A bit of a change of flavour over the next few days.

I’m currently in San Antonio, Texas, with 18,000 other teachers who try to lead the use of technology in their schools, for the annual National Educational Computing Conference.

I’m going to blog about each session I attend to give you an idea of what I find out. Hope its of some use!

The conference theme is “Convene. Connect. Transform.”

The opening keynote was introduced with the exhortation that we all be advocates of change, showcase our students work, and always think BIG.

James Surowiecki, author of ‘The Wisdom of Crowds’ – who is known for his talent for connecting political and social trends together – gave the opening keynote.

Part One.

First he noted that ‘the wisdom of crowds’ might be seen as a counter intuitive expression.

We often think of crowds firstly as crazy – rioting mobs, creators of boom and bust markets – and secondly as stupid – ruled by the lowest common denominator.  The traditional (elitist) view is that the crowd (made up of a few very intelligent people, some experts, but also many average, non-experts) will get it wrong.

In the opening keynote for Necc 2008, Surowiecki challenged the traditional view of crowds with a number of anecdotes, some reference to research studies, and to the success of google, wikipedia, flicker and del.ic.ious.

He opened with the story of a village market competition – guessing the weight of a cow – where the average of the whole group’s guesses was 1197 pounds. The correct weight was in fact 1198 pounds! Not bad.

The Wisdom of Crowds idea is that under the right conditions, the group can be smarter than the smartest person in the group -  If you can make the group work well together and Surowiecki believes that in the last 10 years technology has enabled us to do this on a bigger, more efficient and dramatic way than ever before.

He related several more examples …

1. Jelly Beans Jars guesses – where no one person is smarter than the group as a whole in predicting the amount of beans in the jar.

2. How to be a millionaire – where you may phone a friend or poll the audience. 2/3 of the time the expert friend gets it right. But 91% the audience poll will get it right. Perhaps the pop culture nature of many of the questions help here!

3. The race-track as the perfect prediction machine. The race crowd is made up of experts, cranks, and people just there for fun and enjoyment. It has been founf that the horses with predicted odds of 33%  chance to win – actually win 33% of the time.

He then discussed a number of Web 2.0 phenomena and how they harness the wisdom of crowds ..

Wikipedia – incredibly valuable resource and a collective work.

Flicker (and del.ic.ios)- where the unstructured tagging of photographs (and bookmarks) construct a very workable taxonomy. There is no one in charge.  But the very difficult question of “How do you categorise millions of photographs in a useful and organised way?” is handled very well by the wisdom of crowds.

Surowiecki also sees the key to Google’s success is its ability to harness the wisdom of crowds in the display of its search results. Any search engine could give you hundreds of thousands of results but only Google comes up with an amzingly useful rank order – usually presenting you with what you want in the first ten or so results. Google does this by surveying the crowd where weblinks on individual pages are seen like votes in a marketplace of information. Google has built a multi-billion dollar business out of applying search algorithms to reveal the hidden order and intelligence of crowds. Woah!

Part Two.

What does it take to build a ‘wise crowd.’? Two things.

1.   A way to aggregate and average many individual judgements. (Not simply a suggestion box, gone through by a boss, who edits and filters with toom many judgements of their own).

2.  Diversity. (Diverse groups do better than homogenoeus groups). Not just sociologically or demographically, but cognitively diverse, that look at and frame problems in different ways. People who use different types of tools.  Different Age, Experience, Geography, and Disciplines makes a group less likely to make the same mistake.

To further the point that diverse groups work better, he relied on some research by James March. March developed hundres of computer agents that go about solving math problems and each agent had different competencies in doing math. March found that a ‘random group’ of agents performed better than a ‘cream of the crop’ group that had a ‘higher IQ’ than the random group.

The thesis goes that the random group has a lower IQ but access to more perspectives – so they can see their blind spots. For Surowiecki, adding inexperienced and slower learners to groups makes them perform better – it adds what he calls ‘cognitive intelligence’. The mistakes, errors, of poor performers cancel themselves out against the calls of brilliant performers and you don’t have the downside of groupthink.

Part Three.  GroupThink …. and its danger to the wisdom of crowds.

The principle benefits of of diverse crowds over homogeneous groups is that you avoid ‘everyone making the same mistake’.

Generally people like to work with people like themselves – its more comfortable. But homogeneous groups succumb to group-think too easily. Everyone hears their own opnion refelcted back like an echo chamber. Consequently, it can be harder for the people in the group to see where they are wrong. Harder to spot their biases, blind-spots, and they are likely to become over-confident that they are ‘right’.

Solution. Appoint someone to be devils advocate. Invented by the catholic church. Studies of small group decision making – show that the presence of a devils advocate helps it make decide better. But the devils advocate must change each time. If one person remains ‘devils advocate’ they progressively get less and less effective. So a constantly changing person in this role can work.

But building a diverse team from the start is the ideal. It helps avoid the power of peer-pressure. You need diversity to get people to think for themselves. You want people to rely on their own intuitions, judgements. You want independent thinkers.

But it can be hard to get people to tell you want they really think. For two reasons:

Humans are imitative. And its no wonder, because imitiation works a lot of the time. But you have to get people to move beyond imitation .

Independence is difficult. Going out on a limb is dangerous in many workplaces/classrooms- people might question your ability/intelligence if you get things wrong. The punishments for getting it wrong are often greater than rewards. Our herd instinct – is protective but not effective.

So for the wisdom of crowds to triumph a classroom/workplce has to be willing to welcome real arguments. But alot of people are not confortable with disagreement.

Perhaps we aim for consensus to early when the best group decision emerge out of conflict.

The Paradox – groups are smarterst when people in them are acting as individualistic as they can.

Part Four. Application for Education and the classroom.

In teaching and learning, argument can get you closer to the truth than further away from it.  People must trust people each other – then you can have a group fight it out and be confident that the wisdom of the crowd with come out of it and that that wisdom will be greater than any one person’s view.

Assist students by getting people to talk in reverse order of seniority. Teacher should talk last. You effect students if they hear your view first. You are dictating. You want to hear what people are thinking.

Assist students by finding talkative people. They have an inordinate influence. They speak alot – not just their views, but people talk back to them. They are a hub. We overestimate our ablity to know who to listen to in a group discussion.  Information is often hidden and unspoken.

Tech can be helpful in two ways- 1. casts nets wider and 2. lets more people speak/participate. Online or ‘Non face-to-face’ can let people offer up their real feelings.

A final anecdote. When the USS Scorpian submarine disappeared. The nvy could not find it. A diverse team of specialists was established. They each bet on scenarios for the disappearnce. Using Basiers theorum, the new group info was incorporated with the old guess. No one member chose the spot where the sub was found but their ‘betting’ came up with a collective guess. It was no where near where the navy believed it to be. But the sub was found 20 yards from their guess.

The lesson: the team as a whole could come up with the correct answer but no individual working alone could give it.