I am analyzing the specific household and individual characteristics of a limited number of counties in New Jersey. I downloaded the 2016 household and individual variables I wanted and tried to check my population counts to verify everything. However, when I apply the HHWT or PERWT to the PUMA counts I get a population of 8.9 million. According to the FactFinder the population should be closer to 3.1 million. I double checked my PUMAs and made sure the sample was limited to the geography in question, but it still seems way too big.
What am I doing wrong?