Will Lowe

Can confirm

It's a Sobering thought really: there was no selection for academic Will, only selection of academic Will.

Joke's on them. I'm a spandrel.

Thing is, I'm not trying (or succeeding) to like or dislike anything in particular here. I'm just trying to figure out what you're saying in the blog post, from apparently quite some conceptual and disciplinary distance. So I'm happy to wait for the next one to see what I'm missing.

Budget seems pretty clear in the dating case, no? An individual has propensities to be kind and hot in a ratio empirically identifiable by watching their behaviour over any given a fixed observation period. e.g. what proportion of N texts are hot vs kind? N is a budget, but its value doesn't matter.

😈

🇪🇺 This could be us but you playing

Rubio seems to ponder his shoes

So all this time I've been interpreting "clown shoes" wrong?

a woman wearing a red coca cola shirt is blowing her mind ALT: a woman wearing a red coca cola shirt is blowing her mind

#BongRipCounterfactual

Note to self: get restaurant recommendations from Ruben.

Today in bong-rip counterfactuals:

"Hey, what if you had, you know, your fourth kid... first?"
"Woooah!"

Just remembering when I first heard this term 🤣

Ha. That question is rather good way to check on whether someone has principles or affiliations.

A super simple way to see how you get ideal points from more factor-y structures is to set a budget & demand a tradeoff, e.g. "we all like more hotness & kindness but nobody's perfect, so would you prefer those at 2-to-1 over 1-to-3?" That implies an ideal point. (In my world, it's "guns vs butter")

The second point is just that in an ideal point structure, the distance of relevance is, to put this in IRT terms, not between my Likert response and your Likert response (discrete, with edge effects) but between my theta and your theta (our positions in a continuous preference space).

(This latter question was the second thing I was reading into Ruben's thing.)

On the causal question imo a key question is whether to think of interventions and effects in the measured space or the item space. At this point all the fields I know choose the first. Which means you skip edge effect problems but then have to grapple w the measurement issues.

Happy to hear it.

Inferring ideal point structure is a core problem in community ecology, and causal inference on 'quadratic ideal points' is key to the political science of party systems, voting behaviour, and bunch of other poli subfields. You might enjoy these if you want to read out of field.

Two dose-response functions. When probabilities of exposure are not homogeneous across units, we can only partially identify the expected average outcomes from the average outcomes by exposure — even when the exposure map is correctly specified; see Corollary 11.2 — as shown here using data from Cai et al. (2015). AFEOs by exposure (left) only partially identify the EAO curve (right). Lines indicate bounds on the EAO, with red lines being the bounds when outcomes are assumed to be monotonic under exposure levels. Error bars and bands are 95% confidence intervals.

I'm giving the upcoming Online Causal Inference Seminar, this Tuesday 11:30am Eastern.

I'll be talking about different dose–response functions you might want to estimate when treatment effects may spill over from one unit to another.

Tune in & ask questions!
sites.google.com/view/ocis/home

Mustang 815 Dark Horse Power.

I'm impressed it fit in the box though.

Yup, big time availability bias.

As it happens, Mike Oldfield's 1975 third album is named after how you pronounce Oldham when you get into the taxi after too many beers in Rochdale.

You're going to tell me I don't get a tank or a kangaroo boyfriend either, aren't you? 😔

Second, if someone scores higher than your ideal on some dimensions, the Euclidean distance counts that against them. This might be plausible for a few traits like religiousness, where some prefer low and others prefer high levels. But it seems less plausible for physical attractiveness or kindness, where most would happily take "more" as a nice bonus.

Probably that.

I imagined this section said that preference for religiosity was like the item response function on the left (the 'ideal point' would be the value at the peak) but preference for hotness was like the one on the right (no ideal point bc more is better). Is that about right?

ok, but did I understand the two points right or did I miss the thrust of it?

Lucy holds the football

Incredible stuff from the Commission President

Hmm. Something truly weird happening then.

Per below, I assume the journal actually *created* it. Which is why I have no problem making fun of it.

A question for Margit I think, but it rolled out across Estonia so I imagine there's a stockpile somewhere. Probably even in the philosophy department.

I think I read this phrase somewhere a long time ago and I loved the mental image so much I just can't resist dropping it in.