I have considerable data on prisoners that includes sex, religion, age, offence and self-harm.
I want to see if males self-harm more than females and whether religion and age are contributing factors.
As the population is always changing I would like compare the incidence of self-harm during each period taking account of the population in each period (years and months) and be able to say whether there is statistical significance to the results e.g. there are 1000 prisoners in custody, 250 of whom are female and the population changes +- each day. There are 70 self-harm incidents each month some of which are done by males and the remainder by females but the number of prisoners and the number of self-harm incidents constantly vary.
What form of analysis should I use?
I also presume that I need to include records on all those in custody during the relevant time periods irrespective of whether they self-harmed or not; and that it is better to analyse hundreds or thousands of records rather than a small sample size of e.g. 60 – am I correct?
Thanks for your help.