PyMICE Vignette Reports

Generated: 2026-07-08T17:12:06.615983+00:00 · pytest: passed

Python walkthroughs of the official R mice tutorials (V01–V08), with deterministic console checks and labelled visual/RNG differences.

Recommended learning path

#	Title	Notes
V1	Ad Hoc MICE	Start here — first imputation workflow
V2	Convergence Pooling	Convergence traces and pooling
V3	Missingness Models	Missingness patterns and models
V4	Passive Imputation	Passive imputation and post-processing
V5	Multilevel Data	Multilevel / clustered imputation
V6	Sensitivity Analysis	δ sensitivity and survival
V7	Ampute Simulation	Simulate missingness (appendix)
V8	Parallel MICE	Parallel imputation (appendix)

Status legend

Match — Console output matches the R reference.
Info — Informational or PyMICE-only (no R snapshot to compare).
Partial — Algorithm or plot differs in documented ways (often RNG or matplotlib).
Skip — R-only section (e.g. wall-clock benchmarks).
Mismatch — Output differs from the R reference — needs investigation.

Verification catalog

#	Title	Exact %	Parity status	Links
V1	Ad Hoc MICE	89%	Compliant (24/27 blocks match R)	HTML · MD
V2	Convergence Pooling	62%	Compliant (16/26 blocks match R)	HTML · MD
V3	Missingness Models	69%	Compliant (20/29 blocks match R)	HTML · MD
V4	Passive Imputation	35%	Compliant (8/23 blocks match R)	HTML · MD
V5	Multilevel Data	41%	Partially compliant — 23 match, 6 info, 26 visual, 1 skipped (R-only)	HTML · MD
V6	Sensitivity Analysis	41%	Compliant (9/22 blocks match R)	HTML · MD
V7	Ampute Simulation	67%	Partially compliant — 10 match, 1 info, 2 visual, 2 skipped (R-only)	HTML · MD
V8	Parallel MICE	42%	Partially compliant — 5 match, 6 info, 0 visual, 1 skipped (R-only)	HTML · MD

Test suite (pytest)

........................................................................ [ 28%]
........................................................................ [ 56%]
........................................................................ [ 84%]
........................................                                 [100%]