Friends, this is the promised thread about my and Anthony Fowler's just published:
Thinking Clearly with Data: A Guide to Quantitative Reasoning and Analysis
It has been a decade from conception to finished product. And we are very proud of it.
1/48
I don't agree with this thread.
Scientists are entitled to political and policy views. But calling those views scientific is confused.
Scientist's political & policy views, like anyone's, reflect facts and values.
They don't have any special claim to deference wrt values.
1/
Prof. Ethan Bueno de Mesquita has been appointed dean of the
@UChicago
Harris School of Public Policy, President Paul Alivisatos and Provost Katherine Baicker announced. Read more about
@Ethanbdm
's appointment at
@cblatts
My point is, this tweet is misinformation that gives the impression that we are ideologically closed faculty when we are not.
And young conservatives are reading it too.
Hi folks. A few months back I gave a talk at Caltech.
It was the best campus visit experience I've had in quite some time. And it has made me want to change campus visit norms a bit.
So today at lunch
@sgehlbach
told me I should tweet about it, which I'm now doing.
1/
I have some thoughts on the Chetty et al. papers and attendant public discussion.
First, the data collection is cool and the analysis interesting. This seems like valuable social science to me.
1/
@cblatts
We have also made enthusiasticc senior job offers that have been turned down (I don't think because of our uninviting ideological culture) to senior, public facing Republicans during your time at Harris.
I went to a frustrating informational session for parents of high schoolers getting ready to apply to college yesterday.
The person presenting went on and on about how schools no long care about standardized tests because "the data show they are not predictive."
1/
For those of you starting to think about your classes for next year, your occasional reminder that two of my books--Thinking Clearly with Data and Political Economy for Public Policy--have extensive teaching materials freely available online.
1/4
My (purely introspective, evidence-free) view on smoothing the classes to research transition for PhD students is this:
Encourage PhD students to be trying to write papers basically from day one.
The papers will (mostly) be bad. That 's fine. There are several benefits:
1/3
But, speaking as a visitor, I way preferred learning about everyone else's work than talking more about my own.
And I think if we all adopted the Caltech approach, we'd learn more and have more fun on/ our campus visits.
That's all.
8/8
Watching a group of economist attempt to correct an externalities-driven market failure with moral suasion alone, the humble political scientist can't help but wonder whether the behavioral turn in economics hasn't perhaps gone too far.
With the goal of promoting a fair and inclusive job market, the AEA is asking employers to consider following a specific timeline for interviews and flyouts for the upcoming job cycle. Seeย ย for more information.
@itaisher
@georgeawarner
@DavidDecosimo
@cblatts
In truth, the vast majority of what goes on among the vast majority of faculty and students at the vast majority of universities is the quiet, unglamorous, non-ideological work of learning, studying, and researching.
Their ongoing assault on us is neither deserved nor accurate.
The thing that I know now that I most wish I'd known when I started my PhD is:
If you dip the eggs in the boiling water for 10-15 seconds before cracking them, the whites don't spread out and you get perfect poached eggs every time.
So it turns out my professional duties are taking a somewhat unplanned turn for a while.
If you must select on the dependent variable or misuse "begs the question", your window of opportunity opens March 1.
We've been dreaming about making this happen for at lest the 14 years I've been at Chicago. We are just thrilled that it is finally a reality.
Please consider recommending our program to your promising students and advisees. We'll do right by them.
Big news: The Department of Political Science and the Harris School of Public Policy at the University of Chicago are launching a new Ph.D program in political economy. We have been working on this for over a year...I am thrilled to share some details. 1/
Life is full of trade-offs.
Reasonable people can disagree about how to balance them.
We ought not pretend that part is science.
An important lesson for *all* policy-adjacent researchers.
All I do is run a public policy school with a couple hundred employees and I cannot for the life of me figure out how presidents of the United States have time to golf.
For Holocaust Remembrance.
The painting that hangs in my living room.
By the artist Max Bueno de Mesquita (), one of the few members of my grandfather's family not murdered at Auschwitz.
The painting depicts mother and child going to the gas chamber.
For those of you starting to think about your classes for next year, your occasional reminder that two of my books--Thinking Clearly with Data and Political Economy for Public Policy--have extensive teaching materials freely available online.
1/5
Writing a serious external review takes a lot of time.
The only decision which is sufficiently important to merit asking for that much of 12 people's time is tenure.
Eliminate letters for non-tenure reviews, including junior both faculty reappointments and promotion to full.
Imagine it is August 1 and you aren't ready to teach.
Then you learn that lecture notes, in class activities, video lectures, and quizzes for my book Political Economy for Public Policy are freely available on my website.
What a good day it would be!
Following
@itaisher
's lead because I too can't reply.
The first and last tweets are in tension. Optimal and second best build in utilitarianism as the social goal.
There is no such thing as a value free optimal or second best policy.
You can't just follow the science.
Moreover, it is not great for science or society when scientists elide this distinction.
It makes others (both the policymakers they advise and the public) worry the scientists would fudge the evidence if it didn't suit their values.
