I sometimes talk about using communities to assess learning. People ask, how can you be sure this will result in better evaluations. Certainly a part of my response consists of, "how could it be worse?" In the case of SAT essay tests, for example, we have this story: "there was an exceedingly high correlation between number of lines in the essay and its score, so high that the probability of its being random was effectively zero." See also this.
Today: 5 Total: 5 [Direct link] [Share]
