TX-2012 Primary: 48% Santorum, 17% Gingrich, 16% Romney, 14% Paul (R) (Tribune/University of Texas/YouGov)

Texas Tribune/University of Texas/YouGov
2/8-15/12; 800 adults, 3.5% margin of error
529 likely voters, 4.3% margin of error
298 likely Republican primary voters, 5.7% margin of error
Mode: Internet
Tribune release

Texas

Obama Job Approval
39% Approve, 55% Disapprove

2012 President: Republican Primary
48% Santorum
17% Gingrich
16% Romney
14% Paul (R)

2012 President: General Election
49% Gingrich (R), 38% Obama (D)
49% Romney (R), 36% Obama (D)
51% Santorum (R), 39% Obama (D)
44% Paul (R), 35% Obama (D)

Also on HuffPost:

Comments

[AaronTX]

02:48 PM on 02/21/2012

Another problem:

"METRO. Do you live in the Houston, Dallas-Fort Worth, San Antonio, or Austin metropolitan
areas?
1. Yes, I live in the Houston area 19%
2. Yes, I live the Dallas-Fort Worth area 31%
3. Yes, I live in the San Antonio area 7%
4. Yes, I live in the Austin area 11%
5. No, I live in another part of Texas 32%"

This is a SIGNIFICANT undersample of Houston and big oversample of DFW. To a lesser extent they oversampled Austin, undersampled San Antonio and "rest of state." Here's what the numbers really are:

Population of Texas 2010 census 25,145,561
1) Dallas-Fort Worth-Arlington, TX MSA 6,371,773 25.3%
2) Houston-Sugar Land-Baytown, TX MSA 5,946,800 23.6%
3) San Antonio-New Braunfels, TX MSA 2,142,508 8.5%
4) Austin-Round Rock-San Marcos, TX MSA 1,716,289 6.8%
5) Rest of state 8,968,192 35.6%

Don't know how much adjusting the poll would affect the results, probably only marginally, but that just rubbed me the wrong way.

[dpearl]

06:24 PM on 02/21/2012

Remember that this is just an internet poll of the self-selected YouGov panel that is weighted to match certain statewide characteristics while ignoring others. It's interesting that you spotted one way that they are poorly aligned with the population of interest. My reading of YouGov national surveys finds them to generally track about 3 to 5 points low on Presidential approval but are roughly similar to other polls in their trends.

[AaronTX]

04:06 PM on 02/24/2012

DFW and Houston metros are roughly similar in population, but DFW region is quite a bit more conservative, especially in its northern and southern suburbs. Houston 'burbs are somewhat more working class and more hispanic/black, except for Montgomery county. So by oversampling DFW and undersampling Houston, it probably skews the poll.

Oversampling Austin probably doesn't do too much, since it's a smaller metro.

[AaronTX]

02:30 PM on 02/21/2012

You can tell how old the sample is from this:

"Please indicate how often you use the following devices to read, watch, or listen to
news and information, “never”, “hardly ever”, “sometimes”, or “regularly”.
Never Hardly Ever Sometimes Regularly
1. Tablets 75% 7% 9% 9%
2. Smartphone
or Cell Phone 48% 12% 16% 25%"

[AaronTX]

02:26 PM on 02/21/2012

If we ever have a primary.... polls of TX are pretty useless at this time.

[StatsNo1fan]

12:50 PM on 02/21/2012

TX turning Blue. Finally!!!

[AaronTX]

02:29 PM on 02/21/2012

The adult samples vary somewhat significantly from the LV ones. But I would agree, hispanics in TX aren't LVs.