
Originally Posted by
MatthewT
all of the pollsters that oversampled democrats oversampled democrats. based on what, God only knows. was it an encouragement to democratic voters? you bet your ass it was.
for example, let's say 10000 people polled, 2 man race, actual numbers come out 3,334 Republican, 3,333 Democrat, and 3,333 Independent (whatever the fuck that means).
so if you report the actual numbers, republicans are ahead. still with me?
now massage those numbers for party affiliation. assume, what, 2008 numbers, approx. so 39% Democrat, 32% Republican, and 29% Independent (probably embarrassed Republicans).
39/33 x 3,333 Democrats = 3,939
32/33 x 3,334 Republicans = 3,232
29/33 x 3,333 Independants = 2,929
Poll Original (without massaging for party affiliation): R 33.34%, D 33.33%
Poll Fixed (massaging for party affiliation): D 39.39%, R 32.32%
so with no new data whatsoever, this supposedly random sampling that showed the narrowest of Republican wins originally now shows a comfortable 7.07% Democratic lead.
same numbers. same people who said they were voting how they were voting.
do you get it now?
Bookmarks