General
CORT
Warehouse Supervisor Construction
Komatsu Equipment Co
Mechanic Health Care
Rio Salado College
PA's/Online Instructors General
CORT
WAREHOUSE/DRIVER Education
Assessment Technology, Inc
Social Studies Content Writer

Throughout most of my lifetime, Americans have been rather casual about voting. Voting percentages have been low, even in elections about real things that affect real people. No one ever worried about voter fraud, buying votes or throwing elections.

The 2000 presidential race changed all that. Not only did the winner of the popular vote lose the election, but it is far from clear that the electoral winner actually won the electoral college. In effect, George W. Bush was elected by one Supreme Court justice, giving the term "one man, one vote" an entirely new meaning.

Since the mishandling of the 2000 election in Florida, we have read a great deal about electronic voting and the problems that result from turning the counting of votes over to computers. Independent investigations seem to show that computers (both hardware and software) are hackable, meaning that the results can be manipulated.

This has spawned legislation to require paper ballots, either instead of or in addition to electronic voting. But it is unclear whether any of the solutions will work, and I think this is because the focus has been on the wrong problem.

The fact is that you don't need computers to miscount votes. Legend has it that in Richard Daley's Chicago election, fraud was quite manageable with paper ballots. The essence of the problem is that the vote counting happens secretly. The moment this task is turned over to people who know they are not being watched, the possibility of fraud arises.

It does not actually make much difference how professional or responsible the counters are, because with computers it is also possible to make honest mistakes in programming or data management, which can easily go undetected.

Those of us who do data analysis know that there are ways around this.

One simple solution is to establish a system of "sentinel voters." These are people who are willing to waive their right to a secret ballot and allow their votes, as recorded in the electronic database, to be turned over to one or more certified commissions, who will compare the electronic record with their actual votes, and report to the public if there are significant discrepancies. In many ways, this is the acid test of election validity, that the computer database contains the correct votes.

The second test is that the entire database also be made available to the commissions, for their independent tallying of the votes. This is actually much easier than it sounds, and it could be done on a cheap laptop with existing off-the-shelf software. It would provide the needed check on the software in the voting machinery.

Of course this scheme is potentially hackable, too, but it is an order of magnitude more difficult than it is now. The hackers would have to find out in advance who the sentinel voters were so that they wouldn't allow their hack to change the votes of these individuals.

By revealing the identities of the sentinel voters only after the final election database was closed, this kind of fraud could be made exceedingly difficult.

Software errors (deliberate or not) would be much more likely to be discovered, due to using two entirely different programs.

We will probably never be able to stamp out illegal actions that deprive some citizens of their right to vote, nor will we ever completely eliminate incompetent ballot design, as happened in Florida in 2000.

But we can take steps to make the recording and counting of votes as visible to the public as these other abuses, and where light is shed, fraud and error become less likely. If we are willing, we can recapture our elections.