I remember once the University of Alberta proudly displaying a survy which showed a 95 percent student satisfaction rate. The catch? The survey was taken of students on their graduation day. This is the sort of statistical fallacy educational institutions sould know better than to commit. But they can't help themselves, it seems. Witness this latest in the genre, a survey from Coursera touting a 91% satisfaction rating among students who had completed the course. We are not told the actual number of people completing the survey, but we know it is a very small percentage of the 210,000 total registered students, a self-selected sample from a very biased pool. The survey is part of a full report available from Coursera on request, but with data like this, I wouldn't bother.
Today: 3 Total: 98 [Direct link] [Share]
