On Sun, 05 Nov 2000 11:41:39 -0500 (EST) [EMAIL PROTECTED] wrote:
>
> 25 item multiple choice section 1 for correct; 0 for blank; ¼point deduction
> for wrong answer.
>
> (All completed) 15Correct; 10 incorrect ( 2.5 points deducted) 15 points
> minus 2.5 = 12.5 raw score.
>
> (All not completed) 15 correct, 5 guessed wrong. 5 left blank…15 points,
> minus 1.25 deducted = 13.75 raw score
>
Here is a second problem with your analysis. Your presumption
is that these two scores represent equivalent states of
knowledge. This boils down to the assumption that amount of
knowledge is equal to the number of right answers *alone*!
The point of a "correction" factor is that a score of 15
represents a mixture of correct answers + lucky guesses. The
number of lucky guesses depends on the number of answers you
make. Hence, someone with a score of 15 right and 10 wrong is
presumed to know less than someone with a score of 15 correct
and 10 blanks.
A correction factor uses the number of wrong answers to estimate
the number of times the correct answer was chosen by chance
alone. This is actually a generous rule to the test-taker
because usually at least one answer can be eliminated and hence
the actual chances of getting an item correct by guessing is
higher.
Ken
----------------------
Kenneth M. Steele [EMAIL PROTECTED]
Appalachian State University
Boone, NC 28608
USA
