Azerbaijan: Debate Over Exit Polls Threatens Confidence in Election Results
Plans to conduct exit polls during Azerbaijan's November 6 parliamentary poll elections are fueling controversy. Questions have arisen about the impartiality of selected polling organizations and about a paucity of information concerning the sponsors of some of the exit polls.
When the US Agency for International Development (USAID) announced plans in early 2005 to fund an exit poll, the project was seen as a key component in a broad strategy to ensure free-and-fair elections and maintain stability. Violent clashes occurred following the country's 2003 presidential election, which opposition leaders say was rigged. [For background see the Eurasia Insight archive].
Representatives of President Ilham Aliyev's administration announced that discrepancies between official and exit poll results from individual polling stations during the November 6 vote would lead to an investigation of the polling site. Representatives of both the Azerbaijani government and the US-government funded organization traveled throughout Azerbaijan in recent months to raise awareness on the meaning and importance of exit polls.
On September 2, Washington, DC-based PA Consulting Group, a frequent USAID contractor, won a $1 million tender to run the poll. The contractor, along with its local partner, SORGU, plans to run exit polls in 65 randomly chosen constituencies, slightly more than half of Azerbaijan's total 125 election districts. Fifteen polling stations will be surveyed in each constituency.
Two other companies have also announced plans to conduct exit polls during the elections: Mitofsky International, a New York City-based company that conducted polling for five US television broadcasters and the Associated Press during the 2004 US presidential elections, and Saar Poll, an Estonian company. Both Mitofsky International and Saar Poll will conduct exit polls in all 125 constituencies. Saar Poll has announced that it will cover more than 1,000 polling stations, while Mitofsky's pollsters will work in more than 1,500.
Local media outlets are debating the role of these two companies in Azerbaijan's elections, and their reason for conducting the polls. Attention is currently focused on the polls' sponsors, as both pollsters' clients remain relatively obscure. Mitofsky says that it was asked by the Russian Institute of Comparative Sociological Analysis to conduct the polls at the request of a company called Renaissance, based in Switzerland. Additional information about the company, so far, has not been made public by Mitofsky or located by journalists.
Saar Poll's president, Professor Andrus Saar, told EurasiaNet that Santo Communications, a British financial institution, had commissioned the Saar Poll survey.
Both Mitofsky International and Saar Poll say their clients ordered the exit polls because they are interested in obtaining information that would help them judge political and economic risks, business opportunities and the degree of stability in Azerbaijan.
However, opposition leaders suspect that both Mitofsky and Saar Poll were recruited by representatives of the Aliyev administration to supply exit polls whose results would compete with those of the US-funded poll. "Nobody knows the companies which invited these pollsters [to conduct the exit polls]. I would not be surprised if one of our oligarchs would appear as the client of these two pollsters," said Fuad Mustafayev, deputy chairman of the Popular Front Party of Azerbaijan. "Anyone can register a company in either Switzerland or Great Britain."
The Council of Europe's Cabinet of Ministers has recommended that exit poll results published by the media should include information about the organization or person who ordered and paid for the poll. "This requirement is designed to ensure the transparency of polls," said Alesker Mammadli, a lawyer who trains journalists on legal issues related to election coverage.
PA Consulting Group, as a US government contractor, has not drawn complaints. But opposition members have expressed strong doubts about its choice of a local partner, the research firm SORGU (Survey). The oldest polling organization in Azerbaijan, SORGU has experience working with the World Bank, the United Nations Development Fund and other international organizations in Azerbaijan. But it is best known to ordinary Azerbaijanis for a scandal surrounding an earlier survey it conducted in May 2005 that gave President Ilham Aliyev a 77-percent approval rating. SORGU claimed that polling giant Gallup International had ordered the poll, but Gallup denied any association with the survey. The poll was commissioned by the market research and public opinion surveyor GORBI, a Georgian firm.
"The scandal showed how accurate SORGU is with information," Mustafayev charged. "While authorities were using Gallup's brand name for their own propaganda, SORGU never refuted the claim that the survey had been provided by Gallup International."
At a September 16 press conference, USAID and PA Consulting Group defended their choice. "All our partners meet four requirements: anonymity, transparency, neutrality and professionalism," said David Hoffman, representative of PA Government Services. Working on the exit poll in Azerbaijan will be Sheldon Gawiser, a long-term senior pollster for the US National Broadcasting Company Network, and Dr. Evans Witt, who conducted the United States' first national exit poll in 1976.
A source at PA Consulting Group, who requested anonymity, said that SORGU was selected as the company's local partner based on recommendations from US-funded program managers in Azerbaijan, diplomatic personnel and USAID. The decision was motivated by the fact that the firm has had past experience working with the opposition, claimed the source, saying that one opposition party has been among SORGU's list of past clients. Main opposition groups the Azadliq and YeS blocs, along with the Liberal Party of Azerbaijan say they have never worked with SORGU.
At a September 29 meeting, opposition leaders urged both USAID and PA Consulting Group to reconsider their local-partner selection. Representatives of USAID stressed that SORGU will only receive instructions from PA Consulting Group and will not have the opportunity to manipulate poll results.
The two other exit pollsters have also teamed with local partners. Mitofsky's partner is the Association for Development of Civil Society, a group of non-governmental organizations largely seen as pro-government, and one of the association's members, The Center for Political Technologies. Saar Poll's partner is the other member of the association, The Center for Regional Development.
All three of Mitofsky's and Saar Poll's local partners formerly worked closely together and with another non-governmental organization, the Transparent Elections Center. The Center provided an exit poll during the disputed 2003 presidential elections that confirmed the official results.
With three exit polls for the elections, some observers believe that the number of pollsters in polling stations and the results of their findings may confuse voters, as well as undermine chances for a close examination of discrepancies between official results and the exit polls.
Andrus Saar, the Estonian pollster, however, does not anticipate problems with multiple exit polls. "If there will be a discrepancy between our figures, we will discuss [it] and prove our position," Saar said. The exit poll will be Saar's first outside of Estonia. Nonetheless, the pollster affirmed, "[o]ur prognoses on elections are considered to be the most precise."
Baku residents seem cautious about the exit poll plans. "First of all, people do not feel comfortable with openly expressing their opinions. Second . . . the government has only promised to start an investigation in polling stations where there are discrepancies," said Irada Huseinova, a 42-year-old historian. In the end, she argued, the polls may only "cause another conflict."
Latest from Azerbaijan
We would like to hear your opinion about the new site. Tell us what you like, and what you don't like in an email and send it to: firstname.lastname@example.org