I recall writing multiple-choice tests as professor and penalizing students for wrong guesses. If they didn't know the answer or couldn't narrow down the choices, they should just leave the answer blank. Students didn't like it.
Richard Hart of Nine-Patch Multiple-Choice, Inc. has turned the idea around. You get no points for a wrong answer, and you get points for not guessing — just not as many points as you'd get for the correct answer.
Here's his description of the process from his web site.
- Read the question and see if you can use it to report what you know or can do.
- If yes, then compare the answer you have in mind with the printed answers.
- If you find a match, you are probably right. Mark it.
- If there is no match, you may want to omit, to avoid making a wrong mark.
- You get one point for a right mark and one point for not making a wrong mark.”
At this point, you may think that all that's happened is a shift of grading emphasis. Look again. The idea is that students must report their self-confidence by marking only what they know and admitting what they don't know by leaving those answers blank.
You could even expand that concept in the multiple-choice domain by allowing students to mark more than one answer. Suppose a student has eliminated two of the four answers and still can't decide. If the student marks both answers, then that student in communicating more information to the instructor. If one of those answers is correct, then the student gets some credit.
Imagine a multiple-choice exam where all questions have all answers marked before the students begins to answer the questions. (This is not Richard Hart's approach and only a hypothetical extension.) The students' task is elimination of incorrect answers. Every incorrect answer indicated adds to the student score. Erroneous marks result in zero points. A 25-question quiz with four choices per question would have a maximum of 75 points. Making the entire quiz blank would give a zero because every right answer would have been marked as wrong. Leaving the quiz with all answers marked would give the student 25 points or 33.3% of the maximum possible score because all correct answers are marked correct.
Now, imagine a multiple-choice test where questions may (or may not) have more than one correct choice. Suppose that every question on a 25-question quiz has all four answers correct. Then, the students begin unknowingly with a score of 100. Every answer that they mark as being incorrect reduces that number. If, on average, the number of correct answer per question is two, then the students begin (again unknowingly) with a score of 50. Taking the marks off of all incorrect answers results in a maximum score of 100. In this scheme, every choice of every question counts.
Ideas like those of Richard Hart and the extensions that I've suggested may seem very small in the overall scheme of building better education for our students. Everything is important in learning. Details count. Every positive innovation is a step forward. A version of this sort of multiple-choice scoring is the basis of a company with a patent on its "Confidence-Based Learning." Their market is corporate training and seems to be paying well if their website is any indication.
Even small ideas can have big outcomes.
© 2008 by Smart Science Education Inc., U.S.A. www.smartscience.netFollow this author on ETC Journal.
5 comments:
The biggest problem I see in my classes is that many (probably most) students have never really developed critical thinking skills and don't have much comprehension of why quiz answers are right or wrong, or how the facts fit together in the big picture. I like your idea of starting with all the answers marked right and then eliminating the wrong ones. The end result is the same, but by approaching it differently it forces the student to think about it in a different way.
Rosemary, I tried this once. Students had to mark out all the wrong answers and then record the remaining answer on the answer sheet. Out of a 120 less then 10 did this. It was "too much work". To them multiple-choice was a game to be played with a #2 pencil they brought to the test. They had been trained to "look for the right answer" rather than to use the question to report something they brought to the test that they actually knew or could do. Or as one student put it, "I don't memorize wrong answers".
It is only after a couple of tests that students release they will do better if they learn some topics to the point of understanding rather than rote memorize bits of nonsense over the entire assignment. This begins the transition from passive pupil to self-correcting scholar. Then they like and understand why they get rewarded for not making wrong marks.
Ah, the other problem! Most of my problems in getting innovation into school centers around reluctant teachers. They say they're too busy, etc.
Some students also find that thinking is hard work and resist it. They're used to memorization as a substitute for thinking and these particular students, who do well on memory-oriented tests, resist any change.
Someday, maybe in college or in work, they'll hit a situation where memory doesn't cut it. It'll be too late then.
Changing the way multiple-choice tests are scored is a small step. Memory still can succeed. However, any step is progress.
I'd like to add that the deleting of wrong answers appeals to me as a scientist. I'm used to considering a range of explanations for a particular behavior and eliminating those that I can through literature review and experimentation.
Obviously, I can't have memorized an answer that no one knows. I could go with my gut, but that way doesn't work for scientists. (It doesn't seem to work too well in government either.)
So, the good old science approach is what we scientists must do. Despite student complaints, it wouldn't hurt them to be confronted with situations requiring the same sort of logic.
Post a Comment