Take it for what it’s worth. I have no idea how you get a representative sample, but it’s better to try than not.
Prosecutor Kelly Siegler and former Judge Pat Lykos lead police Capt. Doug Perry and defense lawyer Jim Leitner in the race for district attorney, but about half of voters likely to participate in Tuesday’s primary were undecided last week, the Zogby International poll indicated.
If no candidate gets a majority of the vote, the top two finishers will advance to an April 8 runoff that will crown the Republican nominee against Democrat C.O. Bradford, the former Houston police chief.
In the primary for chief of Harris County government, former District Clerk Charles Bacarisse and incumbent Ed Emmett are running evenly, and a third of likely Republican voters are undecided, according to the survey. On the Democratic side, businessman David Mincberg appears headed for the nomination over businessman Ahmad Hassan.
The summary is here (PDF), and the breakdown with crosstabs is here (PDF). I’ll summarize the summary:
GOP DA Candidate Pct ========================= Kelly Siegler 20.6 Pat Lykos 14.0 Doug Perry 7.4 Jim Leitner 6.0 Not Sure 52.0 GOP County Judge Candidate Pct ========================= Charles Bacarisse 34.3 Ed Emmett 31.3 Other 3.7 Not Sure 30.8 Dem County Judge Candidate Pct ========================= David Mincberg 30.5 Ahmad Hassan 11.2 Other 3.5 Not Sure 54.8 GOP President Candidate Pct ========================= John McCain 64.3 Mike Huckabee 20.6 Ron Paul 4.8 Other 4.0 Not Sure 6.3 Dem President Candidate Pct ========================= Barack Obama 49.8 Hillary Clinton 41.4 Other 1.1 Not Sure 7.7 "County Leadership" Party Pct ========================= Democrat 41.9 Republican 40.1 Other 6.0 Not Sure 12.0 "Local Judges" Party Pct ========================= Democrat 41.1 Republican 37.2 Other 6.8 Not Sure 15.0
Couple points to make: One, I don’t know why you bother including an “other” choice for races like DA and County Judge, as there are no other choices than the stated ones. Two, the samples for the GOP DA race came from the subset of people who said they were very or somewhat likely to vote in that race. Oddly, that total of 346 voters is significantly bigger than the total that said they were likely to vote in the GOP primary, which was 192 of the inital 530 (out of 604) who said they were voting in one or the other. 304 people answered Democrat to that one, with 34 “not sure” (it reports 35, but the numbers don’t add up that way). As such, the samples for a lot of these races are very small, and should be treated with large measures of salt due to very high margins of error. Honestly, I don’t know why you wouldn’t simply keep calling till you found enough likely voters in each party before you bothered asking about these other races. In fact, this whole post seems like a waste of effort now. But there you have it anyway. Greg has more.
On a side note, this Pollster.com study of the different sample size and composition assumptions by various outfits polling in Texas is fascinating. Check it out, and remember that this is even more guesswork than usual this time around.