Abstract Correlations between past biotic diversity and climate can inform humanity’s response to predictions of future global climate change, e.g., extinction risk with global warming. Paleodiversity studies, however, frequently use genera as a proxy for species diversity, a practice that has often been questioned. Tests using actual data of the impact of using species-level versus genus-level taxonomy in paleodiversity-paleoenvironmental studies are also lacking. We conduct such a test, based on a recent study that showed a strong correlation of Cenozoic marine planktonic diatom species diversity to geochemical climate proxies. Using an updated version of the data set, we still find a strong correlation of Cenozoic diatom species diversity to environment. Using identical data but instead genera results in the loss of all significant correlations of diversity to environmental change. This occurs due to the earlier rise and later stability in genus versus species diversity data, a pattern known to be common between taxonomic ranks in the fossil record, and in general models of diversification. We conclude that studies of paleodiversity, particularly those addressing biotic responses to future environmental change, need to demonstrate the adequacy of genera as a proxy for species diversity, or use species-level data.