Some lovely work by Peter Enns 👇 A major political science finding overturned by: 1) Simpson's paradox (i.e. collider bias) caused by 2) Dichotomization In this case, measurement & causal inference are closely linked--as they often are. medium.com/3streams/one...

One of the Most Prominent Findings in Political Science Might Be Wrong

Which means your voice matters more than you’ve been told!

medium.com

Interesting paper on how publication metrics distort research effort. The author exploits a reform in Thailand in 2019, which nudges researchers to publish in Q1 rank journals. Bunching at the cutoff shows that researchers follow incentives, at the expense of quality docs.iza.org/dp18680.pdf

Was today years old when i learnt that the second argument of the base mean() function in R is level of trimming, AND values outside the accepted range {0,.5} will default to the nearest endpoint. As such you CAN specify mean(x,2,na.rm=TRUE) and it will not give an error, and yours truly DID do that

This is a correct argument, but in sociology it is abused by ASA Council to prevent its journals from having policies for quantitative research.

Jess Calarco@jessicacalarco.com · 4mo ago

NEW from me in @pnas.org on the risks of data sharing with qualitative research: "Today, confidentiality risks are amplified by social media, surveillance, and AI tools, which make ample data about people publicly available, thus making deidentification less effective" www.pnas.org/doi/10.1073/...

I’m very happy to share that my first paper has been accepted for publication 🎉 Together with my great co-author Nayeon Lim we look at early public childcare in Germany. More details and the paper link below 👇 www.sciencedirect.com/science/arti...

Can early public childcare reduce child penalties? Evidence from Germany

This paper studies the effects of public childcare expansion for children under age three in Germany on mothers’ child penalties. Exploiting county-le…

sciencedirect.com

`grepl('^[A-Z]', commit_message)` is a simple indicator on whether the work was done by me or AI. If a git commit message starts with a lowercase letter, you can know for sure the work was done by my own fingers because I'm often too lazy/tired to press Shift when writing a commit message.

Dear Google, please allow us to take daily snapshots of Google Scholar!!

Matthew Archer@matthewarcher.bsky.social · 8mo ago

i might actually cry when they inevitably ruin google scholar

I could have sworn I created this before on our Previous Parish, but couldn't find it so made it fresh. I present: OUR BLESSED mixed models // THEIR BARBAROUS fixed effects

Two nice, among so many negative, things about screening abstracts for systematic reviews: 1) You sometimes run into some strand of literature or facts that you wouldn't otherwise have found 2) You sometimes find some gems that make your officemates ask you why you're chuckling all of a sudden

More cites for data and package development -> more arguments to convince funding bodies that collecting data and developing scientific software deserve funding on their own!

Hans Henrik Sievertsen@hhsievertsen.bsky.social · last yr.

Cite the data and cite the programmers!

Update: We submitted the comment to the Journal of Population Economics, and today we got a desk reject with this motivation. However, we do not know their conclusion about the main results and why the paper was not retracted. Clearly, they do not hold. 1/4

Johanna Rickne@johannarickne.bsky.social · last yr.

🚨 REPLICATION REPORT UPDATE: One year ago, a tweet by John Holbein alerted me, @ollefolke.bsky.social, and @jopieboy.bsky.social to a paper with a shocking result about Sweden’s law criminalizing the purchase of sex.🧵

I created a shiny web tool to play around with OL/TWFE so you can teach how this stuff can get so screwy. You can mess with temporal and cohort heterogeneity, treatment timing, whether you have any controls, etc. Share your worst plots! #econsky cannoncloud.shinyapps.io/TWFE_OLS_Pla...

When your statistical model "fits the data well" ...

Claim also doesn't hold up very well. Pager: firms that discriminated against Black & Hispanic applicants in audit study of hiring in a low-wage labor market were more likely than non-discriminators to be out of business 6 years later but 2/3 still around. sociologicalscience.com/articles-v3-...

Are Firms That Discriminate More Likely to Go Out of Business?

