Did a little analysis of how the polls performed this year.

Polling estimates understated the Democratic margin by at least 2 points in 23 of 27 races for governor and U.S. Senate. The mean understatement across all 27 races was about 4 points.

pewresearch.org/short-reads/20…

But polls did a good job of telling the story of the election. Polls consistently documented the electorate’s dark mood this year, including deep dissatisfaction with President Donald Trump, even among some of his 2024 voters. And they consistently found a wide enthusiasm gap

Here's a cool new quiz my colleagues developed. Where do you fit? pewresearch.org/journalism/qui…

Can AI models replace traditional polling? New report from colleagues at Pew Research Center finds that AI-generated survey results struggled to accurately reproduce the findings from high-quality public opinion polls in a variety of ways. Read more at pewresearch.org/?p=549336

Almost 70% of national polls in 2024 were weighted on party affiliation or past vote. This kind of political weighting is controversial. But evidence shows it has made polls modestly more accurate. New report: how it works, how we do it and its limits: pewresearch.org/methods/2026/0…

Three ways to correct a poll’s political balance: weight on self-reported party affiliation, self-reported past vote, or voter file data such as party registration. Each has strengths and weaknesses. No one approach is clearly superior in all contexts.

Bogus or fraudulent respondents constitute one of the most urgent problems in online opt-in polling. This new report by my colleague Andrew Mercer looks at three methods toidentify bogus cases but finds no surefire solution pewresearch.org/methods/2026/0…

Do you feel like you don’t fit in as a ‘Democrat’ or a ‘Republican?’

Our new political typology shows this complexity, sorting the public into nine distinct groups based on their political and cultural values.

The result is a picture of American politics with far more than just red and blue.

The AAPOR Task Force on 2024 Pre-Election Polling released its final report today aapor.org/wp-content/upl…

This AAPOR report is amazingly comprehensive. It's based on a massive data collection effort that compiled polling across elections at various levels and geographies extending back to 2000.

Major conclusions:

Today's release of the 2025 NPORS survey and data wasn't the only thing we were up to. Here's a piece on why we decided to add past vote to our survey weighting and how it's done pewresearch.org/decoded/2025/0…

Calling all data nerds! Pew Research Center has just released its 2025 edition of NPORS, with estimates of U.S. party affiliation, religious affiliation and frequency of internet use. See the fact sheet and links to the dataset at pewresearch.org/methods/fact-s…

And here are party affiliation trends from six years of NPORS data pewresearch.org/politics/fact-…

We took a broad look at social trust in America today. With a survey of almost 37,000 people we were able to look at levels of trust across the states and even in metro areas. Check out our new report here pewresearch.org/SocialTrust

My colleague @katschaeffer and I contribute a chapter about covering election polls for this new edition of The IRE Journal. Check it out!

We put together a little grab-bag of things to know about election polling in the US, for example that polling methods are quite different today than in 2016, that the margin of sampling error captures just one of several kinds of error, and much more pewrsr.ch/3iay5d5

In our new survey, only 24% of voters say the phrase “mentally sharp” describes Biden; 58% say the same of Trump.

Meanwhile, about twice as many voters describe Trump as mean-spirited (64%) than say that about Biden (31%). pewrsr.ch/4bzF6zH