Thursday, August 1, 2013

Correlation, Causation, Math, and Weed

Scary story from today's Denver Post:
Nearly one-third of the crimes committed in Denver occur within 1,000 feet of a medical marijuana dispensary, according to new statistics provided by the city Wednesday.
Yikes! Sounds like those pot clinics are a cause of crime in Denver. Are they? Let's look at some basic math:
  • Number of medical marijuana clinics in Denver, as of 2010: 390.
  • Total area of 1,000-foot-radius circles around all 390 clinics (remember πr2?): 1.23 billion square feet.
  • Total area of Denver: 4.32 billion square feet.
  • Proportion of Denver's area that is within 1,000 feet of a medical marijuana clinic: 0.28, or nearly one-third.
In other words, the proportion of crimes committed near Denver marijuana dispensaries happens to be the same as the proportion of the city that exists near Denver marijuana dispensaries.

Now, okay, this little analysis ignores a few things. There are probably more clinics today than there were in 2010, some of the 1,000-foot radii probably overlap due to the proximity of clinics to each other, etc. But still, a little math might remind us that's there's not much of a story here.


  1. Interesting analysis. I would think that both crime and clinics are more likely to occur in the more more heavier populated areas, strengthening your argument.

    I think you would find something I wrote about why the news media does that nonsense interesting:

  2. This kind of analysis is also relevant to guns within 1000 foot of schools and sex offender housing ordinances.

    In Denver, in particular, a very large share of the land area is in the dogleg of territory that goes to Denver International Airport which has very low population. The proportion of Denver's population that lives within 1000 feet of a dispensary is much higher than the proportion of Denver's surface area that is within 1000 feet of a dispensary, and a lot of those dispensaries are in relatively low rent areas that have higher than average crime to start with. Studies have shown no elevation from pre-dispensary crime levels in the vicinity of dispensaries.