BNP got seven per cent there in 2010 and 'constituency polling oversamples the politically engaged' continues to be a good heuristic so I would say...probably about right given that.
May 23, 2026 at 8:00 PM
BNP got seven per cent there in 2010 and 'constituency polling oversamples the politically engaged' continues to be a good heuristic so I would say...probably about right given that.
I think this poll oversamples the moms of Congressional Democrats
New Quinnipiac poll—
Congressional Republicans
Approve 40%
Disapprove 52%
Congressional Democrats
Approve 21%
Disapprove 68%
Elon Musk
Favorable 38%
Unfavorable 50%
Donald Trump
Approve 45%
Disapprove 49%
Trump plan to take Gaza
Pro 22%
Anti 62%
DEI policies
Good 53%
Bad 38%
Congressional Republicans
Approve 40%
Disapprove 52%
Congressional Democrats
Approve 21%
Disapprove 68%
Elon Musk
Favorable 38%
Unfavorable 50%
Donald Trump
Approve 45%
Disapprove 49%
Trump plan to take Gaza
Pro 22%
Anti 62%
DEI policies
Good 53%
Bad 38%
February 19, 2025 at 7:49 PM
I think this poll oversamples the moms of Congressional Democrats
state level oversamples:
October 30, 2024 at 3:48 PM
state level oversamples:
The flaw in the Lord Ashcroft poll is it undersamples (after weighting) the least Reform-friendly parts of the seat by 21pts. And oversamples the most by 20pts.
The Gowy villages & Helsby make up 22% of the seat's electorate. But only... 1% of the poll's sample.
Detail 👉 BRITAINELECTS.COM
The Gowy villages & Helsby make up 22% of the seat's electorate. But only... 1% of the poll's sample.
Detail 👉 BRITAINELECTS.COM
March 15, 2025 at 10:20 AM
The flaw in the Lord Ashcroft poll is it undersamples (after weighting) the least Reform-friendly parts of the seat by 21pts. And oversamples the most by 20pts.
The Gowy villages & Helsby make up 22% of the seat's electorate. But only... 1% of the poll's sample.
Detail 👉 BRITAINELECTS.COM
The Gowy villages & Helsby make up 22% of the seat's electorate. But only... 1% of the poll's sample.
Detail 👉 BRITAINELECTS.COM
I think this is spot on, and while some of it relates to the NYT being the NYT, I think it mostly comes down to a boomer-centric view of American history which vastly oversamples the Reagan 80s in its understanding of who Americans are and what they want.
November 7, 2025 at 12:00 AM
I think this is spot on, and while some of it relates to the NYT being the NYT, I think it mostly comes down to a boomer-centric view of American history which vastly oversamples the Reagan 80s in its understanding of who Americans are and what they want.
it's worth mentioning here that the polls were, in general, not bad in 2022, but the one place where they really failed is in overestimating the support for election deniers, both in House races and in SecState races.
my guess is that this was the prodrome of Trump supporter oversamples backfiring
my guess is that this was the prodrome of Trump supporter oversamples backfiring
also if that poll is anywhere close to accurate undecideds are FLOCKING toward harris. that MSG rally might have been the single most disastrous own goal ever seen in presidential politics
November 2, 2024 at 11:59 PM
it's worth mentioning here that the polls were, in general, not bad in 2022, but the one place where they really failed is in overestimating the support for election deniers, both in House races and in SecState races.
my guess is that this was the prodrome of Trump supporter oversamples backfiring
my guess is that this was the prodrome of Trump supporter oversamples backfiring
polling is so flawed. there are tons of phishing scams and worthless push polls that are offered via text so i never answer them. and random cold calls? definitely not. the poll oversamples gullible and naive people so it’s not surprising it seems way too high for trump
March 17, 2025 at 3:11 PM
polling is so flawed. there are tons of phishing scams and worthless push polls that are offered via text so i never answer them. and random cold calls? definitely not. the poll oversamples gullible and naive people so it’s not surprising it seems way too high for trump
...in the mid-2000s when 'Geek Blockbusters' were just starting to blow up because they were A.) already online and B.) had the ready-made "expertise," so ALL the foundations for how Marvel/DC/StarWars/LOTR/etc material was/is "covered" massively oversamples continuity-brain mindset.
