Page II: India reports success, while Venezuela fears fraud. What can the world learn?
Without a reliable audit, the groups that sponsored the referendum to remove President Chavez pointed to two apparent statistical anomalies as proof that the machines had been tampered with. About 400 precincts, or mesas, had two or more machines that registered the same number of "yes" votes, and 380 had two or more machines with the same number of "no" ballots. Other precincts had fewer "yes" votes than the number of signatures on their local recall petitions.
A statistical analysis of the first anomaly by three computer science professors -- Johns Hopkins' Rubin and Adam Stubblefield, and Princeton University's Edward Felten -- calculated that the pattern of identical votes could happen but noted that the process by which the votes were counted and recounted was not designed well. "Venezuela is an example of how not to do voter-verified paper trails," Rubin said.
Two Venezuelan economists believe that they have shown that it is unlikely that the elections were fraud-free. An analysis of the mismatch between the demographics of people that signed the recall petition and the actual votes concluded that there is a 97.5 percent chance that someone tampered with the recall vote, said MIT's Rigobon, who penned the second analysis along with fellow Venezuelan citizen Ricardo Hausmann, an economics professor at Harvard University's John F. Kennedy School of Government.
"The most important lesson from Venezuela is that doing fraud in electronic voting is extremely easy and very hard to track," Rigobon said.
The Carter Center has responded to the Rigobon-Hausmann study with its own analysis, showing that the election continues to appear fair.
Inspiring trust in India
Indian officials say the issue of voter trust and confidence is of paramount importance to their country as well, given its history of protests and violence on election days. Tampering is so common that the country even has a name for the seizure of polling stations from poll workers and police: "booth capture."
8%
1%
