The Wason Selection Task
As Adil certainly knows, the world of psychology features many iconic experiments. We know of Pavlov, the Stanford Prison Experiment, the Milgram experiment, and countless others. But what about the experiment once described as “the single most investigated experimental paradigm in the psychology of reasoning?” Have you head of the Wason Selection Task?
Peter Wason was a famous English psychologist in the 1960s. He was a bit of a goofball as famous psychologists go — he specialized in the psychology of reasoning, but thought of the mind as an enigma that he could dissect both academically and playfully.
He also had a fascinating way of conducting experiments. Most academics at the time would read their contemporaries’ literature, find claims they deemed interesting or ripe to be disproved, and then craft experiments around confirming or disproving these hypotheses. Wason had a different mindset. He would tell colleagues that he’d happily read their work about a given subject, but only after Wason had himself conducted experiments in that space. His thought was that for something as abstract and inviting of confirmation biases as the psychology of reasoning, we should create observations first, and then test their fit with a hypothesis, rather than constructing a hypothesis and “cherry-picking” observations that fit with it. In fact, Wason was fond of saying “researchers should never really know why they are running a given experiment. They should figure that out after.”
Anyway, this guy is a little kooky, right? He also had a bit of a troll streak to him. In his obituary, two of his students said that the purpose of his experiments was never to test a hypothesis or theory, but rather produce an unexpected phenomenon of thinking or reasoning, and leave others the task of finding out how or why that happens. #coast
In 1966, this goofy experimental style created a groundbreaking study of deductive reasoning. He posed a puzzle to his subjects, which I will re-create below. I ask that for each puzzle, you stop yourself after 15 seconds of thinking and note what your answer is (or best guess at the time), and then continue to think about as long as you’d like (though Wason limited his participants to 1 minute of thinking). I also would love if y’all share your answers to each puzzle, both at 15 seconds and your final answer, as I’m curious about the results.
(Probably makes sense to wait to share your results until all four of you have read so that no one’s puzzling experience is compromised.)
The puzzles are as follows:
Puzzle #1
Each card shown here has a number on one side, and a patch of color on the other. If a card shows an even number on one face, then the opposite face is red. Which card(s) must you turn over in order to test the truth of the proposition, with no unnecessary cards being turned over?
Puzzle #2
Each card shown here has a number on one side, and a patch of color on the other. If a card shows blue on one face, then the opposite face is a perfect square (think exponents like 4 squared is 16, not shapes). Which card(s) must you turn over in order to test the truth of the proposition, with no unnecessary cards being turned over?
Puzzle #3
Each card shown here represents a person at a bar, with one face representing their drink and the other representing their age. The legal drinking age is 21. Which card(s) must you flip over to verify the bar is serving drinks in accordance with the law, with no unnecessary cards being turned over?
*Puzzles Complete*
I’ll place some fun, topical space-filling pictures here to put some distance between the puzzle and the answer so none of you accidentally see it prematurely. Scroll below to see the solutions.
Solution: Puzzle #1
The correct answer is that you would flip the 8 and brown card. The reason being that the 8 card must have a red face, and the brown card must not have an even number. If either is false, the proposition is false. The 3 and red cards are not necessary to flip — no matter what they have on the opposite side, the proposition could not be contradicted. The proposition says nothing about what a red faced card can have on the other side, nor does it say anything about what an odd number can have on the other side.
Solution: Puzzle #2
The correct answer is just the blue square. If this square does not have a perfect square on the other side, the proposition is wrong. All three other squares are unnecessary to flip, because no matter what is on the opposite face, the proposition could not be contradicted. The two numbers (324, 441) are both perfect squares, so they can either have blue on the other side and fit with the proposition, or have any other color and be logically irrelevant (like UNC). The green square is also irrelevant because it could have an perfect square or not and wouldn’t affect the proposition.
Solution: Puzzle #3
The correct answer is the card with a beer on it and the card with 19 years old on it. The person at the bar drinking coke could be above or below 21, and it doesn’t matter (it’s probably Adil). The 22 year-old can similarly drink whatever they want (except absinthe), and it would be legal.
Discussion
How’d y’all do? For Puzzle #1, empirical studies have shown that only 10–25% of people get it right. Puzzle #2… I made up. So I have no empirical evidence as to how people do, but I would imagine it has similar outcomes to Puzzle #1. Now, Puzzle #3 is the interesting one — 75–85% people get it right in controlled experimental settings.
Did you notice anything peculiar about the puzzles? Puzzle #1 is Wason’s original Selection Task, the one that earned him worldwide fame. Puzzle #2 is one I made up, namely to clear your minds before doing Puzzle #3. Puzzle #3 is a famous extension on Wason’s initial work. Any guesses why it’s famous? It’s actually, logically, the exact same problem as Puzzle #1.
What do I mean by that? All three puzzles take the form of “If P then Q.” To disprove that statement, you have to test the normal form (“If P then Q) and the contrapositive (If not Q, then not P). So in Puzzle 1, you have to test the normal form (if even, then red), so you flip the 8. You also test the contrapositive (if not red, then not even) and flip the brown card.
