Skip to main content

Reading list: some recent Ed evaluation and Stata articles/links of interest


Image from @AcademiaObscura
#EdEval #EdResearch #Stata
Quick roundup of some recent links to some education evaluation/research and/or Stata-related things I've been reading (or have bookmarked to read soon).

>>From the latest DeptEd NVCS 'data point' report:  "Students who reported repetition and power imbalance were components of the bullying they experienced were also more likely to agree that bullying had an impact on various aspects of their lives” https://nces.ed.gov/pubs2018/2018093.pdf

>> Matt Welch (AIR) says "Need to align #edeval criteria to standards for student learning #teachereval"
https://www.air.org/sites/default/files/downloads/report/Teacher-Evaluation-Common-Core-Alignment-October-2016.pdf

>> Chaisemartin and D’Haultfœuille have a new paper about relaxing the treatment effect homogeneity assumption in difference-in-difference analyses (dubbed Fuzzy DID).
Ungated version here: https://dornsife.usc.edu/assets/sites/1003/docs/fuzzy_did.pdf

>> Ratkovic and Tingley on direct estimation of TE following a spline regression. A highlight from this article:
"Our larger lofty goal is aligning machine learning, particularly high dimensional nonparametric regression, with causal inference, tightly and at a conceptual level."    (strikethrough edit/emphasis mine)
https://scholar.princeton.edu/sites/default/files/ratkovic/files/mde_18.pdf

>> Scott Cunningham at Baylor (author of -permute-) has updated his Stata-centric MS  "Causal Inference: The Mixtape" on his site http://scunning.com/stata.html (pdf accessible here:  http://scunning.com/cunningham_mixtape.pdf).   It's a great read and I appreciate that most of the examples have at least some of the Stata code to produce the results/figures and some annotated output/results. More importantly, it uses the 'tufte-latex' scheme (of which I'm a huge fan) and includes such jewels as this opening page in the chapter on DID:

Click to embiggen.

See also this work in progress: https://www.hsph.harvard.edu/miguel-hernan/causal-inference-book/


>> From Andrew Wheeler and co-authors, two studies using Dallas PD data to examine racial bias in decisions to shoot. Of course, these studies don't (cannot) control for decisions to actually unholster their weapons, but this spurs some interesting discussion about disparities in geographic policing and police call rates by race category and how they relate to officer incidents.



>>Simulating/visualizing the Central Limit Theorem with Stata!  https://marshalltaylordotnet.files.wordpress.com/2018/01/sdist_sj_rr2.pdf



Comments