Nobody knocked on every door in America to find out where everyone was born. They asked a slice of the country — and that changes what the number is.
The government asks the country about itself in two very different ways. The census tries to count every single person — and it happens only once every ten years. The American Community Survey goes out instead to a sample of addresses — about 3.5 million of them — and it is conducted every month, every year.
Basis: U.S. Census Bureau, “American Community Survey and the Decennial Census”: the ACS is “Sent to a sample of addresses (about 3.5 million)… Conducted every month, every year”; the decennial census “Counts every person living in the 50 states, District of Columbia, and the five U.S. territories… Conducted every ten years.”
Because it asks a sample, an ACS number is not a head-count — it is an estimate of what a head-count would have found. The Census Bureau says so itself.
Basis: U.S. Census Bureau, “Understanding and Using American Community Survey Data: What All Data Users Need to Know”, ch. 7: “The data in American Community Survey (ACS) products are estimates of the actual figures that would have been obtained if the entire population—rather than the chosen ACS sample—had been interviewed using the same methodology.”
That is why every published ACS figure comes with a margin of error printed beside it — the give-or-take. It measures one particular thing: how much the answer could wobble just because a sample was asked instead of everybody.
Basis: Same chapter: “Sampling error is the difference between an estimate based on a sample and the corresponding value that would be obtained if the estimate were based on the entire population,” and “The margin of error is the measure of the magnitude of sampling error provided with all published ACS estimates.”
The give-or-take is not an apology for sloppy work — it is instructions. Add it and subtract it, and you get a range the true value is expected to sit inside 90 times out of 100. Two states a hair apart may be too close for the survey to separate at all; a state far out in front is a different story.
Basis: Same chapter: “The MOEs for published ACS estimates are provided at a 90 percent confidence level,” and “By adding and subtracting the MOE from the point estimate, we can calculate the 90 percent confidence interval… a range expected to contain the true or population value of an estimate 90 percent of the time.”
So why do some states hold on to the people born there while others fill up with arrivals? The reasons usually offered are family staying close, the cost and trouble of moving, and where the jobs opened up — plus the plain arithmetic that a state which grew fast recently grew mostly from newcomers. Those are explanations people find reasonable, not things this survey measured, so treat them as hypotheses.
※ A hypothesis is an idea that isn't proven yet.
📊 Sources for the rankings mentioned in this note (links to the original data and retrieval dates) are on each quiz page below.