I want to replicate the statistics on the poverty rate for families and individuals. I would like some guidance on how to do this.
This is what I think I should do:
Calculate the poverty rate for families first. Comare family income to the poverty threshold for their given size. Find the ratio of families below the poverty line to those above. Do I need to do any weighting?
Next, calculate the poverty rate for individuals. Basically assume that every individual in a poor families is poor themselves. For every family below the poverty line, add up all the individuals in that family. Compare the total number of individuals below the poverty line to the total number above.
What you outlined is the basic idea. Keep in mind that the variable OFFPOV already assigns the official poverty status to respondents included in the poverty universe, if you are concerned more with the final poverty designation rather than recreating this status yourself. When you are calculating the poverty rate, either for families or individuals, you will want to use the household or person weight, respectively. This will give you an estimate representative of the population poverty rate, instead of the sample poverty rate.
For more details on replicating the official poverty rate using IPUMS-CPS data, please see this User’s Note. In particular, you will want to use the variable OFFCUTOFF for the poverty threshold values and OFFTOTVAL for the family income value.