3 ms·
It's not valid to take the frequency of occurrence in a small sample size (320), and multiply it up to a large population (10 million). I still agree it's not w
by scott_s 10y ago
It's not valid to take the frequency of occurrence in a small sample size (320), and multiply it up to a large population (10 million). I still agree it's not worth the risk, but it's also not likely to be an additional 31,250 suicide attempts among men.
(Which is how my math worked out - I'm not sure where your 50,000 number of successful suicides came from. The study had 320 participants with one suicide attempt. But, again, this sort of multiplication from small sample sizes to large sample sizes is not actually predictive, so it's kinda moot.)
- pc86 10y agoHow did the math change from 50,000 successful suicides to 31,250 suicide attempts?
- scott_s 10y agoExplained in my parenthetical.
- Meegul 10y agoAgreed. But one attempted suicide and one successful suicide in a sample of 320 is not insignificant. This could be an anomaly, sure, but given that the average rate of suicide is 12.93 per 100,000 people [1], and it on average takes 25 attempts to produce a single death, this study should've expected no attempts at all. The fact that there was not only an attempt, but another successful attempt is not insignificant. The researchers were not mistaken when they decided to cut the study short, lest they be responsible for more deaths. 1. https://afsp.org/about-suicide/suicide-statistics/ https://afsp.org/about-suicide/suicide-statistics/
- scott_s 10y agoI agree on the potential significance and caution, but the study had only one attempted suicide. None of the participants killed themselves. There was a separate incidence of severe depression.
- Meegul 10y agoTo get the 50,000, I mistakenly, based off another comment, used 200 as the sample size. Your number, 31,250, would be correct (assuming the extrapolation holds) for the actual sample size.