- Mar 28, 2005
- 7,164
- 0
- 0
I did some checking of the results against the polls. They turned out to be biased in favor of Romney to a moderate degree.
Here are the results against the RCP pre-election poll averages. I am using RCP because they used raw averages rather than adjusting pollsters based on past accuracy. Silver did somewhat better than this because his judgments on these weighting adjustments were pretty good. Silver also predicted the PV perfectly, in spite of very few polls showing Obama winning the PV by that margin. Still, even Silver's state by state poll averages were off in Romney's direction, because the polls were biased in that direction. Ohio appears to be the singular exception.
NC result: R2.2 RCP:R3.0, .8 Romney bias
Fla result: ~O+.6 RCP:R1.5, 2.1 Romney bias
Oh result: O+1.9 RCP: O2.9, 1.0 Obama bias
Va result: O3.0 RCP: O.3, 2.7 Romney bias
Co result: O4.7 RCP: O1.5, 3.2 Romney bias
PA result: O5.1 RCP: O3.8, 1.3 Romney bias
Iowa result: O5.7 RCP: O2.4, 3.3 Romney bias
NH result: O5.8 RCP: O2.0, 3.8 Romney bias
NV result: O5.8 RCP: O2.8, 3.0 Romney bias
WI result: O6.7 RCP: O4.2, 2.5 Romney bias
Minn result: O7.6 RCP O5.2, 2.1 Romney bias
Michigan result: O7.6 RCP: O4.0, 3.6 Romney bias
Popular Vote: Result: O+2.5, RCP: O+.7, 1.8 Romney bias
This was actually a moderately bad year for the accuracy of state by state polling. I suspect it will be overlooked because the aggregators were able to use the averages to predict the outcome in 49 or 50 of the states. But the margins were off in Romney's favor by an average of ~2.5 points. This qualifies as a systemic bias in the polling in favor of Romney. 11 of 12 state poll averages plus the PV being biased in one direction is not a coincidence. A few pollsters were pretty accurate, the ones which appeared to be O leaning outliers.
What is it with these pollsters, most of whom are MSM, and their pro-Romney bias?
Here are the results against the RCP pre-election poll averages. I am using RCP because they used raw averages rather than adjusting pollsters based on past accuracy. Silver did somewhat better than this because his judgments on these weighting adjustments were pretty good. Silver also predicted the PV perfectly, in spite of very few polls showing Obama winning the PV by that margin. Still, even Silver's state by state poll averages were off in Romney's direction, because the polls were biased in that direction. Ohio appears to be the singular exception.
NC result: R2.2 RCP:R3.0, .8 Romney bias
Fla result: ~O+.6 RCP:R1.5, 2.1 Romney bias
Oh result: O+1.9 RCP: O2.9, 1.0 Obama bias
Va result: O3.0 RCP: O.3, 2.7 Romney bias
Co result: O4.7 RCP: O1.5, 3.2 Romney bias
PA result: O5.1 RCP: O3.8, 1.3 Romney bias
Iowa result: O5.7 RCP: O2.4, 3.3 Romney bias
NH result: O5.8 RCP: O2.0, 3.8 Romney bias
NV result: O5.8 RCP: O2.8, 3.0 Romney bias
WI result: O6.7 RCP: O4.2, 2.5 Romney bias
Minn result: O7.6 RCP O5.2, 2.1 Romney bias
Michigan result: O7.6 RCP: O4.0, 3.6 Romney bias
Popular Vote: Result: O+2.5, RCP: O+.7, 1.8 Romney bias
This was actually a moderately bad year for the accuracy of state by state polling. I suspect it will be overlooked because the aggregators were able to use the averages to predict the outcome in 49 or 50 of the states. But the margins were off in Romney's favor by an average of ~2.5 points. This qualifies as a systemic bias in the polling in favor of Romney. 11 of 12 state poll averages plus the PV being biased in one direction is not a coincidence. A few pollsters were pretty accurate, the ones which appeared to be O leaning outliers.
What is it with these pollsters, most of whom are MSM, and their pro-Romney bias?
Last edited: