Abstract Anthropologists often assert that modern hunter-gatherer societies have been relegated to marginal habitats compared to their agricultural neighbors, with the implication that modern social organization and behavior provide little insight into Paleolithic hunter-gatherers. We refer to this idea as the marginal habitat hypothesis (MHH). Despite widespread use of the term ‘marginal,’ there is little consensus as to what comprises a low quality habitat for humans. Here we reassess the MHH by comparing the net primary productivity (NPP) of habitats occupied by, and the population density (PD) of, a sample of 186 pre-industrial societies (foragers, horticulturalists, intensive agriculturalists, and pastoralists). We found that the nature of the NPP-PD relationship varied by subsistence type, and that foragers did not occupy significantly lower net primary productivity habitats compared to other subsistence types. These results do not support the MHH. We conclude by discussing the limitations of using modern ethnographic datasets to address the MHH and suggest alternative ways in which it may still be relevant.