The p-value is not the probability that the null hypothesis is true, a common misconception. Rather, it is the probability of observing data as “unusual” as the actual data, computed on the assumption that the null hypothesis is true. The last step of hypothesis testing is to compare the p-value to a (pre-determined) threshold of statistical significance, such as 5%, 1%, or 0.1%, to decide whether to conclude that the election results reflect irregularities (which occurs if the pvalue is sufficiently small). If the null hypothesis is true (i.e., if the election results are accurate) and many repeated samples are taken and inspected, the null hypothesis will be falsely rejected for a known expected fraction of samples.