Brian Schaffner

Brian Schaffner

@bfschaffner.bsky.social

Political scientist ~ Newhouse Professor of Civic Studies at Tufts ~ CES Whisperer ~ #COYS #GoDawgs

3.4K followers528 following189 posts

Brian Schaffner reposted

Nice post from Elliott here walking through why this screen is not a perfect LV screen but it's useful. The long and short is screening out all people who say they won't vote undersamples those people (lots of them vote) and oversamples people who say they will definitely vote (lots of them don't)

G Elliott Morris@gelliottmorris.com

Likely voters for this year's midterms lean more Democratic than registered voters or all adults, a change from last cycle. Democrats are +9 among self-described likely voters in this month's Strength In Numbers/Verasight poll, vs +7 among RVs. www.gelliottmorris.com/p/2026-07-22...

Brian Schaffner reposted

Brian Schaffner reposted

“Americans have been told for at least two decades that they share less and less, while the data show the actual overlap holding roughly steady,” the political scientists Brian Schaffner and Stephen Ansolabehere write.

nyti.msOpinion | No One Wants to Hear It, but Maybe We’re Not That PolarizedAmericans have been told for at least two decades that they share less and less, while the data show the actual overlap holding roughly steady.

the centerpiece of the essay is a question we started asking our respondents recently: What is the most important thing that happened in your life during the past year? striking similarities across the partisan divide which you can see for yourself here cooperativeelectionstudy.shinyapps.io/miplife/

cooperativeelectionstudy.shinyapps.iocooperativeelectionstudy.shinyapps.ioResponses from Democrats, Independents, and Republicans rotate every few seconds.

somebody should run an adaptation of our inauguration crowd experiment asking Trump supporters what color the water is...

the perspective that "Dems should be Dems" begs the question, but which Dems should they be? goodauthority.org/news/secular...

goodauthority.orgSecular Democrats are on the riseFewer Democrats identify as religious. That’s creating tension with Black voters.

There is no tradeoff--the job of Dems is to be Dems, not to try to appeal to a far, far right GOP. As always, asking a Dem to try to split the difference between GOP and Dem is asking them to move to the far right. Let's not. saideman.blogspot.com/2025/09/prof...

i also dig into the numbers from PA to look at how Shapiro and Fetterman differ...everyone sees Fetterman as more moderate than the Democratic Party, but only Republicans view Shapiro as more moderate than Dems (who they see as quite extreme, of course) 2/3

Brian Schaffner reposted

I don't know what the phrase is when your nephew is a lot like you, but my nephew was in @bfschaffner.bsky.social's methods class this semester and did an interesting study on ideological extremity and politicians raising money and being mean on the internet. tufts-pol.medium.com/who-is-drivi...

tufts-pol.medium.comWho is driving polarization? The role of ideological extremity in communication and fundingby Fatima Eisa (‘26), Shayna Levy (‘28), Chloe Walker (‘28), and Eliot Wertheimer (‘29)

Brian Schaffner reposted

This feels like a pretty significant development that's not likely to reverse.

Brian Schaffner@bfschaffner.bsky.social

there's a new report out from YouGov using CES data to plot trends in partisanship over the past two decades. i've got a piece up today in @goodauth.bsky.social summarizing some key findings. check out both! GA piece: goodauthority.org/news/long-st... YouGov report: yougov.com/en-us/articl...

in conjunction with this report, we are also today releasing the data for the 2025 CES common content. includes interviews with 17,000 American adults conducted in late 2025. dig in! doi.org/10.7910/DVN/...

doi.orgdoi.org
Brian Schaffner@bfschaffner.bsky.social

there's a new report out from YouGov using CES data to plot trends in partisanship over the past two decades. i've got a piece up today in @goodauth.bsky.social summarizing some key findings. check out both! GA piece: goodauthority.org/news/long-st... YouGov report: yougov.com/en-us/articl...

Brian Schaffner reposted

Brian Schaffner reposted

I have a new paper. We look at ~all stats articles in political science post-2010 & show that 94% have abstracts that claim to reject a null. Only 2% present only null results. This is hard to explain unless the research process has a filter that only lets rejections through.

Brian Schaffner reposted

i have a new article out in @psrm.bsky.social with Scott Blatte exploring respondent attentiveness in the Cooperative Election Study. we find fairly low rates of inattentiveness but our intervention failed to nudge respondents to pay more attention (thread) www.cambridge.org/core/journal...

cambridge.orgSurviving the screens: the problem of hidden inattentive respondents in online surveys | Political Science Research and Methods | Cambridge CoreSurviving the screens: the problem of hidden inattentive respondents in online surveys

i have a new article out in @psrm.bsky.social with Scott Blatte exploring respondent attentiveness in the Cooperative Election Study. we find fairly low rates of inattentiveness but our intervention failed to nudge respondents to pay more attention (thread) www.cambridge.org/core/journal...

cambridge.orgSurviving the screens: the problem of hidden inattentive respondents in online surveys | Political Science Research and Methods | Cambridge CoreSurviving the screens: the problem of hidden inattentive respondents in online surveys
View full thread

we also tested an experimental intervention where we flagged respondents' contradictory responses to one set of items, but that intervention did not produce more attentiveness on subsequent items in the survey compared to the control group

tldr: the fairly low rate of inattentiveness is unlikely to affect estimates with the full sample or large subgroups, but may bias estimates for smaller subgroups. this was Scott's pet project & everyone on the CES team is grateful to him for helping us learn more about inattentiveness in the CES!

i have a new article out in @psrm.bsky.social with Scott Blatte exploring respondent attentiveness in the Cooperative Election Study. we find fairly low rates of inattentiveness but our intervention failed to nudge respondents to pay more attention (thread) www.cambridge.org/core/journal...

cambridge.orgSurviving the screens: the problem of hidden inattentive respondents in online surveys | Political Science Research and Methods | Cambridge CoreSurviving the screens: the problem of hidden inattentive respondents in online surveys
View full thread

while these low rates of inattentiveness are ignorable for analyses using the full sample, inattentive respondents tend to become a bigger share of small subgroups and can affect those subgroup estimates more.

it is important to note that the inattentiveness we document occurs in the final sample provided by @today.yougov.com, after they have applied their own extensive quality control checks. these rates are much higher in raw samples obtained from lower quality vendors

i have a new article out in @psrm.bsky.social with Scott Blatte exploring respondent attentiveness in the Cooperative Election Study. we find fairly low rates of inattentiveness but our intervention failed to nudge respondents to pay more attention (thread) www.cambridge.org/core/journal...

cambridge.orgSurviving the screens: the problem of hidden inattentive respondents in online surveys | Political Science Research and Methods | Cambridge CoreSurviving the screens: the problem of hidden inattentive respondents in online surveys

we utilize two pairs of items for which it would be logically inconsistent to indicate support for both items. about 5% of respondents gave logically inconsistent responses to these items and just 2% gave inconsistent responses on both sets of items

respondents who were consistently inattentive moved through the survey much quicker, were more likely to straight-line and take the survey on a mobile device and less likely to answer the post-election wave and be validated voters

Brian Schaffner reposted

If you need a cite backing up that issues commonly associated with African Americans (like criminal justice and poverty) *are actually perceived to be* associated with African Americans and we aren't just all talking out of our @$$ assuming that criminal justice=Black, Tatishe and I did the work:

#OpenAccess from @journalrep.bsky.social - Representing Black and White: The Role of Candidate Issue Priorities in Perceptions of Representation - https://cup.org/3KSlfEw - @miacosta.bsky.social & Tatishe Mavovosi Nteta #FirstView