You can read find more about details at www.journals.uchicago.edu/doi/10.1086/... Thanks to all that helped along the way!

Omg I did this in graduate school and it was the worst.

Met @annaleen.bsky.social and got to spend some good time talking science and spec-fic and they were even more amazing and brilliant than I expected. They had just thought deeply about such interesting and important things and knew how to say things in exactly the right way.

📌

If your research involves RD designs, check out this important new working paper from Ghosh, Imbens, and Wager: "PLRD: Partially Linear Regression Discontinuity Inference" arxiv.org/pdf/2503.09907

Watching Jurassic Park, where a computer nerd with a debt problem and delusions of grandeur tears down all the safety systems, with no understanding of the consequences, so he can better facilitate his planned espionage and theft.

If Vladimir Putin had a plan to foul our air and water, wreck public health and drive America over the cliff of irreversible lethal climate change, it would look exactly like Lee Zeldin’s plan. This is a plan for self-inflicted environmental disaster.
www.theguardian.com/us-news/2025...

Here's what that image was supposed to look like:

We went for roads since it's easy to see how that measurement error could arise. Often we have no idea why RS + ML errors occur

@bstewart.bsky.social and co-authors explore the same issue in text measurement models like LLMs and find something similar--even small measurement errors can lead to large biases in downstream causal tasks when they aren't orthogonal to treatment

@sandysum.bsky.social and co-authors test a different method to correct for the same source of bias across many remotely sensed variables and ground truth data and find this everywhere (www.nber.org/papers/w30861)

Imagine you run a land tenure reform RCT, DV is tree cover. It turns out your treatment also causes more irrigated ag, which is mis-classified as treecover more often than rainfed ag (year-round greenness). Estimated treatment effect will be > that true treatment effect.

While that's our running example for the paper, definitely a broader issue here. We think assuming no correlation between measurement error and treatment is akin to the selection on observables assumption we usually require extraordinary evidence to believe. A couple examples below:

@jonproctor.bsky.social @vivianodavide.bsky.social @bstewart.bsky.social and others I can't find tags for

#css #econometrics #EconSky #poli-sci #eo #remotesensing

9/9
Other great work in this area: www.nber.org/papers/w30861, arxiv.org/abs/2501.18577, arxiv.org/abs/2411.10959, arxiv.org/abs/2306.04746 focus on “predict-then-debias”—the right move if using off the shelf data. But if you’re training the ML model yourself, give our adversarial approach a try!

8/9
Reach out if you want to debias some measurements in a particular application!

7/9
It’s easy to plug in any causal variable that might bias your ML-driven proxy. The adversary directly leverages your labeled data—so if you’re building custom measurement models with large-scale images (or text), you just tack on the adversary, retrain, and your bias vanishes.
8/9

6/9
We then use a labeled forest cover data from high-resolution imagery. When comparing the ML predictions to ground-truth labels, a naive model under-estimates forest cover near roads. Our adversarial model, by contrast, recovers unbiased estimates, giving more reliable coefficients.

5/9
We induce measurement error bias in a simulation of the effect of roads on forest cover. We show that a naive model yields biased estimates of this relationship, while an adversarial model gets it right.

4/9
We also introduce a simple bias test: regress the ML prediction errors on your independent variable. If nonzero, you have measurement error bias. If you run that test while gathering ground-truth data, you can estimate how many labeled observations you’ll need to reject a target amount of bias.

3/9
Here’s how: a primary model predicts the outcome, while an adversarial model tries to predict the treatment using the prediction errors. As the adversary learns how to predict treatment, the primary model learns to make predictions where the errors contain no information about the treatment.

2/9
We are inspired by the algorithmic fairness literature. There, adversarial models force models to have balanced prediction errors across the dist. of a protected attribute (e.g. race). We adapt that approach, but instead of race, our “protected attribute” is the independent variable of interest.

1/9
We are excited to share our new working paper:
arxiv.org/abs/2502.12323
If you use ML predictions (like remote-sensed data) as outcomes, the resulting regression coefficients can be biased by measurement error. With @megan-ayers.bsky.social @mdgordo.bsky.social @eliana-stone.bsky.social