Lucie Hazelgrove Planel and Roger Stern consider the importance of multilevel data in agricultural research. They discuss the concept of measurement units at different levels, such as individual, household, community, and village, and the challenges faced when dealing with multilevel data in practice. Using an example of a student’s experiment on tadpoles, they illustrate key principles in determining appropriate experimental design and data analysis.