Designing a controlled experiment

A controlled experiment changes an independent variable and measures a response while managing other important factors. Random assignment reduces systematic group differences. Replication helps estimate variation and reduces reliance on one unusual observation. To test a fertilizer, use comparable plants, randomly assign treatment, and keep light and watering comparable. Include a control group without the tested fertilizer. If treated plants also receive more light, the fertilizer's effect is confounded with light.