The origins of the egregiously erroneous Figure 1-4 in a recent Academies report illustrates how easily can factual falsehoods propagate to mainstream media and even esteemed scientific institutions.
Dr. Stein's recalculations are correct and should be used by other authors; congratulations for catching the error. However, there is another big caveat. "Teen suicide" rates prior to 2000 are suspect for a different reason. The 1950-1990s era has seen a massive trend toward reclassifying teenage deaths formerly ruled as "accidents" or "undetermined as to intent" as suicides as medical examiners' forensic examinations have improved. For example, in 1970, more than 2,000 teenaged deaths from firearms, poisonings, and hangings were classified as accidents or undetermined; by 2000, that total had dropped by two-thirds as certified suicides among teens rose in tandem. Pre-1990s teens had considerably higher true suicide rates than official figures reflect; see: https://onlinelibrary.wiley.com/doi/abs/10.1111/j.1943-278X.1991.tb00948.x This, of course, does not detract from Dr. Stein's point regarding the 2010s National Academy's misreported suicide rates.
Thank you for the link to your paper; it would be nice to see the analysis extended to the 1990s, as your 1991 paper stops at 1987 data.
I analyzed the 2007-2018 rise for evidence of a shift from accidental (or undetermined) to intentional self-inflicted deaths in The Rise and Child Suicide Stigma (https://theshoresofacademia.blogspot.com/2020/04/the-rise-and-child-suicide-stigma.html) and concluded that a small but substantial part of the 10-14 rise -- but NOT the 15-19 rise -- may indeed be due to decreased stigma (increased willingness of families and coroners to classify child death as a suicide).
The rise in 10-14 deaths, however, is so steep that even if we subtract the maximal stigma shift allowable by the data, the increases are still extreme.
As to comparing teen (15-19) current and early 1990s rates, I've yet to see evidence that those three decades ago were substantially higher.
If I find the time I'll look into this while updating my old series. One complication is changed classification (ICD 9 to ICD 10) and a less detailed query possibilities at WONDER for 1990s data. I might have to ask for a database download from CDC, if I can get it.
Dr. Stein's recalculations are correct and should be used by other authors; congratulations for catching the error. However, there is another big caveat. "Teen suicide" rates prior to 2000 are suspect for a different reason. The 1950-1990s era has seen a massive trend toward reclassifying teenage deaths formerly ruled as "accidents" or "undetermined as to intent" as suicides as medical examiners' forensic examinations have improved. For example, in 1970, more than 2,000 teenaged deaths from firearms, poisonings, and hangings were classified as accidents or undetermined; by 2000, that total had dropped by two-thirds as certified suicides among teens rose in tandem. Pre-1990s teens had considerably higher true suicide rates than official figures reflect; see: https://onlinelibrary.wiley.com/doi/abs/10.1111/j.1943-278X.1991.tb00948.x This, of course, does not detract from Dr. Stein's point regarding the 2010s National Academy's misreported suicide rates.
Thank you for the link to your paper; it would be nice to see the analysis extended to the 1990s, as your 1991 paper stops at 1987 data.
I analyzed the 2007-2018 rise for evidence of a shift from accidental (or undetermined) to intentional self-inflicted deaths in The Rise and Child Suicide Stigma (https://theshoresofacademia.blogspot.com/2020/04/the-rise-and-child-suicide-stigma.html) and concluded that a small but substantial part of the 10-14 rise -- but NOT the 15-19 rise -- may indeed be due to decreased stigma (increased willingness of families and coroners to classify child death as a suicide).
The rise in 10-14 deaths, however, is so steep that even if we subtract the maximal stigma shift allowable by the data, the increases are still extreme.
As to comparing teen (15-19) current and early 1990s rates, I've yet to see evidence that those three decades ago were substantially higher.
If I find the time I'll look into this while updating my old series. One complication is changed classification (ICD 9 to ICD 10) and a less detailed query possibilities at WONDER for 1990s data. I might have to ask for a database download from CDC, if I can get it.
Coming up, thanks for interest. I'm going to upload some trends soon that are deeply disturbing.