Weight for multiple years of ACS

Hi everyone,

My PI only wants to study the 2021-2024 data not any five-year ACS data. For each individual year’s result, should I still use PERWT and then for the pooled result, I shall use PERWT / 4?

Thanks,

Miso

When pooling multiple ACS samples together, you should divide the weight by the number of samples. This will affect estimates of frequencies and those derived from frequencies. Dividing the weight by the number of samples won’t affect other estimates, like averages or percentages.

Note that the Census Bureau recommends using replicate weights to analyze ACS data. You can add these weights to your IPUMS extract. We provide more information on replicate weights and sample code for using replicate weights on this page: Replicate Weights in the American Community Survey/Puerto Rican Community Survey.