StellarRandomized Block Design
A randomized block design reduces the effect of known variation.
Why use blocks?
- When there is a variable (like gender, age, pre-existing conditions) that you know will affect the response
- Blocking on that variable reduces variability within groups
- Makes it easier to detect the effect of the treatment
How it works:
- Divide units into blocks based on a characteristic (e.g., block by sex: male/female)
- Within each block, randomly assign treatments
- Analyze the results for each block separately, then combine
Example: Testing a new teaching method on high school students — block by grade level (9, 10, 11, 12) because academic performance differs by grade. Within each grade, randomly assign students to new vs. traditional method.
Block