March 16, 2025 at 3:00 PM
...in the mid-2000s when 'Geek Blockbusters' were just starting to blow up because they were A.) already online and B.) had the ready-made "expertise," so ALL the foundations for how Marvel/DC/StarWars/LOTR/etc material was/is "covered" massively oversamples continuity-brain mindset.
This is how the Manhattan Institute describes the people surveyed: "emerging multi-ethnic, working-class GOP: nearly 3,000 voters—including large oversamples of black and Hispanic Republicans and/or 2024 Trump voters—were asked about a wide range of policy issues"
January 1, 2026 at 11:20 PM
This is how the Manhattan Institute describes the people surveyed: "emerging multi-ethnic, working-class GOP: nearly 3,000 voters—including large oversamples of black and Hispanic Republicans and/or 2024 Trump voters—were asked about a wide range of policy issues"
Generic ballot for young men LV (18-29). 61D, 31R. Whoopsie. www.thirdway.org/memo/where-d...
February 11, 2026 at 6:01 PM
Generic ballot for young men LV (18-29). 61D, 31R. Whoopsie. www.thirdway.org/memo/where-d...
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)
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...
www.gelliottmorris.com/p/2026-07-22...
Likely voters lean more Democratic for November's midterms, a change from past cycles
In every wave of our poll this year, screening for likely voters has nudged the congressional generic ballot a point or two bluer
www.gelliottmorris.com
July 22, 2026 at 1:40 PM
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)
havent oversamples indicated it's not really a thing
October 3, 2024 at 5:19 PM
havent oversamples indicated it's not really a thing
Among other advice on how to vet and think critically about polling data, a quick primer on avoiding some of the polling terminology mix-ups that I tend to see most often
March 24, 2025 at 2:23 PM
Among other advice on how to vet and think critically about polling data, a quick primer on avoiding some of the polling terminology mix-ups that I tend to see most often
And why aren't we noting more that likely Republican oversamples of nine points in Texas are producing slim Talarico leads and narrow Hinojosa losses even without Spanish as a sampling language? Why did we take a Paxton lead seriously when it came from a poll that didn't use Spanish??
It's wrong
It's wrong
September 23, 2026 at 10:26 AM
And why aren't we noting more that likely Republican oversamples of nine points in Texas are producing slim Talarico leads and narrow Hinojosa losses even without Spanish as a sampling language? Why did we take a Paxton lead seriously when it came from a poll that didn't use Spanish??
It's wrong
It's wrong
I'm thrilled to announce the publication of the #CMPS2020 #Oversamples Spotlight in @pspolisci.bsky.social. To start us off, trailblazers Lorrie Frasure, Janelle Wong, Edward Vargas & Matt Barreto provide the "Introduction to the Collaborative #Multiracial #Post-Election #Survey (CMPS) Oversamples"
Introduction to the Collaborative Multiracial Post-Election Survey (CMPS) Oversamples | PS: Political Science & Politics | Cambridge Core
Introduction to the Collaborative Multiracial Post-Election Survey (CMPS) Oversamples
www.cambridge.org
February 28, 2025 at 2:52 PM
I'm thrilled to announce the publication of the #CMPS2020 #Oversamples Spotlight in @pspolisci.bsky.social. To start us off, trailblazers Lorrie Frasure, Janelle Wong, Edward Vargas & Matt Barreto provide the "Introduction to the Collaborative #Multiracial #Post-Election #Survey (CMPS) Oversamples"
A new ChimeraX tool Geopickr from Ben Barad's lab oversamples dense protein arrays on spheres, tubes, filaments, and arbitrary surfaces for geometry-assisted subtomogram averaging. Available from ChimeraX menu Tools / More Tools....
July 14, 2026 at 2:59 AM
A new ChimeraX tool Geopickr from Ben Barad's lab oversamples dense protein arrays on spheres, tubes, filaments, and arbitrary surfaces for geometry-assisted subtomogram averaging. Available from ChimeraX menu Tools / More Tools....
