One of the stats profs dropped by for something else and I managed to inveigle him into analyzing a particular set of stats. It looked to me as if there was a difference between accredited departments and those which weren't. Sure enough a Chi-square showed significance way out to the p=.0001 level.
What was interesting because I don't yet understand it was that calculating Chi-square using a spreadsheet was half the value generated by SPSS. The stats prof said it was because of the way Chi-Square stats treat a 2x2 matrix. Beyond me, but I passed the good news along to the prof who wanted to find something significant to say about her survey.
Only later I found that "significance" is not as significant as it once was: statisticians seem to prefer describing confidence levels rather than specifying the probability of significance.
A random mental walk.
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment