Bluffing on Exams

I came across an interesting article, published in the New York Times on June 11, 1950, that discusses a series of experiments examining how likely it is that college students will bluff their way through exams. For instance, when Professor Samuel Fernberger, of the University of Pennsylvania, gave his students their final exam, in one of the questions he asked them to define "psychoterminality." It was a meaningless term, but the students didn't know that. According to the NY Times:

Only two students honestly stated they did not know what the term meant. Six left the question blank. But the other twenty-one handed in expositions, ranging from one-half to three pages long, in which they solemnly described it as, among other things, "automatism," "vitalism," "hypnosis" and the "behavior of the lower animals." It was astonishing because, of course, Dr. Fernberger had just coined this mythical word for the occasion.

Professors Ernst F. Thelin and Paul C. Scott of the University of Cincinnati conducted the most thorough investigation of bluffing. They gave 147 college students a test that included numerous trick questions. For instance, they asked the students to indicate the authors of nonexistent books or to define made-up words:

Bluffing was defined by the investigators as "pretending to have greater knowledge than is actually possessed." Some bluffing was done by all students, varying from 5 to 81 per cent. Freshmen bluffed most; seniors least. The average bluffing score of the men (45.8 per cent) was slightly higher than that of the women (43.4 per cent).

Finally the article refers to a study that examined other members of society. An investigator visited bakery shops and asked for "scroofles":

Instead of saying they'd never heard of this mythical product... a surprising number of bakers bluffed they were just out of 'scroofles,' or were not baking 'scroofles' currently because of the lack of demand.

My hunch is that all the figures for the prevalence of bluffing would be even higher today than they were in 1950. But today we'd be more likely to call it bullshitting than bluffing.


Posted on Thu Nov 29, 2007


The last example may not have been bluffing. Maybe there really was a bakery product called scroofles, or something that sounded similar, but it wasn't very popular and the investigator just hadn't heard of it. Or maybe the bakers realized they were being mocked and went along with the joke.
Posted by Big Gary  on  Thu Nov 29, 2007  at  03:47 PM
Or maybe the bakers had been trained to say such things when they didn't have the item (made up or not) so that the customer wasn't offended and didn't come back.
Posted by Christopher Cole  on  Thu Nov 29, 2007  at  04:54 PM
If you come across an exam question that asks for a definition of a term that you have never heard before you might as well take a shot in the dark. What have you got to lose? Maybe the experiment would be better with a type of test where you lose additional marks for giving the wrong answer. I bet most people would not answer if they really didn't know what the term meant.
Posted by Captain Al  on  Thu Nov 29, 2007  at  05:07 PM
I think psychoterminality is a perfectly cromulent word.
Posted by K. Eberhardt  on  Thu Nov 29, 2007  at  05:21 PM
When I attended Fordham University back in the '70's, the final exam for my film course (taught by a Jesuit priest) included the question, "What is a Sacher torte?"

The looks on the faces of my fellow students were priceless as they wracked their brains, trying to recall any reference to Sacher tortes.

I knew that the priest was something of a joker, so I didn't panic and wrote "From the use of the word 'torte,' I'm guessing it's a kind of pastry."

I was correct, but I suspect the question caused much consternation around the room.
Posted by Cranky Media Guy  on  Thu Nov 29, 2007  at  06:24 PM
Captain Al said:

"Maybe the experiment would be better with a type of test where you lose additional marks for giving the wrong answer."

The problem with that is that announcing it would tip off the test-takers that there was going to be a trick question on the test. That would cause them to be on their guard, meaning they wouldn't attempt to bluff an answer, screwing up the experiment.
Posted by Cranky Media Guy  on  Thu Nov 29, 2007  at  06:26 PM
When I was studying for the bar exam, we were specifically told to bluff.

"If you don't know the answer, write something that sounds like a lawyer."

"If you don't know the rule, make one up."

"If you're feeling really ambitious, make an exception to your own rule."

Now that I think about it... This probably isn't the best advertisement for new lawyers...
Posted by Kevin  on  Thu Nov 29, 2007  at  06:41 PM
So who got the higher marks, those who bluffed or those who admitted they didn't know? I'd really like to know how he scored that.
Posted by Nona  on  Fri Nov 30, 2007  at  05:23 AM
The validity of the "test" studies is questionable. Student are routinely asked to try to answer questions even when they are not sure of the answer. As Capt. Al said, why not take a shot in the dark. Indeed, several methods for learning how to read rely on students parsing the word to gain meaning from the pieces that may illuminate the whole. Anyone fluent in Latin can attest to the language's utility in determining the underlying meaning of English text.

