Can you quantify bias?
Part 2 of my rant agaisnt the Columbus Dispatch and its unreliable polls:
If we compare the Dispatch's poll numbers for last years RON initiatives to the actual results I think we can actually find somewhat of an average by which that paper underestimate's the organization and motivation of Republicans in Ohio. The initiaves - Issues 2 thru 5 on the ballot - were all opposed by between 25 and 45% in the CD poll. In reality, however, voters opposed them by between 64 and 70%. The average underestimation of opposition comes to 31.25%. So if one applies the same underestimation to the current governor's poll, Blackwell goes from a 20 point deficit to an 11.25 point victory. While I realize the comparison is flawed, it is no more flawed than the poll itself - and every bit as amusing and relevant to the race at hand.
Blackwell will not win by 11 or more points nor will he lose by 20 points. The realistic margin will more than likely be within 5 points and I still predict that Blackwell will get more of the black vote than any GOP candidate in the state's history and will use that advantage to win by about 3 points.
And I didn't need to have 20 interns opening up 1600 envelopes to figure that out. After the election in November lets revisit these numbers and see who was closer - the Columbus Dispatch and their fantasy poll numbers or mine made up out of whole cloth.
BORU
