembly of people (I don't like the word crowd) is the most trusted mechanism for taking important decisions. It's called democracy and basically it's statistics.

The individual cannot be trusted in that respect. Its role is to challenge the 'common opinion' and drive history forward if he can.

On the other hand rarely the 'crowd' is capable to drive history or to change the order of things because it is to hard for it to get self-orginized. It must be driven by the individual or a group of individuals. In other words the 'crowd' is not creative and if it is not inspired tends to conserve the existing structures due to its inertia.

So everything is good. Things are the way they should be or could be.

Bee said...

Hi Giotis,

Things are the way they should be only in the sense that democracy is a good idea. Unfortunately, as Phil also said, its implementation suffers from not paying sufficient attention to the circumstances necessary for people to make good decisions. Not only are de facto too many decisions made by individual people or small and too homogenous groups, but the availability and integration of information into the decision making process is crucial, and the way it currently takes place it is almost a guarantee that the decisions people reach are not smart.

(I don't like the word "crowd" either. The title of Surowiecki's book is I think a play on Mackay's 1841 book on The Madness of Crowds.)

Best,

B.

Giotis said...

My last sentence was ironical. I should have included a smile at the end :-)

Bee said...

I see, sorry for the misunderstanding.

:-)

Plato said...

I mean sure, we can follow the crowd and fall under misinterpretations of what was happening in a "truly scientific process," was past off part and parcel of something truly fallible within one's character. It happens all the time.:)

Newton’s “Chymistry” of Metal Solubilities

Introduction and Background
Sir Isaac Newton (1642-1727) is a famous name, of course, but few people realize that the man who discovered the law of universal gravitation, unraveled the compound nature of white light, and invented most of the mathematical techniques associated with the calculus, was also a “chymist” (seventeenth-century language for a chemist or alchemist). In fact, Newton wrote about a million words on the subject of early modern
“chymistry,” a subject that combined topics ranging from the refining of ores and the
production of pigments to the elusive goal of transmuting metals into gold. It appears, in
reality, that Newton spent more time on his alchemy than he did on his physics and math
combined! Unfortunately, most of Newton’s “chymical” papers were considered disreputable by his heirs because of the stigma attached to alchemy, and they remain largely unpublished to the present day. He did publish a few things on the subject though,
one of which we will look at below.
See:Newton’s “Chymistry” of Metal Solubilities and more here

Now of course the periodic table has room for predictions of new elements?

Best,

Anonymous said...

A crowd is an ensemble, and sampling the whole ensemble obviously gives better precision for any estimate of the mean.

But Arrow's theorem destroys any illusion that collective decisions is a wise strategy.

Bee said...

and sampling the whole ensemble obviously gives better precision for any estimate of the mean.

You don't want the mean. You want a good decision. It will only be good if individual misconceptions and information is uncorrelated.