Puzzle #2 is a bit different because the contrapositive doesn’t exist in the set. To find that, you would have to have a non-perfect square, and verify that the opposite face isn’t blue. Both 324 and 441 are perfect squares, so you only have to test the normal form (flip the blue card to verify it has a perfect square on the other side). This puzzle was included as a decoy so the connection between 1 and 3 wasn’t as obvious.
Puzzle #3 is the exact same form as Puzzle #1. You test the normal form (if drinking alcohol, then age ≥ 21) by flipping the beer card, and test the contrapositive (if age <21, can’t be drinking beer). So the beer card is a perfect analog to the 8 in Puzzle #1, the 19 is an analog to the brown card, the coca-cola is an analog to the 3, and the 22 is an analog to the red card.
So these problems are logically the same, but empirically, they have wildly different success rates. This isn’t a function of the order in which the puzzles are solved either, as empirical studies only offer one of Puzzles 1 and 3 to random samples and compare those results. Wason’s original aim to create a phenomenon that invites others to dig in is perfectly demonstrated here — he made no effort to qualify his work or understand the underpinnings of what was going on. He left that to others.
There are essentially two schools of thought as to what is going on such that Puzzle #1 (especially under the time constraint) is difficult, while Puzzle #3 is easy. The first comes from Nobel Prize-winning psychologist and economist, Daniel Kahneman. In his 2011 bestseller Thinking Fast and Slow, he describes the “dual-process theory” of decision-making. Kahneman’s view is that we all have two cognitive systems that process problems and decisions. The first, called System 1 or the “old system,” moves quickly and automatically. It’s behind what we would call “snap judgments.” Kahneman gives examples of typical tasks that System 1 accomplishes as localizing the source of a specific sound, doing simple math like 2+2, driving a car on an empty road, or judging the relative distances of two objects you see. The second, System 2 or the “new system”, is deliberate, slow, and conscious. It deals with reasoning for things like combing through your memory to place a song you’ve heard in a movie before, parking in a tight parking space, or thinking of the right chess move.
Whenever we are faced with a problem, these two systems compete to make the decision or process the thought. Typically, when constrained to 15 seconds, System 1 wins out, and your answer comes from there. System 1 is susceptible to things like “matching bias,” where someone would likely see Puzzle 1, register that even numbers and red faces were named in the problem, and cut corners to arrive at an answer of 8 and red (instead of 8 and brown), which happens to be the most common mistake.
But in Puzzle 3, the way the problem is framed is much more familiar to us. It’s presented in the context of a law against underage drinking that we all know, and therefore System 1 races to the intuitive conclusion that you’d check the person drinking beer’s age and the person under 21’s drink. If you noticed that Puzzle 3 felt much cleaner and swifter (even if you got the right answer to both), Kahneman would say that is because the mind gravitates towards System 1 when it feels familiar enough to reliably depend on it. In Puzzle 1, on the other hand, System 1 was off to the races, but because it’s an unfamiliar, abstract, and meaningless context, your mind is unsure if that’s really the right way to solve this problem, and the whole process feels more pained.
In some sections of the GMAT, for example, I’ve noticed that there are questions that prey on this exact concept. I often found that the problems that I got wrong were ones that I spent little time on. For most of these, I remember the feeling of thinking at the time of solving that I had nailed that problem and it was way easy, only to realize later that I had missed some key, tricky bit of information. Purely anecdotal (though I wonder if Dave had similar experiences), but, to me, it shows our vulnerability to System 1 at times.
Not everyone agrees with Kahneman’s take exactly, though. Many evolutionary psychologists, in particular, disagree with this premise. The most well-known critic is a woman named Leda Cosmides, a noted evolutionary psychologist. Her view is that the human brain’s ability to quickly and decisively apply logic is highly dependent on the context. In Puzzle 1, we’re presented with a total abstraction, or a “pure” logic game scenario. We have no concept of why or how even numbers and the color red are linked together. On the other hand, Puzzle 3 presents the problem in a sociologically intuitive context. Cosmides would describe it as a type of “social contract” — we know we’re not supposed to drink under the age of 21. And her evolutionary psychology perspective would say that humans have not developed logical or deductive reasoning capabilities over time in a pure logical vacuum, but rather in “socio-cognitive niches.” This theory argues that, evolutionarily, it was advantageous for humans to excel at reasoning as it related to social norms, much more so than excelling at reasoning in a vacuum on abstract logical problems. One of these niches Cosmides describes is called the “cheater detection module.” Cosmides would say that humans are particularly successful at logical reasoning that deals with figuring out if others are breaking rules or norms, and that Puzzles 1 and 3 show this. A logical problem when framed abstractly is difficult and time-consuming (Puzzle 1). But when framed in a way that draws on our “cheater detection module” of cognition, it becomes routine and easy (Puzzle 3). Cosmides and her colleagues have used this theory of socio-cognitive niches to test human decision-making in various settings, and their thinking has seemed to stand up to nearly 30 years of work since her original study in 1989.
All in all, it’s hard to say who’s right, or whether it’s a combination of the two. One thing is clear, however — Wason’s goal of rocking the boat and letting others deal with the waves has been accomplished. After he retired in the 1980s, he vowed to never weigh in on the fierce debates between the evolutionary psychologists and “dual-process theory” acolytes. Some people just want to let the world (of psychology) burn, I guess.
Sources:
