Wondering, like me, whether the polls in the Democratic primary can be trusted? Over at the Social Science Statistics Blog, statistics whiz Kevin Bartz has posted two very interesting models for predicting results of the Indiana and North Carolina primary. Bartz mines a huge amount of county-level data, including likely voter turnout, and his models purport to explain between 84% and 99% of prior results from Ohio, Pennsylvania, and Virginia. (Because his S.E. is 5%, Bartz qualifies that the "results could be shaky at the county level.")
Here are Kevin's projections:
- Indiana: Clinton 53.5%, Obama 46.5%; turnout 950,000
- North Carolina: Obama 58%, Clinton 42%; turnout 1,200,000
16% seems like a huge margin for Obama--indeed, if the above numbers are even close to right, Obama (who I had for Con Law in the fall of 2000) will likely add to his lead in pledged delegates. If so, then perhaps we can wind up a protracted primary season that is no longer serving the interests of the Democratic Party (granted, that is just my opinion). Kevin explains the differences between his models and the polls:
How do my results stack up against the current polls? In Indiana, the RealClearPolitics average has Clinton +6%, only a point from my prediction. In North Carolina, the RCP average has Obama +8%, significantly below my predicted 16% victory. Two factors shed light on this discrepancy:
- In neighboring South Carolina, the polling average had Obama +11.6% and he won by 28.9%.
- In neighboring Virginia, the polling average had Obama +17.7% and he won by 28.2%.
- So perhaps my analysis isn't so crazy putting Obama above what the polls say in NC.
Here is what Kevin's predictions look like using GIS mapping: