Ariel Edwards-Levy
@aedwardslevy.bsky.social
Polling editor at CNN, keeping (cross)tabs on public opinion and the news. I like puns.
Pinned
How do you read a poll? CNN polling editor Ariel Edwards-Levy walks you through it yourfirstbyline.substack.com/p/how-to-rea...
How to Read a Poll
CNN polling editor Ariel Edwards-Levy walks you through everything you need to know.
yourfirstbyline.substack.com
Excited to share some tips for writing about polls for anyone coming in cold to the subject!
(Pictured: a thing on which pollsters are probably not calling you these days).
(Pictured: a thing on which pollsters are probably not calling you these days).
personally I do not feel like The Discourse -- or our understanding of society -- would be improved without macro measurements of e.g. President Trump's popularity, economic dissatisfaction or partisan motivation, but hey
My estimation of the usefulness of political opinion polling just declines & declines ...
November 20, 2025 at 8:50 PM
personally I do not feel like The Discourse -- or our understanding of society -- would be improved without macro measurements of e.g. President Trump's popularity, economic dissatisfaction or partisan motivation, but hey
"I bet the media won't even cover this" watch, part ♾️
November 20, 2025 at 7:36 PM
"I bet the media won't even cover this" watch, part ♾️
the perennial temptation to abuse autoresponder to just send everyone who emails me an Interesting Poll of the Day they might like
November 20, 2025 at 6:00 PM
the perennial temptation to abuse autoresponder to just send everyone who emails me an Interesting Poll of the Day they might like
Reposted by Ariel Edwards-Levy
OH! speaking of blasts from the past, it is Thanksgiving season and I can’t let it go by without resharing what is perhaps my magnum opus wapo.st/3LRYFfi
Opinion | The 9 best Thanksgiving songs I definitely didn’t just make up
The best Thanksgiving songs with which everyone is familiar and which we didn’t entirely write and record this week!
wapo.st
November 20, 2025 at 5:36 PM
OH! speaking of blasts from the past, it is Thanksgiving season and I can’t let it go by without resharing what is perhaps my magnum opus wapo.st/3LRYFfi
sobbing in genpop response rate
The delay in the data did have one benefit: more survey responses! The collection rate for the establishment survey topped 80% in September, far above the recent average for the first release. Though still below the ~90% that is typical for the second release.
#NumbersDay
#NumbersDay
November 20, 2025 at 2:54 PM
sobbing in genpop response rate
not going to engage with this morning's brand of headline criticism that appears predicated on the line of thinking that "maybe other people don't know that 'the White House' currently refers to the Trump administration" but I expect people reading articles about polling do generally know that
November 20, 2025 at 1:36 PM
not going to engage with this morning's brand of headline criticism that appears predicated on the line of thinking that "maybe other people don't know that 'the White House' currently refers to the Trump administration" but I expect people reading articles about polling do generally know that
have just realized that pia zadora is not one of the sisters gabor
November 20, 2025 at 3:11 AM
have just realized that pia zadora is not one of the sisters gabor
unfathomable why anyone would choose to turn on "your followers can reply" when they'd be missing so much truly inexplicable content
November 19, 2025 at 10:45 PM
unfathomable why anyone would choose to turn on "your followers can reply" when they'd be missing so much truly inexplicable content
"phishing" is right there
November 19, 2025 at 9:53 PM
"phishing" is right there
now just imagining the state of Rhode Island singlehandedly attempting a coup
scholars.unh.edu/survey_cente...
scholars.unh.edu/survey_cente...
Rhode Islanders Divided on Sinking of Boats Off Venezuelan Coast, Oppose Overthrowing Maduro 11/19/2025
Rhode Islanders are about evenly split on the recent sinking of suspected drug trafficking boats off the Venezuelan coast but clearly oppose attempting to remove Venezuelan President Nicholas Maduro f...
scholars.unh.edu
November 19, 2025 at 9:27 PM
now just imagining the state of Rhode Island singlehandedly attempting a coup
scholars.unh.edu/survey_cente...
scholars.unh.edu/survey_cente...
