Congressional Research Service claims that Mass Public Shootings have been increasing, but it fails to provide evidence that this increase is statistically significant

The Congressional Research Service (CRS) has just released a new report titled: CRS Mass Murder with Firearms Incidents and Victims 1999-2013.  USA Today’s headline provides a typical reaction: “Report: Mass public shootings on rise.”  While the report points to increases in the number of shooting incidents over this period of time, the CRS does not discuss whether any of these results are statistically significant.  Indeed, if one looks at the period of time that the title of their paper refers, there is no statistically significant increase in the number of incidents or the number of people killed and/or injured.  In fact, the changes per capita are not even close to being statistically significant and the changes are very small.  For the number of incidents per 10 million people, the change is not only statistically insignificant, it implies that there were 0.142 incidents per 10 million in 1999 and this rose to 0.154 in 2013.  To put it differently, that is only a 0.57% increase per year.

While the report’s title refers to the period from 1999 to 2013, the CRS also provides data going back to 1970.

Previous work by the CPRC has shown that there was no statistically significant increase in the death rate from attacks since 1977 where mass public shootings were defined as 2 or more people killed.  We have requested the CRS data for the years prior to 1999, but we have yet to receive it.  But even if the CRS data is correct, it seems doubtful that there has been a statistically significant increase in the rate of attacks or deaths from them since at least the early 1980s, particularly if the very low rate of mass public shootings in 2014 are included.

The other regression results for mass public shootings over the 1999 to 2013 period are as follows (in both cases t-statistic on the change in deaths or injuries per year is less than 1.