Show response level names for clmm models with the emmeans backend

[jmgirard]

For ordinal::clmm models, emmeans codes the response categories as 1, 2, 3 and the thresholds as 1|2, 2|3. So estimate_means(m, backend = "emmeans") showed these codes in the Response and Threshold columns, not the level names. For levels like 2, 4, ..., 10, the codes looked like real levels but were wrong.

.clean_names_emmeans_ordinal() now replaces the codes with model$y.levels, and with <level>|<next level> for thresholds. The threshold names come from the levels, because names(model$alpha) holds parameter names for threshold = "symmetric", "symmetric2" and "equidistant". The mapping runs only for clmm models, and only when the values are exactly the codes. Other models give the same output as before, and a new test checks this for polr, clm and glmmTMB.

test-brms.R:26 fails at random for me. It compares the medians of two sets of posterior draws from the marginaleffects backend, which this PR does not touch. This PR and #673 both bump the version to 0.17.0.10, so the PR that merges second needs a new bump.

[Copilot]

🟢 Approval recommended

All reviewed changes are covered and no approval-blocking issues were identified.

0 open findings

What changed in this PR

Updates the emmeans backend to show response and threshold names for ordinal::clmm models.

Changes:

• Maps response and threshold codes to readable level names.
• Adds ordinal-model regression tests.
• Documents the fix and bumps the development version.

File	Summary
tests/​testthat/​test-ordinal.R	Adds comprehensive ordinal-model tests.
R/​get_emmeans.R	Converts clmm codes to readable names.
NEWS.md	Documents the bug fix.
DESCRIPTION	Bumps the development version.

🧠 Review effort: Lite

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.