By Jim Ellis — Thursday, March 26, 2026
Senate
We are about to see a new wave of Texas Senate polling, but will the accuracy rate be better than what was produced for the March 3 primary election?
A new Change Research survey conducted for the Agave Social Welfare Fund (March 17-19; 807 registered Texas voters; online) finds Attorney General Ken Paxton leading Sen. John Cornyn 42-39 percent in the upcoming May 26 Republican runoff election, but if polling accuracy from the primary is an indicator, this result means little.
As you may remember, Sen. Cornyn placed first in the GOP Senate primary with a 42-41-13 percent spread over Paxton and Rep. Wesley Hunt (R-Houston). According to the last six polls conducted during the final 11 days before the primary, the period covering Feb. 19 through March 2, Paxton was leading the race by seven points according to the mean average figure from the concluding six polls.
The six research firms conducting the surveys were: Peek Insights (Feb. 19-23; 800 likely Texas Republican primary voters), Blueprint Polling (Feb. 23-24; 529 likely Texas Republican primary voters), Victory Phones (Feb. 24-25; 600 likely Texas Republican primary voters), Quantus Insights (Feb. 26-27; 939 likely Texas Republican primary voters), Emerson College (Feb. 26-27; 547 likely Texas Republican primary voters), and YouGov (Feb. 26-March 2; 1,659 likely Texas Republican primary voters).
The one closest from this group conducted their study furthest from the election, Peek Insights, over the Feb. 19-23 period. Peek found a 36-36-14 percent result, which is very close to the actual 42-41-13 percent.
Blueprint Polling (Feb. 23-24) was furthest away. The firm found Paxton holding a twelve point, 42-30 percent, advantage. In actuality, the Blueprint result was close regarding Paxton’s preference figure, but well away from Cornyn’s support level and margin.
Looking at the aggregate result from the six aforementioned pollsters, the Paxton average lead of seven points means the group cumulatively missed the final tally by eight points when adding Sen. Cornyn’s actual one-point edge.
Therefore, the new Change Research 42-39 percent ballot test conclusion should be viewed with a wary eye since the cumulative primary data, by and large, badly missed the mark.
On the Democratic side in a Senate nomination race where state Rep. James Talarico (D-Austin) defeated US Rep. Jasmine Crockett (D-Dallas) outright by a 52-46 percent margin, four pollsters conducted surveys prior to the March 3 primary. The results were mixed.
The four research firms conducting surveys during the Feb. 23 – March 2 period were Blueprint Polling (Feb. 23-24; 472 likely Texas Democratic primary voters), Public Policy Polling (Feb. 25; 599 likely Texas Democratic primary voters), Emerson College (Feb. 26-27; 850 likely Texas Democratic primary voters), and YouGov (Feb.26-March 2; 2,408 likely Texas Democratic primary voters).
Emerson College’s 52-47 percent ballot test result with Talarico leading was virtually right on target. Public Policy Polling was also close with their 48-42 percent split, a six-point spread that proved correct.
The two missed surveys were from Blueprint Polling (52-40 percent; correctly projecting the Talarico percentage but badly underestimating Crockett’s support), and YouGov (53-40 percent, basically the same ballot test result as Blueprint). YouGov, as well, correctly determined the Talarico support level but severely undercounted the Crockett support base.
Because Talarico won the Democratic primary without a runoff means we won’t see any significant general election polling until the electorate breaks closer to the November vote. Some polls will be released, of course, but they are likely to show what early cycle Texas studies commonly project: a tight race. Until October, expect close polling but allow for a wide accuracy curve.