Sunday Times 2
Guns, lies and opinion polls in Lanka’s elections
Elections, particularly in countries such as ours, are not confined only to ideas and policy issues. Violence, threats, intimidation and deceptions have been widely used techniques, since universal franchise was introduced to the country.
The higher the stakes, the greater was the intensity of the extra democratic acts of coercive persuasion. These practices have become standard Sri Lankan electoral practices to an extent that voters are not only reconciled to them but also regard them as ‘common sense’ for candidates to resort to these malpractices.
Opinion polls are the latest addition to the toolbox of the politicians, to influence the electoral choice of the Sri Lankan voter. Though it is new to Sri Lankan elections, introduced only after mid 1990s. In mature democracies such as the United States, the practice goes back to as early as 1930s. Although election polls have been conducted and published at every major election since then, as a pollster, I find the bona fides of most of those polls gravely questionable.
This time too, polling results started appearing in newspapers. The recent poll results on the imminent elections published by two universities have triggered interest and concern among many Sri Lankans. By virtue of the fact that two state universities conducted them, they received much attention; as the universities are expected to be the authority of knowledge. Given the contradicting results produced, many have become sceptical about the legitimacy of both opinion polls. In this context, how would an average reader discern the objectivity of these poll results? What to accept and what to dismiss? It is in this context that I decided to write this article with the aim of raising some awareness among journalists, scholars and average readers on election polls.
An opinion poll measures the direction and intensity of public opinion. Public opinion is highly unpredictable and sensitive. In the old days, monarchs consulted ‘pundits’ to learn about public opinion or approval of their subjects on different policies. Even in modern politics, expert opinion is sought to predict the direction of public opinion. However, accuracy only depends on the capacity of the ‘expert’ and there is no way to verify it. On the contrary, opinion polls, an offshoot of behaviourism, has the capacity to provide an accurate estimation of public opinion on any given issue.
Conducting an opinion poll is a scientific exercise. From the point of conceptualisation to the point of reporting the results, the researcher should adhere to a rigorous methodology. A questionnaire is designed to interview the respondents. And field researchers are trained on the questionnaire and the data collection method to reduce biases and ensure quality. When data collection is completed, pollsters use statistical analysis techniques to analyse the data. In developed countries, modern methods such as telephones, e-mails and the Internet are used to collect public opinion. However, in countries such as Sri Lanka, pollsters rely on in-person interviews (Face-to-Face), though it is expensive and time consuming, as there is no infrastructure for our pollsters to adopt new techniques. When selecting the people to be interviewed, pollsters follow a procedure where every individual has the same chance to be chosen for the poll.
Role of opinion poll
Public opinion polls at election times are used to achieve multiple objectives. As in the case of technology, public opinion polls are also a double-edge sword. It can be used as a tool of strengthening, as well as of weakening democracy. However, if one uses it as it is meant to be, opinion polls encourage wider public participation in elections and provide a channel through which the public can express their opinion, and thereby influence the election debate.
Public opinion polls also assist candidates in drafting the best campaign strategy. However, opinion polls are also an instrument of capitalism. Hence, polls have become a powerful tool for TV channels to attract audiences for their programmes. On the other hand, funders of political campaigns are also direct beneficiaries of opinion polls. However, irrespective of its main objectives, opinion polls in mature democracies can help the election campaigns to focus on issues and policies that matter to the voters.
Can polls be harmful?
There were instances where pre-election polls and exit polls were banned, or brought under government control in some countries. However, in most of the instances, pollsters managed to get a fair verdict from their respective courts. Some countries have introduced regulations and guidelines for polling agencies to prevent any adverse effects on the election process.
The most famous argument against opinion polls on elections is that poll results can influence the decision of the voter in favour of the leading candidate or party. Technically we call this the “Bandwagon effect”. Some elites argue that the undecided vote can swing in favour of the candidate who runs as the ‘number one’ in the race. However, pollsters argue that in the same way, sympathy votes might swing in favour of the loser (Underdog effect) at the same time and balance off each effect. In the 2000 election, one political party attempted to take advantage of the Bandwagon effect by misinterpreting a stolen set of data from a poll that was conducted by a Sri Lankan University. However, it proved that the argument that polls influence the voters is not true — at least in the Sri Lankan election context.
Can opinion polls be misused?
Public opinion polls are also like every other tool. There are good polls that are conducted impartially and scientifically. And then there are bad opinion polls that are conducted with a political agenda or without a proper scientific approach. Sometimes political parties or candidates conduct polls amongst their loyalists and in their strongholds to show that they are leading the election race. Sometimes, interested groups also could use these types of polls to persuade the state that the majority of the country supports their cause.
Therefore, it is important that the reader is able to differentiate scientific polls from non-scientific ones conducted to push forward a particular political agenda. There is a code of professional ethics and practices issued by the World Association for Public Opinion Research (WAPOR) and other professional setups like the European Society for Opinion and Marketing Research (ESOMAR). The code prescribes the standards for minimal disclosure in the publication of poll results. With this information, readers can judge the quality of the poll that they refer.
Before reading the findings
Pollsters have a professional obligation to reveal the sponsor of the poll, the exact wording of questions asked, a definition of the study population, sampling procedure, size of the sample, sampling error and the method, location and dates of data collection in their publication of results. From such minimal information the reader can make a fair judgment of the accuracy of the poll findings. However, most of the time, laymen do not have access to opinion poll reports and they get to know the results of polls through newspapers or the electronic media. In this context, journalists have to publish opinion poll data with greater care.
A scientifically conducted poll does not mean that the findings are accurate and are not subject to any biases. In this case, the credibility of the pollsters is important to judge the neutrality of the study. Disclosing the sponsor and the objectives is crucial to discern the motive behind the study. Size of the sample, sampling procedure and areas covered in the study assist to find out whether this poll is a scientific or non-scientific one or whether it contains any biases.
The dates of the data collection period are important too. Without the dates, readers will not have any idea whether what they read has any relevance to the current context. Professional pollsters usually provide something called the “Margin of error”, which means the degree to which the poll results can differ from the actual situation due to the selection of the sample. This is not the only error that a poll result is subject to. There could be many errors due to the wording of the questionnaire, problems of field interviewers when collecting data etc. However, an experienced pollster knows how to minimise these human errors.
Strengthening democracy
I strongly believe that public opinion polling could be a valuable instrument to strengthen democracy. Of course, we need pollsters who have the necessary knowledge on this science to conduct reliable polls. We should not undermine the role of journalists and readers in bringing about such a public opinion polling culture. We also need responsible journalism when publishing poll results. Unless journalists provide the minimum information on the poll as prescribed by the ethics bodies of professional pollsters, the reader would not be able to judge the bona fides of the poll. However, it is important to encourage more polling at elections than to merely dismiss them. If there is greater awareness and open debate on election polling in the country, polls can do much good, rather than harm democracy.
(Dr. Peiris is the Sri Lankan country representative for the Asian Network for Public Opinion Research (ANPOR) and he is also a senior researcher and council member of the Social Scientists’ Association.)