Measurement & Experimental Science
Sampling Hierarchy & Pseudoreplication
Many leaves measured on one plant are not the same as many independent plants, and many plants in one room may not be independent room replicates. The experimental unit must match the treatment assignment.
Key concepts
- Subsamples improve measurement of one experimental unit but do not create new independent replicates
- Treatments applied to whole rooms make the room an experimental unit
- Nested sampling should preserve plant, zone, and treatment identity
- Pseudoreplication can make evidence look stronger than it is
What to measure or observe
- State the experimental unit before collecting data
- Label subsamples by their parent unit
- Replicate treatments at the correct level
- Analyze or summarize data without pretending subsamples are independent
Common mistakes
- Calling ten leaves from one plant n=10 plants
- Treating two zones in one room as two independent rooms
- Pooling samples before recording identity
- Reporting a large sample count without the hierarchy
Visuals this lesson still needs
- Plant-leaf sampling hierarchy
- Room-level experimental unit diagram
- Pseudoreplication example
- Nested sample labels