Among Texans over 60, 68% are white and 20% are Latino
Among Texans 40-59, 56% are white and 28% are Latino
Among Texans under 40, 41% are white and 42% are Latino
I'm a little skeptical of the exit polls of Texas Latinos as my polling of registered voters consistently showed Obama with a smaller lead in this group, but it was still a significant lead. As the numbers above indicate, the passage of time bodes well for the Democrats. 2012 is probably too soon to count on the electoral votes, but there were enough gains in the Texas Lege in 2008 to make for a more balanced Congressional map in the 2010's.
Studies have shown that within a demographic, eg white males under 30, there is minimal difference. OTOH, recent polling that includes cell phones has given Obama a greater lead than polling that is only landline. Now, this may be due to an enthusiasm gap skewing which cell-only voters agree to be polled, or it may be a cultural factor that breaks earlier patterns. Either way, a lot of pollsters will be studying the results to see how much they need to change their methodologies in the future.
I do a weekly economic sentiment poll in Texas and the Wall Street crisis didn't have that much impact until last week when the markets dived AFTER the bailout got signed.
$4 gas was what got people worried, and it's under $3 now locally. If the media changes their doom & gloom and starts gushing about the Dow up 11%, expect the economy to dip a bit as the top campaign issue.
Not all pollsters use RDD. Calling from voter files is very common. Calling the telephone book for a presidential poll is -- odd. I use this type of list for economic tracking polls where the movement is more important than the result on any given day, but it's not appropriate for a horse race poll.
Demographically, the younger, the lower the percentage of 'white,' but also the lower the percentage of 'registered to vote' and/or 'actually show up to vote.'
Change elections aside, the natural progression is to vote more regularly as you age, so while 'minorities' are less likely to vote, that is somewhat explained by the age factor. The 'minority' voting block will be growing on two fronts - being a higher percentage of the population, and moving towards higher turnout.
In Texas, according to the state Demographer, there are 29% as many Hispanics aged 60+ as there are 60+ whites. There are 50% as many Hispanics aged 40-59 as there are whites in that age group. There are 102% as many Hispanics aged 18-39 as there are whites in that age group.
I was pretty skeptical of the Rasmussen and SUSA polls saying that Obama and Noriega were within 4 points in Texas. It didn't match my polling, and well, they blew the Texas primary polls when I nailed them (neener neener neener).
I don't think there has been a shift in sentiment as much as the primary enthusiasm skewed the RDD surveys for a while.
Yeah, it's just raw data that I can't actually use, so I thought I'd put it out there for the people arguing about polls late at night. These are all smaller samples - total poll was 1100 for 67 counties - and not corrected for demographics, so any one county may be out of balance.