How to Defend Yourself from Statistical Lies: Part 2
How to Lie with Percentages, Using the Example of School Shootings.
--
COMING UPDATES: If you’d like to support my continuing to do this, you can send an Amazon giftcard to hollymathnerd at gmail or CashApp me at $HollyMathNerd. Thank you!
My statistical self-defense series has many other topics: go here for the list.
Lying with percentages is very easy because small changes can result in huge percentages. The size of the whole matters a lot! Think of it this way: 1 is 50% of 2 but just 1% of 100. Here’s an example of lying with percentages by ignoring the context of the whole:
Using the very best numbers I could find (sources below), I gathered data on school shootings for the year 2010 and the year 2018. The choice of those two years was to try to replicate a typical journalism scenario — “School Shootings This Decade” being the sort of thing we’re used to hearing about in the news, especially in the context of campaigns (“Did School Shootings get Better or Worse Under This President?”) I used Wikipedia to cross-check and help me determine how many victims were students vs. others.
Data summary:
Here are just some of the ways you could use statistics to write a headline about “School Shootings This Decade” from this information, depending on your agenda:
547% Increase in School Gun Danger
This looks at 15 incidents going up to 97 and ignores that some of these incidents were brandishings, suicides, parking lot brawls between neighborhood fans after football games, or happened when the building was empty.
School Shooting Deaths Up 150%
This looks at 6 incidents with at least 1 death going up to 15 such incidents.
Student Gun Deaths Go From 1 in 19.1 million to 1 in 2.4 million
This ignores that many of these deaths were concentrated in mass incidents, exaggerating the risk.
3 People Will Be Struck By Lightning for Every Student Who Dies in A School Shooting
This ignores that the risk of lightning strike is concentrated in a very small group (mountain climbers and others in dangerous environments, most likely to be struck) and just looks at 1 in 700,000 vs 1 in 2.4 million, the latter also being misleading since many school shooting deaths are highly concentrated in massacres.
Students Dying in School Shootings up 8%
This computes that 17% of victims in 2010 were killed students vs 25% of victims in 2018 being killed students.
School Gun Incidents 25% Less Deadly
This takes 6 of 15 incidents involving at least 1 death and compares it to 15 of 97 such incidents resulting in at least 1 death.
Student Deaths Up 699%
This takes the percentage of students who were killed in 2010 and compares it to the percentage who were killed in 2018 without mentioning that those numbers are incredibly tiny: .0000000522% in 2010 and . 000000417% in 2018.
Teacher Gun Danger on Campus More than Doubles
This takes the 5 non-students killed in 2010 and compares it to the 12 in 2018.
Student Gun Deaths Increase 700%
This compares the 4 student deaths in 2010 to the 32 in 2018.
I could go on and do many more, but I trust the idea is clear: someone with an agenda can easily write a scary headline about school shootings this decade or under that President, and depending on your agenda you can go either way with it.
By the way, I hate guns, have witnessed gun violence, and it’s a test of my commitment to being an honest statistician to post this. And I don’t have a boss breathing down my neck expecting a story that fits a certain narrative or news cycle, do I?
In Part 3 of this series, I break down some actual news stories about school shootings wherein the journalists lied with statistics.
Sources:
https://www.chds.us/ssdb/category/shooting-incidents/
And I used this handy calculator to calculate percentage increases:
