Democrat James Talarico leads Republican Ken Paxton by a razor-thin margin in the race for a U.S. Senate seat in Texas, with two recent polls showing the contest inside the margin of error as early voting runs. A University of Houston poll gives Talarico 48% to Paxton’s 46%, FOX 26 Houston reported.
The seat belongs to Republican Sen. John Cornyn. Paxton, the state attorney general, and Talarico, a state representative, are fighting to replace him.
What the polls show
The Hobby School of Public Affairs at the University of Houston surveyed 2,000 likely voters from Oct. 10 to 14 and reported a margin of error of 2.9 points. Talarico also holds a six-point edge on favorability, 47% to 41%. A University of Texas poll from the Texas Politics Project, taken between late September and Oct. 5, had Talarico at 50%, Paxton at 47%, and Libertarian Ted Brown at 1%. Its margin of error was 3.34 points, so the three-point gap does not clearly separate the two.
The Houston survey sampled more Republicans than Democrats, 42% against 31%, with 23% independents. Even with that mix, Talarico won independents 60% to 32%.
Where each candidate is strong
Paxton leads among rural voters by a wide margin, 66% to 30%, and among men, 54% to 42%, according to the UT poll. Talarico has gained as more voters in his own party, and more independents, line up behind him. In both polls, majorities said the country and Texas are headed in the wrong direction, mostly along party lines.
Why the money and the stakes are so high
The race is also drawing heavy spending: about half a billion dollars in ads are slated for Senate contests in Texas, as we noted in our report on midterm ad spending passing $10 billion.
The campaign has also turned theatrical. Talarico suggested that the two could debate on TMZ, an offer the Paxton campaign has not publicly accepted in the reports reviewed.
What to watch
Early voting is underway in Texas, and Election Day is Nov. 3. Polls from different groups can differ because of how they sample voters and word their questions, so analysts usually look at the average of several surveys, not one result.
