Ordinal logistic regression

When to consider it

• Ordered categorical outcome; proportional-odds model.

Assumptions and scope

• Observations must be independent between subjects; repeated rows from one subject are not independent samples.
• Use integer outcome codes in their substantive order. Check proportional-odds plausibility and sparse categories.

Proportional-odds logit; assumption not automatically tested.

Required inputs

• Ordered outcome, predictors — Numeric observations; preserve the row and column structure described above.

Review missing values and data coding before running. Sample inputs are synthetic teaching examples.

Implementation

Outcome first: 3–10 integer categories in substantive ascending order, then numeric predictors. Proportional-odds logit; no predictor intercept. Threshold estimates use the statsmodels transform, not literal cutpoints. Assumptions are not automatically tested. Synthetic example: 240 simulated ordinal outcomes (0/1/2) and two continuous predictors.

• statsmodels 0.14.4: statsmodels.miscmodels.ordinal_model.OrderedModel

These are pinned library versions; saved results record the versions actually used at runtime.

Library references

• SciPy statistical methods
• statsmodels models and statistical tests
• scikit-learn clustering
• lifelines survival comparisons