Thanks for all the help so far, we are looking pretty much set to go. Any further comments, ideas, etc. are still very much appreciated. Especially in the odds department. What would be the equivalent of a guesswork in this case: 10 guesses, two separate trials?
I'm not quite sure what you mean by, "the equivalent of guesswork." The subject could, potentially, achieve any result by guesswork, some results are simply less likely than others.
Score >=
|
Probability 1 trial
|
Probability both trials
5 |0.6230 |0.3882
6 |0.3770 |0.1421
7 |0.1719 |0.0295
8 |0.0547 |0.0030
9 |0.0107 |0.0001
10 |0.000977|0.000000954
So the odds of the subject guessing correctly 10 times out of 10 in a given trial are about 1 in 1000
The odds of the subject scoring
at least 9 out of 10 in
both trails are about 1 in 10000.
The odds of the subject scoring
at least 8 out of 10 in
both trials are about 1 in 300.
I would suggest that you set the pass mark for the trail as being
at least 8 out of 10 for
both trials. This leaves some room for error and does not allow for an excuse along the lines of "I was under too much stress to be 100% accurate."
If she scores 8 out of 10 in one of the trails but fails the other you should be sure to point out that the odds of scoring at least 8 out of 10 in a single trial are about 1 in 20 and so, not particularly impressive.
I have a spreadsheet for calculating these odds but, if you google "binomial calculator" there are quite a few java apps on university websites which will allow you to double-check my calculations.