Greg and Paul,
You can take a measure of comfort in knowing you both apparently beat the pants off Pinnacle Point.
Jim,
Thank you for the effort. I appreciate the number crunching. I don't know anything about statistical models like probit -- does it assume a normal distribution? Given we all agree the contributors have biases, my guess is we don't have a normal distribution. Does it assume normal distribution of error? Given that we could see many contributors' votes while the ballots were open, and could "adjust" our voting accordingly, I'm not sure we have normal distro of error, either.
Enough of that.
Would it be possible to just post the average number of "stars" each course received? If you want, you could round the numbers so we got four simple groupings of ***, **, *, and 0 -- you could set the roundings relatively high to err on the side of conservatism:
Average stars
>= 2.75 ***
>= 1.75 **
>= 0.75 *
< 0.75 0
Frankly, I find ordinal rankings frustrating and almost "number-free" -- all they do is provide qualitative comparisons, eg this course is better than than that course or courses.
Mark