The perils of polling

Indeed, the fact of the matter is that despite all of their expertise and alleged insight, the vast majority of pundits are no better than the average publican when it comes to calling elections or other political outcomes.

2018-05-27T02:59:29+05:00 Hassan Javid

The past few years have not been kind to the political polling industry. The election of a conservative government in the UK in 2015, the election of Donald Trump in the United States, and the Brexit vote are the most visible and high-profile examples of how pollsters have failed to forecast momentous political events. This is even more the case with when considering the record of political pundits in the print and electronic media; unlike polling organizations, which for all their flaws nonetheless aspired to collect and use data to make their predictions, many pundits simply engage in the art of speculation, and have time and again been proven to be wrong when making political predictions. Indeed, the fact of the matter is that despite all of their expertise and alleged insight, the vast majority of pundits are no better than the average publican when it comes to calling elections or other political outcomes.

The reasons behind the failings of both pollsters and pundits are manifold but can be summed up briefly. For one, particularly when it comes to polling organizations, the results of a survey and the analysis that follows from it are only as good as the assumptions that underpin them. The types of questions asked, the manner in which respondents are chosen, and the technical capacities of the organization in question, all affect polling outcomes and, therefore, the predictions that are made. Some organizations deal with these issues better than others, which is why some have better reputations for accurate prediction than others. In some instances, such as the election of Donald Trump, it is even possible for large numbers of organizations to make similar kinds of mistakes when attempting to gauge public opinion, resulting in a collective failure of political prediction.

Secondly, it is important to not underestimate the role played by ideological bias when it comes to predicting politics. Polling organizations can fall prey to this kind of bias (due to the nature of their funding, or the predilections of their leaders and workers), but this is even more evident when it comes to pundits. The main reason why pundits fail to accurately predict elections is because most pundits are blinded by ideological zeal. While objectivity is a virtue that pundits and experts should always strive for, the reality is that everyone has biases, and not everyone is capable of looking beyond them. Things are not helped by the hyper-partisan nature of the news media, which is the primary forum within which pundits opine on the future; ‘senior’ analysts on television are often given their position precisely because they take ideological stances that make for good debate and discussion, but which offer less in the way of meaningful political analysis. This was evident in the coverage of the 2013 elections in Pakistan; at that time, a significant number of experts on television and in the print media predicted that Imran Khan’s PTI would sweep the polls, even as more cautious voices suggested that there was little reason to believe that would be the case given the PTI’s organizational failings and shortcomings at the grassroots level. The stance taken by the former demonstrated the triumph of ideology over fact.

Earlier this week in a television interview, the head of Gallup Pakistan argued that the data collected by his organization showed that while support for the PTI had grown since 2013, the PML-N was still the most popular party in Pakistan and that it would probably go on to win the most seats in parliament in 2018. Following from this, he also suggested that the vast majority of voters vote for parties rather than leaders, implying that the reputations of both Imran Khan and Nawaz Sharif may be marginal to the final electoral outcome. Consequently, it was predicted that while the final tally of votes and seats remains up in the air, it seems likely that the PML-N will emerge as the largest party in parliament once the dust of the elections settles.

As far as polling goes, Gallup has a reasonably good track record in Pakistan although it has to be remembered that political polling of the kind described above is relatively new to this country, and that the absence of high quality alternatives to Gallup makes it difficult to corroborate its findings. Nonetheless, as Gallup itself has stated, the principal reason why its findings might make sense is because the majority of people it surveyed were of the opinion that the country was headed in the right direction. Given that the PML-N has spent much of its term cultivating a reputation for good governance that has resonated with many of its supporters (even though its achievements on the ground do not match its rhetoric), there may be some merit to the idea that many voters will continue to support the party despite its recent travails. Indeed, some might even argue that the average voter is not particularly concerned about what Nawaz Sharif has or has not said about Mumbai as long as the party continues to build roads and make visible shows of its commitment to ‘development’. Similarly, if the PTI’s great failing in 2013 was an inability to cultivate a coherent party apparatus at the local level in Punjab, which is arguably still the case today. Once more, the party relies on defected ‘electables’ to win votes rather than a more positive campaign for support.

Having said that, there are a couple of additional caveats that have to be borne in mind. While historical data unequivocally shows that defections alone do not win elections, since voters will not support defecting politicians unless they switch to the ‘right’ party, it will be interesting to see if the anti-PML narrative being crafted by the PTI and its allies will be sufficient to finally loosen the ties that bind that party to its voters. If that happens, and if the PML-N is seen to be a party in crisis that is on its way out, then its defeat will become a self-fulfilling prophecy as voters and candidates both abandon it for its rivals. Amidst all of this, it will also be necessary to see what, if anything, the establishment is able to do working behind the scenes. As many are current predicting, the most likely scenario in August 2018 is one in which multiple coalition governments at the provincial and national level vie for power, without any one clear political winner. How such an outcome could be engineered depends, again, on the extent to which defections and splits are able to play a decisive role amidst broader narratives of defeat and victory. The data from Gallup is both useful and interesting, but there is still a long way to go before the country goes to the polls this summer.

 

n          The writer is an assistant professor of political science at LUMS.

Indeed, the fact of the matter is that despite all of their expertise and alleged insight, the vast majority of pundits are no better than the average publican when it comes to calling elections or other political outcomes.

View More News