2/
I find it more likely that a relatively unsophisticated algorithm, like selecting on grades and scores, but conditioning on hard-to-game markers of disadvantage like family income, first-gen status, or high school resources will better achieve the outcomes Fryer endorses.
21/21
@mpfix1
Here's a discussion of a graph we find misleading, and what we think you can learn from the misleadingness, from Chapter 15 of Thinking Clearly with Data.
Today I returned to that problem with fresh eyes.
And reader, I am here to report that I am still completely unable to make any progress on it at all.
And for the exact same reasons as before vacation.
Vacation is a scam.
3/3
Knowledge creation is slow, deliberate, and unglamorous.
Adam Davidson, perhaps unintentionally, explains why pushing academics to be ever more accessible, relevant, impactful, or public facing is corrosive to that mission.
Research and story telling are different endeavors.
So it is disappointing to me that the dictates of public facing social science result in over-interpretation of the sort others have flagged.
I think such over-claiming is not good for public discourse, for social scientists' role in society, or for future research.
14/
Moreover, this sort of public position makes certain people with values that differ from the author's worry that scientists won't truthfully report the facts about climate change.
I just don't see how science or society benefits from this sort of muddying of the waters.
5/5
That means, among admitted students at a selective school, there is typically going to be a negative correlation between test scores and grades.
Notice, this is exactly the opposite of the correlation between test scores and grades in the larger population.
4/
I'd like to believe that when people say we should teach statistics in high school they mean Thinking Clearly with Data, which would be a delightful high school textbook.
I'd also like to believe they are joking when they say it should replace calculus.
Alright AEA folks, here's a thread on eating in New Orleans.
I'm far from an expert. But I've had some great meals this week.
I focus on places not requiring reservations. You don't need me to tell you about Commander's Palace, Pesche, or places belonging to Top Chefs.
1/n
The Harris School is searching for assistant professors in political science and/or political economy. Open with respect to subfield or areas. Please apply and send us your best students!
EOE/Vet/Disability
@womenalsoknow
@POCalsoknow
@dylanmatt
Sorry to self-promote, by Fowler and I wrote wrote our book hoping it would be useful for people with exactly this goal.
We also taught this material in a course for journalists developed jointly with WBEZ, though the effort was interrupted by COVID.
I wrote some thoughts on the congressional hearings, free speech, leadership, and campus culture
@BostonReview
.
Much appreciation to Scott Ashworth,
@itaisher
,
@jcohen570
, and
@mlordlordm
who offered much needed feedback, pushback, and advice.
Delighted to find that
@dieworkwear
has solid causal inference intuitions.
Here we see him getting the fundamental problem of causal inference correct in its essentials.
@KenRiesz
In this specific case, the idea would be to encourage gun owners to secure their firearms. If the father's gun was secured, perhaps the shooter would have to go out and buy his own gun, or lie and tell his father he needed the gun for something else.
Would it have definitely
The 17 year old is heading to college in August.
So, to my friends
@VandyPoliSci
and
@VanderbiltEcon
, I am available to give a talk, attend a conference, or serve as a discussant absolutely any time.
Thus, if you find no relationship between test scores and college grades among *enrolled students*, that doesn't mean that there is no correlation between test scores and college grades among *applicants*, which is the predictive question college admissions cares about.
7/
This week in academic twitter we learned that among people who earned tenure, those who gave up tenure for a different job and those who kept their tenured job disagree on the value of tenure vs other job amenities.
Reassuring that everyone seems to be making good life choices.
We
@HarrisPolicy
are proud to launch our Public Sector Scholarship guaranteeing a minimum of $40,000 in scholarship to those with 3 years of public sector service. People dedicated to serving society should have access to the world's best policy education.
Every single person I met with showed up at the meeting and said, "I've got a paper I'm working on that I want to tell you about, do you mind if I write on the white board?"
I spent the day hearing 30 minute presentations of people's most interesting current ideas.
3/
This is wrong for many reasons.
But the most straightforward reason is the selection problem.
Schools admit on several criteria. For instance, if you have good test scores, you can get in with worse grades and vice versa.
3/
The person accepts a scientific fact about climate change.
Now suppose they believe "regulator restrictions on liberty can't be justified on consequentialist grounds."
Does the author believe that value judgment is unscientific? If not, what's science got to do with it?
4/
@itaisher
@georgeawarner
@DavidDecosimo
@cblatts
I don't disagree. But I sympathize w/ those who are fed up w/ being ridiculed by the likes of Bari Weiss, who self-servingly reports on a handful of cases to make universities seem like ideological trash heaps and herself (and friends) appear the sole defenders of truth seeking.
President Biden has unlimited access to the public attention.
It's easy to provide hard evidence of being sharp as a tack.
The unraveling theorem does the rest of the work.
My father and I co-authored a paper about how the political economy of Church-Lay Leader relations in medieval Europe was changed by the institutions created by the Concordat of Worms. I'll do a thread in the coming days, but for now here it is
@The_JOP
.
There may be good justice-based or other reasons for schools to go test optional. But doing so because tests are non-predictive among their enrolled students is a basic failure to think clearly with data.
9/
I have concerns, mostly related to skepticism about whether "sophisticated machine-learning algorithms" can solve hard social problems.
