Andrew Mercer
@awmercer.bsky.social
6.9K followers 1.3K following 4.1K posts
Principal methodologist at Pew Research Center. He/him
Posts Media Videos Starter Packs
awmercer.bsky.social
Just finished Wayward on Netflix and my main takeaway is that Nirvana’s version of Where Did You Sleep Last Night is unsurpassed.
Nirvana - Where Did You Sleep Last Night (Live On MTV Unplugged Unedited)
YouTube video by NirvanaVEVO
m.youtube.com
awmercer.bsky.social
That’s interesting. What about the apple ecosystem do you think it is?

I can totally see it with phones and iOS, but it’s harder to figure for PCs. iCloud maybe?
awmercer.bsky.social
Like… is it *most* students? Only some? One or two per class?
awmercer.bsky.social
The idea that the youths don’t understand directories is one of those things that I’ve heard from enough different professor friends that I’m sure it’s true but at the same time it seems like it can’t possibly be.
awmercer.bsky.social
I have the same problem. Otterbox makes a pretty low-profile case with a little rubber clip that prevents this while still being easy to open and close. Highly recommend.

www.otterbox.com/en-us/airpod...
AirPods Pro 2nd gen case | OtterBox
Apple AirPods Pro (2nd gen) case - Protect your AirPods with trendy cases in exclusive designs - Shop now!
www.otterbox.com
awmercer.bsky.social
I feel like this happens all the time with pilots. It’s like when they green light the show they also hire a show runner who knows what they’re doing.
awmercer.bsky.social
Ok, I’ll bite. What the hell is aura farming?
Reposted by Andrew Mercer
colincarlson.bsky.social
Epidemiologists of Bluesky: do you know of any papers that developed statistical models to infer Covid (or something else) incidence or mortality at scale based on survey data about contacts (i.e., “has someone you know died?”)
awmercer.bsky.social
Also, FWIW, it seems like this could, in fact, be both a reference to a poem AND a penis joke at the same time. These things are not mutually exclusive.
a little girl is asking why not both while standing in a kitchen .
Alt: Meme of a girl shrugging and asking "Why not both?"
media.tenor.com
awmercer.bsky.social
By some unfortunate twist of bad data entry or poor record linkage, someone on the Virginia voter file is listed as having my phone number.

This person also seems to have been deemed “persuadable” by both parties.

It’s going to be a very long month.
a man in a suit and tie is sitting in front of a window and says " i just want it to stop "
Alt: Jim from The Office saying “I just want it to stop.”
media.tenor.com
awmercer.bsky.social
It’s always 5pm somewhere.
awmercer.bsky.social
When will it be available for preorder?
awmercer.bsky.social
I guess you’ve also Lion El’Jonson and Guilliman with weird random names instead of comic book superhero names.
awmercer.bsky.social
He might be the only primarch whose name isn’t completely on-the-nose.

Looking at you Ferrus Manus…
Reposted by Andrew Mercer
awmercer.bsky.social
I keep sharing this post by @jcheng5.bsky.social about how using LLMs for data science doesn’t eliminate the need for statistical expertise. It’s focused on Posit’s own Databot tool, but the principles are pretty universal.
Posit
Databot is an exciting new LLM tool for exploratory data analysis, but to use it safely and effectively, you still need the critical skills of a data scientist.
posit.co
awmercer.bsky.social
I keep sharing this post by @jcheng5.bsky.social about how using LLMs for data science doesn’t eliminate the need for statistical expertise. It’s focused on Posit’s own Databot tool, but the principles are pretty universal.
Posit
Databot is an exciting new LLM tool for exploratory data analysis, but to use it safely and effectively, you still need the critical skills of a data scientist.
posit.co
awmercer.bsky.social
Say what you will about Russ, he's very consistent when it comes to branding.
awmercer.bsky.social
But the kinds of polls Conrad's talking about are what we call "opt-in" samples which means anyone who wants to can go sign themselves up to take them which can indeed create problems with representativeness.
awmercer.bsky.social
For polls that use random samples, the odds are much, much smaller than people think. I do this for a living and I have never been selected for a national political poll. If you live in a swing state or competitive congressional district, the odds are higher but still, most people won't get surveyed
awmercer.bsky.social
I haven't read the paper yet, but "professional" as it's traditionally been used in the literature doesn't necessarily imply dishonest or low quality (though I'm sure there's some overlap). "Bogus" respondents, on the other hand, create huge problems for data quality.
awmercer.bsky.social
Agree with everything you say Conrad, although one thing to note is that there's a difference between "professional" respondents, who are people who take lots of surveys, and what we call "bogus" respondents, who are essentially just clicking through and not answering honestly.
awmercer.bsky.social
Thank you for doing it! It's such a great resource!
awmercer.bsky.social
Even if you don’t “manage” data but just work with it, there’s a lot of good stuff here.