PCM4 Mixed-Model Evidence Boundary

PCM4 deserves a direct boundary page because it is easy to over-read. The lecture uses a Bayesian mixed-model workflow whose main outputs are posterior summaries, convergence checks, and phylogenetic variance estimates.

flowchart TD
    A[Documented source lecture] --> B[Checked-in datasets and bundles]
    B --> C[Reviewable study boundary]
    C --> D[Posterior workflow still external]

What Is Source-Backed And Deployed

• the cichlid dataset and tree provenance
• the species-level summarization step used by the lecture
• the decomposition of the lecture into smaller evidence slices
• governed bundle IDs for each checked-in evidence unit
• the Brownian PGLS comparator named at the end of the lecture

What Is Not Yet A Repository-Owned Parity Claim

• Gaussian MCMCglmm posterior sampling
• chain-convergence diagnostics equivalent to the lecture
• governed posterior ledgers for phylogenetic random effects
• governed credible-interval summaries for the mixed-model parameters

What Readers Can Safely Conclude

• the repository has preserved the lecture family in a reviewable, checked-in structure
• the repository has made the dataset, decomposition, and evidence units inspectable
• the repository has not yet reproduced the lecture's posterior engine under a governed native or parity-equivalent claim

Why This Boundary Is Useful

This page separates two different questions:

1. Has the repository documented the lecture family well enough to review it?
2. Has the repository already reproduced the lecture's Bayesian posterior outputs under governed evidence?

For PCM4 today, the answer is yes to the first question and still no to the second.