Comparing mortality rates? 2-sample t-test vs Chi sq vs Poisson regression vs others

Hi there,

Sorry for my ignorance but been a while since I did stats.

I am comparing mortality rates between two breeds of pigs. My ultimate goal is to figure out whether the difference in the rates can be explained by random variation or whether there is a breed effect (assuming all other management parameters were kept same)

Breed A has had 7 different populations each of which have their own mortality rates.
Breed B has had 19 different populations each of which have had heir own mortality rates.
For each of the rates, I am unaware of the total number in the population. But should be relatively similar.

I have done 2-sample t-tests. The results of which do not allow me to reject the null hypothesis that the proportion averages are the same. ( i have assumed that the number of pigs in each population and between each of the groups are same in order to do this test).

I read in a thread that Poisson regression is appropriate when comparing rates, unfortunately, I cant seem to get my head around this method and thus cannot identify whether it would be suitable in my case.

A few studies I have looked at on comparing mortality rates in animals have had very rudimentary stats in them but one of them used chi-sq for association. I dont know how I can use this method since I only have rates.

If someone can point me in the right direction I'd be incredibly grateful.


Active Member
Re: Comparing mortality rates? 2-sample t-test vs Chi sq vs Poisson regression vs oth

If you can get the actual counts of deaths in each population, you should run Poisson regression... If all you know is the mortality rate in each population, which is [# of deaths] / [population size], you should run a generalized linear model where beta distribution is the distribution of the dependent variable. For both of those tasks, R is the most convenient package, to the best of my knowledge. R is available for free from here