Proposing New Questions for ACS

In addition to what Stas shared about the content review, here is a helpful visualization on how a question gets incorporated into the ACS: How a Question Becomes Part of the American Community Survey (census.gov)

Please note that requests for edits to the survey come from federal agencies. I hope this helps!

Right -- basically the only reason something is on ACS is because there is a federal program that needs to allocate money or resources based on the statistics computed out of ACS. My pet peeve has been phone use: through the 2010s, before the advent of the call blockers, the survey community was in the dire need of local estimates of the relative use of landlines and cell phones for sampling design and weighting purposes. ACS would have been a very logical survey to provide that, but the timeline of the content review did not allow to incorporate something within say a year or two, and there were no federal surveys that really demanded that statistic. You'd think -- hm, there's the National Immunization Survey is a phone survey that dials maybe 2M phones a year, and Behavioral Risk Factor Surveillance System may be dialing even more, but it exists at the state level. So none of these seemed worthy enough to push the detailed phone use question onto ACS. The basic question of "any phone" use seems to find way in some surveys like the National Health Interview Survey, and had its valid reasons in early 2000s when ACS was taking off the ground, and is probably informing something; but this is a missed opportunity for the rest of the survey world. (We have been getting by with the NHIS-based state-level estimates, and the phone surveys are pretty much dead by now anyways, with the survey world mostly moving to address-based, mail-push-to-web designs.)

Right -- basically the only reason something is on ACS is because there is a federal program that needs to allocate money or resources based on the statistics computed out of ACS. My pet peeve has been phone use: through the 2010s, before the advent of the call blockers, the survey community was in the dire need of local estimates of the relative use of landlines and cell phones for sampling design and weighting purposes. ACS would have been a very logical survey to provide that, but the timeline of the content review did not allow to incorporate something within say a year or two, and there were no federal surveys that really demanded that statistic. You'd think -- hm, there's the National Immunization Survey is a phone survey that dials maybe 2M phones a year, and Behavioral Risk Factor Surveillance System may be dialing even more, but it exists at the state level. So none of these seemed worthy enough to push the detailed phone use question onto ACS. The basic question of "any phone" use seems to find way in some surveys like the National Health Interview Survey, and had its valid reasons in early 2000s when ACS was taking off the ground, and is probably informing something; but this is a missed opportunity for the rest of the survey world. (We have been getting by with the NHIS-based state-level estimates, and the phone surveys are pretty much dead by now anyways, with the survey world mostly moving to address-based, mail-push-to-web designs.)