Researchers at two of the nation’s most esteemed universities, Cal Tech and MIT, have confirmed what many Americans suspected: A significant number – between 4 million and 6 million, it turns out – of votes cast in the 2000 presidential election went uncounted due to faulty voting equipment, and even faultier voting procedures. Further, while Florida got all the attention, and blame, for the debacle, three states – South Carolina, Georgia and Illinois – did worse.

With almost precisely 100 million votes cast last November, this means the margin of error for the most important election the world’s leading democracy holds is above what would be acceptable for an unscientific telephone survey of consumer preferences. We are likely to have a better idea of what Americans look for in a toothpaste than what they look for in a president.

By no means should this study, or the previous evidence of disproportionate concentration of electoral problems in minority and low-income neighborhoods, be used to question the legitimacy of President Bush’s victory. He won under the existing conditions and there is no reason to believe that the dreadful margin of error did not hurt both sides equally.

The shame of it is that those conditions existed for so long and were so well known – not to the average American voter, who had every expectation of fairness and accuracy, but to elections officials, from the federal level down to the precinct. As long ago as 1975, the agencies now known as the National Institute of Standards and Technology and the Federal Election published a report citing significant and growing technical and management problems in the electoral system that had clear and substantial effects upon the integrity of the vote counting process. This led to the codifying of voluntary standards for states to adopt – such as, happily, Maine – but in many others little or no progress has been made.

The Cal Tech/MIT study is more than a highly educated guess on what ails American elections; it is a remarkably precise diagnosis. On the technical side, punch card ballots and, to a slightly lesser degree, lever machines are prone to uncorrectable problems. Optical scanners, such as used in this state, are by far the most accurate and, because mismarked ballots are rejected on the spot, offer the opportunity for correction.

At a cost of $5,000 and up, optical scanners are expensive, certainly far more expensive than continuing to use the old punch card and lever machines paid for decades ago. Since the actual conduct of elections is determined at the local level, that is one reason why the miscounts were so high in poor districts. That’s a good reason for Congress to begin serious consideration of the several election reform bills that include money – the leading contender includes $357 million – to help local jurisdictions make the change.

On the management side, the university researchers and the National Association of Secretaries of State, now meeting in Little Rock, agree that simply plugging in new machines is useless unless there are corresponding improvements in registration processes and polling place techniques. That is a view supported by the chronic troubles some Maine communities have in keeping track of registered voters. Both groups suggest giving election officials computer access to the complete voter rolls at the polling place to facilitate checks, and proposed that states develop qualified voter files. The Cal Tech/MIT report is especially critical of unrestricted absentee voting as being especially susceptible to coercion and fraud, and took a very dim view of Internet voting, saying the security issues are far from resolved.

The biggest impediment to reform is not money, however, but attitude. The belief persists, in Congress and elsewhere, that elections are a local matter and local officials should set the rules. This ignores the overarching federal role in everything from setting the voting age to guaranteeing civil rights. Mostly, it ignores that we’re trying to elect a president, not select a toothpaste.