Reciprocal health dynamics: a guided tour

This compendium documents, in order, every step from an empty data/ folder to the tables in the article. Each vignette is self-contained; together they are deliberately exhaustive.

1. Data acquisition with lissr. Registering, authenticating, browsing the archive, and downloading exactly the files the study needs.
2. Building the panel. The historical extract route that produced the published inputs, and the recipe-driven lissr route that replaces it going forward.
3. Sample selection with weasel. The minimum-three-of-seven-waves rule as an explicit scenario, with sensitivity and selectivity audits. Contains a fully executed demonstration on synthetic data.
4. Measures and cleaning. Construct definitions, the physical-activity plausibility ceiling, the education mapping behind the SES strata, and income equivalisation and cleaning.
5. Models and inference. The CLPM and RI-CLPM specifications, estimation settings, the boundary-constraint rationale for the model comparison, invariance checks, and the moderation test.
6. Results and reproduction. Every released table rendered from tables/, plus the exact reproduction ladder. Renders without any LISS data.

Chunks that require LISS microdata are marked eval: false and mirror the pipeline verbatim; everything else executes at render time.