| I
would propose that the algorithm for counting incorrect attempts be
changed. Currently on a multipart question, if a student say chooses to
answer only the first part of the question and leaves the other parts
blank, that is counted as an incorrect attempt. I think this leads to
misleading statistics about problem difficulty, particularly on the
types of questions I like to write, where the sequence of questions is
supposed to walk the student through a solution. Not to mention the
more egregious case of a conditional problem, where the second part of
the question doesn't even appear until the first part is answered
correctly.
So would propose that an attempt be only counted as incorrect when a
student enters some nonblank answer to some question which is
incorrect. Or alternatively that there should be some flag variable
which the problem author could use to control this behavior, eg.
blankIncorrect = 0;
Zig Fiedorowicz
<| Post or View Comments |>
|