Friday, 16 April 2010

The Prime Ultimatum

No, it's not a Matt Damon film...

Richard Wiseman has an interesting experiment over at his blog this week. Have a look at it before you read on - or if you do read my post first, please don't participate in his experiment in case you bias the results.

Back now?

The experiment combines two really interesting cognitive principles. It's an example of the ultimatum game, a game theory experiment which measures our attitude to fairness. In a rational world, the person making the split should offer their partner just £1 (or even £0) and the partner would have no reason to refuse. In the real world, people do tend to refuse offers less than about £3, and knowing this, the offerer tends to offer an average of about £4.40, with a strong peak on the even split, £5.

One thing that interests me is that many of the commenters - even though they are on the blog of a leading popular psychologist - have evidently never heard of the ultimatum game, which is one of the most famous experiments in the whole discipline. It's refreshing to see people enthusiastically discussing something about which professionals are so jaded.

But what also surprises me is that not a single person comments on the other aspect of this experiment - it's a powerful test of priming.

Richard specifically and visibly pre-selects participants into male and female groups before having them make their offer. This is very different from asking your sex after the offer is already made. Why? Because asking the question can influence the offer you make.

An experiment by Steele and Ambady at  demonstrated that when students are asked to fill in a questionnaire about their sex before doing a test, they are more likely to conform to their perceived stereotype: women are better at the humanities, or worse at maths (there was a similar experiment by Shih, Pittinsky and Ambady (1999) with racial priming). The very action of classifying ourselves into a group which we believe has an effect on our performance, does have an effect on our performance. (This is my excuse for the picture of Danica McKellar, a mathematician who focuses on encouraging girls and young women to enjoy and become better at mathematics).

Thus, I speculate that the results of Richard's experiment will show a greater difference between men and women than if the experiment were done without priming. Women, I suspect, will make fairer offers, and men will make more selfish offers, because that's what we think we're meant to do.

Any thoughts, Richard? Is this what you're testing? Do you have non-primed statistics to compare to? Have I spoiled the whole experiment with this posting? I expect most of the results are probably in now, so I hope it's relatively safe.

No comments: