Scott Barkowski

1.8K posts

Scott Barkowski banner
Scott Barkowski

Scott Barkowski

@ScottBarkowski

Health and Labor Economist, @UBuffaloEcon @SUNY. Aspires to draw his graphs on the board perfectly, every time.

Buffalo, NY Katılım Ocak 2020
446 Takip Edilen388 Takipçiler
Scott Barkowski
Scott Barkowski@ScottBarkowski·
@ryancbriggs I demonstrated in this thread that I know this literature better than you. So why assume I am confused just because I don’t agree? Maybe you were confused about the critique? Consider this: if Ioannidis’ work is correct, how do we have AI? How do we have covid vaccines?
English
1
0
0
59
Ryan Briggs
Ryan Briggs@ryancbriggs·
@ScottBarkowski I think you are probably confused about what the papers are doing. I don’t know if you “know what you’re talking about” or not because I don’t know you. The former does not imply the latter.
English
1
0
0
37
Ryan Briggs
Ryan Briggs@ryancbriggs·
I see some economists are jumping to claims of econ supremacy in empirical work from one especially audacious example of a superspreader of bad research. Fwiw when we actually look at replication, bugs in code, or power, empirical econ looks a lot like psych or quant polisci.
English
10
11
143
10.9K
Scott Barkowski
Scott Barkowski@ScottBarkowski·
@ryancbriggs No, it just doesn’t make sense logically. There’s no new information from post hoc power analyses. People use it to argue we shouldn’t believe some effect. But it isn’t certain, because low power is not zero power. Besides, there is much more to empirical work than power.
English
1
0
0
47
Ryan Briggs
Ryan Briggs@ryancbriggs·
@ScottBarkowski You’re misunderstanding what both papers do. The power analysis they do is likely biased, but it’s likely biased up (and power vs the effect sizes from meta analysis is already so low)
English
1
0
0
37
Scott Barkowski
Scott Barkowski@ScottBarkowski·
@ryancbriggs Ex post power calculations really shouldn’t be a thing. We already have standard errors.
English
1
0
0
57
Ryan Briggs
Ryan Briggs@ryancbriggs·
Econ and PS look very similar here too. The largest difference is Econ articles had more coding errors (but also more lines of code, so possibly a similar rate of bugs per line of code) papers.ssrn.com/sol3/papers.cf…
English
3
1
19
1.8K
Scott Barkowski
Scott Barkowski@ScottBarkowski·
@cremieuxrecueil The papers conclusions might be wrong, but just quickly adding FE might not be enough to undermine the analysis. You have to be sure your interpretation after adding the FE is correct.
English
0
0
0
16
Scott Barkowski
Scott Barkowski@ScottBarkowski·
@cremieuxrecueil I’m not very familiar with the GBTM approach the paper uses, but are you sure it’s actually compatible with individual/family fixed effects? Gemini says it is trying to estimate group effects, the FE might just end up absorbing those effects.
English
1
0
0
109
Crémieux
Crémieux@cremieuxrecueil·
Checked: Nothing here survives within families. "controlling for all the right stuff" evidently did not include using reasonable fixed-effects like family-level ones. Social scientists have one job and they consistently fail to do it properly.
Jonathan Haidt@JonHaidt

Powerful new longitudinal study finds that adolescents who increase their social media over a 2 year period show lower cognitive performance, compared to those who did not increase. Fom @jasonmnagata's team, using ABCD data, and controlling for all the right stuff.

English
15
52
835
75.6K
Michael
Michael@feline_feelins·
@cremieuxrecueil Sorry if this is daft, but does this mean specifying a regression including a predictor variable with a ton of levels (for each family unit)? Would a mixed model with random effects for family not resolve the confounding?
English
1
0
0
292
Wojtek Kopczuk 🇵🇱🇺🇦 and 🇺🇲
Wojtek Kopczuk 🇵🇱🇺🇦 and 🇺🇲@wwwojtekk·
They are screwed though because there is nothing better than A. They should've picked an increasing numerical scale as some countries do. Not too smart, Harvard.
Peter Schmidt-Nielsen@ptrschmdtnlsn

I kind of think that Harvard should just target 2% grade inflation per year forever, and start giving out letters above A. It's fine if the average GPA is a 7.0 at Harvard in 30 years, so long there's still variation and therefore signal.

