A lot of people have been confused by the disparity in the state and national polling. Many explanations are given: Bad national pollsters, bad state pollsters, a certain candidate doing better/worse in solid states yet are relatively unaffected in swing states. I decided to analyze the polls and do a state by state analysis for where each state "should be" based on the national polling, and where it actually is in the state polling.
Obama won the popular vote 53-46 (I will be using rounding for all these numbers). The relative consensus of the national polls is that if the election held today, Obama would win by 3 points (51-48), a dropoff of 4%. Therefore, we can assume that 4% is distributed equally throughout the nation until we begin to delve into the data. I take the current RCP average (or if there is no average, the most recent poll, except in fishy instances) and compare it to where Obama "should be" to see his relative position within a state.
For example:
Alabama
2008 vote: 60-39 McCain (McCain +21)
Polling average: 51-36 Romney (Romney +15)
Where Obama SHOULD be: 62-37 Romney (Romney +25)
Disparity: Obama +10
Instead of boring you with that long chart for every state, I will simply show you the disparity.
Notes:
- I will not be excluding any pollsters besides internals. Although I find We Ask America extremely sketchy, and Ras is Ras, excluding pollsters only corrupts the data into telling me want I want to hear. On the other hand, not excluding what may be bad pollsters could also corrupt the data, but it's a problem that cannot be solved. Feel free to adjust these predictions more favorably for Obama as you see fit.
- A few states have the last poll being extremely sketchy (SC, KS, off the top of my head). Instead of just using that one, I will average with an older one to get a more accurate picture.
- Obviously there are other factors at work here (home state effects such as McCain in Arizona, demographic changes, state by state issues, etc.) In order to get an easily gathered set of data I must gloss over these. Feel free to address them in the comments though.
Read More