So You Think You’re Logical
In case anyone wants to find out about the Wason test along with PM, here it is in one easy click.
[Note by Jerry S (Sorry OB, I’m invading your entry!)]: I programmed this four years ago; I’d do it slightly differently if I was programming it today – there are a couple of problems with it. However, it is a pretty rigorous experimental design (on the analysis page, there’s a link with technical details about the ‘between-subjects’ and ‘within-subjects’ aspects of the design). And the results, right at the end, are interesting.
Actually all these questions involve the same kind rule in logic. That is:
If A then B implies the following:
1) If A then B
2) If not A then (B or not B)
3) If B then (A or not A)
4) If not B then not A
cases 1 and 4 tell you definitely if the the rule is broken.
Yup, that’s the point. It’s a basic IF rule, not an IFF rule.
What’s really interesting is how poorly people do, even in a very familiar context – although I guess you have problems with click through (could have done with an easy control question I reckon). And when you look at cheating contexts that are alien (like the mushroom one) people are pretty awful.
Now I’m happy to put that down to evolutionary reasons – probably don’t often need to evaluate if-then rules in this artificial way, particularly since we seem to be much more inductive than deductive as a species. Although I’ve seen some evidence that people just don’t interpret the rules and tasks in the way you want them to, ‘no, look, I said if-then, not if-and-only-if damn it! And I don’t care if you want to look at three cards!’
PM
I noticed in the other thread that you suggested that the explanation for our doing relatively well with the cheating context might be social rather than evolutionary (due to the fact that cheating is where we’re most likely to come across these kinds of rules).
I’d need to be persuaded. Partly because there is evidence that being exposed to this kind of reasoning in a formal way doesn’t result in our performance increasing (I can’t find the evidence now, but I remember reading – I think in a book by Helena Cronin – that philosophy graduates don’t do any better at the Wason than a control group).
But also because of the phenomenology of the response to the cheating context. There is a “pop-out” experience – we just know the right answer – which it is implausible to suggest is socially conditioned. (Of course, that won’t convince you, but it is something which is repeatedly found; and also people find its opposite, a resistance to accepting that the formal questions have been answered incorrectly).
Click through. Yup, that is a problem. I thought about getting people to answer an additional question about the scenario just to demonstrate that they had read it. But I figured that people might deliberately answer incorrectly, if they felt they were being checked like this.
Having said that, these results aren’t out of line with the results which psychologists find generally. And also, of course, people get the cheating context questions right.
The other thing to bear in mind is that the statistics are generated only taking the first set of people’s results into account (to avoid order-effects), so boredom is less likely to have set-in at that point!
That’s very interesting about the phenomenology and the “pop-out” experience. I know absolutely nothing about all this stuff. I don’t think I have that pop-out experience very often, if ever. It must be like playing 20 Questions and somehow just knowing what the other person is thinking, without even one question. Would feel quite strange.
I took the little quiz and I didn’t see anything different between the questions. I thought the answers were obvious, which caused me to second guess as to whether there wasn’t some trick being played. Such anxiety turned out to be wrong, as all the answers were the same and I got them all right. But doesn’t that suggest that there is an converse bias, insofar as the imposition of a test provokes the anxiety that one is being cheated?
Why do you oppose the biological to the social in your above comments? Consider that human beings did not develop a language and thereby become sociable. Language emerged out of natural evolution only on the basis of a long prior evolutionary history of the intensification of sociality and its increased requirements of analog/relational communication, without which language would not have been possible. This intensification of social behavior began with pre-simian primates.
Given that this was devised in 1966, it would be interesting to see whether the rate of correct responses has grown as it has become increasingly the norm for students (who are usually the lab rats for this sort of thing) to have some experience of elementary computer programming. The hypothesis would be that as formally logical activities become increasingly commonplace in the population, the number of respondents who (like Halasz and me) find the test a little facile should increase as well.