[yk]

The main problem of this paper seems to be 'publication bias.' I mean by this, given a number n of statisticians who look at a certain fair election. Each using n novel method. Then we expect ß n papers from it, given that ß is the chance of falsely rejecting the election as fair. Since each statistician uses a number of tests, which may or may not correlate with each other, it is very much impossible to determine sufficient level of certainty, without committing first to the statistical tests used.