Originally posted by: PJ Stroh
There is such a thing as "Mean average" ...but its almost always considered a bad way to gather data. A "weighted average" is more accurate.
Example : School district "A" has 100 students with an average score of 1000
Schoold district "B" has 500 students with an average score of 800
The mean average would be 900....which is the average if the two scores regardless of the number of students
The weighted average would be 833 - which takes into account that school district "B" has more students
Boilerman used a suburban middle class school distirct with a few thousand students and compared it to an inner city schoold district with 30x as many students. And like I said - you can do with that what you want. My IQ isnt low enough.
Right---but not for two data sets, which is what Boiler was Boilerbabbling about. A "mean average" would be the midpoint of a meta-set of averages of independent data sets. Since Boiler was comparing, or pretending to compare, two data sets, the concept of a mean is meaningless...there is no such thing as the "mean" of only two data points.
To be fair to Boiler--something he doesn't deserve--a weighted average wouldn't speak to the foolhard point he was trying to make, which is that students in his imaginary Indiana town receive better educations than students in Chicago. That Chicago has several hundred more data points than his Indiana town doesn't necessarily invalidate the comparison. If you "weight" two widely disparate data sets on the criterion of sample size, it reduces the smaller set to the status of background noise.
If we wanted to continue this ludicrous Boilerboob discussion--which we don't--we can find a whole raft of cities with Republican administrations (and voters) and compare their students' performance to that of students in any middle-class suburb--red or blue--and we'd get the same differences--since the determining factor is not political ideology, but poverty vs. affluence. (DUH.)