I think online mostly oversamples freelance writers, an industry that is doing incredibly poorly for most involved by any objective metric ... what has been disappointing is seeing how many freelance writers assume their experiences are universally generalizable
April 14, 2026 at 8:19 PM
I think online mostly oversamples freelance writers, an industry that is doing incredibly poorly for most involved by any objective metric ... what has been disappointing is seeing how many freelance writers assume their experiences are universally generalizable
For my PhD, I gathered some really good data with oversamples of ethnoracially minoritised citizens of France, Germany and the Netherlands.
With qs about discrimination, political attitudes, voting and feelings of representation.
Here is a thread with all the datasets. Please share widely!
With qs about discrimination, political attitudes, voting and feelings of representation.
Here is a thread with all the datasets. Please share widely!
November 27, 2024 at 8:47 AM
For my PhD, I gathered some really good data with oversamples of ethnoracially minoritised citizens of France, Germany and the Netherlands.
With qs about discrimination, political attitudes, voting and feelings of representation.
Here is a thread with all the datasets. Please share widely!
With qs about discrimination, political attitudes, voting and feelings of representation.
Here is a thread with all the datasets. Please share widely!
lol this poll is also one of AARP's oversamples of older voters too, which makes both this result and the GA-GOV race being D+2 extra funny
July 23, 2026 at 6:34 PM
lol this poll is also one of AARP's oversamples of older voters too, which makes both this result and the GA-GOV race being D+2 extra funny
there are some GOP oversamples here
July 1, 2026 at 4:30 PM
there are some GOP oversamples here
"better than normal" can still be very bad, and there's likely bifurcation in "better than normal"
Online might do so as an aggregate, but online *also* oversamples demographics that are *not* doing well. So less low income cishet white people, more high income cishet white people.
Online might do so as an aggregate, but online *also* oversamples demographics that are *not* doing well. So less low income cishet white people, more high income cishet white people.
April 14, 2026 at 7:55 PM
"better than normal" can still be very bad, and there's likely bifurcation in "better than normal"
Online might do so as an aggregate, but online *also* oversamples demographics that are *not* doing well. So less low income cishet white people, more high income cishet white people.
Online might do so as an aggregate, but online *also* oversamples demographics that are *not* doing well. So less low income cishet white people, more high income cishet white people.
New Emerson poll is even worse for Stevens: 39-54.
However the poll being text-to-web polling oversamples college graduates 87%-13%.
You can easily activate a motivate base of CE voters to respond to text polling where older voters and minority voters might not respond.
However the poll being text-to-web polling oversamples college graduates 87%-13%.
You can easily activate a motivate base of CE voters to respond to text polling where older voters and minority voters might not respond.
July 30, 2026 at 11:17 AM
New Emerson poll is even worse for Stevens: 39-54.
However the poll being text-to-web polling oversamples college graduates 87%-13%.
You can easily activate a motivate base of CE voters to respond to text polling where older voters and minority voters might not respond.
However the poll being text-to-web polling oversamples college graduates 87%-13%.
You can easily activate a motivate base of CE voters to respond to text polling where older voters and minority voters might not respond.
Alright who wants to do an IRA defense survey that oversamples voters in these states. DMs are open.
As the GOP’s budget megabill moves to the Senate, let @emilypont.bsky.social introduce you to the 10 senators who’ll be crucial to the Inflation Reduction Act’s survival ⬇️
The 10 Senators to Watch as the IRA Fight Picks Up Again
Plus two potential spoilers.
heatmap.news
May 29, 2025 at 6:00 PM
Alright who wants to do an IRA defense survey that oversamples voters in these states. DMs are open.
Maybe it's time for liberals to do some introspection instead of blaming everyone else for their failures? If Harris loses, this will be a big part of the reason why.
New poll: Trump is getting record levels of support among young Black and Latino men
The GenForward survey included oversamples of voters of color.
www.politico.com
October 23, 2024 at 9:51 PM
Maybe it's time for liberals to do some introspection instead of blaming everyone else for their failures? If Harris loses, this will be a big part of the reason why.