ACS Data Users Group

I'm new to ACS and trying to use state estimates as denominators to estimate rates with confidence intervals over time, specifically for under 5 years by race. I found the necessary population estimates but they provide an estimate and error for the total population and a percentage and error for the <5 population. The data I'm using for numerators is complete data and while I have calculated rates with CIs for estimated numerators, I have not calculated them using a denominator with error before. Does anyone have a recommendation for how to do this? Or a place to look for the methodology?

What geographic levels are you using?

Hello! I'm using state level data.

Hi Jordan -

I think the process you need can be found on pages 2-3 of the Instructions for Applying Statistical Testing document. It sounds like the formula for a proportion is the one you want (but depending on how your numerator is related to your denominator, you may need to use the formula for ratio).

See www2.census.gov/.../2016StatisticalTesting1year.pdf

Hello! Thank you for replying!
This pdf is very informative, but it looks like it assumes a standard error on the numerator as well. I have complete counts for the numerator.
Thank you for the pdf link!