We have our first published poll of the season.
Mayor Annise Parker, leading the city during an era of budget cutbacks and high unemployment, has the lowest approval ratings of any Houston mayor in decades.
That’s the striking headline popping out of an exclusive poll conducted less than a month before the city’s Election Day. The mayor faces only token opposition, but the survey conducted by KHOU 11 News and KUHF Houston Public Radio indicates that a well-financed candidate could have seriously challenged Parker’s bid for re-election to her second term.
“She is down in almost every demographic and geographic area of the city,” said Bob Stein, the Rice University professor who supervised the poll.
The poll indicates fully half of likely Houston voters — 50 percent — rate Parker’s job performance “fair” or “poor,’ while 47 percent rate her “good” or “excellent.” That’s an unusually low approval rating for a first-term Houston mayor.
The mayor blames her low ratings mainly on general discontent with the economy. She also points out that fate has dealt her a difficult hand, forcing her to make politically unpopular decisions, like cutting services during an unprecedented budget crisis and imposing water conservation rules during an unprecedented drought.
Putting aside any specific disagreements one may have with the Mayor, I think there’s a lot to that. Every officeholder is less popular in bad times than they would be in good times. I’m not saying Mayor Parker would have Bill White/Bob Lanier levels of popularity, but she wouldn’t be upside down in a stronger economy. I’m sure she’d love to have the chance to be Mayor in some flush years.
As for election numbers, my advice is to reach for the salt shaker:
Perhaps as a result of that discontent, a whopping 50 percent of likely voters say they still haven’t made up their minds how they’ll cast their ballots in the mayor’s race. If the election were held tomorrow, the poll indicates the winner would be “Don’t Know.” Parker wins the support of 37 percent of voters. Her five opponents — little known and little funded — split 11 percent of the vote, but they’re all mired in single digits.
You can see some questions and their occasionally mismatched answers here. The critical numbers:
GENERIC BALLOT Frequency Percent Vote to reelect Parker 143 19% Vote for another candidate 118 16% Don't know 101 14% Refused 3 .3% Total 364 49% With Names Frequency Percent Kevin Simms 6 .8% Amanda Ulman 4 .6% Dave Wilson 6 .8% Fernando Herrera 17 2.3% Annise Parker 132 17.6% John 'Jack' O'Connor 7 .9% Do not know 178 23.8% Refused 4 .6% Total 353 47.3%
I presume what this means is that they asked about half of the sample the generic question, and for the other half they named names. Parker led the generic sample 39-32, and had 37% on the named ballot; Fernando Herrera was next with a shade under 5%.
The question is how to interpret these numbers. As Greg notes, one way to look at it is to take the generic Parker/not Parker result, which translates to a 55/45 split for the Mayor. This basically assumes that a lot of the “Don’t Know” respondents will stay home, which strikes me as a very reasonable assumption. Another thing to consider is that while “not Parker” got a fairly hefty total, the sum of her named opponents garnered a total of 11%, barely a third of the generic opposition. If you assume that most of the people who expressed some preference are likely to vote, then the question is what to the people who don’t care for Parker but don’t know any of the opponents do? My guess is some will randomly choose an opponent, some may have heard enough about one of them to enable them to make a choice – having someone handing out literature at all of the early voting locations might pay a dividend – and some of them will simply skip the race. That is to the Mayor’s benefit, and it suggests her actual level of support is higher than the generic re-elect total. If you combine all of the Parker/not Parker totals, you get a 275-158 split for her, which is 63.5% in her favor. Let’s call that the opening over/under line from Vegas.
(Yes, it’s possible that some people who say they support the Mayor are among the unlikely voters, and that this could shift the real percentages the other direction. But then the same might be true for some of the non-supporters of the Mayor, and who’s to say which group is greater? My assumption, as I said above, is that most of those who expressed a preference are likely to vote. I’ve got to assume something.)
Of course, all of this follows from another critical assumption, which I am not prepared to make, that this sample is made up of actual likely voters. Here’s the key question from the poll:
How likely are you to vote in the November City of Houston election? Would you say you are very likely to vote, somewhat likely to vote, or not likely to vote in this November’s election.
Very likely to vote 620 83%
Somewhat likely 127 17%
Total 748 100%
Funny how nobody answered “not likely”, isn’t it? Pollster Bob Stein says in the KHOU story that he “expects about 17 percent of voters to cast ballots, putting the turnout at about 125,000″. I’d say that puts him on the optimistic side of the equation, but regardless of that, how many of these people are really in that 125,000? I don’t have the crosstabs, though I have asked for them, and I don’t have any information about whether a pre-screen was done to narrow the sample down to those who really do tend to vote in odd numbered years. A poll of plain old registered voters is next to meaningless in a low turnout context, as we saw in 2009. For my money, even being a 2009 voter isn’t enough this year. If you were eligible to vote in Houston in 2007 and failed to do so, I don’t consider you a “likely” voter for polling purposes. Turnout in 2007 was 125,856, right in line with Dr. Stein’s projection. Polling any larger sample for anything other than a read on general attitudes is a waste of time, in my opinion. KUHF has more.