American ginseng (Panax quinquefolius) is a globally desired medicinal plant that is becoming increasingly difficult to study due to harvest-induced rarity. Thus, this species' conservation could be greatly improved via species distribution models, making it a model organism for studying sampling bias. In an attempt to refine a state-derived distribution model for ginseng in Virginia, we conducted additional surveys in a biologically diverse yet under-sampled region of the state—the Cumberland Mountains—thereby increasing the number of documented ginseng occurrences in this region thirteen-fold (N₁ = 16, N₂ = 214). Our surveys resulted in the model predicting an increased probability of American ginseng occurrence not only statewide (µ₁ = 0.099, µ₂ = 0.104) but particularly so in the Cumberland Mountains (µ₁ = 0.170, µ₂ = 0.278), highlighting a consistently overlooked hotspot for biodiversity in Southern Appalachia. We suggest that more geographically balanced surveys and reduced overrepresentation of heavily protected and managed areas such as National Parks—in addition to heeding local knowledge—can be an effective method of mitigating geographic bias in predictions from species distribution models.