A military report on suicide rates broken down by occupation finds the highest rates among categories of troops that often work and train around weapons blasts.
This correlation seems to have nothing to do with blast exposure as the title suggests From the article
The rates for these combat occupations are roughly twice those of service members who work in noncombat jobs like data processing or food service.
The article even mentions there isn’t any implied causation between blasts and non-blast exposed combat troops either in the data presented (except maybe a 4% difference in suicide rate between artillery crews and combat troops with less exposure). The data they are presenting* shows the largest drop in rate is between combat and non combat troops How can the author have this information and come away with this conclusion?
Explosives ordinance disposal team members, who disable roadside bombs and routinely train and work around very large blasts, had the highest suicide rate — 34.77 deaths per 100,000 people per year — followed by infantry and special operations forces; armor crews; and artillery troops; whose rates are closer to 30 deaths per 100,000.
That’s a 15% bump in suicide for the most blast exposed troops when compared to less blast exposed combat troops.
Obviously combat exposure vs not is going to have a large effect, but blast exposure seems to have a further effect on top the horrors of combat.
Also:
In the Air Force, where blast exposure is rare, there were no significant differences in suicide rates among different military occupations. But among Army and Marine Corps troops, the rates are elevated wherever blasts are part of daily work
The report presented the data, but made no attempt to grapple with the implications:
The report released on Wednesday does not mention blast exposure as a factor, and offers no insights into what may be contributing to the different suicide rates. Still, the correlation between deaths by suicide and levels of blast exposure is a common theme in the figures
Alright, now here’s where i am ready to be wrong… How is an increasing suicide rate from 30 per 100k to 35 per 100k statistically significant? Am i even worse at math than i thought?
It’s the same reason that an increase from a 0.0001% chance of something happening to a 0.0002% chance is a 100% increased chance of the thing happening.
More relevantly, 0.0003477% is ~ 15% bigger a number than 0.0003000%. The overall numbers are still low, but the increase is significant.
Well then i guess i am wrong, cuz i have to concede 15% is a significant difference. Thanks for explaining that. It still feels like artificial inflation of stats to my mind, but that’s just like, my opinion, man
I remember reading over some logic - statistics “puzzles”, where the logic is sound if you follow step by step on paper, but for some reason your mind just has trouble grasping the concept which leads to confusion.
This correlation seems to have nothing to do with blast exposure as the title suggests From the articleThe rates for these combat occupations are roughly twice those of service members who work in noncombat jobs like data processing or food service.The article even mentions there isn’t any implied causation between blasts and non-blast exposed combat troops either in the data presented (except maybe a 4% difference in suicide rate between artillery crews and combat troops with less exposure).The data they are presenting* shows the largest drop in rate is between combat and non combat troopsHow can the author have this information and come away with this conclusion?The paragraph before the one you quoted says it:
That’s a 15% bump in suicide for the most blast exposed troops when compared to less blast exposed combat troops.
Obviously combat exposure vs not is going to have a large effect, but blast exposure seems to have a further effect on top the horrors of combat.
Also:
The report presented the data, but made no attempt to grapple with the implications:
Alright, now here’s where i am ready to be wrong… How is an increasing suicide rate from 30 per 100k to 35 per 100k statistically significant? Am i even worse at math than i thought?
It’s the same reason that an increase from a 0.0001% chance of something happening to a 0.0002% chance is a 100% increased chance of the thing happening.
More relevantly, 0.0003477% is ~ 15% bigger a number than 0.0003000%. The overall numbers are still low, but the increase is significant.
Well then i guess i am wrong, cuz i have to concede 15% is a significant difference. Thanks for explaining that. It still feels like artificial inflation of stats to my mind, but that’s just like, my opinion, man
Stats can be very unintuitive. You’re not alone.
I remember reading over some logic - statistics “puzzles”, where the logic is sound if you follow step by step on paper, but for some reason your mind just has trouble grasping the concept which leads to confusion.
Yep, NP!
The New York Times, that’s why. It’s a paper which drastically lost credibility.