Politics & Government · The Record
Texas Southern University poll puts Talarico at 47 percent, Paxton at 45
Sen. John Cornyn predicted in February that Paxton could lose the seat, or win it only narrowly. Republican primary voters made Paxton the nominee by about 29 points in May.
A new statewide survey from Texas Southern University's Barbara Jordan Public Policy Research and Survey Center finds Democrat James Talarico narrowly leading Republican Ken Paxton in the U.S. Senate race, 47 percent to 45 percent, according to the Center's published release. The same release reports Republican Gov. Greg Abbott holding a steadier 6-point lead over Democrat Gina Hinojosa.
That two-point gap is the number in circulation this week. Here is the record it should be read against.
The claim
Sen. John Cornyn (R-Texas), then seeking the Republican nomination against Paxton, said on Feb. 19, 2026 that "Ken Paxton will be the kiss of death for Republicans on the ticket in November of 2026," according to a report by WFMD.
Cornyn put the rest as a pair of alternatives rather than one flat forecast. He said he thought the attorney general, if he were the nominee, could very well lose the seat - and that if he did not lose it, he would not win "except by the hair of his chin." He added that the result would not help the down-ballot races. Those are different scenarios and should not be merged: one has Paxton losing the seat, the other has him holding it by a small margin.
The record: what Republican voters did
The first group to rule on Cornyn's argument was the Republican primary electorate, and it ruled against him. In the May 26, 2026 runoff, Paxton took about 64 percent of the vote to roughly 35 percent for Cornyn, KUT Radio reported, citing returns as of 6 a.m. on election night. Those are preliminary election-night figures reported as approximate, not certified totals.
KUT also reported that more than 873,000 Texas Republicans cast early ballots in the race, citing data from the Texas Secretary of State, and that more than 1.3 million Republicans participated in the runoff.
Measured against the February warning, that is a gap of roughly 29 points between what Cornyn told Republican voters and what Republican voters did.
The record: what the polling says
The Jordan Center's release is the survey behind this week's headline: Talarico 47, Paxton 45.
A separate pollster, Texas Public Opinion Research, surveyed 1,048 likely Texas voters from July 15 to July 17, 2026, with a margin of error of plus or minus 3.4 percentage points. In that survey, the firm wrote, "Talarico holds the best statewide net favorability rating of any figure we tested (+12) while Paxton holds the worst (-13)."
Those two figures do not travel together, and readers should not carry one onto the other. The plus-or-minus 3.4 belongs to Texas Public Opinion Research's own July survey. The Jordan Center release reviewed for this article publishes no margin of error, no sample size, no field dates and no description of the population sampled, so no sampling error can be attached to the 47-45 result here. Gaps of that size are commonly within the sampling error reported for statewide surveys; on the material available, a two-point separation is not a demonstrated lead.
Analysis
What follows is analysis, grounded only in the records cited above.
Cornyn's February statement now has two records to be scored against, and they point in opposite directions. On the first - whether Republican voters accepted the electability argument - he was rejected by roughly 29 points by the electorate he was addressing. On the second - whether the general-election numbers would be close - the summer polling in this evidence base is consistent with the narrower of his two branches rather than the broader one.
The down-ballot piece is where the two halves meet. The Jordan Center reports Republicans leading in all five down-ballot races named in the release, though by narrower margins and with double-digit shares of voters still undecided, and it reports the Senate race as the one statewide contest in that release where a Democrat is ahead. Cornyn's own formulation was that the outcome would not help the down-ballot races; describing that as Paxton dragging the ticket down is this article's inference, not his phrasing. On the release's own numbers, the down-ballot Republicans it names are ahead and Paxton is not.
No poll here is a forecast. Both are estimates of opinion at the moment they were fielded.
A checkable prediction
Daily Pol predicts that whoever wins the Texas U.S. Senate seat in the November 2026 general election will do so with less than 52 percent of the vote. This is checkable against certified statewide returns published by the Texas Secretary of State by Dec. 15, 2026. A Paxton win with 52 percent or more would falsify both this prediction and Cornyn's February scenarios; a Talarico win at that share would falsify this prediction alone.
What would settle it
The Jordan Center's full methodology - sample size, field dates, whether respondents were registered or likely voters, and the margin of error - would settle whether 47-45 is distinguishable from a tie. That material is not in the release reviewed for this piece. Certified November returns settle the rest.
Right of reply
Daily Pol has not yet put these findings to Paxton's campaign, Cornyn's office or Talarico's campaign, and none of the three has commented. Under Daily Pol's charter those requests must be made and documented by a named editor of record before a piece at this risk tier publishes; this article carries a right-of-reply hold until that happens.