The wisdom of the crowd works on the principle that everyone is smarter than you. This doesn’t mean you’re the dumbest person in the room; just that averaging answers from a large group of individuals increases the accuracy of the answer over any one persons guess. But what do you do if you’re alone? How can you improve your chances of making a good decision when you are a crowd of one?

Two psycologists, Stefan M. Herzog and Ralph Hertwig from the University of Basel, have designed a method to mimic the wisdom of the crowd by simply polling yourself and averaging those results.

Why does the wisdom of the crowd work?

When averaged, large numbers of predictions about everything from sports results to the weight of a cow can turn out to be startlingly accurate.

A recent demonstration of the predictive power of pooling opinions manifested itself last year in the election website fivethirtyeight.com. In a nutshell, they took polling data from many sources and aggregated them to predict the final electoral collage vote better than any of the individual pundits out there. Fivethirtyeight used slightly more complex statistical methods than simple averaging, but the underlying principle remains the same – averaging data from many sources yields better results.

But why does this work? By gathering estimates from different people you are drawing on the different levels of knowledge individuals have, and most importantly – the different ways in which they are inaccurate.
Simply averaging answers eliminates any random error between people’s answers and it can greatly reduce the effect of systematic error. Systematic error is the killer here, think of it as your own personal bias: do you always under estimate how long it will take to do some task or how long it will take to walk somewhere?

Get a large enough group together and you will get enough over estimators and under estimators together to cancel out and get an answer that is close to the truth.

Poll yourself

The method Herzog and Hertwig have developed is called dialectical bootstrapping where you average your own conflicting opinions to mimic the varied opinions from a large group. But of course the crowdsourcing approach works because everybody has a slightly different amount of background knowledge they are using to get their answer, and consequently slightly different sources of error.

To get a conflicting opinion from nothing but the brain that generated the first answer requires a little mental tweaking to get us thinking about the problem in a different way.

To test if this was possible, the study asked people to identify the dates of historical events like the invention of electricity – not something that many of us would know off the top of our heads.
After making their best guess, participants were then asked to give a second estimate using a consider-the-opposite strategy:

First, assume that your first estimate is off the mark. Second, think about a few reasons why that could be. Which assumptions and considerations could have been wrong? Third, what do these new considerations imply? Was the first estimate rather too high or too low? Fourth, based on this new perspective, make a second, alternative estimate.

Surprisingly this worked brilliantly. The result of averaging the first answer with the second was a guess that was closer to the right date than either of the original answers alone. The key though is the introspective-dialectical framing of the second answer. When people were merely asked to make a second guess without considering why the first answer might be wrong, the average wasn’t any more accurate.

I’m in two minds

So does dialectical bootstrapping really offer us a better way to make decisions? Well, maybe.

Making any decision about something you’re not sure of requires discipline. Making a decision using dialectical bootstrapping requires that we be uncertain and “in two minds”, this of course can be incredibly frustrating. Purposely vacillating between options isn’t a natural inclination of our brain – it really prefers to keep track of one reality at a time, and so requires conscious effort to do more than that. But the benefits of second guessing yourself are real and tangible, especially when you don’t have a large crowd handy to solve your problems for you.

There is also evidence that you can make this process slightly easier by separating your two answers in time. It’s still the same principle, but with the idea that you are more likely to be using a different knowledge base at a future time than now, and hence still increasing the accuracy of the average.

It’s still early days yet for this mental tool so it’s not clear if everyone is capable of being disciplined enough to use it. It is however, not limited to guessing dates you could easily look up in Wikipedia. In principle you can use dialectical bootstrapping to predict just about anything. We have an office pool going here and now I’m asking myself “Why might my team might not win this weekend?”

Here’s hoping close enough is good enough.

Right now I’ve predicted that if you subscribe to the instant Email updates or RSS feed that you’ll get a mental kick out of all the articles published here at Very Evolved.

March 19th, 2009