Preventable Deaths and the Need for Data-Driven Journalism

In the wake of the recent mass shootings in Texas and Ohio that claimed the lives of more than 30 people, Neil deGrasse Tyson drew significant controversy by posting a tweet which compared the death toll from the shootings to the (larger) numbers of people who died from other preventable causes over a 48-hour time period. Dr. Tyson concluded his message with a warning: “Often our emotions respond more to spectacle than to data.”

The negative responses to his tweet were swift and numerous, with many users voicing outrage, disappointment, and disgust. The next morning, he issued an apology on his Facebook page and acknowledged that he “got this one wrong.” But did he? I don’t disagree that his timing could have been better, but I think his point is important: We shouldn’t let our emotional reaction to shootings inflate our perception of their impact or their probability. Sounds straightforward, right? Well, it’s not as easy as it seems, and one of the primary reasons for the difficulty is a cognitive bias known as the zero-risk bias.

Zero-risk bias is the tendency to prefer the complete elimination of a single risk even when alternative risk reduction strategies lead to a greater decrease in overall risk. People tend to feel better when a given risk is eliminated rather than mitigated. Imagine the following scenario: I present you with two magic buttons. One of them will reduce mass shooting deaths by 100 percent for an entire year, the other will reduce drug overdose deaths by five percent. If you could only press one, which would it be? If you’re like me, your first impulse will be to end mass shooting deaths. It seems more consequential; that impulsive preference is the zero-risk bias in action.

Let’s go over that thought experiment again, only this time imagine that the buttons have numbers on them instead of proportions. The first button will save 372 people from being killed in mass shootings; the second will save 3,425 people from being killed by drug overdoses. This time, I immediately gravitate towards the second option. The numbers are the same in both situations, but they feel very different. They also reflect approximate mortality statistics for mass shootings and drug overdoses from 2018. Several other causes of death could stand in for overdose deaths to illustrate a similar difference in magnitude, but this specific comparison goes deeper than a simple comparison of number of deaths. Both mass shooting and overdose deaths are multiplicative at the societal level in a way that dying in an elevator accident is not.

Mass shooting deaths are also well suited to zero-risk thinking due to how much more preventable they seem to be compared to other types of premature death (which is partially a result of relative rarity). This form of reasoning can lead us to endorse statements such as “We can’t possibly stop people from overdosing on drugs, so let’s focus on something we can stop.” But, as outlined above, this sort of logic doesn’t appropriately engage with societal risk factors. It also ignores the significant progress being made on the overdose front, and the potential for future intervention.

Journalism’s Death by a Thousand Tweets

Sydney. London. Toronto.

QuilletteThomas Moller-Nielsen

Recognizing this bias, and its effect on how we understand preventable death statistics, contextualizes the conversation around mass shootings, both societally and in the realm of debate. Many contend that such comparisons are faulty, but they are helpful when it comes to combatting these problems. Every dollar, minute, or ounce of energy spent addressing a specific problem is one that, almost by definition, cannot be spent on one of the others. Resources, including time, are zero-sum; when we allocate those resources, we should endeavor to place them where they will achieve the most good. These considerations become even more important when we consider that average life expectancy in the US has been in decline¹ for the past three years—the longest sustained period of decrease since WWI. Preventable deaths, specifically those related to drug overdose and suicide, are central drivers of our life expectancy problems, and that should alarm us.

Does this mean that we should ignore mass shootings? Certainly not. I do not mean this comparison to trivialize mass shootings or to diminish efforts to curb the risk of future violence. Updating the laws surrounding firearms and their use is a worthwhile goal that could help to decrease violence more generally and put a dent in the suicide epidemic. There are also important reasons to focus on mass shootings that extend beyond the death toll, or actual level of risk, and this analysis should not detract from those considerations. Ideological factors, such as white nationalism, that have driven a significant portion of recent mass shootings have the potential to damage our system in ways that are not properly reflected by the death toll alone. They should be a central part of the public conversation. But that conversation needs to maintain a sense of perspective.

Many have noted that the effect that a mass shooting has on a community (and wider society) goes well beyond the number of people killed. The same, however, is true of other forms of preventable death, such as suicide.² Overdose deaths, and the opioid crisis in particular (47,600 overdose deaths involved opioids in 2017), have had a dramatic effect on communities which also extends well beyond the number of people who overdose—those who die likely represent a small proportion of those affected. Considering estimates that as few as one in 10 of those with a substance use disorder can access the treatment that they need, this seems like an area which merits additional attention.

It is also worth noting that these arguments address preventable death from a societal perspective and are aimed at those whose concern is to save lives in a general sense. There are rational, from a personal risk perspective, reasons that someone may be more concerned with mass shootings than the other types of preventable death mentioned here. If you feel that you, and the people in your social group, are not at risk of falling victim to a drug overdose or suicide, then you might want the government to focus on risks that you cannot mitigate yourself, rather than on strategies designed to halt the maximum number of preventable deaths, even when those risks are extremely low (as is true of mass shootings).

My primary goal—and Dr. Tyson’s if I read his tweet correctly—is to provide context. It is perhaps inevitable that our biases will continue to inform the responses of the general public on social media. But responsible journalism could do a better job of marshaling the available data so that we can keep the threats to public welfare and safety in proportion and allocate our resources most effectively. I agree with those who have pointed out that we are facing a crisis in the United States. But that crisis has multiple fronts, and facts, rather than feelings, should inform our response. So, give Dr. Tyson a break, he was only trying to help us keep things in perspective.

References:

¹ Dyer, Owen. “US Life Expectancy Falls for Third Year in a Row” BMJ (December 2018);363:k5118 https://doi.org/10.1136/bmj.k5118
² Cerel, Julie, Myfanwy Maple, Judy van de Venne, Melinda Moore, Chris Flaherty, and Margaret Brown. “Exposure to Suicide in the Community: Prevalence and Correlates in One U.S. State.” Public Health Reports 131, no. 1 (January 2016): 100–107. doi: 10.1177/003335491613100116.