The article appears to be drawing conclusions, apparently incorrectly, from a flawed study.
A reader comment that illustrate the flaws in the study...
There are really 3 problems I see in how the authors presented their findings which make them very misleading.
1) The authors make it appear that 6.4% of the sample voting electorate were non-citizens. This is NOT the case. Let's look at 2008. The authors say 6.4% turned out to vote. This is only 6.4% of the 339 non-citizens they say are listed as registered voters, NOT 6.4% of the entire sample of voters which total 32,800. So to put this into a more accurate (less misleading) context we take the 2008 “adjusted estimate” from their chart and divide it by the total sample 32,800 which comes out to 0.0006 or 0.06%. This means that non-citizens supposedly voting made up ONLY 0.06% of the total sample. This is even less in 2010, sitting at about 0.01% of the total were supposedly voting non-citizens.
2) They don't even know how many of these “registered non-citizens” voted. By their own admission these are “best guesses”.
3) Then they take those non-citizen estimated turnout percentages (already misleading) and apply them to particular races 2008. This, in of itself, is a very questionable application of the data because it makes the assumption that a significant portion of those non-citizens who voted lived in either North Carolina or Minnesota in 2008. The problem is that the total sample number (32,800) are a NATIONAL sampling, not samples taken in one state. To put this into perspective, for 2008 only 21 out of 32,800 people were supposedly voting non-citizens. That is 21 PEOPLE spread out across the nation. Most likely these non-citizens who supposedly voted were 1 in Texas, 3 in California, 2 in Minnesota. HARDLY significant.
~mpbt
