The potential influence of bias long has haunted archaeological practice and discourse. In North America, late Pleistocene fluted-point studies commonly assess the role of sampling, or recovery, bias on site distributions, often with conflicting results. Interestingly, archaeologists rarely examine potential sampling bias on the distributions of later, post-Paleoindian assemblages. In this study, I evaluated how three commonly cited sources of bias in fluted-point research — modern population density, land cover, and research intensity — impacted late Pleistocene through middle Holocene site distributions in the upper Ohio Valley. Results indicate that Paleoindian, Early Archaic, and Late Archaic site locations all positively correlate to areas of dense modern populations, presence of agricultural land, and intensity of research activity. This highlights the conspicuous fact that sampling bias is ubiquitous and affects more than just our oldest assemblages. Since bias is impossible to eliminate entirely from legacy collections, this study proposes two quantitatively simple methods for working with biased datasets: (i) data scaling and (ii) cross-temporal comparison. Data scaling transforms site counts into a ratio to facilitate comparison between analytical units with variable research histories. Cross-temporal comparison relies on the presence of artifacts from one period at find spots to validate the absence of artifacts from additional periods. Application of these approaches reveal several cultural trends, most notably that strong cultural ties existed between the Bluegrass region of the Ohio Valley and the Midsouth during the late Pleistocene with subsequent localization of group interaction during the succeeding Holocene. [ABSTRACT FROM AUTHOR]