Amanda Weiss

@amandaweiss

Assistant Professor, Cornell Department of Government. Political methodology, causal inference, ethics, and criminal justice policy. www.amandakweiss.com Opinions my own, reposts not (necessarily) endorsements, etc.

498 Following1,451 Followers353 Posts

Bluesky

Is this your profile?Claim it and amandaweiss.dirty.social is yours.Claim profile

Amanda Weiss@amandaweiss· May 4

I'm a quant methods professor and a millennial, I need someone to save me from myself.

Amanda Weiss@amandaweiss· May 4

Y'ALL I was reading up on the history of the Kitagawa-Oaxaca-Blinder decomposition, which is already wild. Then I discovered that THIS is how young Evelyn became Mrs. Kitagawa (even before becoming Dr. Kitagawa).

Wikipedia screenshot of a biography of Evelyn Kitagawa, which reads: "She was born as Evelyn Mae Rose, in 1920[4] in Hanford, California, to a family of Portuguese Catholic descent. After earning a bachelor's degree in mathematics from the University of California, Berkeley in 1941, she began working for the War Relocation Authority, which ran the internment camps of Japanese-Americans during World War II, as head of its statistics unit. In one of the camps, she met her future husband, Joseph Mitsuo Kitagawa, who had come to the US in 1941 as a divinity student and became an Episcopalian minister while interned. After marrying him, her family disowned her and she lost contact with them."

Amanda Weiss@amandaweiss· Mar 31

I made an over-engineered Qualtrics survey to handle my wedding RSVPs (name look-up, custom font packs, etc.) Now I'm like, Should I think of a survey experiment to run on family and friends?!

Screenshot of a Qualtrics survey collecting RSVPs for a wedding.

Amanda Weiss@amandaweiss· Mar 28

I just found out, via reviewer request, about AJPS' new submissions stream for reanalyses. I'm intrigued. Philosophers have written about creating groups of dedicated replication scientists (see, e.g., work by Felipe Romero) and I've wondered about how that model would work in poli sci.

Screenshot of a post from the American Journal of Political Science Editors-in-Chief, reading: "This post describes new policies at AJPS regarding Correspondence and Corrections.

Maintaining an accurate scholarly record is one of the most important responsibilities of a journal. When findings published in AJPS contain significant errors, the discipline is best served when those errors are identified, evaluated, and addressed publicly.

Until now, AJPS has not had a dedicated channel for publishing critical reanalyses by scholars other than the original authors. We are introducing one.

AJPS will now publish short critical essays called “Correspondence” in which scholars who are not the original authors challenge the central claims of a previously published AJPS"

Amanda Weiss@amandaweiss· Sep 22

Just dug up this banger of a @mattblackwell.bsky.social post to use in my instrumental variables slides. Pumped to be teaching the next generation about choosing estimands based on identification assumptions. 😂🙏

Amanda Weiss@amandaweiss· Sep 15

Okay, follow-up on the students' first grades of the semester. On the one hand: *So many* Calendly sign-ups. On the other hand: it feels like a huge teaching win that one of my students told me this morning that they were disappointed in their grade, but also understood it based on my comments.

Baby clenching handful of sand; image known as the "Success Kid" meme

Amanda Weiss@amandaweiss· Aug 13

What about the exclusion restriction? Vignettes were quite specific w.r.t. moving the target emotion more than other emotions. But image sets always moved at least one other emotion more than the target emotion (e.g., anger images had big effects on disgust). AEMTs were in the middle. (11/17)

Coefficient plot comparing the effects of different anger manipulations on anger to those manipulations' effects on related emotions.

Coefficient plot comparing the effects of different fear manipulations on fear to those manipulations' effects on related emotions.

Coefficient plot comparing the effects of different gratitude manipulations on gratitude to those manipulations' effects on related emotions.

Amanda Weiss@amandaweiss· Aug 13

We were wrong about the strength of more personalized instruments! Vignette instruments had the largest treatment effects on emotions that we studied, except for anger. Images (we looked photographs from affective image inventories) tended to work well, too. (9/17)

Grid display of 6 coefficient plots showing the effects of different emotion manipulation instrument types when used to manipulate six different emotions.

