NOI polls: Not wuruwuru but ignorance…

Many Nigerians have been understandably startled by the NOI polls on the 2023 elections which put Peter Obi as the preferred choice of voters if the elections were held on the day the polls were taken. While the Obi camp and its supporters are expectedly excited, Tinubu’s campaign is shocked, embarrassed and angry. Atiku’s camp is yet to utter a word, but I’m sure that the third-position ranking of the PDP presidential candidate was not totally unexpected. We’ve always known that Obi will take more votes and support from Atiku than Tinubu. So far, neither Peter Obi nor his Labour Party has responded to the survey, but his supporters who call themselves ‘Obidients’ have been noisily animated on social media.

Published last Wednesday, the poll showed a significant lead for Obi with 21% of voters proposing to vote for him if the elections were to be conducted today (the day the poll was taken); and 13% each proposing to vote for Bola Tinubu and Atiku Abubakar who are both tied in the second place. The explanatory note that follows the poll states that: “Peter Obi’s 8 percentage point lead at this early stage is significant but not sufficient to separate him completely from a leading pack of candidates scoring 21%, 13% and 13% respectively”.

The response from the Tinubu Campaign (or APC Campaign Council) immediately after the poll was published betrayed an apparent lack of understanding of the objectives of the survey and a misinterpretation of the statistical presentation. I should note that the respondents were asked: “If the presidential election was held today, whom will you vote for?” So, the answer given simply reflected the preference of the respondents as of when the poll was taken, which probably was a few months ago. The presidential poll is still six months away, so the poll is not predicting or forecasting what the 2023 results will be. The result of the presidential election to be held in February 2023 would definitely be different from the result of this NOI poll, if not in the arrangement of the candidates, but in the percentage of votes recorded. In other words, the NOI poll is not a predictive exercise, and so there was no reason for such a tendentious statement from the Tinubu camp.

The presidential candidates and their strategists should rather pay attention to the undecided voter size of 32%, and those who refused to reveal their choice (15%). The relatively big size of undecided voters six months to the election means that the winner of this election is yet to be known. It could swing anywhere depending on many factors that would play out between now and the election day. The NOI poll is therefore not a predictive one in the sense of forecasting the 2023 election result. No. That is not the aim of the poll. The aim is to study the behaviour of Nigerian voters over time – between now and February. As the explanatory note says, the survey will be conducted again and again so it will help us understand how the voters change their minds over time till the election date. The fact that the poll would be repeated a few more times means that this one was not predictive. If the study were a one-time survey, it is referred to as a cross-sectional survey. But if it is conducted again and again, it becomes a longitudinal survey. Longitudinal surveys help us understand the behaviour of our electorate. For example, as we move close to the election date, we should expect to see the percentage of the undecided should tend to zero, but it will never be zero.

So, instead of issuing angry and incendiary statements, each campaign should be more interested in the movement of the undecided numbers and certain details of the undecided voters: Where do they live, their demographics, preferences, gender, income, and lifestyles among others. In many competitive and unpredictable elections as 2023 would be, the winner emerges by convincing the undecided voters through effective and targeted messages and messaging. The fact that the bulk of the undecided voters (38%) are in the north-central region and the least (19%) are in the south-east is a fascinating revelation. How then should a creative campaign strategist approach the campaigns in the next six months? As is obvious, the Igbos are very clear in their minds that they want Peter Obi.

Let me conclude by explaining a few statistical concepts which were not mentioned in the explanatory notes, but which I think were likely used in the survey. It would be helpful for NOI to explain their applications in subsequent surveys. These are validity, reliability and accuracy. Validity explains how the instrument used in the polls (set of questions) was vetted by experts. Reliability, on the other hand, shows the ability of the instrument to measure what it was meant to measure. It is the closeness of a set of repeated results to each other. Again, I should stress that the instrument was not designed to measure who will win the election in 2023, but rather who would win the election if it was conducted on the day the survey was made. Accuracy is another important concept and it explains the closeness of the measured value to the actual.

The novelty of polling in Nigeria and the unfamiliarity of many with the statistics and the intricacies of polling are the main reason the NOI poll has triggered so much misunderstanding and vitriolic emotions. This poll is not wuruwuru as Bayo Onanuga, the director of media and publicity of the APC Campaign Organization, termed it. As I said earlier, the poll was not meant to predict the winner of the 2023 election. Onanuga misinterpreted or misunderstood this point. I’m sure that subsequent polling will address that.

Finally, there are many reasons why voters did not choose Tinubu as their number one choice in this poll. I will not, however, go into those details in this piece. But I will rather urge our political strategists and communicators to brush themselves up on the basic rudiments of statistical analysis and the science of polling. I thank NOI for this initiative.

Etim writes from Abuja