Using a remarkably (if perhaps unconsciously) biased survey sample, a recent study has shown that the DACA program does good things for the economics of its beneficiaries and, indirectly, for those of state governments.
The newspapers, of course, picked up the cheerful news and ignored the bias — which is, I must admit, a bit harder to understand. (DACA is the Obama administration's by-fiat temporary legalization program for relatively young illegal aliens who came to the United States before their 16th birthdays.)
The study was conducted for two left-leaning organizations, the National Immigration Law Center and the Center for American Progress, as reported here. The results were based on a survey sample of 526 DACA beneficiaries.
Of course, the newly and partially legalized aliens are doing better than before; they are, after all, three years older than they were at the time of legalization (which boosts earnings at this time of life) and now can work legally (which can also be expected to increase wages). But the problem is not whether DACA was helpful to its customers, but whether it should have happened at all.
As I was reading the to-be-expected findings I ran into this sentence:
In Texas ... the survey data showed that 33 percent of the respondents bought a car at the average cost of $10,346. ... [T]he survey goes on to calculate the average state tax payment to be $647 per car, based on a 6.25 percent tax rate.
Then I did a rapid calculation; current USCIS statistics show that about 16 percent of the DACA approvals to date are in Texas. If 16 percent of the sample population was from Texas, there should be 84 respondents in the state, and 33 percent of them would be 28, a modest number.
A population that buys cars at these prices is not exactly poverty stricken, which is the unstated suggestion of the survey. Further, I know from previous surveys, some of which I conducted, that illegals in Texas always earn less than those in the rest of the country, suggesting even more prosperity for the entire DACA population. I noticed, however, that the survey did not report the national data on either the purchase of cars or their expense, nor did it show the average amount spent on cars by all respondents in Texas.
In surveys of this kind one always gets better, more representative data from the universe, not from a fragment, and 28 aliens is no more than a fragment. Then, concerned about the appropriateness of the sample, I turned to this in the report:
The survey, with a sample size of 546 respondents, was conducted online in June using a peer-to-peer sampling strategy to identify DACA recipients and advertisements on social media sites such as Facebook to create a wider respondent base.
Now a peer-to-peer sampling technique produces more responses, to be sure, but they tend to be from the same kind of people that you started with; and a sample of the illegal-alien population that is based on social media is highly likely to be skewed toward the better-educated and the more prosperous subset of that population.
It is a little like talking to members of the Harvard class of 1980 and then extrapolating the results to the population of American college graduates generally.
A similar fatal error is well known in the polling business. Back in 1936, there was a prominent publication called the Literary Digest, which was well regarded for its polling. That year it predicted a comfortable margin of victory for the GOP's Alf Landon over FDR. It turned out that Landon did, in fact, carry Maine and Vermont, but FDR got the other (then) 46 states.
Largely because of this error, the publication soon went out of business.
What the publication had done was to conduct its polling among those who had telephones, and they missed the massive support for the Democrats among the then large population without telephones. Those conducting the DACA study have repeated the Literary Digest error nearly 70 years later. (Many DACA aliens are barely literate and do not have computers.)
No wonder the study found a small subset of the DACA illegals paying so much for cars!