This study used archival data from three different research groups and case file data from three independent community organizations to explore how well research samples reflect cases of hoarding that come to community attention. Using data from 824 individuals with hoarding, we found that research volunteers differ from community clients in several ways: community clients are older, more likely to be male and less likely to be partnered; they have lower socio-economic status and are less likely to demonstrate good or fair insight regarding hoarding severity and consequences. The homes of community clients had greater clutter volume and were more likely to have problematic conditions in the home, including squalor and fire hazards or fire safety concerns. Clutter volume was a strong predictor of these conditions in the home, but demographic variables were not. Even after accounting for the influence of clutter volume, the homes of community-based clients were more likely to have squalor. These findings suggest limitations on the generalizability of research samples to hoarding as it is encountered by community agencies.