if i was to describe a "high quality poll", it would have random sampling and be conducted with phones. online sampling is not polling, it's political manipulation. averaging out 100 bad polls (which is what silver does) and trying to weight it by applying a pile of arbitrary personal biases is the worst approach possible. snapshot polling is preferable to aggregate analyses.

i guess i need to do this.

there's a list of recent polls at the texas senate wiki site. let's cut it off on oct 19th, the day before two weeks before the election. which polls are high quality and which aren't?

poll source	landline or internet	end date	outcome	valid poll?
https://www.foxnews.com/politics/fox-news-poll-texas-senate-race-galloping-toward-photo-finish	landline	oct 19	in the margin	yes
https://emersoncollegepolling.com/texas-2026-poll-toss-up/	internet	oct 19	in the margin	no
https://cdn.votehub.com/pdfs/VoteHub_Verasight_Oct_2026_Texas_Report_D8A925D0.pdf	internet	oct 20	in the margin	no
https://www.univision.com/noticias/politica/poll-univision-texas-latinos	internet	oct 21	in the margin	no
https://www.activote.net/polls/2026/tx-u-s-senate/2026-10-21/	internet	oct 21	in the margin	no
https://www.rasmussenreports.com/public_content/politics/elections/election_2026/election_2026_texas_senate	internet and landline	oct 21	in the margin	low quality mixed sampling
https://www.wfaa.com/article/news/politics/elections/new-poll-talarico-abbott-lead-as-early-voting-in-texas-nears-close/287-7f4cca13-6e1d-433c-9041-727931b591cd	internet and landline	oct 25	in the margin	yes
https://254labs.io/news/2026-10-tx	internet	oct 26	in the margin	no
https://www.documentcloud.org/documents/28856724-cnn-poll-conducted-by-ssrs-texas-october/	internet and landline	oct 27	in the margin	low quality mixed sampling
https://drive.google.com/file/d/1UACnmnM84Dd1hU0l6iqHNgXkyh5oCelO/view	internet	oct 27	in the margin	no
https://www.activote.net/polls/2026/tx-u-s-senate/	internet	oct 28	in the margin	no
https://quantusinsights.org/polling/f/texas-survey-paxton-up-one-abbott-ahead-by-five	internet and landline	oct 28	in the margin	low quality mixed sampling

if there is further polling in texas, it's not at wikipedia.

all of the polling, both real and fake, is in the margin, and it is all mixed sample or worse and consequently of low quality. the only high quality poll was done for fox news two weeks ago and reported a result in the margin of error. the cnn poll intentionally generated a higher than normally acceptable margin of error by using weird methodology and then put the democrats on the outside fringe of that large margin, which inflated the spread through bad sampling practices, and is suspected political manipulation for that reason.

given the myriad of real world factors working against the democrats in texas on election day, it is highly unlikely that they will win this seat. that will not be a polling error. the polling is inconclusive.

i consequently see no reason to expect this seat to flip.

 

we've been here before, as well. remember beto o'rourke?