Some recent studies have shown that states with more medical marijuana availability have lower rates of opioid overdose and young male suicides. This was interpreted as meaning that people who use medical marijuana are at lower risk of overdose and less likely to take their own lives. If you think that constitutes good reasoning, you should also believe that smoking and being exposed to radon reduces your risk of cancer because in the aggregate, those variables are negatively correlated with cancer rates!
The recent medical marijuana studies have fallen into a seductive logical error called the ecological fallacy, which my colleagues and I explain in detail at a post at PLOS Mind the Brain. Once you understand the ecological fallacy, you will realize that many, many news stories about research make claims that just are not true (we give examples in our post).
Perhaps surprisingly, whether medical marijuana availability at the state level correlates with some other state-level indicator actually tells us nothing about how medical marijuana affects individuals. If a state-level correlation with some indicator is positive (e.g., states with more medical marijuana have higher rates of violence) the individual level relationship can still be negative (e.g., medical marijuana use makes people less violent).
Some medical marijuana activists have argued to me that it is inappropriate to point out that the studies in this area are methodologically flawed because doing so harms “the cause”. I don’t sympathize with such Lysenkoism. I believe scientists should seek and report the truth regardless of whether it concords with a political agenda (The scientist’s own or anyone else’s). If we give up on that regulating ideal of scientific inquiry, we can’t really complain when other people deny the evidence of oceanic acidification change or assert that the MMR vaccine causes autism.