Quote:
|
Originally Posted by Jason
. . . I suppose we would come up with new criteria for rating cars. . . .
|
A radical approach might be to use a statistical method. Calculate the mean and standard deviation. Everyone at or above one standard deviation from the average would get a star. For a normal, bell-curve distribution, that would be the upper third.
Another approach would be count and order all values. Then take the upper 20% ( "B" and above when grading on a curve) to get a star.
Both approaches avoid the problem of EPA 'moving the goal posts.' Unfortunately it doesn't do much for the ratio of city/highway MPG but then that is also an EPA artifact.
Bob Wilson