Background In 2009, policy changes were accompanied by a rapid increase in the number of medical marijuana cardholders in Colorado. Little published epidemiological work has tracked changes in the state around this time. Methods Using the National Survey on Drug Use and Health, we tested for temporal changes in marijuana attitudes and marijuana-use-related outcomes in Colorado (2003-2011) and differences within-year between Colorado and thirty-four non-medical-marijuana states (NMMS). Using regression analyses, we further tested whether patterns seen in Colorado prior to (2006-8) and during (2009-11) marijuana commercialization differed from patterns in NMMS while controlling for demographics. Results Within Colorado those reporting “great-risk” to using marijuana 1-2 times/week dropped significantly in all age groups studied between 2007-8 and 2010-11 (e.g. from 45% to 31% among those 26 years and older; p=0.0006). By 2010-11 past-year marijuana abuse/dependence had become more prevalent in Colorado for 12-17 year olds (5% in Colorado, 3% in NMMS; p=0.03) and 18-25 year olds (9% vs. 5%; p=0.02). Regressions demonstrated significantly greater reductions in perceived risk (12-17 year olds, p=0.005; those 26 years and older, p=0.01), and trend for difference in changes in availability among those 26 years and older and marijuana abuse/dependence among 12-17 year olds in Colorado compared to NMMS in more recent years (2009-11 vs. 2006-8). Conclusions Our results show that commercialization of marijuana in Colorado has been associated with lower risk perception. Evidence is suggestive for marijuana abuse/dependence. Analyses including subsequent years 2012+ once available, will help determine whether such changes represent momentary vs. sustained effects.