Discussion: (85 comments)
Comments are closed.
A public policy blog from AEI
View related content: Carpe Diem
The chart above was inspired by a similar one featured by Max Ehrenfreund in his recent Wonkblog post titled “We’ve had a massive decline in gun violence in the United States. Here’s why.” In contrast to the widely embraced narrative, perpetuated by liberal politicians and the media, that gun violence in America is getting worse all the time, the data reveal that the exact opposite is true. According to data retrieved from the Centers for Disease Control, there were 7 firearm-related homicides for every 100,000 Americans in 1993 (see light blue line in chart). By 2013 (most recent year available), the gun homicide rate had fallen by nearly 50% to only 3.6 homicides per 100,000 population.
Ehrenfreund says that “Even as a certain type of mass shooting is apparently becoming more frequent, America has become a much less violent place. Much of the decline in violence is still unexplained, but researchers have identified several reasons for the shift.” He then points to factors explaining the decline in violent crime in general and gun homicides in particular, including more police officers on the beat making greater use of computers, a decline in alcohol consumption, less lead exposure, and an improving economy.
But there’s another possible reason for the decline in gun violence overlooked by Ehrenfreund – the significant increase in the number of guns in America , illustrated above by the dark blue line in the chart. Based on data from a 2012 Congressional Research Service (CRS) report (and additional data from another Wonkblog article “There are now more guns than people in the United States“), the number of privately owned firearms in US increased from about 185 million in 1993 to 357 million in 2013. Adjusted for the US population, the number of guns per American increased from 0.93 per person in 1993 to 1.45 in 2013, which is a 56% increase in the number of guns per person that occurred during the same period when gun violence decreased by 49% (see new chart above). Of course, that significant correlation doesn’t necessarily imply causation, but it’s logical to believe that those two trends are related. After all, armed citizens frequently prevent crimes from happening, including gun-related homicides, see hundreds of examples here of law-abiding gun owners defending themselves and their families and homes.
In a December 2013 Breitbart article, “Congressional Study: Murder Rate Plummets as Gun Ownership Soars,” Awr Hawkins referred to the CRS report referenced above and connected the two trends:
So after all the pro-gun control grandstanding and the relentless focus on how the so-called easy availability of guns drives up crime, the CRS report shows that more guns–especially more concealable guns–has actually correlated with less crime.
Bottom Line: Even if you’re not convinced that increased gun ownership reduces violent crime and gun homicides, you should be totally convinced of this indisputable fact: Gun violence has been decreasing significantly over time, not increasing as you’ll frequently hear from anti-gun politicians and progressives. The gun-related homicide rate of 3.6 deaths per 100,000 population in each of the years 2010, 2011 and 2013 makes those recent years the safest in at least 20 years, and possibly the safest in modern US history, since “older data [before 1993] suggest that gun violence might have been even more widespread previously,” according to Ehrenfreund.
Update: In the comment section below, David says that “The [top] chart you created is misleading. If both values were either per capita or per 100,000 people and plotted on the same scale, the lines would not cross as depicted. You manipulated the range of the y-axis to create that artificial intersection.” In response, I created the second chart above, showing the percentage changes in both variables (guns per person and the gun homicide rate) since 1993 as an alternative way to graphically display the relationship between those two variables over time using a single, unit-free scale measured in percentage changes of both variables. That should avoid the criticism (sometimes justified, sometimes unjustified) that is sometimes leveled against “dual scale” charts, like the example below provided by David in his comment.
That chart was correctly criticized by Politifact and others as being misleading and “ethically wrong.” Politifact wrote that “dual-axis charts are particularly susceptible to showing spurious correlations.” Actually the chart above isn’t really even an example of a true dual-axis chart, it’s simply a graphic with two arrows crossing. That is, it likely wasn’t produced by a graphic software program, it was likely created “free-hand” with two arrows crossing and is really a “no scale” chart.
To understand why “dual-axis scale” graphs can be a legitimate way to graphically display two variables that are measured in units that depart significantly in scale, let’s consider an example. We know (or suspect) that there should be some statistical relationship over time between: a) US corporate profits and b) stock prices measured by the S&P 500 reflecting what we frequently hear from Larry Kudlow that “profits are the mother’s milk of stocks.” But the problem charting both variables simultaneously is that corporate profits are measured in trillions of dollars and the S&P500 is based on an index scale that is equal to a value of 100 in the base year of 1968, and now has a value of about 2,000. Plotting those two variables on a single-scale graph would be meaningless and would look like this:
Corporate profits are measured in units (trillions) that are so significantly larger than the S&P500 units (thousands), that there is no way to realistically compare the movement (co-movement) of those variables over time unless they are plotted on a “dual axis scale.” The chart below shows a “dual axis scale” chart of the S&P 500 Index and Corporate Profits, where the S&P 500 Index is on the left scale (measured in thousands) and corporate profits are on the right scale (measured in billions of dollars. We can see graphically the close statistical relationship over time between corporate profits and stock prices by using a “dual axis scale” graph.
To verify that the “dual axis scale” chart above hasn’t presented a misleading, distorted or spurious relationship between corporate profits and stock prices, consider the chart below that shows the unit-free percentage changes in both variables since Q1 1993. You’ll notice that the patterns and trends are exactly the same when both variables are expressed as unit-free percentages on a “single axis scale” as the “dual axis scale” above that displays both variables in their original units.
Bottom Line: For those troubled by the “dual axis scale” chart at the top of the post, you can ignore that one, and instead focus on the second “single axis scale” chart that shows both variables (guns per person and the gun homicide rate) in a single, unit-free measurement: percentage changes since 1993. And the manually created Planned Parenthood graphic with two manually drawn arrows (really a “no axis scale” chart) shouldn’t be a reason to condemn all “dual axis charts” – they are a legitimate and accepted way to compare economic and financial variables over time when they are measured in widely disparate units.
Comments are closed.
1789 Massachusetts Avenue, NW, Washington, DC 20036
© 2018 American Enterprise Institute