Thread by @tylerblack32

*************
Variance 🧵:
Why the media (and non-experts who "dabble" in mortality statistics) particularly suck at reporting the numbers of suicide
**************

Quite frequently, someone will send me an article like this.

#epitwitter #statstwitter

1/ Suicides are up 67% between the ages of 12-17 in Pima County.

By "mid-Nov 20", there have been 43 teen suicides, compared to 38 in total last year!

Without context, it certainly seems that the pandemic or the lockdown is to blame.

2/ Sure enough, i go to @CDC Wonder and fire up Arizona suicides for 2019 between 12-17 and I see there were 36 suicides in 2019 (not sure why there is a discrepancy between AZ DOH and CDC, but this is actually common by about 5%ish).

3/ There is a media motto: "If it bleeds it leads." The media is FAR more likely to report something upsetting, dire and fearful, than it is uplifting, because of our human behaviour. We tune in. We read it.

Thus, the media has a blind spot to the context of death numbers.

4/ Because 36 out of 567,579 kids is about 6.3 per 100,000, this is officially a tragic yet VERY RARE event. What this means is that variance will have a tremendous impact on the relative rate.

This reporter could have reached out to an expert in statistics, but didn't.

5/ Sure enough, when we take a look at the actual numbers, we see TREMENDOUS variance. In fact, 2018 was at 55 by the end of the year and it is VERY UNLIKELY that 2020 will catch up. (It extrapolates to 51 in the year)

6/ Plotted year by year, we can see that 2020 does not appear out of line with any of the previous 10 years, with the error-bars for high-variance comparison overlapping.

(The bars represent 95% CI's - important for comparing two rates, even if whole population)

7/ Also, in suicidology, raw suicide #'s are not helpful to compare without knowing the denominator (number of people and for how long). If I adjust all of the previous years to the 2019 12-17 Arizona population (37% increase in 21 years!), we see 3 pvs yrs that eclipse 2020:

8/ These headlines tell a narrative that simply isn't true. Most data from almost all jurisdictions show that 2020 suicide rates are WELL WITHIN YEARLY VARIANCE from previous years. There is no "wave" or "increase" or "pandemic-caused" suicide wave as a whole.

9/ Now, I am concerned about disproportionate effect on racialized and marginalized groups, but unfortunately, this data is very sparse and difficult to determine.

10/ So the next time you see an article comparing a NUMBER IN 2019 to a NUMBER IN 2020, ask two questions:

"what is the normal variance year by year"

and

"what is the denominator (population) each year?"

11/ And what about that "Suicides are up 67% in 12-17y in Pima" stat?

Well, again, variance matters.

AZ statewide, the % change in the past 10y suicides (12-17) is:

'11: ***75%***
'12: -17%
'13: -14%
'14: 30%
'15: 25%
'16: -27%
'17: 44%
'18: 11%
'19: -37%

12/ That's right - up MORE THAN 67% has occurred already in the past 10 years. without that context, you could be led to believe that a 67% swing is completely anomalous, when in fact it is most likely right after a significant dip (like in 2019).

13/ If the article does NOT address both points, the article is failing in its task of comparison.

/end

14/ 👆👆
🧵Suicidology / variance / media suckage 🧵

@jeremyfaust @meganranney @drjessigold @DrJenGunter @DoctorYasmin @MDaware @DrJonGoldin @amybarnhorst @BartAndrews @CaulfieldTim @evolutionarypsy @StanKutcher