"Quick counts" for verifying election outcomes

There is an interesting short Reuter’s article about the Peruvian elections which discribes very briefly the way in which they determine who is likely to win elections without using exit polling.

LIMA, Peru, June 4 (Reuters) – Alan Garcia, who left Peru in economic ruin after his 1985-1990 term as president, was heading for a dramatic comeback as he lead the country’s election on Sunday, according to a quick count by Datum pollsters.

With 50 percent of the vote sampled nationwide, Datum said Garcia had 55.4 percent, compared with 44.6 percent for nationalist ex-army commander Ollanta Humala.

The quick count is based on early vote tallies at a sample of polling stations around the country.

Quick counts are based on using only valid votes cast in a sample of precincts in a jurisdiction–a county, state, or nationwide–to determine the likely winner. The IFES ACE project has a nice explanation of the quick count methodology, noting that quick counts are not the same as exit polls, which is a common misperception

As an AP story regarding the Indonesian election explains:

The “Quick Count” system employed by NDI predicts the results of national elections by counting about half a million cast votes from 2,500 selected voting stations. Similar polls by the same organization have accurately predicted results in dozens of elections around the world.

The National Democratic Institute has issued a report The Quick Count and Election Observation. The excerpt from the executive summary explains why quick counts are so valuable.

Political parties and candidates must develop skills to monitor the large variety of processes and institutions surrounding elections, and they must learn to mobilize public support and use complaint mechanisms to seek peaceful remedies for their grievances. Civic organizations and others committed to democratic governance also must engage directly in comprehensive monitoring efforts to help ensure electoral integrity. Elections simply cannot be separated from the broader political context of a country, and efforts to ensure electoral integrity must also be cast widely.

Nonetheless, all election processes come to a critical point on election day—and that is where reliable quick counts play a crucial role. A highly accurate and rapid report on the quality of the voting and counting processes from a random statistical sample of polling stations can serve to reassure political competitors and the citizenry alike that they should have confidence in the elections. Identifying irregularities can lead to timely corrections and proper assessments of their potential impact on electoral outcomes. A highly accurate and rapid projection of electoral results collected and reported from a sample of polling stations can deter fraud, calm tensions and allow those who assume office as a result of the elections to do so based on public confidence in their democratic mandate. On the other hand, systematic, impartial and accurate verification of results and the quality of election-day processes can also reveal widespread irregularities and attempts to hijack elections….

As the authors stress, not every election requires a quick count, at least not in its most comprehensive form. Moreover, quick counts only speak to electionday processes and say nothing in and of themselves about whether pre-election or post-election developments uphold or negate the democratic nature of an election. Quick counts are best understood as a critical element of comprehensive election monitoring, but they are unique in their impact and sometimes essential to determining the warranted degree of confidence in election results. NDI is therefore pleased to offer this handbook as part of a series of resource materials for election monitoring.

Given the concerns that exist today in the United States about election processes and outcomes, such a system may be beneficial for interested groups to study here. It is standard fare internationally and is a different beast than exit polling.

In the future, we’ll write more about election monitoring tools like quick counts and the research that now exists regarding the use of these election verification techniques. One resources in the academic literature for interested readers is Chapter 13 of Eric C. Bjornlund’s book, “Beyond Free and Fair: Monitoring Elections and Building Democracy”; this chapter is titled “Verifying the Vote Count: Quick Counts, Parallel Tabulations, and Exit Polls in Macedonia and Indonesia” and it discusses the use of quick counts and other verification methods as applied in these cases.