Not all elections should be American-style

On Sept. 15, Macedonia held its first parliamentary elections since recovering from brief but violent ethnic clashes last year. But before Macedonian authorities could announce the results of the election, a US government-supported organization beat them to the punch.

Three hours after the polls closed, an organization based in Washington and funded by US aid, announced the results of its own exit poll, and then touted the poll as "an important step forward in the country's democratic development."

But it is far from obvious that a tool such as exit polling could in any way be an advance for Macedonian democracy. Failing to distinguish which features of American politics can be adopted abroad, and which are inappropriate, hardly helps the US in its cause to assist with free and fair elections.

Exit polls can be problematic and controversial even in the US, where they have a long history. Remember Florida in 2000? Television networks relying on exit polls first forecast an Al Gore victory and then later reversed this judgment, only to finally conclude that they didn't actually know. There also continues to be concern that the early release of exit polls will influence those still on their way to vote.

The validity of any exit poll relies absolutely on the willingness of voters to tell a stranger how they voted. In Macedonia, violence, intimidation, and extreme nationalist rhetoric plagued the preelection environment. In the last three weeks before election day, two policemen were murdered, opposition supporters were blocked from entering the capital of Skopje, and party rallies had to be canceled for fear of violence. The interior minister publicly threatened to arrest the leader of the most popular ethnic Albanian party, and many politicians and voters expressed fears about special security forces and paramilitaries.

Indeed, intimidation was so pervasive that, despite the country's population of only 2 million, the international community mobilized the largest international monitoring effort ever. Hardly the conditions needed for an appropriate and successful exit poll. The use of a "pretend" ballot box into which the exit poll survey was placed was little guarantee of anonymity.

Had the voters felt safe to express their political preferences, a valid exit poll relies statistically on knowledge of voting patterns that identify key predictive precincts.

Researchers in Macedonia, however, had little data, much less analysis, of previous precinct-level results to guide their choice of representative polling stations. Rather than accept that this made it problematic to carry out the exercise, the exit poll was, instead, based on the highly dubious statistical assumption that polling stations with average turnout rates in a previous election would be predictive of current choices.

The results of the questionable survey were then announced nationally, first on the Internet and then at a press conference. But this countrywide tally was essentially meaningless, as there were no national candidates. The only outcomes that mattered were party results from each of the six parliamentary districts.

This unfortunate exit poll exercise was unnecessary and detracted from other, more legitimate efforts. A nonpartisan Macedonian election monitoring group, Citizens Organization MOST, conducted its own "parallel vote tabulation" based on statistically drawn samples of actual results and reported these findings for all six districts. Their data provide a genuine basis by which to assess the credibility of the official count. Nevertheless, these valid data were initially ignored because an American group provided the first numbers, albeit inaccurate and meaningless ones, to the media.

It hardly seems an "important step in [a] country's democratic development" for foreigners to scoop local electoral authorities and nongovernmental groups with misleading electoral information in order to satisfy the curiosity of the international community.

Not every tactic borrowed from the US contributes positively to democratic development abroad. Perhaps the failings of this exit poll will inspire critical thinking about the use of such advanced tactics in developing democracies. As we debate the merits and lapses of our own democratic system, we should be cautious about exporting any – much less all – of it to others.

• Eric Bjornlund and Glenn Cowan have advised election monitoring organizations in more than 50 countries.