@JLeslie The billions upon billions of things we count in a complex civilization like ours all depend on the population data we use to put them into perspective. Are 400,000 cancers a little or a lot? It depends on how many are expected which is largely a matter of how many there were last year the age-race-sex-adjusted rate. How do you know how many people there are in each age-race-sex cell in the years in between censuses? Well, this is what demographers do: they keep track of how each cohort changes due to baby booms, baby busts, epidemics, and migrations—and the only chance they get to calibrate their statistical models is once every ten years.
Knowing how the age-race-sex structure of a population differs from place to place allows demographers to model migration, birth and death rate information that greatly increases the inter-census estimates, and that helps to lower the margin of error of any estimate.
When a pollster calls a few thousand people at random and gets 20% of the respondents answering “yes” to a survey question, can he take that finding at face value? Well, no, not exactly. It depends on whether the people at home at the time he survey taker called are representative of the general population. The findings have to be statistically adjusted to discount the overrepresented segments and vice versa—which you cannot do unless you have census data to tell you what a representative sample looks like.
Every once in a while you hear an alarmist story in the news about how illegal immigrants are causing a crime wave as young Mexicans pour across the border, ravaging our white women and filling our prisons with their illegal selves. The problem, however, is that this “crime wave” is a statistical mirage. Once you correct the age difference between the two populations the “crime wave” disappears—and along with it a lot of hot-headed rhetoric and the “need” to build an expense wall to keep out those ravening brown hoards.
Likewise billions upon billions of dollars of decisions depend on opinion polls, sales and marketing data, school enrollments, economic statistics like GDP, to disease incidence counts, and the consumer preferences of demographic groups, all of which only come into focus with accurate “denominator” (Census) data.
@Ron_C The Constitution, by the way, does not say the SOLE purpose of the Census is “to set up congressional districts and allocating federal resources in proportion to the population. Nor does the 14th Amendment render racial questions moot. Gerrymandering still occurs, which is why somebody still needs to collect this information, so it can be proven in cases where it occurs.
The Census is not about collecting “medical” data. It is about interpreting medical data. If you know that Native American ancestry increases your risk of diabetes four-fold, if you run into a case of 2 or 12-fold it is a cause for further investigation.
If you don’t like talking to “the government” once every 10 years; don’t worry, your neighbors don’t have any such inhibition in talking about you. Or, perhaps you would prefer a national ID card that follows you from birth. The odd thing is that you think nothing about blabbering the most damaging and embarrassing information about yourself to complete strangers on the Internet and yet quibble over divulging information can mostly be obtained by taking a good look at you.
The main thing that worries demographers is whether the people who refuse to answer certain questions are just your random dickhead, or whether they have some characteristic in common. If it’s random, no problem. But if it isn’t, it throws their calculations way off. So, they will often go to great lengths to get you to change your mind. The more stubborn a “refusal conversion” you are, the more information you potentially hold about all the people the Census misses, so the more effort they will expend to get you to cooperate.