Many of these concerns are informed by the kind of ideas Anthony Fowler and I discuss in Thinking Clearly with Data.
8/
I believe university professors should strive to do good work (which these folks have definitely done) and talk about it in a straightforward manner.
And we should be prepared to live with it if that means big media outlets don't want to quote us.
15/15
This is a really nice paper. I'll describe my favorite result:
Suppose there's a minority of voters who observe policy choices and a majority of less informed voters who only observe something like their own allocation or outcome.
1/
๐ฃI'm on the Econ Job Market this year, and it's JMP time! ๐ค
My JMP "Voter Information and Distributive Politics" is available here ()
More details on my work can be found on my website:
@prisonrodeo
What if you've never used (or even heard of) tidyverse, but the only things you've done with R involve modifying (over and over again) a function called 'ethans_little_function' that
@WashUChancellor
wrote for you in approximately the year 2004?
The argument, from the thread, that science implies that it isn't ok for a person to conclude "climate change may be real, but I don't think we should have government regulation to deal with it," illustrates the point.
3/
I'd like to share an story that taught me something about research and from which, I flatter myself to think, my followers might also learn something.
For the past month, I struggled to prove a final result in a model I've been working on for years.
1/
That this is getting so much attention and is being retweeted by important academics is not great.
It is a straightforward implication of a simple spatial model of legislation, even if MTG is an extreme outlier.
It does not imply the average Republican agrees with MTG.
1/
This really says it all: Almost every single House Republican votes with Marjorie Taylor Greene *over 90%* of the time.
The idea that sheโs the โoutlierโ is just wrong. She IS the Republican Party. (h/t
@CAPAction
)
Anyway, as I say, not revolutionary. Just academics talking about ideas.
But I feel like, in these 30 minute meetings, we often think the polite thing is to ask the visitor about their work even though we are about to spend the seminar learning about their work.
6/
@liorjs
@UChicagoLaw
I did a lunch discussion with 50 MPP students yesterday: half our time on the sources of the gender-wage gap and half on the Israeli-Palestinian conflict.
Exact same experience.
People showed little model derivations, research designs, pictures of equilibria, etc. on the board.
It was awesome.
I learned what everyone was thinking about, asked lots of questions, got to offer thoughts or suggestions to the extent that I had them.
4/
IMPORTANT PSA!!!
LaTeX changed how \input works in a way that messes with direct import of Stata tables through estout.
SO STOP DOING EMPIRICAL WORK!!!
Or, alternatively, in your preamble, replace
\let\estinput=\input
with
\makeatletter
\let\estinput=\@
@input
Sometimes I tell junior faculty I was happier as an assistant professor.
They roll their eyes and say I'm remembering the past through rose-tinted glasses.
Then I explain to them that, when I was an assistant professor, I'd never heard of a Bartik instrument.
Here's the promised thread on the paper my father and I wrote together on the political, religious, and economic consequences of the resolution of the investiture controversy in the cocordats of Worms (and Paris and London).
1/
My father and I co-authored a paper about how the political economy of Church-Lay Leader relations in medieval Europe was changed by the institutions created by the Concordat of Worms. I'll do a thread in the coming days, but for now here it is
@The_JOP
.
It puts students in the right mindset.
1. They start thinking of idea generation as their main task.
2. They process the methods and ideas they're learning through the lens of "how could I springboard into my own ideas?"
2/3
@EpiEllie
Is your argument that there are no reasonable trade-offs to consider here? That any risk of COVID infection is lexicographically more important than any amount of the other stuff?
If not, what position is this meant to support?
If so, why do you think that?
If it's helpful as we enter another remote teaching scramble:
Slides, in class activities, data/code, and exercise solutions for two of my books are freely available:
Thinking Clearly with Data
Political Economy for Public Policy
3. It also can help with the classroom learning. For me, learning and internalizing technique went way faster and deeper when it was necessitated by a challenge posed by a paper idea rather than by a problem set.
3/3
Fancy machine learning algorithms do not solve these fundamental inference problems, no matter what their boosters claim.
Economists have been leading the discussion about such issues for decades and should continue to educate the public about them.
11/
This is a terrible loss.
There are few scholars I admire as I do Bob.
He was the clearest of thinkers. His papers are just a delight.
He worked on the most pressing and important problems, making deep, fundamental contributions.
And he was kind, generous, and encouraging.
My friend and colleague Bob Powell passed away today.
Bob was an unequalled scholarly force, an incredibly generous colleague, and an intellectual hero to many of us including me.
His memory will always be a gift to our field and to Berkeley. I will miss him immensely.
Everyone knows that the right way to find a research questions is to read the literature from a discipline 1-2 rungs down the prestige ladder from your own and answer their questions "the [insert your discipline here] way."
Rather, kids with higher test scores had lower high school grades on average and kids with lower test scores had higher high school grades on average.
If both are predictive of college grades in the population, there are offsetting effects in the sample of enrolled students.
6/
The negative correlation among enrolled students (which is who gets analyzed in a school's internal study) means that when you compare the college grades of students with higher and lower test scores, you aren't holding high school grades (or other admissions criteria) fixed.
5/