Abstract:
Judges or members of Election Management Bodies (EMBs) of many
less developed democracies are often faced with reviewing electoral
complaints and coming out with a decision that is acceptable to all
parties involved. Some judges have ruled for complete cancellation of all
votes in the affected polling stations. Where the electoral irregularities
are widespread, it is very difficult to declare a winner without incurring
the additional cost of a re-run. In this article, we use the hypergeometric
model to provide an alternative method to direct cancellation or
proportional deduction approach used by many EMBs in the settlement
of electoral disputes. To be specific, we provide a bound on the margin
of victory when either a re-run or fresh election is not necessary to
determine the winner of the polls, i.e., the non-overturn condition. We
use this method to investigate whether or not the decision by the
Electoral Commission of Ghana to declare the NDC candidate the
winner of the presidential run-off of 2008 is statistically correct given the
data available to the Commission on the day to declare the winner. Using
this approach in the review of electoral complaints could help developing
nations to save some money for development that would otherwise have
been spent on election re-runs