# Documentation Data

## Ding Chapter Examples

Seven files in `ding/` are unmodified copies of Peng Ding's
[Replication Data for: A First Course in Causal Inference](https://doi.org/10.7910/DVN/ZX3VEV),
Harvard Dataverse, distributed under CC0 1.0. The Dataverse file IDs and published
MD5 checksums were verified on 2026-09-09.

| File | Dataverse file ID | MD5 |
|---|---|---|
| card1995.csv | 7440266 | 102b8856ed50ba53696879f0ed584917 |
| cps1re74.csv | 7440281 | 984ef72e2187423feedc018c0ddd92a8 |
| jobsdata.csv | 7440269 | e249939c9b58c51e52ea964482fd085b |
| nhanes_bmi.csv | 7440222 | 8e5d1e1c1dfb44ecc64d2de5cb0f857d |
| Penn46_ascii.txt | 7440282 | 84233c5c8a31ac08d235a49ceb117fc2 |
| resume.csv | 7440245 | a09b2afb2a0a6271c39a9bef98d60281 |
| star.dta | 7440219 | dff69b726e37eb1f44a31065aae6d348 |

Download individual files from `https://dataverse.harvard.edu/api/access/datafile/FILE_ID`.
Metadata: [Dataverse JSON export](https://dataverse.harvard.edu/api/datasets/export?exporter=dataverse_json&persistentId=doi:10.7910/DVN/ZX3VEV).

`ding/ZeaMays.csv` is the 15-pair Darwin maize dataset, exported from
`HistData::ZeaMays` by [Rdatasets](https://github.com/vincentarelbundock/Rdatasets/blob/master/csv/HistData/ZeaMays.csv).
See the [dataset documentation](https://vincentarelbundock.github.io/Rdatasets/doc/HistData/ZeaMays.html)
for attribution. SHA-256:
`d98c8de259364a8b5d4666fd01d2a4cadc440dd7a54bf1a562687e5c3a108f73`.
This CSV was byte-checked against the public export on 2026-09-09.

## Other Examples

The existing `basque.csv` and `california_prop99.csv` support the
[panel case studies](../ablations/same-root-panel-cases.html); their study
descriptions and source references are on that page. `harman_holzinger.csv`
and `small_msq.csv` support [sparse factor rotations](../examples/sparse-rotations.html).
The staggered-adoption vignette uses seeded synthetic data, not an external CSV.

Benchmark measurements and host metadata are kept separately in
`../ablations/data/` and are regenerated by `benchmarks/scaling/run_grid.py`.