Dan Malinsky

US Latinos mobilize to monitor – and improve – local air quality: ‘We have to fix it’ As Latino communities ramp up air quality monitoring, Trump’s EPA moves to weaken pollution regulations

Inspiring to see communities doing both the science and advocacy needed to protect their neighborhoods, as the federal government rolls back environmental regulations

www.theguardian.com/us-news/2025...

by @thaliajuarez.bsky.social

www.aclu.org/news/free-sp...

www.theguardian.com/commentisfre...

Mediation Analysis in the Presence of Sample Selection Bias with an Application to Disparities in Liver Transplantation Listing The study of disparities in the liver transplantation process may focus on quantifying causal effects, particularly the average, direct, or indirect effects of various social determinants of health on...

In case you want to take a break from ~everything else going on~ to read a lil bit about causal mediation analysis, sample selection bias, and disparities in liver transplantation decisions... here's a new paper: arxiv.org/abs/2509.01969

Real “is there a doctor on this plane??” moment

Ooof can't believe I didn't catch that error on edit, my bad!

If a Leverhulme prize-winning philosopher calls it "philosophy" on the social constructionist account of "what is philosophy"...

I wrote something about statistics under authoritarianism

Authors briefly mention tetrad constraints but then move on to discuss goodness-of-fit tests (not reliable for this) and a more informal "plausibility" considerations rather than formal model search procedures with consistency guarantees

Curious how you think this connects to the large literature on discovering latent causal models from data (eg www.jmlr.org/papers/volum...) or at least testing the causal status of a hypothesized model rss.onlinelibrary.wiley.com/doi/full/10....

In addition to the below I think this may be helpful: academic.oup.com/ectj/article...

Do you know any papers that make the case for this well? Bc I think it is an ill-advised idea in general

Mohan and Pearl (and Tian) give results for various special cases of m-DAG models, but this work outlines (part of) a general identification theory more broadly applicable. Also I would say there are some philosophical differences

Causal and Counterfactual Views of Missing Data Models It is often said that the fundamental problem of causal inference is a missing data problem -- the comparison of responses to two hypothetical treatment assignments is made difficult because for every...

Agree! I think the view is expressed nicely here: arxiv.org/abs/2210.05558

This paper I contributed to may be of interest to both public health folks and philosophers of science — we outline a view of how to understand and operationalize “intrinsic health” as distinct from “no disease”

(ffo of words like “emergence” and “operationalize”)

Zohran's victory is a glimmer of light in a very dark time <3

ADD YOUR NAME: Sign the Open Letter in Support of NIH Staff Join the courageous and committed National Institutes of Health (NIH) public servants by adding your name now.

Support the brave staffers at @NIH! #BethesdaDeclaration actionnetwork.org/forms/add-na...

Join the AAUP and your dues go to support lawsuits like these against the Trump admin!

Now is the time for scientists to stand up against Trump’s repressive agenda | Daniel Malinsky The administration is attacking research, health and the environment. We might seem unlikely activists – but we have a duty to dissent

A powerful piece by Daniel Malinsky on why scientists must resist Trump’s authoritarian agenda:

“In these times … scientists have a duty to dissent.”

We completely agree. Ready to act? Join the #SummerFightForScience now (1/2)➡️

www.theguardian.com/commentisfre...

Graphical Models for Inference Under Outcome-Dependent Sampling We consider situations where data have been collected such that the sampling depends on the outcome of interest and possibly further covariates, as for instance in case-control studies. Graphical models represent assumptions about the conditional independencies among the variables. By including a node for the sampling indicator, assumptions about sampling processes can be made explicit. We demonstrate how to read off such graphs whether consistent estimation of the association between exposure and outcome is possible. Moreover, we give sufficient graphical conditions for testing and estimating the causal effect of exposure on outcome. The practical use is illustrated with a number of examples.

This article by Didelez et al talks about (non)-collapsibility of OR and outcome dependent sampling in an illuminating way, imo: projecteuclid.org/journals/sta...

Now is the time for scientists to stand up against Trump’s repressive agenda | Daniel Malinsky The administration is attacking research, health and the environment. We might seem unlikely activists – but we have a duty to dissent

I wrote something to say that scientists cannot afford to be 'apolitical'

Scientists actually have the skills and the duty to join in the fight against authoritarianism

www.theguardian.com/commentisfre...

My NIH-funded friends: please take 1min to post a public comment opposing this awful policy change that would expedite political interference in federal science funding

www.federalregister.gov/documents/20...

Good thing you were convinced by your judicious advisor to make this paper about fraud in “science” and not fraud in metaphysics so as not to enrage the wrong epistemic community 👀

👋

Tag yourself!

I'm the "NA" in the NIEHS budget line

OpenAI's "most powerful system" makes shit up more than half of the time. For its "o4-mini" model, the rate is ***79 percent***

Ezra Susser, CC ’74, Public Health ’82, Public Health ’92, professor of epidemiology and psychiatry at the Mailman School of Public Health, said that hearing that two of his grants were cancelled was “soul-crushing.” He said the cuts are affecting a “whole legacy of work.” Susser’s grants were for his research on the genomics of schizophrenia in South Africa and for a psychiatric epidemiology training program. “We had two more years, really, to finish it,” Susser said about his study on schizophrenia. “And we’ve been working on it for 10 years or 12 years, officially, but for at least 15 years, and all that could be lost.”

Decades of research that American taxpayers have invested in is being imperiled at universities like mine. Our paper wrote about some of the projects affected www.columbiaspectator.com/news/2025/04...

To Columbia’s trustees: Protect the future of higher education Even before coming to power this year, President Donald Trump’s administration made its plans for targeting American higher education abundantly clear. <a href="https://www.youtube.com/watch?v=0FR65Ci...

Columbia faculty (700+ signatures!) urges our Trustees to do 6 things:
1. Stand up for higher ed
2. Fight back against federal attacks (sue!)
3. Protect our students
4. Preserve academic freedom & fac governance
5. Affirm tenants of DEI
6. Protect research & provide backup funding
shorturl.at/6OLTJ

To Columbia’s trustees: Protect the future of higher education Even before coming to power this year, President Donald Trump’s administration made its plans for targeting American higher education abundantly clear. <a href="https://www.youtube.com/watch?v=0FR65Ci...

Columbia faculty (700+ signatures!) urges our Trustees to do 6 things:
1. Stand up for higher ed
2. Fight back against federal attacks (sue!)
3. Protect our students
4. Preserve academic freedom & fac governance
5. Affirm tenants of DEI
6. Protect research & provide backup funding
shorturl.at/6OLTJ