Re:Changes in midlife death rates across racial and ethnic g...
by hyperpape on Thursday August 16, 2018 @09:52AM
(#17774586)
Yes, I read this sentence.
This is a problem with common ways of doing age adjustments.
The issue is somewhat subtle: if you partition people into age bands (45-56), at different times in history, the average person who is in that age group may be older or younger, because different years had higher or lower birth rates. This was a particularly strong effect when the baby boomers were in the 45-56 age range.
My initial comment was that this paper didn't make clear whether they addressed the issue. Having now followed through to the CDC link they have, it looks like it also uses somewhat coarse age bands.
This is a problem with common ways of doing age adjustments.
The issue is somewhat subtle: if you partition people into age bands (45-56), at different times in history, the average person who is in that age group may be older or younger, because different years had higher or lower birth rates. This was a particularly strong effect when the baby boomers were in the 45-56 age range.
My initial comment was that this paper didn't make clear whether they addressed the issue. Having now followed through to the CDC link they have, it looks like it also uses somewhat coarse age bands.