It’s a Zogby poll, so don’t get too excited, but that’s still four out of four since Election Day.
While the race for Houston mayor remains too close to call, Parker’s 5.5 percentage point lead stems from advantages among several demographics, including women, whites, Hispanics and self-identified independent voters.
Parker leads with 41.9 percent of the vote, followed by Locke’s 36.4 percent, according to the poll conducted last week by Zogby International. Nearly 18 percent of likely voters remain undecided in the contest, a sign of how fluid the race remains just days before the campaign will come to an end.
“There’s a huge pool of undecided voters and the real question now is which way they split or whether they vote at all,” said John Zogby, president and CEO of Zogby International, which conducted the poll last week for the Chronicle.
The results are drawn from a survey of 601 likely Houston voters selected randomly from purchased telephone lists of registered voters. The margin of error is 4.1 percentage points.
OK, here’s the thing. By this point, we have a pretty strong idea about who is actually participating in this election, and who is not. According to the analysis Kyle Johnston has done, 92% of the people who had voted through the first five days of early voting were people who had voted in at least two of the last three general elections. As such, any sample that doesn’t match this just isn’t going to be accurate. I seriously doubt there are that many undecideds among those who really are going to vote.
Now, I don’t know which candidate would benefit from a truer sample of likely voters, though I’m sure both of the campaigns themselves do. It may be that it all comes out in the wash. But I just don’t get the reluctance, if that’s what it is, of pollsters like Zogby to pre-screen in a more realistic manner. I mean, it’s not like this runoff is out of line with others in terms of who is voting in it. Even if I’m wrong about it not exceeding the general election turnout, we’re still talking something like 20% participation. Why wouldn’t you try to be more selective in who you poll? I just don’t get it.
The crosstabs for this poll are here. As it happens, again going by the Johnston numbers, Zogby is reasonably accurate with some subgroups, less so with others. He’s got about the right number of Republicans and African-Americans, for example, but he’s oversampled Hispanics and Independents, and undersampled Democrats and Anglos. Again, I can’t really say how that might affect this result, but I do think it’s skewed the other race they polled:
In the city controller’s race, City Councilman M.J. Khan leads with 35.4 percent of the vote to his fellow Councilman Ronald Green’s 29.5 percent, with 34.5 percent of voters still undecided.
Khan dominates among Republicans, Green is somewhat less dominant but still strong among Dems, and Khan has a tiny lead among indies. Having more Dems and fewer indies would make this race appear closer, perhaps putting Green in a slight lead. Zogby has it at 43.5D/35.5R/21.0I, when in reality 57% of early voters have a Democratic primary voting history, 32% have a GOP primary voting history, and 11% have no primary voting history. It may be that the runoff is like the general, in that a greater share of Republicans turn out on Election Day than they did during early voting. But I think that was caused in part by the late push from the Harris County GOP for Roy Morales, which I believe turned a number of undecided voters who may have otherwise stayed home into Morales supporters. I say that because of Roy’s third place finish on Election Day itself, where he surpassed Peter Brown. Without a Republican candidate in the Mayor’s race, will there be a similar surge for the runoff? Maybe, but it seems doubtful.
The effect on this in the Mayor’s race is more nebulous. Locke actually led by a tiny amount among Dems, due to his strong lead among African-Americans. Parker led among Anglos, Hispanics, Republicans, and Independents. Replacing some indies and Hispanics with Anglo Dems would likely leave her in about the same position. Hard to say for certain, though.
Anyway. The poll that really matters is going on right now, and we’ll know soon enough whether or not Zogby did any better guessing this outcome than he did the one in November. Martha and Erik have more.