SoylentNews is people
SoylentNews
U.S. Life Expectancy Fell By 1.5 Years In 2020, The Biggest Drop Since WW II:
Life expectancy in the United States declined by a year and a half in 2020, according to the Centers for Disease Control and Prevention, which says the coronavirus is largely to blame.
COVID-19 contributed to 74% of the decline in life expectancy from 78.8 years in 2019 to 77.3 years in 2020, according to the CDC's National Center for Health Statistics.
It was the largest one-year decline since World War II, when life expectancy dropped by 2.9 years between 1942 and 1943. Hispanic and Black communities saw the biggest declines.
[...] "The range of factors that play into this include income inequality, the social safety net, as well as racial inequality and access to health care," Curtis said.
(Score: 4, Informative) by Immerman on Wednesday July 21 2021, @06:29PM (1 child)
Not commenting on anything but your own math error - that 2.6% is not the increase in rate at which blacks die - in fact I don't believe it tells you anything meaningful.
As an extreme example - assume some group were only 0.01% of the population, but died at 10x the normal rate - they're such a small fraction of the population that they'd still only be about 0.1% of the total deaths (=0.01% of the population * 10x the death rate). Subtracting the two percentages as you did would get 0.09% - which by your logic would suggest that they were dying at almost the exact same rate as everyone else.
To get the death rate compared to the average you need to divide by their fraction of the population, rather than subtract: 15.1%/12.5% = 1.208, or in other words assuming those numbers are accurate, they're dying at 20.8% higher rate than the average.
Basically, to get the total death rate for a population from the subgroup death rates you'd have to add the proportional rates for each group, e.g.
Group 1 size * group 1 rate
+ size 2 * rate 2 + ...
+ size N * rate N
etc. = total number of deaths
The same math works if you're dealing with percentages rather than absolute numbers - in that case you're essentially just pre-scaling all the sizes (and the total) to a representative population of size 100.
(Score: 2, Troll) by fakefuck39 on Wednesday July 21 2021, @06:41PM
hold on, I get your basic arithmetic, but we're comparing the two links here - that's the whole point of my comment. Did you look at the links? No, so let me paste this here:
1st link:
"Nationwide, Black people have died at 1,4 times the rate of white people."
the numbers it lists are "Deaths per 100 000 people by race"
those are not deaths that need to be standardized to the size of that group - the group size is already 100k people for any race. So why would "assume some group were only 0.01% of the population" matter, since the sample we're looking at is 100k for each race?
>The same math works if you're dealing with percentages rather than absolute numbers
you're doing that math twice. you're taking the number adjusted for that race's % of the total population, and adjusting it again.