First of all thank you for the many emails and instant messages that encouraged me to pursue this. While it is depressing that this situation has arisen, I am encouraged by the fact that so many of you are concerned about it and think that it is wrong.
Unfortunately the issue has not resolved itself. The parties did not come forward. So after further attempts to figure out how this could possibly have happened at random failed, I decided to fall back on statistics.
This afternoon we entered all of the multiple choice answers for everyone who took the same form as the questionable tests (and gee that was fun!--NOT). Once this was done, a correlation matrix was calculated. Because the scores are not normally distributed I will report the median rather than the mean. The median correlation was about a .27 with a standard error of about .07. The highest? Above a .96 on the exams in question. This more or less convinced me there is a problem.
After these calculations, I notified Dr. Peterson and Dean Fischer. We will meet soon to discuss the next steps in the process.
I will keep you informed.
Thanks for your patience on this important matter. Remember it is more important to get it right than to act quickly.
No comments:
Post a Comment