What I have done so far:
If the student was expected to make a 75%, we basically expected he would get 15 of the 20 questions correct. Now that he has misbubbled his answers, we expect those 15 he should've gotten right to now be wrong. For the remaining questions he now has a 1/10 chance of getting each correct (since the scantron has 10 choices). So the expected value of his score would be 5/10. This just intuitively feels wrong to me. His expected number of questions to get right was 15, but obviously there is some variance in there, I'm basically working it that he was for sure going to get those 15 right...any help/hints/suggestions are very appreciated.