Lightnews — Scholar-powered news

Sean Gailmard @sean-gailmard.bsky.social · 9d

This sentiment is more common in my field than I expected when I entered it. In this line of thinking, what is the role of a demos in a democratic system?

Kevin Collins @kwcollins.bsky.social · 9d

I need some people to stop treating public opinion as an un-moved mover of politics, and instead identify strategies not for following public opinion but for leading it.

2

Sean Gailmard @sean-gailmard.bsky.social · Sep 5

Interesting paper, thanks for linking it. I find it hard to know what it means for policy going forward. The wage/employment effects are estimated under a specific immigration regime; I am unsure how to extrapolate them to a different, greatly expanded regime.

1 3

Sean Gailmard @sean-gailmard.bsky.social · Aug 25

Grumbach is external validity pilled, it's happening

Jake Grumbach @jakemgrumbach.bsky.social · Aug 24

This is an excellent point. It’s not really a ‘methodology’ problem because these experiments are very internally valid. It’s that they don’t generalize externally to politics. That requires us to believe actual voting would really be different if Dems used nearly identical message A vs message B

Will Stancil @whstancil.bsky.social · Aug 24

By the way, there’s a fourth page I cut for length, but on it you can see that the best-testing and one of the worst-testing messages are nearly identical. In other words YOUR METHODOLOGY IS TRASH.

2 1 7

Sean Gailmard @sean-gailmard.bsky.social · Aug 21

If we "do politics," it seems undignified to be surprised when they do it back.

4

Sean Gailmard @sean-gailmard.bsky.social · Aug 21

It's unpleasant for our self concept but the university is the agent of the state. Especially the public university, but given their funding model, private research universities too. If they're not useful to the state, we should expect conflict and curtailment.

1 3

Sean Gailmard @sean-gailmard.bsky.social · Aug 21

I'm sympathetic to Dave's points, but also wonder what did we collectively expect? The political actors doing this stuff do not see academia as a non-ideological truth machine, and I'm not sure we make a convincing case that they should. It's easy to find colleagues who explicitly reject that model.

David Darmofal @daviddarmofal.bsky.social · Aug 21

Wait til red states start folding their humanities & social sciences departments into these ideologically-driven conservative centers & making those centers these faculty’s tenure homes. Bc that’s coming next. And when that happens, academia as a non-ideological truth-seeking enterprise is dead.

1 4

Sean Gailmard @sean-gailmard.bsky.social · Aug 20

a man wearing glasses and a black turtleneck is holding a pair of headphones .

ALT: a man wearing glasses and a black turtleneck is holding a pair of headphones .

media.tenor.com

1

Sean Gailmard @sean-gailmard.bsky.social · Aug 20

In that world, if you move Dem candidates further left on L-R position and make no other changes, you will lose more.

1 2

Sean Gailmard @sean-gailmard.bsky.social · Aug 20

Or maybe there is a host of interlocking factors, and if you shift L-R ideology far from center, you have to improve other factors to compensate. Random victory among viable candidates in a primary doesn't address this.

1 1

Sean Gailmard @sean-gailmard.bsky.social · Aug 20

Relevance being, the reduced-form null effect of L-R position is consistent with many different structural models. The tempting inference for left-maxxers is that the r.f. null is a structural 0 on candidate moderation in voter utility functions.

Maybe...

1 1

Sean Gailmard @sean-gailmard.bsky.social · Aug 20

Gon spam you til you reply bro sorry

Sean Gailmard @sean-gailmard.bsky.social · Aug 20

Also your list doesn't include reasons why moderation doesn't matter, it includes reasons why we measure it wrong.

2 2

Sean Gailmard @sean-gailmard.bsky.social · Aug 20

Perhaps you exclude it from a list of "plausible mechanisms" not because are unaware it's a mechanism, but you find it implausible. Not sure why that would be, however.

1

Sean Gailmard @sean-gailmard.bsky.social · Aug 20

Going to say again, another mechanism is that viable candidates maximize Pr(win), and observed positions are close-to-optimally calibrated to their jurisdictions. Null effect = 0 effect of marginal change = first order condition satisfied.

1 4

Sean Gailmard @sean-gailmard.bsky.social · Aug 16

There is an external validity problem built into extrapolating the result of one person's optimal choice to utility that another would experience from the same choice.

Sean Gailmard @sean-gailmard.bsky.social · Aug 16

If you love shrimp and eat 100 at the buffet, you get about the same jollies from eating 99 or 101. If I don't love shrimp but eat 100 at the buffet, I get a definite increase in jollies from eating 99 instead and a definite decrease from eating 101.

1

Sean Gailmard @sean-gailmard.bsky.social · Aug 16

I further don't understand why anyone is comfortable drawing general conclusions from RD or DD models. Candidate position is a choice variable. 0 effect from small changes is what you expect if it is chosen optimally. You are measuring (successful) optimization by the candidates in question.

1

Sean Gailmard @sean-gailmard.bsky.social · Aug 16

I don't understand how you know their model is biased (or they know it about yours). You just know that they are different.

1

Sean Gailmard @sean-gailmard.bsky.social · Aug 9

In your view, would that move lead the Trump Admin to back down and restore funding? If not, how much should UCLA or the system be willing to pay, for the principle of fighting back and saying no?

Sean Gailmard @sean-gailmard.bsky.social · Jul 29

Pluto was once a planet

1 3

Sean Gailmard @sean-gailmard.bsky.social · Jul 28

The student's return on effort to become skilled at our standard assessments is now close to 0. If we want them to put in effort, the return must be higher. We must either teach them to do things that AI cannot do, or teach them to use AI to make their own work better than it would be with AI alone.

Sean Gailmard @sean-gailmard.bsky.social · Jul 28

There's a Turing test point in here. Whether AI bots can think/analyze/etc. is irrelevant. In many standard tasks we use to assess students, they are reliably as good as a college junior at a reputable university. Therefore, bans are a waste of time and, at best, let the professors fool themselves.

1

Sean Gailmard @sean-gailmard.bsky.social · Jul 28

Students are entering a world where Gen AI will be widely available. We must teach them to use it as a complement for the skills we teach, not resist it as a substitute.

A corollary is, we must teach skills that AI cannot reliably replicate reasonably well. Being honest, we often don't.

Itai Sher @itaisher.bsky.social · Jul 28

I have seen some faculty say there should be a total ban on AI use in classes.

Maybe for some classes but I don’t think that is a tenable or desirable policy on the whole.

Also not worrying about it isn’t tenable.

What’s needed is thoughtfulness and a middle way.

3 11

Sean Gailmard @sean-gailmard.bsky.social · May 6

This is true. But it’s also true that SS/HUM defines literatures, fac lines, and whole fields in ways that appeal to progressives and reinforce their worldview. It’s not healthy for academia, intellectually or politically.

Victor Ray @victorerikray.bsky.social · May 6

I’ve said it before but “diversity of thought” is a trojan horse designed to help mainstream unserious crackpots who would be laughed out of any self-respecting graduate seminar. Yaving is only notable because his billionaire backers are even less capable of discerning a good idea than he is.

4

Sean Gailmard @sean-gailmard.bsky.social · Apr 17

It's back! I would appreciate any comments.

1 2