Here, we present a protocol for studying plant growth behavior and especially phenotypes in a reproducible manner. We show how to provide variable and at the same time stable light conditions. Proper analyses depend on sufficient sample numbers and valid statistical evaluations.