In three national elections over the past 13 months, the official count was sharply at odds with an independent national exit poll. As in the former Soviet republics of Georgia and Ukraine, U.S. exit polls projected a clear victory for the challenger. John Kerry was projected to win the national popular vote by a 2 percent to 3 percent margin and was ahead in nearly every closely contested state. Of course, the official counts, as in the other nations, showed an almost mirror image victory for the incumbent party candidate.

The citizens of Georgia and Ukraine refused to accept the official tallies, protested vigorously and , with international support, overturned the election, but U.S. voters have passively accepted the results of their election and gone back to business, oblivious to the discrepancy and blind to the implications.

A 5 percent shift in a poll like this is extraordinary. Exit pollsters do not have to guess about who is actually going to vote, or whether they might change their mind. Exit polls can achieve larger samples cost-effectively: the national election-day sample had more than 13,000 respondents, meaning that it should have accurately forecast the result within plus or minus 1 percent.

Polling error beyond statistical margins of error is possible, of course. That's why we actually count the votes, and why the count determines the winner. But when there are serious questions over how elections are conducted, we look to these exit polls.

So what if an incumbent party controls the election machinery and there are other reasons to doubt the count? Irregularities similar to those found in the Ukrainian election have been documented here.

An investigation by members of the House Judiciary Committee limited to Ohio alone has substantiated:

-- Deliberate vote suppressions (unmailed and lost absentee ballots; obstacles to registration, such as rejection of forms over a technicality; lack of voting machines in Democratic strongholds resulting in waits of more than eight hours, while Republican areas had surplus machines; widespread misinformation about polling places; overuse of provisional ballots, many not subsequently counted);

-- Apparent fraud (undercounts in Democratic precincts where 25 percent of voters reportedly did not vote for president; unreasonably high numbers of votes recorded for third-party candidates in 10 heavily Democratic precincts in Cleveland; extraordinarily high voter registration and turnout inconsistent with records in precincts of Appalachian Ohio) and

-- Secret counts and recounts (Warren County locked out count observers because of a terrorist threat attributed to the FBI, which the FBI has denied; recounts conducted in the absence of observers and in pre-selected precincts, violating state law; testimony that representatives of a voting system supplier improperly participated in the recount).

Beyond these conventional manipulations, the United States has introduced electronic voting, a new system of potential mass and undetectable manipulation. Thirty percent of Americans in this election used electronic voting machines, which produce no confirmation that votes are recorded as cast -- the "paper trail." Stanford University computer scientist David Dill draws the analogy of telling a man behind a curtain whom you want to vote for and trusting that he has recorded it faithfully. Voters using electronic voting machines likewise blindly trust that the programmer has written code that can and will record their votes as cast.

The system is made worse yet by a concentrated electronic voting-machine industry characterized by overt partisanship, conflicts of interest and a lack of transparency in nearly every aspect of operations.

So why is the response rebellion in the former Soviet Union nations but passive acceptance here? It's not that exit polls are reliable everywhere but here. In fact, both of the exit polls in the Ukraine were flawed. One did not adequately cover the strongholds of the government candidate; the other used face-to-face interviews, thus asking respondents to risk retribution. Both polls are alleged to have been sponsored by the West, principally the United States, hoping to install a friendly, pro-NATO government. The U.S. exit poll, in contrast, was independent, well-funded and run by the most experienced exit pollsters in the world.

We may believe that "it can't happen here": After all, we are not only a democracy, but the democracy. Voting is embedded in all our cultural values and institutions. Paradoxically, however, U.S. democratic traditions may have led to unwarranted laxity. Other countries do not take democracy for granted. They know, as the founders of our country did, how vulnerable it is, and that the price of liberty is eternal vigilance.

The purpose of conducting research and questioning the election outcome is not partisan -- it is equally democratic, republican and libertarian. Americans should take up this cause as neither "for Kerry" nor "against Bush." Indeed, one reason resistance to the count has not coalesced is that for the past year, the country has looked to Kerry and George W. Bush as its leaders. But it's clear that neither is taking the lead on protection of voting rights. When I documented the discrepancy between the official count and exit-poll predictions, thousands of people e-mailed me to thank me for stating the obvious. Why weren't others asking these questions?

The absence of questions does not make a democracy function; democratic processes do. It has been a long time since this country has paid a price for liberty. It seems clear now that a large payment of vigilance is long overdue.

Steve Freeman is on the faculty of the Center for Organizational Dynamics at the University of Pennsylvania. To view his 2004 election research, go to www.appliedresearch.us/sf/epdiscrep.htm.

© 2005 San Francisco Chronicle

###