With potentially crucial Democratic primaries on tap tomorow in Texas and Ohio, the polls are showing a close race in both states, but with significant variation. In OH, the polls range from a Suffolk University survey that has HRC up by 12%, to a Reuters/CSPAN/Zogby poll that has Obama up by 2%. In TX, the variation is a bit smaller, ranging from InsiderAdvantage’s 4% lead for Clinton to Rasmussen‘s identical lead for Obama.
As Mark Blumenthal of Pollster.com has shown in separate posts on polling for TX and OH, underlying these different results are very different estimates of both turnout and of the demographic compositition of the primary electorate. If you are at all interested in polling, you should read Mark’s posts in their entirety. But the bottom line is that TX polls are all over the place in their estimates of Latino, African-American, younger-voter, and independent participation, while OH polls vary significantly in estimates of total turnout.
Blumenthal notes at the end of his post on OH polls:
The polls we have before us can tell us a great deal about how preferences differ across the key demographic and regional groups, but the tools of survey research are simply not powerful enough to predict who will vote with great precision.
That’s an important reminder.