ACS1 vs ACS5 Can we even use ASC5 for many measures (Working from home, UI, Mean # in Household etc.) is it better than ACS1?

I always read that the ACS5 is more comprehensive and accurate. However, how can you even assume that percent working from home, income, and some other key measures are more accurate in ACS5? The world has changed a lot over the last 5 years, using all years I would think would NOT be a better representation that ACS1. Yes, the ACS1 has less geography (especially in county where I look the most).

I also think that the economy and many other features of our world are going to be much more variable and have much bigger swings within a 5 years span. And don't forget it takes a year for the numbers to come out.  So I think using many measures from the ACS5 will be fairly worthless (compared to the ACS1), especially from large areas that have data in the ACS1.

Any thoughts, or comments?

Tom

Parents
  • I think you've kinda answered your own question. It's a matter of whether you want precision in time or precision in location. If you're only dealing with larger areas, it's a question of whether you want precision in time or sample size (particularly if your variable is very specific). If one were always better than the other, there'd be no reason for the Census to produce both. So it really depends what your aim is.

Reply
  • I think you've kinda answered your own question. It's a matter of whether you want precision in time or precision in location. If you're only dealing with larger areas, it's a question of whether you want precision in time or sample size (particularly if your variable is very specific). If one were always better than the other, there'd be no reason for the Census to produce both. So it really depends what your aim is.

Children
  • Actually, I think if you work with a large area (again I'm looking at counties, that in ACS1 represent 85% of the US population) The ACS1 is better in precision in time and precision in location for counties that it has. Now that the census has responded to my question about the use of average and mixing and matching mean, medium and average within the same survey results (it's absolutely OK), perhaps I'll just mix and match ACS1 and ACS5 data and use ACS1 for counties and fill in the gaps with ACS5 for the remainder, because why not.

  • Because they're not comparable. The 1-year survey uses a different sample than the 5-year. I suggest you review the Bureau's page, Period Estimates in the American Community Survey, before you "mix and match".