Yeah, I know we have do a census because it’s in the constitution. But in the modern age, does the census actually provide information we didn’t already know? I would expect the records of government agencies like the IRS, the post office, and the driver’s license bureau would already provide a far more accurate view of the population than a survey mailed out to houses.
I saw an article quoting government statistics saying this year’s responses from the census in Texas were low, something like only 8% of citizens were responding to it. Which raised the obvious question – if we already knew how many citizens existed to respond, then why bother doing a census in the first place?
Ignoring the constitutional mandate for a census, in the modern age, does it actually provide any information we didn’t already know?
The census attempts to be a complete count. That makes it different from everything else. Driver’s licenses are for adults and there are no national databases of them and no interstate cooperation. People are slow to update addresses and millions of Americans don’t drive in the first place. Similar objections can be made to every other source.
The census also hires millions of people to go after those who don’t return their forms, so its completeness is beyond anything else.
Nothing compares to the census for accuracy. We know good approximations of populations but the final figures always provide surprises. And they are available in usable form for literally millions of calculations that are accurate down to base levels and provide population information that is slicable in unique ways. Both governments and businesses and non-profits use census data daily.
I can’t stress how wrong the notion is that the data is already available outside the census.
The Post Office has records of street addresses, but no idea about who lives there. (And they don’t care. The only time they look at the name is if there’s a COA associated with the address, and that only lasts for a year.)
The IRS doesn’t have any information about minor children (unless they owe income tax) and may not have information about non-working spouses, if the husband files singly.
The DMV doesn’t have any information about non-drivers, including children, and in any case isn’t a federal agency and there are no nationwide standards for the collected data.
We can get rough ideas of the population of an area easily enough – and in fact the Census Bureau does this by using some data from all the sources you listed (and more.) But the only way to get accurate demographic data in a uniform way across the entire country is to do a bigass survey once in a while, and get enough of a sample to be within a reasonable level of confidence that the data are accurate.
Surveys of course do suffer from selection bias – you’re only counting people who think filling out the census survey is a good idea. That’s why survey data has to be supplemented with additional methods of gathering data, like doing door-to-door canvassing in random neighborhoods.
Despite doing proper address changes, half the State of North Carolina and Franklin County’s agencies still think we live at a site that hasn’t even had a structure on it since last summer when the tree fell on our home. There may well be Federal agencies who still have us on record as being in the house we haven’t owned since 1992.
And our present residence “officially” has three tenants, the woman who rents it and her two biological children – but we’ve been here for nine months (Finally got our own place, due to move in a few days) – and she has a ward, her daughter’s same-age friend and ‘sister’ who is the daughter of the mother’s friend with substance abuse issues. In addition, her former boyfriend needing a place to live and a stepson have stayed here for multiple-day periods while working through their issues. Sort out that count and where they are residents without a Census.
Yeah, that’s an unusual-example anecdote. But multiply that beh however many households there are in the U.S., including undocumented residents, homeless kids and the mentally ill, etc., and you have a serious statistical deficit.
Social Security wants to know who they are paying out to and who they are collecting from, but not necessarily in what local government area they live. Go by driver’s licenses and NYC, Philadelphia, and a few other cities with good public transportation will be vastly undercounted. And so on. The Bureau of the Census functions to fill the need for an objective source of statistics which is not accumulating them in pursuit of another major mission, but whose major mission IS to collect them.
If the estimates that article is using are a few percent off, it might mean that the number of respondents is actually 7.5% or 8.5% of the population, which is still close enough to support the point the article is making. But if the Census number is a few percent off, it could lead to Texas gaining or losing a Representative.
Not in the least bit a GQ answer, but …
Seems to me that if Texans don’t want to fill out the census, that’s their decision.
The Feds should count all the forms they do get, and reapportion all Federal funds and representatives accordingly. Once Texas only has 2 Federal reps & only gets enough Federal highway funds for 75 feet of new 6-lane interstate per year, perhaps they’ll recognize what a good deal it is for the Feds to know they really exist.
After 10 years perhaps they’ll be better at forms-filling out and a little less pushy with the loud but uninformed rhetoric.
Same goes for the other places which claim to hate the Feds but seem to love cashing checks from them.
Children should be treated like children and leave the important stuff to the adults.
I don’t think the census is all that accurate. If you had the cooperaton of the various government agencies sharing info, you could get a much better result.
The fact is, time and time again, we find the census releases numbers and they don’t come near the number a city estimates. So the city will pay for its own recount and what do you know the numbers ARE higher and the estimates were closer.
You have to remember census workers are temps that have nothing to gain or lose for their jobs. They can make stuff up or try half assed and it doesn’t effect them.
I think if you combined the actual count with the other information you get from other agencies you could get a very good result. Of course poltically this isn’t possible.
I know in the 1990 census I didn’t have a permanent place to stay and no one counted me. If they went by my driver’s license they would’ve gotten me at an address I hadn’t lived at in over 9 years. Indeed from 1981 - 1994 my DL had an address I hadn’t lived at.
So there is no one best answer, the key would be combining information, which today would be very easy to do, but unthinkable even 30 years ago
All of those sources have strong motives to cheat, or at least misrepresent the distribution.
I don’t imagine that the census is lily white and perfect, but at least they are trying to count people in a neutral manner.
One great source of error in any counting process will be illegal immigrants. Nobody wants to stand up and be counted by any government agency if they aren’t supposed to be here. The census forms say quite clearly that the data cannot be used for immigration or other law enforcement purposes; they are at least trying to make it so a greater percentage of individuals will step forward.
So, to answer the OP, along with the other problems already stated (e.g. other datasets don’t cover everyone), there is no other source of relatively untainted data on which to base these important political decisions.