Article: Are Firms That Discriminate More Likely to Go Out of Business? | Sociological Science | Posted September 19, 2016

sociologicalscience.com

Trying to figure out how academic publishing works

New interview on the value of exploratory research. "My brief explanation of this crisis is that researchers didn't say ‘I found something interesting’, but ‘I found exactly what my fantastic theory had predicted, heureka’. And then they gave a TED talk." #PsychSciSky #AcademicSky 🧪

The crucial value of exploratory research » Eiko Fried

Exploratory research is vital for discovery, but often undervalued. Here I argue it is key to scientific progress & needs defending.

eiko-fried.com

Imperial propaganda if Warhammer 40k was set in academia 3/∞

Us college educated folks are so diverse in our cultural practices, right? Not when it comes to the high-stakes cultural practice of social reproduction: #parenting! We parent much more alike than parents with fewer years of #education: www.sciencedirect.com/science/arti...

Hidden patterns of inequality: The heterogeneity in parenting within educational groups

When sociology deals with differences within groups of similar socioeconomic status, research and theorizing tend to focus on the heterogeneity among …

sciencedirect.com

My hot take is that saying things that are simultaneously true, novel, and also non-trivial is incredibly difficult, regardless of whether you are a social scientist or a journalist.

I’ve long used FiveThirtyEight’s interactive “Hack Your Way To Scientific Glory” to illustrate the idea of p-hacking when I teach statistics. But ABC/Disney killed the site earlier this month :( So I made my own with #rstats and Observable and #QuartoPub ! stats.andrewheiss.com/hack-your-way/

🏆 Research integrity consultant and image forensics expert @elisabethbik.bsky.social has uncovered fraudulent data in over 7,600 scientific papers and exposed the practices of ‘paper mills’ that produce counterfeit scientific articles. She is honoured with the €200K Individual Award. Congrats!

*Please repost* @sjgreenwood.bsky.social and I just launched a new personalized feed (*please pin*) that we hope will become a "must use" for #academicsky. The feed shows posts about papers filtered by *your* follower network. It's become my default Bluesky experience bsky.app/profile/pape...

I know "generalized LATE for continous instruments" doesn't roll off the tongue as easily, but it would have saved me more time than I dare admit, trying to figure out the difference between MTE and marginal effects.

Paul Hünermund@p-hunermund.com · 2y ago

Not marginal effects, a generalization of LATE for continuous instruments. You can express other treatment effects (ATE, ATT, etc.) as weighted averages of MTE. www.nber.org/system/files...

I'm reading a lot about "data doesn't lie." But I'm an expert in statistics, and I say confidently that is incorrect. It should be "data don't lie."

Most academics are terrible mentors because we've just been lucky enough to have a random publication streak or enter the market at a better time

Here's to hoping that users of danish register data can soon say the same @dst.dk

Olle Hammar@ollehammar.bsky.social · 2y ago

👏 Good news for researchers working with Swedish register data and who want their research to be reproducible 👉 If you include in your ethics application/research plan that replication may be part of the project, then external replicators (e.g. data editors) can be given access to the data!

Ok sociology, what do you think are genuine breakthroughs that our field has made. Contributions that might convince skeptical but sympathetic *academics* (not the public) of the value of our field? I'll brainstorm some of mine in the thread - I treat sociology very broadly

❗New #ESRA Statement on Data Access and Open Science: "ESRA expresses serious concern about recent developments threatening the access to publicly available data and open science in the US with world-wide implications." www.europeansurveyresearch.org/esra-stateme...

ESRA Statement on Data Access and Open Science

The European Survey Research Association (ESRA) is the leading professional organisation of survey researchers in Europe, with members from the academic, government, non-profit and commercial sectors

europeansurveyresearch.org

Great thread! I’d add that in the social sciences, providing a secure computing environment for large and sensitive data is only possible because of indirects.

Elizabeth Wrigley-Field@wrigleyfield.bsky.social · 2y ago

Here are some facts about "facilities and administrative" (F&A) costs, what we in the business call "indirects" and what Musk is calling "overhead" as he tries to convince Americans with being ok with cutting billions on dollars from medical and public health research at universities & hospitals 1/

How cool is this?! A behind the scenes seminar series on how an economics paper came about - all the invisible work that never gets into the printed version. First event on 17th of February. I am also looking forward to using the recordings for teaching. www.bts-seminar.net