Hi,
Some tables use a lower (and perhaps higher??) number as a starting point such as S1701. This is usually stated in the field titles somewhere (which is great! ). Some tables have a population number that matches what I expect (on most tables). Sometimes it is different, I think due to people not wanting to disclose this info (they skip the question). However, I don't think they always state it in the table and sometimes the numbers are different. Some tables just have field elements as a percent etc. and I can't know what the starting point was. Did they use the full sample size or not? Is there any reference or method to know which tables use the full sample or not? A few percentage points diff is OK, but what if its 10% or higher, I've seen 15%?
Any advice or thoughts.
Thanks
Tom
Look for “universe” which should tell you if total population or something else is the denominator.
Hi, yes I looked at that. Sometimes it has universe= Total population (less than 1/2 the time), often its blank. Blanks can = Total population and sometimes not, because the numbers are lower. When they list such as "Occupied housing units" or "White alone, not Hispanic or Latino population" its very helpful. However I'm seeing this on less than half the tables .
What’s your source for the tables? The table you mentioned, S1701, has the universe listed in the metadata when you download here
https://data.census.gov/cedsci/table?q=S1701&mode=results
Hi, Tim if you want an interesting story, wait for Jan when I update my ranking report. Always interesting to see who the big guns are. public.tableau.com/.../PubMed_Ranking
It may be that depending on how you structure your search (level etc.) it lists diff universe? I keep the key notes and URL for each table that I access. I haven't spent any time to notice if the universe changes based on how I'm accessing the table, I usually go for geo = county, year=2021, ACS-1 and then a few other things to narrow down.