Imagine I ask you what the score of the Lakers game is. And you say "Lakers are up by 10" That is not NEARLY enough information to form a narrow expectation of what the final score will be, if the Lakers are favored to win. Now apply that to how we analyze poll data, and you're top 0.01%

Anyone who writes on the subject of "poll error" should probably give this a read. Preprint will be available soon.

This poll said Brooks was up by 8. Then he won by 20. Therefore, the polls were way off. Name the flaw in this logic, and show me how you would explain to literal experts in the field (while avoiding being demeaning or condescending) why their analysis as above is wrong.

A lot of "polls was wrong" commentary tonight. Because up by big number must yield win by that number or else wrong. I am very tired.

This is absolutely remarkable Unlike most stats nerds, I do not proclaim there to be a "right" answer to questions like this. Finances cannot be accurately simplified into expected value calculations. However For any person that is not in legal or literal survival trouble, this is a bad decision

David Elliott Pritt, Republican from West Virginia, was arrested and charged with sexually explicit conversations with a minor The victim, 15, said they had several "physical sexual encounters" He allegedly groomed his victims on the game Roblox, before moving to Snapchat Waiting for the mugshot

No wonder these men are so lonely and hateful Their own family doesn't even love them

This guy claimed vaccines aren't effective because more vaccinated people died than unvaccinated ones. Look how I break them

Declaring which polls were more accurate by which one most closely predicted the final margin requires a fundamental ignorance of what data a poll gives us. No one who understands how polls work would do it. It's equivalent to judging the accuracy of a snapshot by the final race result.

Consider two competing polls: Poll 1 Candidate A 46 Candidate B 44 Undecided 10 Poll 2 Candidate A 44 Candidate B 46 Undecided 10 Given the information below, which poll was more accurate?

Hey everyone I finally figured out what's happening in this video It's called massage a knee. Commonly practiced by far right weirdos like JD.

All these accusations of knives and razors. Did anyone stop to consider files?

"well established facts" Your opinion is meaningless I love putting people like you in your place so thoroughly that the best you can do is claim your position is special and immune to rational inquiry 🤣 Slowly but surely, people like you are learning your lesson "I do not need a source"

"let's run a campaign on looks and masculinity"

Wild things in Peru. Presidential Election votes currently being counted, nearly 98% in, with a virtual tie. The far-right Fujimori is behind by hundredths of percentage points

He thinks this is good for him

Drowsy Don just can't keep his eyes open. Wonder why?

Stop arguing with them about rigged elections. Stop explaining. Shutting them up requires two words. Repeat after me: Prove it

Some humans may be sexless: not male or female Source cited. Now, do you think all humans are either male or female? Yes or no?

Sex isn't binary. And you admitted it. Your ignorance of what "binary" means suggests remedial science courses are in order.

No notes

The "parties never switched" people forget we literally have video There's a reason they don't like history It's a simple question Were the people in that video progressive or conservative? The won't answer

Real people really believe that this guy offered meaningful arguments and insights

Why "40% of the (statewide) vote should yield 40% of the state's Representatives" has nothing to do with "fairness" nor anything to do with #gerrymandering. In two sentences, I explain a concept that the @washingtonpost.com article and graphics actively misinformed readers about.

Elections with any number of undecideds make the "margin" between candidates a bad metric, because there is nonzero uncertainty in how undecideds vote Elections with more undecideds (and/or many candidates polling highly) increase likelihood of an unexpected outcome, if you only consider "margin"

This is one of many reasons why people who work in this field should be ashamed. They teach that you can compare the poll to the election and that the difference is poll error And that's not right Is this poll "Outside the margin of error" @aapor.bsky.social @britpollingcouncil.bsky.social ?

Follow up to yesterday's post Guess in comments Assume this map is divided into 5 contiguous districts of 5 units each without any consideration (or knowledge) of partisanship Of all possible maps, what % of the time do you think: A) Blue wins exactly 3 districts? B) Exactly 4? C) Exactly 5?

Leave your guess in the comments Assume this map is divided into five contiguous districts of five units each without any consideration (or knowledge) of partisanship Contiguous here means they touch at a side - not a diagonal Of all possible maps, what % of the time does blue win exactly 3?

MISINFORMATION GRAPHIC I've seen a lot of people sharing it and you should probably stop. This graphic furthers the incorrect MAGA-popularized notion that gerrymandering is when a state doesn't have representation proportional to its statewide population.