Who’s Smarter?

April 17, 2016
Posted by Jay Livingston

I was not optimistic when I tried a version of the old Quizmaster-Contestant-Observer game* that illustrates the fundamental attribution error. This is the error that occurs when we explain someone else’s behavior mostly in terms of their personal traits and ignore the effects of the situation. My attempts to replicate well-known effects usually fail. My students and I are just bad replicators I guess. But this time it worked.

Here’s the set-up: I asked students to come to class with seven trivia questions – challenging but not impossible, the kind that even when you can’t think of the answer, and then someone tells you, you say, “Oh, of course.” I gave them some samples and added, “This is not a graded assignment. You can get help from your friends. You can use the internet (I did).”

In class, I divided them into groups of three and told them to choose** roles. They could be the Questioner, the Contestant, or the Observer. Then the Questioner asked the seven questions, the Contestant tried to answer and was told if the answer was right or wrong, the Observer observed. They then returned to their regular seats and got this form.

We assume that most Contestants will get a few wrong. We also can be sure that the Questioner knows all the answers. But what are we to conclude from that? That the Questioner knows more than the Contestant? Of course not. It’s easy to ace the quiz if you’re the one who makes up the questions. If the roles had been reversed, if the Contestant had been the one to ask the questions she thought up, she would be the one with more answers.

But we just cannot resist the temptation to draw conclusions about the persons themselves and to ignore the advantages and disadvantages of the positions. The chart shows how participants in the different roles rated the Questioners and Contestants.

As attribution theory predicts, Observers were quick to make judgments about the relative knowledge of the Questioner and Contestant. They ignored the role differences and concluded that if the Questioner knew more answers, that’s who must know more trivia.

Attribution theory also says that when we look at our own behavior, we are more likely to make “situational attributions.” Accordingly, Questioners and Contestants may have taken the constraints of role into consideration. In any case, they did not see so large a difference, though even they could not escape the conclusion that the Questioner was better informed. Contestants were impressed by the knowledge that the Questioner supposedly possessed, even though they knew that the Questioner could have gotten the questions from the Internet.

Sometimes the fundamental attribution error extends beyond what might be suggested by the specific tasks – in this case, knowledge in a trivia game – and gets to more permanent qualities. We see someone trip and stumble. Why? Because he’s clumsy, we say. We have assigned him a more or less permanent trait on the basis of one brief event. And similarly, if we see a rigged trivia game where the person who wrote the questions knows all the answers . . .

---------------------------------------

* See Ross, Amabilie, and Steinmetz,1977 (here).

** The roles should be assigned randomly. That was not the only way in which my experimental design was deeply flawed, and I report the results here not because they are convincing, but merely because I was so pleased that in at least one way they turned out the way the theory says they should.