ACS Data Users Group

I can't make sense of this data. For example consider zip code 94947, a town in Marin County CA.

Population 25,183 with median income (B19019_001E) 121,388. So far so good.

Now B19001_001E (total) is 9,204. And this is the sum of the 16 brackets (B19001_002E - B19001_17E). But what does 9,204 represent? The ACS subject definitions (page 88) says "“Total income” is the sum of the amounts reported separately for wage or salary income..." But 9,204 is far too small to represent that total dollar figure. And it doesn't make sense to report dollar figures in income brackets. I suspect that 9,204 might represent the total number of people who reported their income, and each of the other fields is how many people reported income in that bracket. Across all the zip codes I've sampled, this seems reasonable since B19001_001E (total) is roughly 1/3 of the population most of the time. Yet if this is true, then the documentation is either incorrect or misleading. It means B19001_001E is total people who reported income, not the total amounts reported.

B19001 group breaks out the count of households by income, so B19001_001E is giving the count of all households in the geographic unit, B19001_002E gives the count of households with incomes under $10k, B19001_003E gives the count of households between 10K and $14,999, etc.

That makes perfect sense - thank you!

BTW, how do you know this? I searched the docs and couldn't find any explanation except the one source I quoted, which was wrong.

PS: I've confirmed it is count of households. For every zip code, multiply B19001_001E by BB25010_001E (avg household size) and it is very close to BB01003_001E (total population).

I'm honestly not sure where this is explicitly documented. Just going by my experience, the ACS data in the detailed tables (prefixed with 'B') is almost always count data. The exceptions are usually identifiable because they include some calculation in the table name (i.e. Median Age, Median Income, Aggregate Travel Time). The subject tables (prefixed with 'S') are little better at highlighting when the estimate is providing something other than a count. For example S1901 labels the data 'Median income (dollars)' so you can tell that the units returned aren't a count.