The Government Report That Made Me Stop Trusting Our Statistical Agencies

Even if it was in the ballpark of 24,000 non-citizens, Trump lost the 2020 popular vote by over 7 million. Are we supposed to believe those 24,000 were distributed just right in the swing states to win Biden the electoral college? It's statistically implausible.

Are we also supposed to believe this number (if it were close accurate) is supposed to be a concern for the midterms somehow? That a bunch of non-citizens will happen to concentrate their vote in close midterm elections likely in red states that would flip the Congress?