 Originally Posted by ataraxis
err, you obviously dont get what I'm saying here. Say that Wendylove gave a 1 question test, the question was
"How many states are there in the US"
A) 49
B) 50
C) 124
Of course, SHE would say that its 49, because Texas isn't a state, of course.
So by the logic you are presenting, if 99 people took the test, 33 would get a 100, the rest would fail. Now, this is obviosuly untrue. Most people would say B, 50, so they would get it wrong.
So people can easily get a 0. If they are guessing, as you said, there is a very small chance of them getting a 0. But most people would choose, but choose wrong.
Basically, Xie, probability like that only applies when each answer has an equal chance of being picked. In your "random guessing" example, yes, each answer has an equal chance. However, when there is a universal inclination to pick one answer, (for example, everyone 'knows' that there are 50 states, but in this example test that is the wrong answer), then the balance is greatly shifted and your estimated probability of getting a 0 is 'completely wrong'
That was my whole point: you're just as likely to go for A, B, or C in the test which we're talking about (not the one you just made up, which contains a trick question and is therefore not representative at all), if you don't know the answer, for the vast majority of the questions, because all of the answers were completely reasonable.
Also, I notice you kind of ignored the fact that if some questions are biased then some would be biased towards the right answer and others towards the wrong answer. Therefore you couldn't say there would be a great shift in any direction in particular, so therefore all one can do is ignore this and take the average, which is the figure I gave in the first place.
|
|
Bookmarks