Wednesday, October 10, 2007

Scoring the results

So how did PPP poll results fare in comparison to actual election results from yesterday’s elections in Wake County? I would say pretty well, considering our sample size problems and large numbers of undecided voters in a low intensity election with low turnout, which makes it even harder to predict the demographics of the electorate.

Where will did well…

Cary Mayor
PPP Poll: Weinbrecht 59%, McAlister 33%
Election: Weinbrecht 58%, McAlister 42%

Cary At-large
PPP Poll: Portman 59%, Byrd 24%
Election: Portman 69%, Byrd 22%

Considering the large margin of error for the Cary races we did very well. We did a good job of getting the gist of the election; Weinbrecht would be a big winner and Portman even bigger.

Raleigh At-large
PPP Poll: Stephenson 25%, Baldwin 20%, Anderson 16%, Tart 6%, Williams 5%, Best 3%
Election: Stephenson 29%, Baldwin 28%, Anderson 19%, Williams 12%, Tart 11%, Best 3%

If not the exact percentages, we got the order just right.

Raleigh District A
PPP Poll: McFarlane 48%, Craven 46%
Election: McFarlane 54%, Craven 46%

I wish I could go back and do this poll again. Before weighting the results to match the demographics of District A, Nancy McFarlane was winning by much more than 2 percent. I probably overestimated the number of Republican voters who would turn out this year.

Wake School Board District 6
PPP Poll: Clark 49%, O’Brien 16%, Armogida, 10%, Zal 3%
Election: Clark 64%, O’Brien 26%, Armogida 6%, Zal 4%

Where we could have done better…

Raleigh District B
PPP Poll: Koopman 40%, Taliaferro 40%, Menendez 8%
Election: Koopman 44%, Taliaferro 33%, Menendez 22%

Same goes for District B as went for District A. Before weighting the results Koopman was winning. Whatever weighting I did do to match the demographics of previous elections in District B did not account for the change in District B of who felt compelled to vote this year. Last year, Taliaferro was unopposed and that may have led to lower turnout, but with a competitive election this year more people voted and those new voters swung the election to Koopman.

Where we missed…

Wake School Board District 3
PPP Poll: LaVance 28%, Hill 20%, Wilson 13%
Election: Hill 47%, LaVance 30%, Wilson 23%

In our defense this race had the highest number of undecided voters, so a lot of change was possible. Factor in the margin of error and you could make a case that our numbers weren’t that bad. But the fact is that we didn’t get the order correct and missed a lot of Kevin Hill supporters.

Overall, our polling missed an undercurrent of hardcore Republican support for candidates Angel Menendez, Sean O’Brien and Alfreda Wilson. But we got the big picture correct. That’s about as good as we could do considering the margins of error and the unknown turnout.

In a high intensity general election in even-numbered years it’s relatively easy to predict percentage turnout of Democrats and Republicans within a few points. But in this low intensity, low turnout local election it’s harder to know who is going to vote. For example, I weighted the Raleigh results to match previous municipal elections in the percentage of the vote coming from African-Americans. But this year there was a city-wide black candidate, Paul Anderson. Would that increase black turnout all over the city? Because Anderson didn’t win I don’t think so, but that is the kind of uncertainty we faced.

No comments:

Web Statistics