_______________________
With regard to the problem of using course performance as a measure
of student achievement or learning, Peter Cohen's (1981) oft-quoted
meta-analysis of 41 studies on 68 separate multisection courses
purportedly showing that:
"the average correlation between an overall instructor rating and
student achievement was +0.43; the average correlation between an
overall course rating and student achievement was +0.47 ... the
results ... provide strong support for the validity of student
ratings as measures of teaching effectiveness"
_______________________
Monce's post reminded me of the odd thought I had.
Hake was quoting the Champions of SET as stating that the validity of SET is
shown by the correlation it has with the assigned grades for the course. As
I understand the above.
This immediately brings up the suggestion that a lot paper work and money
spent on administrators could be saved by simply looking at the course
grades as the indicator of teaching effectiveness and eliminating SET's.
(Tongue mostly planted in cheek, but not entirely)