idea for people who celebrate christmas:
1. use tinsel to divide your tree into differently shaped sections with roughly similar numbers of ornaments
2. say it's "treedistricting"
1. use tinsel to divide your tree into differently shaped sections with roughly similar numbers of ornaments
2. say it's "treedistricting"
November 19, 2025 at 9:15 PM
idea for people who celebrate christmas:
1. use tinsel to divide your tree into differently shaped sections with roughly similar numbers of ornaments
2. say it's "treedistricting"
1. use tinsel to divide your tree into differently shaped sections with roughly similar numbers of ornaments
2. say it's "treedistricting"
indictment also invalid because jury were unaware of silent "c"
November 19, 2025 at 8:40 PM
indictment also invalid because jury were unaware of silent "c"
occasionally I think about this absolutely perfect five-question battery
November 19, 2025 at 8:11 PM
occasionally I think about this absolutely perfect five-question battery
my only thought about that UCSD math story is that it appears the statute of limitations of my being kinda mad about getting 5s on the AP Euro/US history tests and Cs in the classes has not yet expired
(was I a Good At Tests kid, yes)
(was I a Good At Tests kid, yes)
November 19, 2025 at 4:14 PM
my only thought about that UCSD math story is that it appears the statute of limitations of my being kinda mad about getting 5s on the AP Euro/US history tests and Cs in the classes has not yet expired
(was I a Good At Tests kid, yes)
(was I a Good At Tests kid, yes)
guys, I am so hopeful that this is the election cycle in which everyone can stop flipping out any one survey result in isolation without considering th[touches earpiece]
November 19, 2025 at 2:56 PM
guys, I am so hopeful that this is the election cycle in which everyone can stop flipping out any one survey result in isolation without considering th[touches earpiece]
Reposted by Ariel Edwards-Levy
Not the strangest newsroom behavior I've seen, tbh
“Why are you pouring water over your head?” asked Frog.
“I hope that if I pour water over my head, it will help me to think of a story,” said Toad.
“I hope that if I pour water over my head, it will help me to think of a story,” said Toad.
November 19, 2025 at 2:42 AM
Not the strangest newsroom behavior I've seen, tbh
More crucial than ever for those publishing survey data to make sure that methodology statements* address not only the mode of the poll (online, text-to-web, phone, etc.) but also the sample frame (where they found respondents -- if panel, which and how panelists recruited)
*which they should have
*which they should have
November 18, 2025 at 8:20 PM
More crucial than ever for those publishing survey data to make sure that methodology statements* address not only the mode of the poll (online, text-to-web, phone, etc.) but also the sample frame (where they found respondents -- if panel, which and how panelists recruited)
*which they should have
*which they should have
Reposted by Ariel Edwards-Levy
I don’t put much stock on message testing polls like this. People don’t know why they like who they like, so “Does this make you more or less likely to vote for x” questions are basically useless.
November 18, 2025 at 7:33 PM
I don’t put much stock on message testing polls like this. People don’t know why they like who they like, so “Does this make you more or less likely to vote for x” questions are basically useless.
¹in a 2003 interview with Karen O, the maps themselves reiterated their intent to "stay the same"
November 18, 2025 at 7:02 PM
¹in a 2003 interview with Karen O, the maps themselves reiterated their intent to "stay the same"
little known fact about total survey error: if the difference between subgroups is wildly within MOE *but* it's directionally correct to a point that you really wanted to make, in that case, overemphasizing it is fine actually
November 18, 2025 at 5:55 PM
little known fact about total survey error: if the difference between subgroups is wildly within MOE *but* it's directionally correct to a point that you really wanted to make, in that case, overemphasizing it is fine actually
oh, I'm extremely familiar with the word "parasocial," we're basically besties
Cambridge Dictionary have declared "parasocial" their word of the year, but 59% of Britons say they've never heard the term before
Confident know what it means: 11%
Vaguely know what it means: 15%
Have heard it, but don't know what it means: 15%
Have not heard it: 59%
yougov.co.uk/topics/socie...
Confident know what it means: 11%
Vaguely know what it means: 15%
Have heard it, but don't know what it means: 15%
Have not heard it: 59%
yougov.co.uk/topics/socie...
November 18, 2025 at 4:54 PM
oh, I'm extremely familiar with the word "parasocial," we're basically besties
Reposted by Ariel Edwards-Levy
[according to reports]
[does not link to any reports]
[does not link to any reports]
November 18, 2025 at 4:29 PM
[according to reports]
[does not link to any reports]
[does not link to any reports]
Reposted by Ariel Edwards-Levy
Some numbers from our new KFF/NYT Survey of immigrants:
- 22% know someone who was detained or deported
- 49% feel less safe since President Trump took office
- 47% struggled to pay for food, housing, or health care
- Still, 70% would come to the U.S. again if given a choice
- 22% know someone who was detained or deported
- 49% feel less safe since President Trump took office
- 47% struggled to pay for food, housing, or health care
- Still, 70% would come to the U.S. again if given a choice
November 18, 2025 at 3:59 PM
Some numbers from our new KFF/NYT Survey of immigrants:
- 22% know someone who was detained or deported
- 49% feel less safe since President Trump took office
- 47% struggled to pay for food, housing, or health care
- Still, 70% would come to the U.S. again if given a choice
- 22% know someone who was detained or deported
- 49% feel less safe since President Trump took office
- 47% struggled to pay for food, housing, or health care
- Still, 70% would come to the U.S. again if given a choice