Random Stuff

Me helping to explain the Barnum effect on Netflix


Firearm Mortality in Red and Blue States

Yesterday (November 5, 2017) there was another mass shooting in America, so today I constructed the enclosed chart. The mortality data come from this page of the CDC website and are for the most recent year available, 2015. The figures are per 100,000 of total population. The difference is statistically significant, t(48) = 4.95, p < .001.

Firearm mortality rates vary substantially across states. Alaska has the highest rate at 23.4 per 100,000. Massachusetts has the lowest at 3.0. As a result, firearm deaths are 7.8 times more likely in Alaska than in Massachusetts.


—November 6, 2017

The Obamacare Effect

That sharp dip starting in 2010 was produced by the Affordable Care Act, as many more people secured healthcare coverage. What will this chart will look like in the future?

Screen Shot 2017-06-25 at 12.44.15 PM.png

Buy A House?

I created this graph on plot.ly while working on the second edition of my 2008 book Going Broke: Why Americans Can’t Hold on to Their Money. It shows that homeownership is now at a historic low, as is the rental vacancy rate. One effect of the Great Recession has been a wholesale move from owner-occupied homes to rentals. Homeownership in the U.S. has not been this low since the 1960s.
Plot 195
Of course, some people lost their homes to foreclosure and were forced to make the transition to rentals, but we are now nine years out from 2008. The foreclosure rate has dropped precipitously in recent years. It looks as though the Great Recession has created a lasting reluctance to sign a mortgage.

Murder Weapon Trends

I was wondering about the methods used in homicides. I found these data for murder victims at infoplease, and plotted them using plot.ly. The numbers for murder exclude non-criminal homicides, such as accidents and legal interventions.

The most striking—though not surprising—thing about this graph is the much higher rate of homicides that involve a firearm, and also the dramatic variation in gun homicides since the mid-1970s. Guns are quite lethal, of course, but they also appear to be sensitive to other factors. For example, during the crack epidemic years of 1984-1989, homicides more than doubled among young black men, which may be a partial explanation for the spike during those years shown here.

The only other weapon that shows any degree of variation is the decline in the popularity of cutting/stabbing. Today this weapon is responsible for fewer than half the number of homicides as in 1980.

(Posted March 14, 2016. Additional text added on March 27, 2016.)