I was recently going down a wikipedia rabbit hole on the current U.S. immigration policy before I came across a very bizarre claim that was stated like an objective fact with no caveat.
The summary table on the page about Deportation in the second Trump administration states "nearly 3 million illegal migrants deported (as of 20 January 2026)". Which is very likely complete bullshit.
The source is this webpage from the DHS website. The actual article itself doesn't provide any citation of an actual formal government report. Doing a little digging, another DHS announcement in August provided a graph from the pro-restrictionist Center for Immigration Studies.
While the DHS didn't bother providing it on their page, you can see the actual source for the self-deportation claim here.
Analysis of the CPS data shows the total foreign-born population of all ages, both in and out of the labor force, declined an unprecedented 2.2 million from January to July – the largest six-month decline ever within the same year.
Non-citizens accounted for all of the falloff in the total foreign-born; the naturalized U.S. citizen population has actually increased some since January.
We preliminarily estimate that the illegal immigrant population declined an astonishing 1.6 million (10 percent) to 14.2 million from January to July of this year.
They do try to do
So it's based off of BLS survey data. Given the current political environment, it's not difficult to imagine that the median noncitizen would be hesitant to disclose their status, no matter what promise of anonymity they were given. This would imply a significantly biased sample. (That, combined with my prior that ~3 million "self-deportations" is just completely fucking absurd, makes me a skeptic.)
In fact, they admit it themselves:
There are a few important caveats. First, given recent stepped-up enforcement efforts, it is possible that the observed decline in the foreign-born was due, at least in part, to a greater reluctance by immigrants to participate in the survey or to identify as foreign-born; 2) Some of the administrative data necessary to estimate illegal immigrants is not yet available, making our estimate for July only preliminary;
One could make the argument that there should be a similar decline observed among naturalized citizens if we assume survey respondents (or rather, nonrespondents) are acting under political pressure. Indeed, figure 2 shows that not to be the case. But as of February 2026, the scale of mass deportation is a much more prominent and present reality vs mass de-naturalization. The two groups are currently operating under very different incentive structures, so it is not surprising to see more non-response bias from one of them. Nowhere is this issue sufficiently addressed.
A side tangent, but also from the same page:
Interestingly, the number of unweighted U.S.-born respondents has declined by 3.5 percent since January. Given the heightened enforcement environment in recent months, an increasing share of foreign-born residents might claim to be U.S.-born, reducing the size of the foreign-born population in the survey. But the decline in the number of unweighted U.S.-born respondents indicates this is unlikely to be the case. Further both the number of unweighted naturalized citizens (2.4 percent) and non-citizens (14.8 percent) respondents declined as well.15 So it also does not seem to be the case that non-citizens are claiming to be naturalized citizens either. However, additional analysis needs to be done to rule out this possibility.
To clarify, they're talking about the raw number of US and foreign-born respondents — not their respective proportion of the sample. They make the argument that, because the number of US-born respondents declined, it follows that it is unlikely that foreign-born respondents opted to lie about their citizenship. Somehow, the possibility that the former offset the latter but both nevertheless occurred just completely escaped them. This shit was cooked on the tomfoolery machine.
To clarify, my claim isn't necessarily that the statistic is wrong (though I personally think it is), but rather that it isn't sufficiently supported to the point where it's basically just pulled out of their ass. Yet the median person, who may even be well-intentioned and not overtly partisan, will read that number and treat it like a fact without knowing any of the background context behind it.
This is just one number, but this happens again and again. Populists thrive in a post-truth environment. Taking the time to look at the data for yourself is the best way to fight back.
Posted by pcqz