Bad Pols or Bad Polls?

Even as party Democrats work in unison to combat Republican efforts to limit voter turnout that tends to favor Democrats at the polls (voter ID, early voting, absentee ballots, shutting down polling places in Democratic strongholds), under the radar accusations are being leveled at the Clinton campaign (actually unnamed alleged bad actors) of election fraud in the Democratic primaries and caucuses this year. In a paper (not a study - no peer review) that has been rushed to the public due to the importance of timeliness  over academic rigor in an election year, two statisticians assert that electoral fraud favoring Clinton has definitely occurred based on the magnitude of the difference between the preliminary exit poll results and the actual vote tallies in states using voting machines with a paper trail vs. those states without the paper trail.

Source: nymag.com

One obvious flaw in this approach is that they do not consider all of the possible relevant differences between those two groupings of states. The authors point to the internet being "littered" with allegations of electoral fraud as support for their findings as if there is something statistically significant about that - the "where there is smoke, there is fire" argument.
Ironically, it was mainly the states under the greatest Republican control in the early 2000's that started adopting the use of touchscreen machines without the audit trail.
At the moment, the mainstream press is avoiding this topic, though a The Nation article debunking the case for electoral fraud followed an earlier article and some debate on the subject.
For now, the Bernie supporters are running with this one, but we can expect the Trump campaign to pick this up and make their "crooked" Hillary allegations later this year if Bernie supporters and the candidate himself fall in line behind Clinton at the convention. Maybe the authors of the paper can subject their paper to critical peer review in the meantime while they continue their search for alternative explanations of the data. At the moment responsible academics do not appear to be taking this paper seriously, but the conclusion is too juicy for this not to nose its way into the mainstream media in the next few weeks. Bernie Sanders is the one person who could nip it in the bud, but he is trying to hold on to as much power as he can at least through the convention, so he is likely to remain quiet until then.

Source: Huffingtonpost