Everyone knows that basic type of test, where they give you a question and give four diffrent answers to pick from. Most people should know the answer and pick the correct one… But in those cases where you have NO idea, I’ve always heard to pick C.
My Question is, Has there ever been a study of any kind about which letter most often coresponds to the correct answer?
I am not aware of any studies, but as a teacher, I have found that if I am not careful, that works on my tests. The idea behind is that the teacher will usually give a couple of wrong answers, then put in the correct answer before they forget to, then finish with another wrong.
On advanced tests like the AP tests (which you shouldn’t guess on anyway) the best bet seems to be “go for the longer answer.”
The reliability of such a strategy really depends on the skill of the test designer as well as the specifics of such a test. Even public school teachers these days have access to test design software that can randomize answer choices making a “Pick C” strategy much less effective than it may have been in the past. The large standardized tests like the SAT and GRE are designed to make such a strategy less effective as well.
OTOH, there is a general human tendency to stay away from the extremes. Teachers generally have a bias against making the correct answer A or the last letter choice as much as chance would dictate. The ones in the middle should get the most play and that is the basic idea behind the advice. Another factor is that the last answer choice is often “all of the above” or something similar and that response requires all of the the preceding answers to be correct which isn’t usually correct so the advice, again, is to play the middle when you have no idea at all.
I swear I have seen some studies with actual percentages but I can’t find them now. However, they won’t be consistent across all types of multiple-choice tests so a blanket statement is meaningless.
Of course, even if the test was made with truly random test-creation software (or if the teacher rolled a die to place the answers, or whatever), picking C is no worse than any other answer, and for most tests, definitely better than just leaving it blank. I remember when I was taking my ACTs (which don’t, or at least didn’t, penalize for wrong guesses), on one section the proctor announced “One minute left”, and I still had a fair number left, so I just went down one column. I presume I got a few points at least from that.
When I was a beginner in the 1980s the consensus was “pick B” as some study (never named) at the time had shown that B had a tiny advantage over the other four letters. This gave us comfort in the situations in which you still had twenty-seven questions out of five hundred left to answer and the proctor droned, “You have five more minutes.” B, B, B, B, B, B, B…
The longest answer was often also the right answer. I know it was when I first started making up test questions, as an inexperienced test writer.
But. Once I got to higher level testing, a whole lot of phenomena occurred that invalidated that. Not just randomizing. Other test question writers would go over your “stem” and answers and look for weaknesses. If your question made it to the medical students, statistics were run on how they did to see if everyone got it right, no one got it right (both kinds of question were toseed from the next exam), or just some people got it right. If the people who were the top scorers on the exam got it right but no one else did, it was said to have a lot of discriminating power, and it was kept in if they wanted a highly discriminating exam. If the people in the top half or third usually got it right and those in the bottom half or third usually got it wrong, it would stay in forever; that’s the kind of question they wanted, the “Did you pay attention in class?” question.
Now that I make up CME credit exam questions, there’s a new wrinkie. The five answers have to go in alphabetical order by first letter of the answer. Now that’s an effective randomizer without requring any math. Suppose I say, “Which of these frequently posting Straight Dope members is an SDSAB? Twickster, Hal, Monty, Copperwindow, Eve,” then C is right. Alphabetize them and suddenly it’s “Copperwindow, Eve, Hal, Monty, Twickster,” and D is right. It’s next to impossible for the test question writer not to achieve randomness when she thinks of the answers by fiendishness qualities, and they get ordered by alphabet.
So the real question to you is, how tough is the exam you are taking? If it’s a friendly little thing one teacher has developed for his/her own class, B or C may be right if you don’t know what else to do. If it’s higher level, give it up, man. You have no hope of a better than 20% chance.
On a non-K-type question, that is.
And with real, real bad randomizers, what you have to do is make sure you have an equal amount of each letter.
I have taken tests designed like that…
When taking advancement exams in the military, this strategy is known as “When in doubt, Charlie out”
‘Charlie’ being phonetic alphabet-ese for ‘C’