Hi,
I'm looking to cross tabulate a few different variables with a dichotomous poverty variable (below threshold or not) in the 1-year PUMS. I derived the poverty status variable by recoding the income to poverty ratio. Unfortunately the statistical package I'm working with doesn't allow me to calculate standard error using replicate weights, so I'm using the design factors to approximate the SE. The Accuracy of the PUMS document (http://www2.census.gov/programs-surveys/acs/tech_docs/pums/accuracy/2014AccuracyPUMS.pdf) lists poverty status and ratio of income to poverty as having different design factors. However, I've so far only been able to find ratio of income to poverty level as a variable. Am I missing a poverty status variable? And, if not, is it valid to use the smaller design factor for poverty status since that's what I'm calculating, or do I need to use the larger design factor from ratio of income to poverty level because that is the variable I manipulated?
Thanks for any advice you can provide.
- Elizabeth