As to the bakery example, a "good" business practice I've often heard is to never tell the customer they can't have something if you can possibly get it for them. While I would personally try to get more information, I can understand why others may take the shorter route in the example.

To truly measure "bluffing" as defined here, you have to remove other reasonable incentives. None of these studies did that.
Posted by Locke  on  Fri Nov 30, 2007  at  10:05 AM
A "Sachertorte" is a real kind of cake.
Which raises the central problem with these experiments, i.e., is guessing the same as bluffing? We probably wouldn't call right guesses bluffs, we'd call them lucky guesses or something similar (or perhaps just "knowing the answer"). So why is a wrong guess necessarily a bluff?
Posted by Big Gary  on  Fri Nov 30, 2007  at  10:56 AM
I vividly remember my high school teachers always suggesting we guess on answers because, in the case of multiple choice questions with four possible answers, we 'have a 25% chance of guessing correctly' and that 'if you leave it blank, it is definitely wrong' and 'if you don't even try, you've failed'.

With that kind of training, I would indeed fail a phony, badly designed 'test experiment' such as is described in this article.
Posted by OriginalSim  on  Fri Nov 30, 2007  at  11:33 AM
Let me see if I've got this right: There's a question on a test. It involves a made-up word. I, the test taker, get points for every correct answer. My goal here is to pass the test with the maximum number of points...

So, why wouldn't I guess at an answer? It would be silly not to. What have I got to lose?

In order to test bluffing in a way that means something, the test would either have to penalize you for incorrect answers or it would have to have no penalities whether you answer the questions correctly or not. There surely is a way to find out what percentage of people will admit "Gee, I don't know" and what percentage would rather bluff, but this ain't it.
Posted by Kathleen  on  Fri Nov 30, 2007  at  03:58 PM
Big Gary said:

"A 'Sachertorte' is a real kind of cake."

It sure is. I wasn't exactly sure of that when I wrote my answer (although "torte" was kind of a hint) so I guess I was bluffing in a way, huh?
Posted by Cranky Media Guy  on  Sat Dec 01, 2007  at  03:44 AM
Yeah, I also disagree with calling making a guess on an exam a "bluff".

I had a (really bad) paleoanthropology professor who invented a word on the first day, and declared that our goal as students was to decide on what that word meant. Our mid-term exam was simply to write an essay on what we'd decided the term should mean, and how we justified that meaning.

He failed every person in his class (aside from me for some reason, to whom he gave the lowest possible passing grade) and said that we all got the meaning wrong, that we obviously weren't serious about learning, that he couldn't see why we just couldn't understand, and so on.

I dropped that class that afternoon.
Posted by Accipiter  on  Sat Dec 01, 2007  at  03:11 PM
>>why wouldn't I guess at an answer? It would be silly not to. What have I got to lose?<<

The test takers had nothing to lose, so it was logical for them to try to come up with an answer. But I think bluffing is a fair term for what they did. If they had simply selected randomly from a list of multiple-choice answers, that would be guessing. But they actually invented answers out of whole cloth -- answers they had to know were wrong. That seems like bluffing to me. They were pretending to know something they didn't.
Posted by The Curator  in  San Diego  on  Tue Dec 04, 2007  at  12:36 AM
I'm not sure I can articulate the difference between bluffing, guessing, making an educated guess (which is what it would be if you knew enough about prefixes and suffixes and Latin roots and so on), faking it, and so on. There's just not enough difference between the meaning of those words, and anyway, when you come right down to it, context is everything. So I can't say with any degree of certainty that "bluff" is the wrong word here.

But what I was really trying to say is that I just don't think the test, as described, is very meaningful. I, at least, don't see the value in discovering how many people sometimes make a guess/bluff/whatever when they are taking a test. That's pretty much all of us, after all.
Posted by Kathleen  on  Tue Dec 04, 2007  at  08:12 AM
Posted by Unfairly Balanced  on  Wed Dec 05, 2007  at  03:31 AM
Commenting is not available in this channel entry.