Using data obtained by the Cooperative Congressional Election Study, the authors offer the following findings:
Fourteen percent of non-citizens registered to vote in both the 2008 general election and the 2010 midterms.
Over six percent of these non-citizens actually voted in the 2008 elections while over two percent voted in the 2010 midterms. (That translates into 1.4 million votes for 2008 and nearly 500,000 for 2010.)
Such voters may well have provided the margin of victory in Al Franken's United States Senate campaign in Minnesota and in Barack Obama's victory in North Carolina.
In a reversal of patterns regarding citizen voters, non-citizens without a college degree reported higher participation than non-citizens who earned a college degree. The authors attribute this pattern to lack of knowledge about laws regarding voting among those non-citizen voters without a college degree. It may also rest upon uneducated non-citizen voters merely disregarding the law. Lower educated citizens disregard all laws in higher proportions than to educated citizens.
Interestingly, the conservative remedy for illegal voting--photo identification--seems to have no effect. Photo identification are often required even in fake elections in some countries.
Democrats oppose voter identification laws. (It is funny that even they assume illegal voters will vote Democratic.) The authors report, however, that such regulations failed to prevent illegal voting. Non-citizen voters reported that in most cases they actually provided photo identification. They were illegally registered--but they were registered nonetheless. Since their photo identification matched the information contained in their illegal voter registration, they encountered no obstacle to illegally voting.