Will Bill White's real approval ratings please stand up (please stand up)?

Burka writes:

The number that jumps out at me is White's favorable/unfavorable. Yesterday, Rasmussen had this number as 42% favorable, 47% unfavorable. Obviously, this would be a disaster for White. Blum & Weprin have White’s favorable/unfavorable number(s) as 45% favorable, 24% unfavorable. This is much more in line with previous polls. At 45/24, White is a viable candidate. At 42/47, he is not. But there is a third dimension to White’s favorable/unfavorable score: 26% of respondents have not heard enough about White to have an opinion.

First of all, when I look at the Blum and Weprin numbers

How do you reconcile the large difference between the Blum & Weprin poll (White approval 45-24) with the upside-down Rasmussen approval numbers (42-47)? Easy. You assume they are both basically true.

B&W asked*, "Do you have a favorable or unfavorable opinion of Bill White or haven't you heard enough about him to have an opinion?"

Rasmussen asks, "I'm going to read you a short list of people in the News. For each, please let me know if you have a very favorable, somewhat favorable, somewhat unfavorable, or very unfavorable impression." Rasmussen does not give a "not sure" option, though if people volunteer one then it is scored as such. In Rasmussen's last poll, Bill White scored 21% very favorable, 21% somewhat favorable, 28% somewhat unfavorable, and 19% very unfavorable, with 11% volunteering an unsure option.

So let's go over the reasons why the numbers are different:

1. Options given. As I mentioned, B&W gives just 3 options, and one is "not sure." Ramussen gives 4 options, and doesn't give a not sure. This is the key difference between the two polls, and is way more significant that the more maudlin factors that I will mention later. Given the option to approve of Bill White, 45 did in the Blum&Weprin poll whereas 42% did in the Rasmussen poll. 24% disapproved of White in the B&W poll contrasted to 19% disapproving strongly in the Rasmussen poll. So, we have 45-24 in B&W and 42-19 in Rasmussen.

The key difference is that Rasmussen gave a "somewhat disapprove" option. You know who picked this option? Republican voters who don't pay that much attention to politics in September. They don't really have an opinion of Bill White yet other than that he's a Democrat. The B&W question makes these voters take a firm stand against White, and if they aren't paying close attention, then they are likely to just pick "not sure." But if they are given an option, like in Rasmussen's poll, then they pick "somewhat disapprove."

2. Question ordering. In the B&W poll, they ask the "ballot test" question first, but don't ask about Bill White until the 4th question. In Rasmussen, they ask the "ballot test" question third, and immediately follow it by asking about Bill White. That's a key difference in how much the subconscious emotional reaction dies down. Rasmussen asks about job approval right after respondents hear that Bill White is a Democrat, so non-political junkie Republicans choose "somewhat disapprove."

In my opinion, it is a slightly more accurate reflection of the approval numbers when the question is immediately after the ballot test. In the ballot box, partisan emotions are as high as they ever are, so a question that replicates that is more likely to be accurate. In the B&W poll, respondents had the 2nd and 3rd question to cool down before deciding on whether not knowing much about Bill White (except that he is a Democrat) puts them in the "somewhat disapproving" or "not sure" camp.

3. Wording. This is way less important than the other two, but the Rasmussen wording of the ballot test question puts a bit more emphasis on the partisanship of the candidates, where the BW poll mentions the parties a bit more off-handedly.

Just to be clear: I'm not criticizing either the B&W poll or the Rasmussen poll. They tell us different things and give us data points to draw our own conclusions.

My conclusions would be that a good number of the people who picked "not sure" in the B&W poll are Republicans or Republican-leaners. Perry's campaign hasn't yet convinced them that Bill White deserves a firm "disapprove." That requires too much commitment when they don't know much about White yet (eg, 11% of the Rasmussen volunteered that they were unsure about White even though they weren't given the option in the script). However, when pushed, a large majority of those give a "somewhat disapprove" based largely on White's party.

White does have an opening still, if he can find a message that really resonates. He's not very well-known, which is why 41% of the B&W sample didn't give an opinion. Perry is under 50% in a red state in a very Republican year, which is good for White. On the other hand, that's also the problem: White is running in an anti-Obama election in a red state. When people know that Bill White is a Democrat, then the numbers skew towards 42-47.

* I'm going to assume that this DMN page is a close approximation of the script and order.

Posted by Evan @ 09/28/10 10:40 PM


Previous Entry | Home | Next Entry


Here's a message that I'm sure Bill White would be willing to stand behind - if he governs anything like the way he did when he was the mayor of Houston.

Well, let's have him run on a platform of a Sanctuary State for Illegal Immigrants. Oh, and don't leave out higher taxes, and bigger government.

Posted by Socialism Doesn-t Work! @ 10/05/10 12:20 PM

Add Comments

No flames or impolite behavior. HTML will be stripped. URLs will be transformed into hyperlinks.