----- Original Message -----From: Margit JohanssonSent: Sunday, February 25, 2007 11:29 AMSubject: statistics and vote fraud, from Cornell UPolitical science
Election forensics
Feb 22nd 2007 | SAN FRANCISCO
From The Economist print editionHow to detect voting fiddles
A WEEK is a long time in politics. And a decade is a long time in political science. Ten years ago, bright young academics would probably have thought that analysing the impact of ballot formats and comparing the merits of voting machines was unworthy of intellectual pursuit. The 2000 presidential election in America, with its butterfly ballots and controversial outcome, changed all that and spurred a more scientific approach to studying voting and how it might be subverted.
One way to detect fraud is to use statistics. Walter Mebane and his team at Cornell University have devised a new method of doing so, which they described to the American Association for the Advancement of Science. It is similar to that of a mathematical curiosity known as Benford's law. This law states that in certain long lists of numbers, such as tables of logarithms or the lengths of rivers, the first digit of each number is unevenly distributed between one and nine. Instead, there are far more numbers beginning with one—about a third of the total—and far fewer starting with nine. For example, a 2km stream is twice as long as a 1km stream; by contrast, a 10km stream is only 11% longer than a 9km stream. So you will find more streams measuring between 1km and 2km than between 9km and 10km.
![]()
The pattern that Dr Mebane has detected concerns not the first but the second digit of lists of election results from different precincts of a constituency, where he also observes a non-uniform distribution of possible digits. The effect is far more subtle, with zero occurring about 12% of the time and nine turning up on 9% of occasions. Dr Mebane has observed this pattern with great consistency for voting results from precincts across the United States in recent elections. The cases where there are significant deviations from the pattern correlate well with places where there have been known problems of election rigging.
In contrast with more sophisticated statistical analyses, which require laborious research to correlate voting patterns with variables such as race, wealth or historical precedent, Dr Mebane's test can be applied to data without further ado. It is a very simple test for fraud.
Dr Mebane is careful to point out that the test is not foolproof. It sometimes fails to detect a discrepancy in a vote that is known to have been problematic, and occasionally detects fiddling where there was none. However, he has managed to develop a mathematical model that explains the distribution of the second digits, putting what might appear to be a statistical oddity on a more solid footing. He has also had some encouraging success using it in practice.
One example concerns an analysis of the last three elections in Bangladesh. The 1991 election showed no strange results. For the 1996 election some 2% of results were problematic. And fully 9% of the results in 2001 failed the test. The 2001 election was fiercely contested. Yet monitors from the Carter Centre and the European Union found the election to be acceptably, if not entirely, free and fair. Tests like Dr Mebane's one could provide monitors with quantitative estimates of exactly how free and fair an election has been, on which to base their qualitative judgment of whether that is indeed acceptable.
As Jasjeet Sekhon of the University of California, Berkeley, points out, the emerging field of election forensics presents opportunities as well as risks. In the best cases, it is encouraging political scientists to team up in an interdisciplinary way with statisticians, economists, psychologists and even software engineers, to understand the complex interplay between human and technological factors that can influence an election. These collaborations force political scientists to question some of their cherished assumptions in the light of insights that experts from other disciplines bring to the problem.
On the downside, the excitement surrounding this sort of research can attract self-appointed experts with a limited knowledge of the realities of voting. For example, statistical discrepancies between exit polls and actual votes cast have been used by some to deduce widespread fraud on statistical grounds. In fact, it has long been known that exit polls, though they may claim to be based on random samples, suffer from biases in the way that they are conducted. In America these tend to skew the results towards the Democrats.
Dr Sekhon notes that, thanks to Karl Rove, the mastermind of George Bush's election victories in 2000 and 2004, the Republicans have been far more effective at exploiting new scientific information about voting and elections. Indeed, several of his colleagues have been surprised to receive hand-written notes from Mr Rove pointing out minor blemishes in their academic articles. "In contrast, the Democrats have had faith-based electioneering tactics," says Dr Sekhon. The party often engages in activities such as contacting people who have voted in the primaries, which can be shown to be a waste of time and money. In 2008 its workers would do well to consult the latest scientific literature instead.
Advertisement
Advertisement
![Click here to find out more!]()