Amanda Weiss@amandaweiss· Aug 13

🚨 Updated working paper! Ekin Dursun and I ask what instruments best manipulate emotions on surveys (osf.io/56h4g). We find that vignettes really work! They have large effects on emotions of interest & smaller effects on emotions *not* of interest. But as always, it's complicated.👇 (1/17)

Abstract of a paper written by Amanda Weiss and Ekin Dursun and titled "Robust Emotion Manipulation for Surveys: Evidence from Three Experiments."

The abstract reads: "A large number of experiments investigate the effects of emotions on critical political outcomes, including policy attitudes, support for authoritarians, tolerance, and political participation. The success of these experiments depends on emotion manipulations: Manipulations must be strong enough to shift one target emotion while also being specific enough shift other confounders only minimally. In this project, we identify emotion manipulations that fulfill these imperatives. First, using causal graphs, we show that in such experiments, emotions are intermediate outcomes of randomly assigned emotion manipulation instruments—not randomized treatments themselves. Then, we present evidence from three experiments (total N = 6, 649) on the effectiveness of vignettes, autobiographical emotional memory tasks, images, and more for inducing anger, gratitude, fear, political anger, political gratitude, and political cynicism. We show that vignettes are reliable instruments in terms of both strength and specificity. We also investigate compliance with emotion manipulation instruments and find that pre-treatment attitudes toward research may moderate treatment effects.

Amanda Weiss@amandaweiss· Aug 4

Dear Ithaca NY, I thought I was an okay runner. But I did not know your game. Just some cute path and then BOOM 15% GRADE. I have been suitably chastened. Respectfully, Amanda

Recreational path surrounded by tall trees in upstate New York.

Amanda Weiss@amandaweiss· Jul 29

New metaphor for the relationship between nonparametric and parametric statistics just dropped @lauretig.bsky.social

Screenshot of two Bluesky posts. The first reads "The bootstrap is basically magic." The second reads "Survival models are basically black magic. I'm convinced only a few people understand them, and they're all in Epi or Biostats departments."

Amanda Weiss@amandaweiss· Jul 21

Polmeth @ Emory this past week was so great (research thread to follow) - but the real highlight of the trip may have been the horse parked on the sidewalk in downtown Atlanta.

Very impressive and ginormous horse on the sidewalk in downtown Atlanta, tied to a tree.

Amanda Weiss@amandaweiss· Jul 14

Oh there are so many layers. U.S. culture wars are so complicated. Under the circumstances, it's completely reasonable that the ad that Buzzfeed served me was for moving to Germany.

Screenshot of Buzzfeed news article about M&M candies and the culture wars, showing an advertisement for "Jobs in Germany."

Amanda Weiss@amandaweiss· May 20

I'm not on Instagram - but I guess I made an appearance!

Screenshot of a graduation day post, with a picture of a woman in PhD regalia, from the the Yale Graduate School of Arts and Sciences Instagram account.

Amanda Weiss@amandaweiss· May 8

'Leo' is the 12th most popular baby boy name is Illinois and 18th most popular nationally, as of the most recent SSA data. So the new pope is right on-trend.

Chart from the Social Security Administration showing the most popular baby names in Illinois as of 2023.

Amanda Weiss@amandaweiss· Mar 24

Dear Yale Diploma Office: I am honored, but...

Screenshot of an email, which reads:
"Dear Amanda,
Your Diploma name and address are now available for your review. If you do not confirm your diploma name, it will be printed with the name we have on file. Reprints will not be made...
Diploma Name: AMANDA WEISS
Degree(s): Doctor of Philosophy
Department(s): MD/PhD Program, Immunobiology"

The line "MD/PhD Program, Immunobiology" is circled in red because the recipient of the email is not doing an MD/PhD in Immunobiology.

Amanda Weiss@amandaweiss· Mar 10

I submitted my dissertation, so it's official! Graduating is bittersweet... But I could not be more excited to join Cornell as an Assistant Professor of Government this fall!

Screenshot of a Yale University Graduate School of Arts and Sciences webpage showing that Amanda Weiss has submitted her dissertation.

Amanda Weiss@amandaweiss· Feb 20

Franklin Delano Roosevelt "Frankie" Weiss is ready for the cold snap.

Tiny white-and-black dog wearing a striped turtleneck sweater, standing next a brick path.

Amanda Weiss@amandaweiss· Feb 12

This obituary has more bangers than my mind can properly comprehend.

Quote from an obituary of Donald Shoup: "To the extent homo sapien falls short of homo economicus, the politics of pricing curb parking are bleak. We struggle to make sense of the time value of money and hate paying for something we had thought of as free—even if doing so would improve our lives. (Not to imply that you should pay for sex.)"

