For example, there is very strong evidence that Scott Brown reached the US Senate as a result of election fraud. Details are in http://electiondefensealliance.org/files/BelieveIt_OrNot_100904.pdf That analysis compared the results in machine count jurisdictions and hand count jurisdictions. The usual disparity between hand count and machine count results (based on prior elections) runs around 0.25%. Coakley led in hand count jurisdictions by 2% and Brown in machine count jurisdictions by 5%. That is a 7% disparity. It also turns out that the company operating the machine counts was Republican-connected, and that the ballots were neither saved nor sampled to validate the accuracy of the machine counts. There are numerous ways to tamper with a machine count of paper ballots, especially in a two-person special election.
The method published in the subject paper could not pick up this kind of election fraud.