# Which stats test to use in this situation

1. Aug 19, 2013

### TrpnBils

This is a two-fold statistical thing I want to pursue and I'm unsure of which test(s) to use. Using data of gender and age at time of death from a large, old cemetery in town I want to compare two things:

1. Average life expectancy of citizens of our town compared to the national average
2. Whether there is a noticeable difference in life expectancy after the development of major vaccines over the past 100 years.

I'm figuring I will lump ages into groups of 5 years (died at ages 0-5, 6-10, etc) so there will be multiple groups, not including separation of genders.

Thoughts?

2. Aug 19, 2013

### Staff: Mentor

To get the current life expectancy, I think you will need data about recent deaths, together with the age distribution* in your town.
You also need the assumption (or data) that "gets buried at the cemetery" is independent of the age at time of death.

Same issue, and there it is even harder to get the age distribution in the living population. It might be a reasonable assumption that moves in/out of town are independent of the age*, in that case you can just count how frequent different ages are.

* a simple example why this is relevant: imagine a small town with a big old people's home. Many old people will move to that town and die there. Does this mean that all inhabitants there reach that age? Of course not.

3. Aug 19, 2013

### TrpnBils

So is it even possible then to compare something like the average life expectancy of a male born in my town in 1930 to the national average for males in that year?

It doesn't seem like that should be that difficult

4. Aug 19, 2013

### Staff: Mentor

It is possible, but it is not easy. It is not sufficient to count gravestones for deaths in 1930 (or any other year).