Season of pollsters and poll-stars By Bell Ihua

 


This article was first published in Business Day Newspaper on Monday 16 January 2023

At every General Election cycle, especially the presidential and gubernatorial elections, predictions, forecasts, speculations, guesses and even prophecies are rife as to which candidates are likely to emerge victorious and crowned winner at the end of the cycle. Politicians often work with polling consultants and media practitioners to conduct and disseminate public opinion polls to help them design and test their campaign strategies, as well as create the needed media buzz in order to gain some public optics for a possible bandwagon effect.

Nigeria’s 2023 General Election cycle is not an exception, as we have seen an array of polls being released in favour of one presidential or governorship candidate or another. If my memory serves me right, there’s been about seven polls or so released so far. As I read through the press releases or news reports on the poll results and findings, some questions usually come to mind as a discerning reader or an expert in this field of opinion research, polling and surveys. The essence of this piece is to highlight a few points and checklist to enlighten readers on how to assess the quality, credibility and authenticity of these election polls and surveys, as well as discerning what they could be representing.

In developed climes, public opinion polling is the job of pollsters. Simply put, pollsters are people who collect and analyse opinion polls. They identify the population of interest, take an appropriate sample off that population, design questions around the subject matter, administer interviews using the questionnaire to gather the opinion of people, then go ahead to analyse the cumulative data collected and use the results to draw inferences on the population. I have just taken you through a complete polling cycle, or if you like, Polling 101. However, as simple at it may seem, it takes a lot at each step of the cycle to ensure that findings of the poll are representative enough to speak for the wider population. The following four questions can help anyone reading a polling report appreciate and discern a good poll from a bad one.

First, who is the pollster or polling company releasing the poll? This question is important to help readers assess whether or not the individual or company releasing the opinion poll are professionals with requisite knowledge and skills to do the job. This is Nigeria, and we know that apart from professional pollsters, there are also people I often call poll-stars. These folks are the fly-by-night pollsters who have other day jobs, but show up only at election cycles to serve as ad-hoc polling consultants. We have seen the harm caused by these unscrupulous folks seeking to deceive the undiscerning public in order to achieve their own pecuniary gains. So, the next time you come across news reports on a new poll be sure to check who is releasing it. If it’s a firm which has been around for a while and known for its past credible works, then you should take the report seriously. But if it’s a firm you’ve never heard about, perhaps you need to read with a bit of caution.

What methodology was applied to conduct the poll? As my PhD supervisor, Professor Andrew Fearne, use to say often, “the rise or fall of any piece of research [and I add, opinion poll] lies in its methodology”. In other words what determines whether or not anyone should give a hoot about your poll rests in its ability to answer some questions – How was the poll conducted? Was it conducted in person or via telephone, web-based or SMS? How was primary data collected? Was it collected manually or with the aid of some form of technology? How were the questions worded? Were they worded objectively or with some element of subjectivity? Who were those interviewed? How representative was the data? What was the sample size? What statistical tests were applied? What was the margin of error and the confidence level etc. Scientifically speaking, all these questions have implications on the outcome of the poll, and ultimately its perceived or actual quality and veracity. As far as Nigeria is concerned, the best two data collection methods for opinion polls are in-person / face-to-face and telephone methods. These two have been proven and tested, especially given the current GSM penetration which the Nigerian Communications Commission estimates at 214 million active lines. However, when it comes to online web-based or SMS polls, I can categorically state that it remains a tough call in Nigeria, as you would need to factor the internet penetration and literacy rates of 38 percent and 78 percent respectively.

What was the sample size of the poll? Even though this question falls under the methodology, I have decided to zoom in on it a bit more given its importance. As a professional pollster I often get asked questions on the appropriateness of sample sizes. What sample size is robust enough to ensure representativeness and enable deductions to be drawn on the larger population? My constant response is that it isn’t necessarily about the sample size, but about the selection of the sample and the needed granularity. Therefore, a 1,000-sized poll selected scientifically may be okay to provide a national outlook or Bird’s eye view of an issue, but if you need a deeper dive into how things look at the state, senatorial district or local government level, it would require a much larger sample size to allow deeper granularity. I believe it was Gallup, one of the world’s leading polling companies, that gave the illustration relating a sample to a blood test. When you fall ill and visit the hospital, the lab scientist wouldn’t need to draw out your entire blood to check what’s wrong. He only needs to take a tiny sample. It is this sample principle we apply in conducting scientific public opinion polls. However, for a country like Nigeria with a population of over 210 million citizens, I wouldn’t recommend any national poll with sample size less than a thousand. Similarly, as it relates to the General Elections and need for deeper granularity, I would recommend samples much more than a thousand.

Who is sponsoring the poll? Polls can be commissioned by individuals, organisations or groups. However, it is important to disclose who is funding the poll, to enable readers assess the level of independence and neutrality, or possible biases, of the polling results. In the United States there are professionals referred to as republican or democratic pollsters, there are also those known as independent pollsters. This upfront disclosure instantly acquaints the audience of the political leaning, or otherwise, of the polling company. In Nigeria, while some politicians commission public opinion polls to help them strategize or test ongoing campaign efforts, others hire unscrupulous pollsters to help them conduct dubious polls, or worse still, merely concoct figures showing they are having the upper hand and likely to win the election. We also find some politicians criticizing and casting aspersions on the results of credible polls that do not favour them. All these antics are usually applied to generate a possible bandwagon effect and sway undecided voters in their favour.

In conclusion, it is worth noting that the culture of actively seeking and engaging public opinion polls and surveys is still growing and not yet entrenched in Nigeria, like it is in more developed climes. Besides, there are not many firms and professionals who fully understand the processes, techniques and nuances of polling. From my experience, only cerebral and forward-thinking politicians value the need to commission credible polls to guide their campaigns, and I’ve been privileged to work with a few of that ilk. The majority believe polls are not as important, and they are better-off saving their money for mobilisation. So, when next you come across news reports of opinion polls or election surveys, don’t be too quick to consume it hook, line and sinker. Read it with an air of skepticism and think of these points discussed above as you read through.

 


Bell Ihua mni, is a pollster and professor of practice in opinion research at Coal City University, Enugu, where he heads the South-East Barometer project. He is also Executive Director at Africa Polling Institute.

Comments

Popular posts from this blog

REFLECTIONS FROM ABIDJAN

VIDEO: Amazing Sights & Sounds of the Ekpechor Wonder Ebu aka Ebu Wonder

Ten Points from Ten Months: Reflections from NIPSS By Bell Ihua