Amanda Weiss@amandaweiss· Feb 9

I'm reminded of coverage of Peter Thiel and J.D. Vance's relationship during the campaign. If you don't believe that women or members of minority groups can contribute to public life, you may well think random dudes are necessarily better than whoever currently runs government/the Fortune 500.

Block quote from a Slate article by Ben Mathis-Lilley. The quote reads:

"Since he was nominated for the vice presidency, Republican Ohio Sen. J.D. Vance has been in the news for his history of comments about childless women and his endorsements of figures on the right who have said (among other things) that the Sandy Hook massacre was staged, that Jan. 6 was staged, and that the United States should become a dictatorship.

There’s “a lot going on there,” as people say, which perhaps explains why there’s been relatively little attention on this old chestnut of a take by Peter Thiel, the venture capitalist and Republican megadonor who has been a mentor figure to Vance for years. (According to Vance, a 2011 speech Thiel delivered at Yale inspired him to become involved in the conservative movement; Thiel later hired Vance to work at one of his companies and unilaterally made him a viable Senate candidate in 2021 by giving $15 million to a super PAC that supported him.) Wrote Thiel:

"The 1920s were the last decade in American history during which one could be genuinely optimistic about politics. Since 1920, the vast increase in welfare beneficiaries and the extension of the franchise to women—two constituencies that are notoriously tough for libertarians—have rendered the notion of “capitalist democracy” into an oxymoron.""

Amanda Weiss@amandaweiss· Feb 4

Having a resist-lib moment.

Screenshot of a rationale for cancelling her subscription to the New York Times, which reads "Ten or so years ago, this subscription was one of my first "real adult" purchases. But NYT coverage of the White House and the takeover of the federal government by Elon Musk is so pusillanimous as to be an abdication of the responsibility to inform. Once the Times stopped covering the news of the day - which today is the destruction of American democracy - I could no longer support the Times."

Screenshot showing a new subscription to Wired magazine.

Amanda Weiss@amandaweiss· Jan 31

E.g.,

Perma-linked Census Bureau webpage providing data on metropolitan statistical areas from December 2009.

Amanda Weiss@amandaweiss· Dec 3

This is some NYT-level understatement. 😂 One would hope.

Headline: "South Korean president's failed attempt at martial law may put his position in peril"

Amanda Weiss@amandaweiss· Nov 5

3rd time voting for president, 1st time doing so in person. Sometimes it hits me anew what a privilege it is to be able to vote for the person who is going to hold the most powerful office in the world.

Woman wearing an "I Voted" sticker outside of a polling place.

Amanda Weiss@amandaweiss· Sep 9

This life stage has a steady stream of wedding & birth announcements. Exciting! I would also like to participate in happy announcements! I will do updates like: It’s a fluffy doggie! 🐾 Amanda is excited to announce that Franklin Delano Roosevelt "Frankie" Weiss has reached a new level of cuteness.

Amanda Weiss@amandaweiss· Sep 6

Come to our APSA panel data panel Friday at 12pm! I'm presenting "How Much Should We Trust Modern Difference-in-Differences Estimates?" on alongside 4 other exciting papers.

APSA Schedule for the Panel "Challenges in Learning from Panel Data"

Amanda Weiss@amandaweiss· Aug 29

(2) All DID-style methods, including TWFE, are at least somewhat underpowered to study many policy roll-outs in the US context. The problem is more severe for more data-hungry, robust modern methods.

Graph showing the power available with different DID methods under different outcome models and parameter combinations.

Amanda Weiss@amandaweiss· Aug 29

(1) Many modern methods produce confidence intervals that do not include the true average effect at the nominal rate.

Graph showing coverage probabilities of different DID estimators under different outcome models and parameter combinations.

Amanda Weiss@amandaweiss· Aug 29

🚨 Job Market Paper alert! "How Much Should We Trust Modern Difference-in-Differences Estimates?" at osf.io/bqmws I assess the fit between modern DID-style estimators and real-world data features, and point to approaches for improving inference. Quick thread ⬇️

Abstract of paper on difference-in-differences methods.

Amanda Weiss@amandaweiss· Apr 3

I love that the state is not submerged in @cdechaisemartin.bsky.social et alia's new R package. (Big shout-out to DIDmultiplegtDYN, and the EU for helping make it happen.)