The source data included a total of 73 countries for which both metrics (per capita ownership “X” & death rate “Y”) were available.
We created a new metric by dividing Y/X and called it “V”. You can think of it as the measure of “Violent Behavior” or “Tendency to Use Guns” by a gun-owner in a country. When all countries are included, the regressed graph does not display a clear relationship between “per capita ownership” and the “death rate”. However, when the extreme
countries are removed (very high and very low violence), we obtain the above graph and the “best fit” line with the formula.
y = 0.0863x + 0.3526
R2 = 0.73 (indicates how well the line explains the data, 1.0 being perfect explanation)
The graph contains a sample of 45 countries, about two thirds of the total, which have relatively close “V” levels (average close to 0.12). The extreme “V” countries removed on the other hand have “V” levels almost 20 times that on average. This regression indicates a much stronger relationship between the 2 metrics, which may indicate effectiveness of gun control.
The chart is meant to provide a general view, as the accuracy of the source is disputed in certain cases, and there are some discrepancies between the dates of x,y pairs, in addition to the possibility of anomalies in the particular year the data was taken. Another important thing to note is that countries with very high “V” values are concentrated in South & Central America. There may be a spillover effect to the US, but the latter’s relatively low value of “V” indicates otherwise.