. Cochran-Armitage Trend Test

Purpose:

• Tests for a monotonic trend in proportions across ordered groups.
• Used when the response variable is binary (e.g., yes/no, success/failure).

When to Use:

• If you have a binary outcome and an ordinal predictor (e.g., different dose levels).
• Example: Does the proportion of people who recover from a disease increase as the drug dose increases?

2. Jonckheere-Terpstra Test

Purpose:

• Tests for a monotonic trend in continuous outcomes across ordered groups.
• More powerful than the Kruskal-Wallis test when a trend is expected.

When to Use:

• If you have a continuous outcome and an ordinal predictor (e.g., increasing drug doses).
• Example: Does blood pressure decrease with increasing drug dosage?

3. Cuzick Test (Trend Test for Medians)

Purpose:

• A variation of the Wilcoxon rank-sum test that assesses trends in medians across ordered groups.
• Similar to Jonckheere-Terpstra, but often used in epidemiology and medical research.

When to Use:

• If you need to test trend in continuous outcomes but with a simpler rank-based approach than Jonckheere-Terpstra.
• It is commonly used in survival analysis and biomarker studies.
• Example: Does tumor size decrease as treatment intensity increases?