Reply #36: Have to check the Jan report again

[Bill Bored]

but I think that when I added up the 1250, which are presented as quintiles of varying size on page 36, I weighted the quintiles by the number of precincts in each and got a weighted average of -6.77, if mem serves me. There are a few other tables that include all 1250 that could be checked in the same way. Page 39 has one "Size of Place." Page 40 has some with 1249. But I've only checked the one on pg 36. They should all come out the same though.

The -6.5 is Mitofsky's number, which included all 1460 precincts. If you then calculate the average WPE of only the 210, it comes out to about -5.

I agree that Red Shift may be different because of varying precinct sizes, but also because the early exit polls on which the Red Shift is based were weighted demographically and the precinct data were not. We know this because the earliest exit poll release had overweighted women which they corrected later on even before adjusting to the vote count.

See if my -6.77 is right.