Tuesday February 16 2016

Opinion polls: Why businesses should care


By Othman Semakula & Jonathan Adengo

An opinion poll is just but an opinion. As such, it remains an opinion whatsoever.
But even then, no matter the status; conducting an opinion poll is a highly risky affair with far-reaching implications, especially if the outcome is perceived to be skewed.
In the United Kingdom (UK), pollsters fell flat last year, predicting a tight race with a possible win for the Labour Party in the country’s general elections.
But at the fold of the election eve, the results told of a different story with the Conservatives taking 331 seats against Labour Party’s 232 out of the 650 seats in the House of Commons.

This retrospectively was serious failure, considering that almost every pollster had handed the win to the Labour Party, and substantial at that.
The public was outraged and scorned the pollsters, who for the first time in so many decades, had failed to make the right projections on an election.
Years of reputational building had been drenched and smeared down the drain and ordinarily it will take pollsters years, if not decades, to regain public trust.

Unhappy with the whole episode, the British Polling Council commissioned an inquiry, whose findings partly blamed the pollsters for failing “to reach enough Conservatives voters.”
However, beyond the UK, Uganda has had its own episodes of opinion polls, which time and again invoke debate and public disdain.
But before we get into the dynamics, one of the biggest challenges of pollsters is, just like any other entity, most of them operate full truth business models with the end target of making a return on investments.

Conducting an opinion poll requires huge investments in terms of human resource and finances. Therefore, at the end of it all, there must be returns, either in form of influencing debate, setting trends or generating income.
For this, it is hard for pollsters to detach themselves from the desire to make money, which according to analysts, drives them into the hands of the highest bidder.
In such circumstances, are we safe to say it is for financial gains that the market ends up with skewed opinion polls results, which in most cases tend to tilt in favour of the piper?
Opinion polls influence business decisions and skewed results might have far-reaching implications not only on the economy but society as well.

But for Virginia Nkwanzi, there is no way a “serious” pollster can skew outcomes of an opinion poll because in this field, professionalism is the glue that keeps the public and the pollster trust relationship.
Therefore, Nkwanzi, the Ipsos Uganda managing director, believes the rationale of any opinion poll is not to show a positive picture of any company or individual because there is no “professional researcher who can be involved in such as survey.”

Ipsos Uganda has had its own share of disdain with the public, especially on social media, calling its opinion polls, particularly those on elections, unprofessional and skewed.
But for Nkwanzi the issue is staying professional and not giving in to pressure.
“If a poll is conducted using the right methodology and questionnaires, results are usually likely to be accurate and will properly reflect peoples’ opinions,” she says.

Polls help to shape public opinion as well as monitoring changes in perceptions. The perceptions might be in an election, a market segment or industry trends.
Therefore, they should be conducted in the most accurate way possible because beyond the public, they have financial implications in terms of making investment decisions.
For instance, last month GeoPoll, a US pollster that captures trends in the television and radio industry in different markets across Africa released a market share survey report that was captured in the last quarter of 2015 running from October to December.

The poll placed NTV in the lead position taking 32.4 per cent of Uganda’s television viewership market share.
Bukedde TV came in second at 22.7 per cent followed by NBS at 10.7 per cent. UBC at 9 per cent and Urban TV at 8.9 per cent came in at forth and firth positions, respectively.
Such data, if accurately captured, not only helps to direct industry trends and market perceptions but also helps advertisers to make informed decisions in terms of where they can effectively have a return on their adverts expenditure.

Do Ugandans believe opinion polls?
But why are opinion polls in Uganda received with a lot of apathy?
Many people will tell you they do not believe in opinion polls because it is unfathomable that a sample space of about or less than 2,000 can be used to determine the will of more than 30 million people.
This, some say, is a very small number. However, Nkwanzi argues: “…generating public opinions does not require interviewing the universe.”

The problem, she says, most people do not bother to understand the methodology, survey implementation and the depth of the results, which in most cases limits the scope of one’s focus to the obvious without looking a little deeper.
On this, Nkwanzi is supported by Research World International’s Patrick Wakida, who says dust risers with hindsight ignore the methodology used and concentrate on finding flaws.
“If you cannot understand the concept, then it is hard for you to believe it,” Wakida says in reference to the apathy that people have towards opinion polls.

Does Nkwanzi and Wakida’s argument hold given that most opinion polls are procured for mass consumption?
According to Max Richman, the GeoPoll chief data scientist, polls serve a wide variety of purposes. But “the key thing is the research company must always be transparent about the methodology and how the poll was conducted.”
“…. well conducted randomised polls can provide useful estimates of population statistics. However, like anywhere else in the world, surveys must be done in a way that is randomised and transparent,” he says.
But beyond the sample space is the concern of accuracy, which in most cases seems to be the point of departure, especially by those who reject the outcomes.

For economist Fred Muhumuza, the most important thing is to understand the purpose of the opinion poll and the methods used.
This, he says, will help those who are favoured and the non-favoured to reach a consensus.
According to Nkwanzi, the accuracy of an opinion poll solely depends on transparency and deploying the right methodology.
“With the right methodology and the right questionnaire, the results are highly likely to be accurate or a proper reflection of people’s opinions,” she says.

But for Richman, it is easy to measure the accuracy of an opinion poll if it is a random selection with an acceptable margin of error.
“A survey that is scientifically random can speak for a larger population, albeit with a margin of error depending on the sample size and confidence level chosen.
But beyond the above challenges is the issue of who commissions and pays for the opinion.
Many people argue that the piper will definitely have more influence on the outcomes and therefore it is difficult to put faith in an opinion that has been paid for by an interested party.
However, to Nkwanzi, “credibility of opinion poll does not depend on who commissions (pays) the poll.”

What is important, Nkwanzi says, is the credibility of the survey design and the other issues are just but feeders in the wider scheme.
Even if there is an interested party, people or companies that “commission opinion polls are interested in monitoring changes in perceptions or informing their strategy”, therefore, as Nkwanzi says, you would be retrospectively cheating your own self if you ask for a skewed opinion.
Therefore, being an opinion makes it hard to generate 100 per cent confidence but opinion polls definitely have a much larger effect, especially in terms of influencing debate and market trends.

Polling & Public Opinion: The good, the bad, and ugly

Pollsters. If you took a public opinion poll about polls, odds are that a majority would offer some rather unfavourable views of pollsters and the uses to which their work is put. Many potential respondents might simply slam down their telephones. Yet if you asked whether politicians, business leaders, and journalists should pay attention to the people’s voices, almost everyone would say yes. If you then asked whether polls are, at least, one tool through which the wishes of the people can be discerned, a reluctant majority would probably say yes to that too.

Several conundrums of public opinion polling are enfolded in this hypothetical tale. People of all kinds, activists and ordinary citizens alike, regularly cite polls, especially those that find them in the majority. But people are deeply skeptical of polls, especially when opinion moves in the “wrong” direction.
Some of their doubts are about pollsters’ methods. Do they ask the right questions? Are they manipulating the wording of questions to get the responses they want? And whom did they interview? Some of the doubts are wrapped up in a mistrust of the political parties, marketers, and media giants that pay for the polls.