English
6
0
18
4.9K
Professor Eric Rasmusen
Professor Eric Rasmusen@erasmuse·
Interesting, but I'm confused. Did quantity demanded go down? (how much?) Then it's just elasticity being about equal to one. Ordinary the minimum wage reduces Q demanded and some workers get fired. Here it would be that it reduces Q demanded and all workers get lower quantity of work. Do relate these simple ideas to the entry-of-workers one.
English
2
0
1
55
Scott Barkowski
Scott Barkowski@ScottBarkowski·
@andy_garin @erasmuse You’d need both, not either or, right? If the supply curve is horizontal, it moves up with the min wage, traces out the shape of the demand curve. You’re estimating a unit elastic demand curve. So the story is gig workers are much more effected by min wage than other workers.
English
1
0
1
21
Andy Garin
Andy Garin@andy_garin·
@erasmuse In short, either a demand elasticity of 1 or infinitely elasticity labor supply could rationalize the result, but we provide some reasons we think free entry is operative
English
1
0
1
37
Andrey Fradkin
Andrey Fradkin@AndreyFradkin·
@paulnovosad But you could have submitted to RESTAT at #3. Also, as an economist, I take social science publications in Science / PNAS pretty unseriously because they often publish bad work.
English
2
0
7
923
Paul Novosad
Paul Novosad@paulnovosad·
Once more unto the breach, dear coauthors, once more
Paul Novosad tweet media
English
10
3
298
31.2K
Scott Barkowski
Scott Barkowski@ScottBarkowski·
@AA_Millsap @Jdavenport723 @JasonSorens No worries! This project is the most exciting research in econ in the last year, IMO. Of course, I think the health results are the most interesting! We owe Sam Altman a huge thanks for funding this! He spent at least $40mil.
English
1
0
2
31
Adam A. Millsap
Adam A. Millsap@AA_Millsap·
Another dagger for #UBI. Work fell, labor income fell, no impact on degree attainment. Money was used to subsidize leisure. We need to improve our safety net, but UBI isn’t the answer. nber.org/papers/w32719
English
33
81
382
62.3K
Scott Barkowski
Scott Barkowski@ScottBarkowski·
@AA_Millsap @Jdavenport723 @JasonSorens Actually they worked in Illinois to get a bill passed that kept the money being given from affecting public benefits including Medicaid, snap, TANF. In Texas there was some effect on benefits, but not for Medicaid. Pretty cool they were able to do that!
English
1
0
4
86
Adam A. Millsap
Adam A. Millsap@AA_Millsap·
Since this was a nongovernmental study I assume their eligibility for other benefits didn’t change. That said, the UBI may have crowded out other benefit programs if the extra income pushed them about the cutoffs. But if that happened its further evidence we shouldn’t just layer a UBI type benefit on top of the current system. The whole thing needs to be fixed so that it actually helps people get jobs.
English
3
0
7
2.6K
Scott Barkowski
Scott Barkowski@ScottBarkowski·
@MelissaDGentry Wow. Congratulations, Melissa. That’s a really fantastic outcome for you. Keep up the good work.
English
0
0
0
11
Melissa Gentry
Melissa Gentry@MelissaDGentry·
Excited to share two big announcements: I successfully defended my dissertation today, and I'll be returning to University of South Florida as an assistant professor in economics this fall!! So excited to be a Bull again! 🤘
Melissa Gentry tweet mediaMelissa Gentry tweet media
English
24
3
148
7.2K
Scott Barkowski
Scott Barkowski@ScottBarkowski·
@Smithdanj Also, to what extent are textbook discussions of hicks really discussions of Keynes? Should those hicks references be allocated to Cambridge?
English
0
0
1
16
Scott Barkowski
Scott Barkowski@ScottBarkowski·
@Smithdanj Or maybe where they did their NP cited work? Becker would probably change affiliations, for example. Or affiliation when getting the Clark medal. Still, it would probably understate the influence of some institutions.
English
1
0
1
19
Scott Barkowski
Scott Barkowski@ScottBarkowski·
Columbia’s rank seems too low. Before 1940 it was arguably 1 of the 2 most important USA depts with Harvard. Home of the institutional school. Friedman got his PhD there, Becker arguably did his most important work there. Stigler was there. Do all get credited to Chicago?
Daniel J. Smith@Smithdanj

Which modern economics department is most represented in history of economic thought textbooks? As measured by page mentions as percentage of total pages of over 70 HOET textbooks, published between 1901 and 2022, looking only at pre-2000 Nobel Prize winners, Chicago. All Souls College (Hicks) and George Mason University (Buchanan) punch above their weight.

English
2
0
4
254
Scott Barkowski
Scott Barkowski@ScottBarkowski·
What about Ken Arrow and Kuznets, who got their PhDs there? Does Columbia get “credit” for them at all? Classification is clearly important here.
English
0
0
1
39
Scott Barkowski retweetledi
Daniel J. Smith
Daniel J. Smith@Smithdanj·
Which modern economics department is most represented in history of economic thought textbooks? As measured by page mentions as percentage of total pages of over 70 HOET textbooks, published between 1901 and 2022, looking only at pre-2000 Nobel Prize winners, Chicago. All Souls College (Hicks) and George Mason University (Buchanan) punch above their weight.
Daniel J. Smith tweet media
English
12
32
211
27.3K

Keşfet

@ryancbriggs @cremieuxrecueil @feline_feelins @wwwojtekk @Tyler_Menzer @erasmuse @andy_garin @elonmusk