GA-SEN 2022 Megathread: Werewolves and Vampires

[coloradocowboi]

(I realize no one is listening)

But in a strong R year, this race is Lean R right now depending upon how fast Georgia is trending blue

Perdue was the only one that would make this Lean R

Nobody is listening because your opinion is completely baseless.

This race is anything but Lean R right now. Warnock won it just a month ago lol.

And Perdue is the only R who could win? He is one of the few Rs in recent history to lose statewidee in Georgia. Not saying it's a slamdunk for Warnock, but I think it's just bizarre to assume he loses to another loser because of a hunch lol

[coloradocowboi]

He's gonna get gay rights taken away from all of us

Honestly if he is what gay rights leads to I’ll give them up.

He apparently does this sh**t when he's out in West Hollywood. One of the many reasons I avoid the bars entirely now

[coloradocowboi]

[tweet]

Question about focus groups, and maybe an experienced user like Pollster can answer here: isn't it almost impossible to get a geographically representative sample in a state even as large as Georgia? I understand that they're not meant to be statistically representative, obviously, but it feels irresponsible to just report the state and not say where exactly this is: Athens? Atlanta?

Yes. Focus groups usually never have more than 12 participants (and even that is large for a focus group) so it's impossible to be representative of anything, let alone geography in even a small state. Focus groups aren't designed to be representative and are intended only to be directional or informative, not predictive, and should never be the sole research tool used by any persuasive campaign (political or not). You will be hard pressed to find a focus group memo ever given to a client that doesn't have the words "qualitative takeaways are not statistically representative and should be confirmed with quantitative research before making high-stakes decisions" or something along those lines. That's part of what makes this focus group podcast (and Longwell's editorialization) so funny. There are a LOT of useful functions that focus groups provide, but what she's trying to do generally isn't it.

You won’t see it the other way around, unfortunately. I think that if you truly want something “representative,” a mix of qual and quant (usually in that order too) is absolutely crucial. Focus groups can help better design polling instruments and provide natural language for answers that more accurately reflects public opinion.