When the National Survey on Drug Use and Health (NSDUH) indicated that marijuana use by teenagers in Colorado rose after the state legalized the drug for recreational use in 2012, prohibitionists trumpeted the results, even though the change was not statistically significant. Drug warriors were notably quieter when subsequent NSDUH data indicated that adolescent consumption in Colorado fell after state-licensed marijuana stores began serving the recreational market.
That change was not statistically significant either, underlining the uncertainty about the impact of legalization on underage consumption. It is plausible that legalization would increase adolescent use by making marijuana more socially acceptable (although probably not cooler) or by making it available from legal buyers 21 or older. But so far there is little evidence that is happening.
The Colorado Department of Public Health and Environment says cannabis consumption by teenagers in the state "has not changed since legalization either in terms of the number of people using or the frequency of use among users." That conclusion is based on data from NSDUH and the Healthy Kids Colorado Survey, which has a much larger sample of Colorado teenagers.
A study published in the February 2017 issue of JAMA Pediatrics covered yet another survey, the Monitoring the Future Study. University of California, Davis, epidemiologist Magdalena Cerdá and her colleagues looked at past-month marijuana consumption among eighth-, 10th-, and 12th-graders in the three years preceding legalization (2010–12) and the three years following it (2013–15). They compared trends in Colorado and Washington, where voters also approved legalization in 2012, to trends in the 45 contiguous states that did not legalize marijuana for recreational use during this period.
Cerdá et al. found no significant differences in Colorado or among high school seniors in Washington. But Washington eighth- and 10th-graders deviated from the national trend. Although the incidence of past-month marijuana use by eighth-graders did not rise significantly in Washington, it fell significantly in the other states. Past-month use among 10th-graders did rise significantly in Washington, from 16.2 percent to 20.3 percent, while falling in the rest of the country.
Assuming that the deviations among eighth- and 10th-graders in the Evergreen State have something to do with legalization, Cerdá et al. say, the mechanism is unlikely to be diversion from adult buyers, since state-licensed pot shops did not open there until July 2014, halfway through the post-legalization study period. But they argue that legalization may have changed attitudes in a way that encouraged adolescent use.
If so, it's a bit of a mystery why there is no evidence of this phenomenon in Colorado. But with only a few years of data to consider, the only safe conclusion is that it's too early to draw any conclusions.