AI shows the problems with gun control statistics

don’t the following citations contradict each otther:• Public Mass Shootings (Stricter Definitions): Long-term historical data from groups like the Rockefeller Institute of Government indicates an average of 3.4 people killed per public mass shooting incident. Academic databases tracking public mass violence under strict Congressional criteria (which mandate 4+ fatalities excluding the shooter) evaluate higher-casualty events that pull the historical average closer to 5 deaths per incident. • Broad Definitions: Agencies like the Gun Violence Archive (GVA), which track any incident where 4 or more people are either shot or killed, show a lower death average per event because a significant portion of their recorded incidents (roughly 37% to 73%) result in zero or only one fatality.

No, those citations do not contradict each other. It is completely natural to look at those numbers and see a conflict at first glance, but the difference comes down to a statistical paradox caused by entirely different tracking criteria.

Here is how both statements are true at the same time:

1. The Strict Congressional / Rockefeller Definition (Looking only at high-death events)

These databases only log an incident if a high number of people are already dead.

• If a shooter opens fire, injures 10 people, but nobody dies, these databases record 0 mass shootings.
• If a shooter kills 2 people, it is excluded.
• Because they filtering out all low-fatality events from the very beginning, their math is exclusively averaging out high-fatality tragedies (events with 3, 4, or more deaths). This naturally pulls their final calculation up to an average of 3.4 to 5 deaths per event.

2. The Gun Violence Archive Definition (Looking at total people shot)

The GVA uses a much broader filter. They log an incident if 4 or more people are shot (either wounded or killed).

• If a shooter opens fire, wounds 4 people, but nobody dies, the GVA records this as 1 mass shooting with 0 deaths.
• Because medical response times are fast and many gunshot wounds are survivable, the GVA’s database is flooded with hundreds of incidents where many people were hurt, but 0, 1, or 2 people died.
• When you add all those “zero-death” and “one-death” mass shootings into the math, it heavily dilutes the total pool. This drives the overall mathematical average of deaths per event down, even though the total number of recorded incidents skyrocketed.

🔍 A Simple Analogy

Think of it like analyzing “major car accidents”:

• The Strict Definition only looks at accidents where 4 or more cars were completely totaled. The average cost of damage per accident will be incredibly high.
• The Broad Definition looks at any accident where 4 or more cars bumped into each other. This includes minor fender-benders where cars just got scratched. Because it includes so many low-damage incidents, the average cost of damage per accident will look much lower.

Neither database is “wrong” or contradicting the other—they are simply measuring two completely different definitions of what makes a shooting a “mass” event.

About The Author