Background: Government-maintained state registries and online listings are the predominant sources information used to research how proximity to cannabis retailers impacts health. The validity of these data with respect to retailer existence and location is unclear. This study compares cannabis retailer location information from online listings and government-maintained state registries overall, within states, and by legalization (medical or medical and recreational cannabis use). Methods: We collected 7,832 cannabis retailer locations from five online listing sources (June 2019) and 7,733 cannabis retailer locations from government registries (May – October 2019) across 26 states legalizing the sale of cannabis for medical or recreational use. We sequentially sampled 10% of retailers in each state for phone validation (n=826 online listings, conducted November 2019 – May 2020; n=856 government registries, conducted February – June 2020). Proportions of observations validated as operating cannabis retailers at listed locations by source were compared (overall, comparing states with both legalized medical and recreational cannabis use against states with medical only use, and by state) using chi-square tests. Results: 74% (n=608) and 75% (n=639) of cannabis retailers were operating and located at listed locations for online listings and government-maintained registries, respectively(p=0.626). Across both sources, states with legalized medical and recreational cannabis had higher proportions of validated retailers than states with only legalized medical cannabis (77% and 69% respectively, [ptextless.001]). Data sources are comparable within states except for California, Florida, Louisiana, New Mexico, and New York which had more validated cannabis retailers when sourced from government registries (p textless.05) and Oklahoma and Oregon which had more validated cannabis retailers from online listings (ptextless.05). Conclusions: There are state-level differences in the validity of online listings and government-maintained registries of cannabis retailers. Researchers should consider which state location-based research on cannabis retailers is being conducted in to direct choices surrounding data source.