SheepDog4Life
Natural Gray Man
So by quickly looking at this article by a source I have no reason to either believe or not believe the following thought comes to mind. This is after all only one source are there any others and how credible are these sources?
Gee, it's not like it's published in a peer-reviewed academic journal about elections or anything...
There were two other published papers, one used Harvard survey data, the other used government survey data. Both concluded that 14-16% of non-citizens register, and a third to half of those vote. The "argument" against those two was that the sample size was too small, though the 95% confidence error did not reflect this criticism (the standard was of expressing possible inaccuracy in a sample).
Anyhow, for this 2014 paper the researchers conducted their own original survey data with a sufficiently large sample size... and they came up with essentially the same result... 15%.
There are three different academic papers out there, each based upon a different survey data source (i.e. the three are independent), and all three conclude that non-citizens ADMIT registering to vote at the rate of 15% +/- 1%.
And, these are likely "floors" as the survey bias is to NOT admit committing a crime.