Our good friends at the inestimable Gun Violence Archive (GVA) have added a new enhancement to their website which allows users to search the real-time gun violence numbers in every state and every Congressional district within every state. This is not only a very important search tool, but it also gives the digital (email and twitter) contact information for each Member of Congress so that someone’s concerns about gun violence in their own neighborhood can be sent directly to the federal representatives who might, God willing, get to vote on a gun bill next year. Because if the unthinkable happens and he becomes Number 45, he will immediately call for a national concealed-carry law, but if HRC moves into 1660 Pennsylvania Avenue, she’ll no doubt want to extend background checks on private gun transfers to all 50 states. Hard to figure that one out, isn’t it?

So being able to check gun violence stats on a daily basis and then being able to send the stats or a comment about the stats, or both, to your elected reps in DC is a very valuable tool for driving the GVP message forward loud and clear. Good job – well done!

But of course being something of a data-head myself, I couldn’t resist the opportunity to take a look at the year-to-date stats because I have mentioned on multiple occasions that the numbers generated by GVA seem to be at variance with what we get from the CDC. The CDC numbers cover every type of gun injury, and even though the data often contains coding errors or just plain gaps, if the government pays the CDC to report all injuries in order to evaluate the health of the American population, then either we depend on their numbers or we don’t. But when I looked at the state-level data in this new GVA search engine, I’m not so sure that what we get from the CDC gives us even a rough idea of the real level of gun violence that we suffer every day.

Before comparing GVA and CDC numbers, let’s understand the limitations under which GVA operates no matter how exact they try to be. GVA is an open-source aggregator, which means that the data comes whichever public sources post information on the web. So GVA is dependent on the various media venues that generate web-based information, which eliminates suicides as well as most unintentional injuries caused by guns or anything else.

In 2014, the most recent year for CDC data, the five states with the highest rates of gun homicides (AL, AR, LA, MS, SC) totaled 1,423 gun deaths; the five states with the lowest rates (CT, MA, NY, RI, VA) recorded 772 gun homicide deaths, together this gets us to 2,195. The GVA count for those same 10 states in 2014 was 2,420, a differential of 10%. Apply this to the country as a whole and gun homicides would go from 11,409 to 12,500 or so. Of the 50 state-level totals, the CDC admits that gun homicide numbers from 12 states are exceedingly rough estimates and might not be real. But of the 10 states whose numbers I counted, only one state’s data – Rhode Island – was too ‘unstable’ to be used.

As I said earlier, we have to assume that the GVA data is probably well below the actual level of gun violence, given the sources on which it is based. I’m not saying the sources are wrong, I’m saying that, by definition, they can’t catch every shooting event. But what we do learn from the GVA’s new search engine is that the number that most of us use to quantify the terrible toll from gun violence is probably much less, at least 10% percent less than it really is. If it were up to me, when it comes to understanding the true degree of gun violence, perhaps we should try to get HRC (assuming there’s good news on November 8th) to substitute the GVA for the CDC.