←
Nick HK17,171 posts
Nick HK avatar

Nick HK

@nickchk

Econ prof @SeattleU. Book The Effect theeffectbook.net out now! Check my pinned thread for all my projects. Substack nickchk.substack.com/

◷ Joined October 2010▧ 1,200 Photos and videos
Nick HK avatar
Nick HK@nickchk·3 Jul 2019

This is a thread of me QTing myself so I can pin all my posted-on-Twitter projects. Causal inference animated graphs:

Nick HK@nickchk29 Nov 2018

As requested, slower graphs! Also added a graph on collider bias, the webpage explanation helps there. These graphs are intended to show what standard causal inference methods actually *do* to data, and how they work. This is what controlling for a binary variable looks like:

Statistical Horizons avatar
⟳ Nick HK retweeted
Statistical Horizons@StatHorizons·Mar 13

"Causal Inference in Econometrics" with @nickchk is on demand! Join April 13-May 11 to learn the how & why of #econometric analysis of observational data using regression (how econometricians evaluate data) & identification (how they link theory to data) at your own pace. statisticalhorizons.com/semi…

Nick HK avatar
Nick HK@nickchk·Jan 22

I am teaching my Causal Inference in Econometrics asynchronous workshop with Statistical Horizons from April 13-May 11 of this year. Learn about the tools of causal inference and go at your own pace. Get more info, and sign up if you're interested, here: statisticalhorizons.com/semi…

Nick HK avatar
Nick HK@nickchk·8 Jul 2025

I got an exciting box in the mail today! Looks like The Effect: Second Edition is live! I've updated the website as well, so you can check out the changes and the new Partial Identification chapter (+ a new mini-chapter on finding material in the book) at theeffectbook.net

Post attachment
Nick HK avatar
Nick HK@nickchk·1 Apr 2025

The second edition of The Effect is now available for preorder! This version has a whole new chapter on Partial Identification, a considerable update on staggered treatment and control variables in DID, and zillions of other little updates throughout. routledge.com/The-Effect-An-…

Nick HK avatar
Nick HK@nickchk·25 Feb 2025

The working paper for the Many-Economists project, which got a lot of juice from this very platform, is available! The Sources of Researcher Variation in Economics looks into researcher degrees of freedom and tries to isolate which ones are important. Data cleaning!

The Many-Economists project with @nickchk, The Sources of Researchers Variation in Economics, is now on SSRN. We had 146 teams perform the same research task three times, each with less freedom, to explore what drives differences across researchers. papers.ssrn.com/sol3/papers.…

Nick HK avatar
Nick HK@nickchk·17 Dec 2024

btw this was a twitter-originated collab

I have been working on something similar but I hadn’t thought to ask an LLM. We have DAGs for our data generated by clinical experts, DAG-trained experts, and ML. Wanna collab on getting an LLM to build one?

Nick HK avatar
Nick HK@nickchk·17 Dec 2024

But if you think we've missed a trick or want to try this again when the next generation comes out, our code and data are very easy to update and are at osf.io/spzbu/

Nick HK avatar
Nick HK@nickchk·17 Dec 2024

LLMs are consistently changing, and who knows what comes next, but for now this is not a task the mass-market LLMs are capable of.

Nick HK avatar
Nick HK@nickchk·17 Dec 2024

Beyond that, the LLM results are very inconsistent. Which prompting methods generate the best results varies significantly across model and setting. And even minor things like "the order of the multiple choice options" greatly changes the results.

Post attachment
Nick HK avatar
Nick HK@nickchk·17 Dec 2024

So if you were planning to do this you'd need to bring in the experts anyway to actually end up with a model, and at that point what work are you even avoiding.