You can derive the other lifetable columns from ex. That’s ehat Sauerberg does here: www.sciencedirect.com/science/arti...

On this Day of the Dead, let us remember John Graunt, Edmund Halley, & the almost unknown Edward Wigglesworth, inventors of the lifetable (or tables of mortality). Life tables find wide application in insurance, annuities, biology, & epidemiology 1/6 🧵

What´s important to know about SDRs❓

🟠Available by sex and for 1-, 5-, and 10-year periods
🔵Only based on abridged period #lifetable mortality rates
🟡Data downloads into R are compatible with the HMDHFDplus package

More in the Methods Protocol ⬇️
mortality.org/File/GetDocu...

Trying to illustrate how life tables are not predictive of actual survival rates, I compared the 1930 US lifetable for White males to the number of US-born White males born in 1930 counted in each Census. Probably more interesting to me than it is useful for explaining things. Fun!

From Bluesky user @timriffe1.bsky.social‬: 'There is an exact procedure to abridge a single-age lifetable, but is there a method to graduate an abridged one that 1. Matches lx and ex in abridged ages 2. Is fully transitive (e.g. mx-ax-qx identity and others) 3. Is reasonably smooth 4. Has […]

There is an exact procedure to abridge a single-age lifetable, but is there a method to graduate an abridged one that 1. Matches lx and ex in abridged ages 2. Is fully transitive (e.g. mx-ax-qx identity and others) 3. Is reasonably smooth 4. Has standard-looking ax values (close to .5 exc tails)?

Have a dog and want to know how long it can be expected to live (in this modern era when dogs no longer 'run loose' and get hit by cars, and where they get much improved veterinary care)?

Here is the relevant lifetable from an excellent 2022 study. www.nature.com/articles/s41...

Calculating how many people were likely born the same minute I was, distributing them by world region, and applying the relevant lifetable mortality rates to figure out how many of us are left

I am unfortunately not aware of any stats! @hmdatabase.bsky.social has the right cause of death data, but you would need to do some work on the lifetable to partial out these causes of death. Maybe some of my demographer friends know? @jschoeley.bsky.social @timriffe1.bsky.social @ikashnitsky.phd

And, not to litter, but here's a function that does this redistribution, where you need some age-structured counts, and the dx column of the lifetable: github.com/timriffe/Tha...

@timriffe1.bsky.social
Much nicer looking (and more informative) with "lifetable redistribution" at period age-specific rates! Is this what you meant?

And just to be clear, ltre() is not limited to lifetable problems; it's general just like the others.

DemoDecomp R package is now updated on CRAN cran.r-project.org/web/packages..., now including lifetable response experiment function, which uses numerical derivatives in case you don't have your own sensitivity function.

by "lifetable redistribution" do you mean calculate a distrib of death ages? As in: Suppose living x year olds face current period m(x), m(x+1), m(x+2) ... How many are expected to live y=0,1,2,... more years? And then count how many people are in each n category? (That's better than what I did👍).

I'd call @cschmert.bsky.social 's pyramid a prospective age pyramid using the term of Sanderson & Scherbov. I tried to get at thanatological age both directly (surveys w mort follow-up) and indirectly (lifetable redistribution). Had a gif about this somewhere once...