Skip to content

PCM4: Cichlid Bayesian Linear And Phylogenetic Mixed Models

PCM4 introduces the cichlid parasite dataset and the first Bayesian mixed-model workflow in the course-derived evidence book.

flowchart LR
    A[Gaussian regression] --> B[Bayesian linear model]
    B --> C[Mixed model]
    C --> D[Phylogenetic random effect]

Why PCM4 Matters

PCM4 is where the repository has to be especially precise about what it does and does not reproduce. The core lecture depends on posterior sampling, random-effect uncertainty, and a phylogenetic inverse covariance passed into MCMCglmm.

Read PCM4 In Order

  1. Mixed-model evidence boundary

Checked-In Study Dossier

The full checked-in record lives under evidence-book/studies/cichlid-bayesian-linear-and-phylogenetic-mixed-models/.

Inputs And Evidence Map

  • lecture source copies: cichlid.csv, cichlidtree.nex, and fishdata.RData
  • catalogs: dataset catalog and provenance guide
  • evidence bundles: evidence-001 through evidence-007

What To Take From PCM4

  • the cichlid lecture family is fully documented and decomposed into governed evidence units
  • the Brownian PGLS comparator is useful context, but it is not a substitute for the lecture's posterior workflow
  • the main scientific boundary is posterior mixed-model reproduction, not basic study intake

What Not To Overread

  • PCM4 does not prove repository-owned replacement of the lecture's full posterior workflow
  • it does not treat nearby regression utilities as if they were MCMCglmm results
  • it keeps posterior summaries and random-effect uncertainty outside the current parity claim on purpose