 |
 |
Gee, I thought we had put this issue to bed! The problem with averaging the averages is you don't take the weight of each group into account. Say you get an average of 40% from one population of 1,000. And you get an average of 60% from a different population of 2,000. If you average the two averages, you get 50%. BUT, if you weight the averages to represent the sizes of the populations accurately, you get 53.3%. (Now, Janet, look what you started up again by finding that nice quote <G>). Jane K. >A cardinal rule of statistics as I remember is that the average of the >averages is the average of all. This frequently simplifies calculations. > >>It is not necessarily true that averaging the averages >>of different populations gives the average >>of the combined population. >> >>Simpson's Paradox