STA240 — Advanced Statistical Methods
Interactive learning resource
1 Welcome
This app covers the three core methodological blocks of STA240:
1.1 CS1: Linear Mixed Models
For clustered and longitudinal data — when observations are not independent.
- Random intercepts & slopes
- REML vs ML estimation
- Inference and diagnostics
1.2 CS2: Multinomial Logistic Regression
For unordered categorical outcomes with more than two levels.
- Baseline-category logit model
- Log-odds ratios & interpretation
- Model comparison & goodness of fit
1.3 CS3: Ordinal Regression
For ordered categorical outcomes — exploiting the ranking of levels.
- Cumulative link model
- Proportional odds assumption
- Testing & inference
2 How to use this resource
Each section follows the same structure:
| Block | Purpose |
|---|---|
| Concept | Mathematical formulation + intuition + interactive visualisation |
| Worked example | Fully solved — read and understand before proceeding |
| Guided practice | Fill in the gaps; hints available |
| Retrieval challenge | Solve without support; model answer revealed on demand |
| Why-questions | Target common misconceptions and deepen understanding |
| Spaced review | Short questions revisiting earlier material |
Code blocks are fully interactive — modify and run them directly in your browser. No R installation required.
3 Prerequisites
- Basic probability and statistical inference (hypothesis testing, confidence intervals)
- Simple and multiple linear regression
- Binary logistic regression
If any of these feel shaky, work through the concept blocks carefully — they are written to be self-contained.
4 Related courses
Covers classification (logistic regression, LDA, decision trees, random forests, XGBoost) and unsupervised learning (PCA, clustering). Same interactive format — concept blocks, quizzes, and browser-runnable R code.
