Lightnews — Scholar-powered news

yiqingxu.bsky.social @yiqingxu.bsky.social · Aug 25

Thanks! Please send us suggestions if you have any!

yiqingxu.bsky.social @yiqingxu.bsky.social · Aug 25

We also highlighted the need to use the leave-one-out approach to obtain pretrend estimates, thanks to Zikai and Anton's cautionary note: osf.io/ngr3d_v1/ @astrezh.bsky.social

2

yiqingxu.bsky.social @yiqingxu.bsky.social · Aug 25

In this new version, we fixed many bugs and introduced more features.

Check it out: yiqingxu.org/packages/fect/

Special thanks to Ziyi, Rivka, and Tianzhu for their incredible work and dedication. More features are on the way!

1 2

yiqingxu.bsky.social @yiqingxu.bsky.social · Aug 25

Thrilled to share that **fect** has won the 2025 Best Statistical Software Award from the Society of Political Methodology. We're honored!
polmeth.org/statistical-...

To celebrate, we've just released fect v2.0.5 on CRAN & Github 🎉

1 6 26

yiqingxu.bsky.social @yiqingxu.bsky.social · Aug 25

Honored by this recognition!

Political Analysis @polanalysis.bsky.social · Aug 8

We are pleased to announce the 2025 Editors’ Choice Award for the paper “How Much Should We Trust Instrumental Variable Estimates in Political Science? Practical Advice Based on 67 Replicated Studies” by @apoorvalal.com‬, @maclockhart.bsky.social, @yiqingxu.bsky.social‬, and @garyzu.bsky.social.

3

yiqingxu.bsky.social @yiqingxu.bsky.social · Jun 9

Glad this paper is finally out in @apsrjournal.bsky.social, five years after we began the project. Hope it proves useful to researchers.

www.cambridge.org/core/journal...

Software support: yiqingxu.org/packages/fec...

Causal Panel Analysis under Parallel Trends: Lessons from a Large Reanalysis Study | American Political Science Review | Cambridge Core

Causal Panel Analysis under Parallel Trends: Lessons from a Large Reanalysis Study

www.cambridge.org

4 11

yiqingxu.bsky.social @yiqingxu.bsky.social · Apr 27

Hi Vincent, thank you! Yes, we're working on a user manual that will incorporate these datasets. It will take a few weeks!

1

yiqingxu.bsky.social @yiqingxu.bsky.social · Apr 5

As a Chinese immigrant living in the U.S., I have witnessed declines in optimism purely due to the drastic erosion of existing institutions—first around 2015 in my home country, and now, in 2025, in the United States. It's devastating.

1 1 7

yiqingxu.bsky.social @yiqingxu.bsky.social · Apr 3

7/ We have prototypes for all the algorithms used in the Element and will soon roll them out in **interflex** for R.

Comments and suggestions are more than welcome!

1

yiqingxu.bsky.social @yiqingxu.bsky.social · Apr 3

6/ Through simulations & many empirical examples, we find AIPW-Lasso & PO-Lasso outperforming competitors with political science data, while DML requires much larger samples to do better.

1

yiqingxu.bsky.social @yiqingxu.bsky.social · Apr 3

5/ We then turn to double machine learning (DML), incorporating modern learners such as neural nets, random forests, and HGB to estimate nuisance parameters and construct signals.

1

yiqingxu.bsky.social @yiqingxu.bsky.social · Apr 3

4/ The core of the Element is robust estimation strategies.

First, we adapt AIPW-Lasso & Partialing-Out Lasso, both w/ basis expansion & Lasso double selection.

We walk through signal construction step-by-step to aid intuition. For smoothing, we support both kernel and B-spline regressions.

1

yiqingxu.bsky.social @yiqingxu.bsky.social · Apr 3

3/ We start by defining the estimand, the CME, and presenting main identification results in the discrete-covariate case.

We then review & improve the semiparametric kernel estimator. The improvements include fully moderated models, adaptive kernels, and uniform confidence intervals.

1

yiqingxu.bsky.social @yiqingxu.bsky.social · Apr 3

2/ We prepared it for Cambridge Element & previewed parts of it in our response to a blog post last month.

Scholars are often interested in how treatment effects vary with a moderating variable (example below).

We hope this will serve as a useful reference for this common task down the road.

1

yiqingxu.bsky.social @yiqingxu.bsky.social · Apr 3

Draft “A Practical Guide to Estimating Conditional Marginal Effects: Modern Approaches” is on arXiv: arxiv.org/pdf/2504.01355

w/ two amazing grad students, Jiehan_Liu & Ziyi Liu 🧵

2 12 42

yiqingxu.bsky.social @yiqingxu.bsky.social · Apr 3

5/ We then turn to double machine learning (DML), incorporating modern learners such as neural nets, random forests, and HGB to estimate nuisance parameters and construct signals.

CME Estimates from AIPW-Lasso vs. DML Methods

yiqingxu.bsky.social @yiqingxu.bsky.social · Apr 3

4/ The core of the Element is robust estimation.

First, we adapt AIPW-Lasso & partialing-out Lasso (PO-Lasso), both w/ basis expansion & Lasso double selection.

We walk through signal construction step by step to aid intuition. For smoothing, we support both kernel and B-spline regressions.

Raw data and signals constructed using outcome-modeling, IPW, and AIPW.

1

yiqingxu.bsky.social @yiqingxu.bsky.social · Apr 3

3/ We start by defining the estimand, the CME, and presenting the main identification result.

We then review & improve the semiparametric kernel estimator by introducing fully moderated models, adaptive kernels, and uniform confidence intervals.

1

yiqingxu.bsky.social @yiqingxu.bsky.social · Apr 3

We prepared this for Cambridge Element and previewed parts of it in our response to a blog post last month.

Scholars are often interested in how treatment effects vary with a moderating variable (example below). We hope this Element will serve as a useful reference for such tasks down the road.

Example: Plot of (conditional) marginal effects against a moderator/modifier.

1

yiqingxu.bsky.social @yiqingxu.bsky.social · Feb 26

I woke up at 5:30 a.m. today, and the first thing on my mind was Stefan Zweig’s The World of Yesterday: Memoirs of a European, a book I loved as a child.

His despair stuck with me and feels especially relevant now.

1 3

Reposted

Dean Eckles @eckles.bsky.social · Feb 24

"The bottom 80% of earners spent 25% more than they did four years earlier, barely outpacing price increases of 21% over that period. The top 10% spent 58% more."
www.wsj.com/economy/cons...

1 11 29

yiqingxu.bsky.social @yiqingxu.bsky.social · Feb 21

We streamlined six new DID-like estimators and created this tutorial for implementation in R.
yiqingxu.org/packages/fec...

Hope you no longer need to spend months figuring out what these estimators are and how to use them.

6 98 330

yiqingxu.bsky.social @yiqingxu.bsky.social · Feb 11

Hah, thanks. You're the reason I'm here, Guilherme

1

yiqingxu.bsky.social @yiqingxu.bsky.social · Feb 11

8/ Finally, we thank Professor Simonsohn for his thoughtful critique, which brings renewed attention to the estimation and testing of conditional relationships.

w/ Jens Hainmueller, Jiehan Liu, Ziyi Liu & Jonathan Mummolo

yiqingxu.bsky.social @yiqingxu.bsky.social · Feb 11

7/ Based on the discussion, we propose a set of recommendations, starting from clearly stating the quantity of interest and the key ID & modeling assumptions.

We provide software support for various estimation strategies: yiqingxu.org/packages/int...

1 1 1