The Wason Selection Task, Part III

I stated in Part II that evidence suggests that people do well on conditional reasoning tasks but not on the Wason Selection Task.

For example, in response to Margolis's suggestion that subjects perform poorly on the Wason Selection Task (WST) because they misinterpret the rule, Prudkov writes

The formal wording of the task is unusual but very clear; most people in an industrial society are able to solve similar problems and are familiar with the reuisite [sic] logical rules. Nevertheless, it is very difficult to find the right solution to the task (Johnson-Laird & Wason 1977). . . . Why do the subjects misinterpret the task if their knowledge and skills allow the correct interpretation to be constructed? (Margolis's version confirms this.)

And Handley (PDF) cites these sources in describing the contrast in performance between other conditional reasoning tasks and the WST:

There is an apparent anomaly in the reasoning literature: In relation to people’s performance on simple conditional inference tasks, performance on abstract versions of Wason’s selection task is particularly poor when tested against the criteria of formal logic (for a review, see Evans, Newstead, & Byrne, 1993; Manktelow, 1999).

Now, of course, what these two writers are describing is only the relative performance differences between participants in the WST and those in other conditional reasoning tasks. It is not at all evident that the general population performs perfectly with any kind of logical reasoning, much less conditional reasoning. Inglis and Simpson (PDF), for example, write

It has been found that many people respond in an apparently irrational, non-normative fashion when given straightforward logical reasoning tasks. For example, experimenters have found that people are much more likely to endorse logical arguments as valid if the conclusions are believable. Conversely, it is much harder to correctly evaluate logically valid arguments when the conclusion is unbelievable (Evans, Barston & Pollard, 1983).

And Thompson and Byrne (PDF) specifically single out simple conditional reasoning as being problematic as well:

For the modus ponens inference (MP), reasoners are given the true antecedent (TA), Sarah went to Moose Jaw, and they are asked to judge the validity of the true consequent (TC), Tom went to Medicine Hat. For the modus tollens inference (MT), reasoners are given the false consequent (FC), Tom didn’t go to Medicine Hat, and are asked to judge the validity of the false antecedent (FA), Sarah didn’t go to Moose Jaw. These two inferences are valid, regardless of whether one interprets the conditional as an implication relation (the antecedent is sufficient but not necessary for the consequent) or an equivalence biconditional relation (the antecedent is sufficient and necessary for the consequent). The findings indicate that reasoners tend to make the MP inference readily but that the MT inference is more difficult and many individuals conclude erroneously that nothing follows (see Evans et al., 1993).

Yet, even though subjects don't always perform optimally on any kind of logical reasoning task, they perform extraordinarily poorly on the WST. Why is this the case?

There are a number of explanations. One of the simplest is something known as the matching bias. The matching bias is understood as the tendency to choose items in the WST that match the names given in the rule. Thus, since the rule given in the formal task I presented in Parts I and II reads "every card that has a D on one side has a 3 on the other," participants are more likely to choose the D and 3 cards. Similarly, if I changed the rule in the "drinking" version of the task from "every person that has an alcoholic drink is of legal age (21)" to "every person that is drinking vodka is 29," participants, falling prey to the matching bias, would be more likely to choose the Vodka and 29 cards.

Another explanation is something known as the confirmation bias (wiki). In essence, this bias, as applied to the WST, would lead subjects to look for ways to confirm the rule, rather than finding ways in which the rule could be violated. Thus, in the formal task, subjects choose the D card to hypothetically confirm that a 3 is on the other side and the 3 card to see whether or not there is a D on the other side.

And Margolis, in his "misinterpretation" explanation, provides the juicy quote that will take us into the next post. I'm putting the funny stuff in red:

The cognitive illusion comes at the stage of interpreting the task, not from the inability to handle modus tollens that is the usual explanation. That claimed inability has always warranted more suspicion than it has received, since anyone who listens to their children will hear them quite readily make what are functional equivalents of modus tollens inferences. And not very surprisingly, since the world provides us with endless occasions to make such inferences. (If I picked my keys off the desk, they would now be in my pocket. My keys are not in my pocket. So they are probably on my desk.)

Wason Task: Part I | Part II | Part III | Part IV | Part V | Part VI