I want to use the survey data to calculate the unemployment rate by sugroup on a county or state level. In the FAQ it is stated that weights are calculated on a national basis.
So my question is: Can I use the ACS samples to calculate countyl-level unemployment rates and will the weights differ for the same for identical individuals in different states?
The associated weights assigned to IPUMS USA variables will calculate representative statistics even at sub-national levels. One detail you should monitor is the sample size of your calculations if you are cutting the sample up by geography and by subgroups of other variables. Generally speaking, smaller sample sizes will increase the standard error of your estimates.
I hope this helps.