Shedding light on the role of pollsters in democratic systems

During the seminar, organised by Konrad Adenauer Stiftung, the panellists exchanged their experiences in polling, which vary according to the socio-political background of various countries. 

Singaporeans are still a long way from appreciating the links between politics and polls, based on the general election result in May 2011 and the subsequent presidential election in August. Conducting exit polls, for example, and publishing the results, is strictly not allowed, according to Kirpal Singh, director of the Wee Kim Wee Centre at Singapore Management University. 

At the same time, in New Zealand and Australia, the focus group approach has been widely used to gauge public opinion. But it is not yet popular in Asia, according to Stephen Mills, principal of UMR Research, Australia/New Zealand.

Political polls are seen as an essential part of a democratic system. They are perceived as an expression of public opinion. This is despite the fact that questions and challenges face the pollsters. For instance, the public may question the accuracy of polls, especially if the predicted figures turn out to be wrong in the election outcome. Elections are the final reality for public opinion polls. It makes them visible and powerful, but at the same time can also make them vulnerable.

The problem is that polls are quite good at measuring opinion and behaviour at the time the poll is taken. But Wolfgang Donsbach, director of the Department of Media and Communications at Technische Universitat Dresden, Germany, says it is difficult to predict future opinions and behaviour, or to make inferences on the views and behaviour of those who have declined to participate in a poll.

Pollsters can also face allegations over skewing the results of election polls, especially from politicians who lost. For instance, polling results can create a “bandwagon” effect if a proportion of undecided voters decides to vote for the expected winner in order to stay among the winners. Conversely, undecided voters may follow the “underdog effect” by voting for candidates who were predicted by the pollsters to be among the losers.

Polls can also create a “defeatism effect” if voters feel their candidates are losing anyway and they choose to stay at home on the election date. Conversely, again, some may not go to vote if they feel that their choice of candidate will win anyway. This is known as the “lethargy effect”. However, Donsbach says there is no scientific proof regarding this effect. Even if there was more evidence, he says, it would still be hard to blame the polls. 

Polling results can also create “populist” politicians, according to Donsbach. Some politicians watch the results of polls to discover what is the majority of public opinion, and then make decisions that can be perceived as “populist” in order to stay in power. These politicians then choose short-term populism rather than better strategies for the country, to ensure they might be re-elected.

Noppadon Kannika, director of ABAC Poll Research Centre at Assumption University, Bangkok, acknowledges that pollsters were widely criticised for polling inaccuracy before and after the 2011 election in Thailand. The big error in projection was in Bangkok, where the exit polls predicted the Pheu Thai Party would get 24 seats and the Democrats 9, out of 33 in total. But the actual voting results showed that the Democrats got 23 seats and Pheu Thai 10.

Noppadon said the major source of errors in the ABAC exit poll result in Bangkok related to both sampling and non-sampling errors. He attributed the errors to the fact that some voters refused to complete the questionnaire. Secondly, the ABAC exit poll was conducted from 8am to noon, while the poll closed at 3pm. Thirdly, he said that the data collectors were under social pressure. For instance, one of them was arrested by the police. Noppadon said all these pressures made it difficult for data collectors to work effectively, and this affected the quality of the exit poll data.

Mills suggests that Asian pollsters explore methods such as focus groups, where the data collectors interview a group of 6 to 8 people to survey their opinions on certain issues. He says that focus groups generally provide more depth of understanding than quantitative research. For instance, Australian media often report the results of quantitative polls which show that a majority of Australians believe that man-made climate change is happening and want to see the government taking strong action. But focus groups tell a different story and, as events have proved, a more accurate story, including that there was considerable opposition to the attempted introduction of a carbon tax in Australia in 2011.