Great news – suicides among military staff are down 22 percent this year! The only problem is that it’s possible that the rate of suicides is unchanged, or possibly even increased. This is a demonstration of how math can flirt with the truth and distort our perspective.
To be fair, they calculated the 22 percent decrease correctly: they compared 2012 January through October suicide count (316) to 2013’s count (245) for the same months. Indeed, 245 is 78 percent of 316. So far, so good? Not really.
We don’t know what the count includes. If they do not count people who un-enlisted, they might be overlooking their biggest risk group. With the end of the Iraq war and withdrawal of troops of Afghanistan, perhaps many people took the opportunity to get out. Disenchanted with civilian life and disconnected from the military, these people may be the most prone to depression.
If they count only those whose death certificates list “suicide” as the cause of death, they may be ignoring social taboos. No one wants to have their local war hero die by his or her own hand. There may be subtle pressure on medical examiners to list events as accidental rather than suicide, when the deceased is a veteran.
This is also a great demonstration in how definitions and data sources can shift, creating false trends. It is, after all, human beings counting other humans and deciding the cause of death. Instead of “We’ve had a decrease in suicides!” the truth is more likely to be “We’ve had a decrease in people reporting deaths as suicides!” The truth is a lot less exciting, obviously.
In addition to possible issues with counting suicides, there is another problem – using the absolute number of suicides is misleading. It’s more revealing and accurate to measure what proportion of the population is affected by suicide. For example, we could calculate the number of suicides per 10,000 active duty staff or per 10,000 active and recently un-enlisted staff. If we did that, we might see that 245 suicides this year makes suicide more frequent than last year’s 316.
The best, clearest measure would be to compare military personnel’s deaths from all causes to the general population’s death rate. Clever mathematicians can make adjustments for the military folk’s more strenuous jobs, educational level, and income. This would show us whether the military folks have a higher risk of death than their neighbors who have similar jobs but don’t serve their country. That would be useful and more interesting than simply counting suicides, where the data is created by hundreds of different people with no set standard.
Don’t get me wrong – I’m glad if fewer military people took their own lives. But I still don’t know if that is what happened. If math and science do anything, they should lead us away from falsehood and to new understanding of truth.