A 69,110-response online exercise is being marketed as political evidence. Its own methodology says it is not a representative poll of Nigerian voters.
There is a problem with the headline declaring that President Bola Ahmed Tinubu has “beaten” Peter Obi and Atiku Abubakar in the Oxford-Metro Analytics National Trust Poll.
It sounds definitive.
The methodology is not.
Oxford-Metro's final figures show Tinubu with 24,469 responses, Peter Obi with 23,402 and Atiku Abubakar with 18,430. That puts Tinubu ahead of Obi by 1,067 responses, or approximately 1.54 percentage points.
But before anyone starts declaring a political victory, there is a detail that should not be buried beneath the headline:
Oxford-Metro itself says this is not a scientific election poll.
The organisation previously described Public Pulse as an open participation platform, not a scientific election poll. It said traditional opinion polls use random sampling, whereas Public Pulse measures people who deliberately come online and participate.
That distinction is enormous.
Tinubu did not “beat Nigeria”
Tinubu received the largest share of responses on the platform.
That is the fact.
But that is not the same as saying Tinubu currently leads Nigerian voters.
The final report itself says the exercise was an open, self-selected participation exercise, that the results were unweighted, and that a conventional sampling margin of error does not apply. It also explicitly states that the results are not a representative survey of Nigeria's electorate and are not a prediction of the 2027 presidential election.
So the phrase “Tinubu beats Obi and Atiku” requires an important qualification:
Tinubu beat Obi and Atiku among the people who participated in this particular online exercise.
That is a much smaller claim.
The 69,110 figure creates a false impression of certainty
The biggest weapon being used to sell the poll is its size.
Sixty-nine thousand responses sounds enormous.
But a large sample is not automatically a representative sample.
The question is not simply how many people responded?
The more important question is:
How were those people selected?
Here, participation was voluntary and self-selected.
That means the people who decided to visit the platform, engage with the poll and submit responses are not necessarily representative of the Nigerian electorate.
Oxford-Metro acknowledges precisely this limitation. Its own explanation says internet access, digital literacy and willingness to participate online create structural biases. It describes Public Pulse as a window into digitally active Nigerians, rather than a census of all voters.
That should immediately stop anyone from presenting the numbers as a national electoral verdict.
More than half of the responses came from Lagos and Abuja
This is perhaps the most revealing statistic in the entire exercise.
Oxford-Metro says Lagos and the FCT together contributed 37,622 responses—54.44% of the entire 69,110-response dataset.
More than half of the national dataset therefore came from two locations.
That does not make those responses fake.
But it makes it impossible to casually treat the dataset as though every part of Nigeria participated proportionally.
Nigeria's political population is not concentrated in Lagos and the FCT.
Millions of Nigerians live outside those two centres, including large rural and semi-urban populations whose political preferences may not be captured adequately by an online platform.
The poll therefore tells us something about its participants.
It does not automatically tell us what the entire Nigerian electorate thinks.
Tinubu's “1.54% lead” is not a conventional polling lead
This is another point that deserves attention.
Tinubu's advantage over Obi was only 1,067 responses.
That translates into approximately 1.54 percentage points.
Ordinarily, people might immediately start talking about a statistical margin of error.
But Oxford-Metro says a conventional margin of error does not apply because this was not a probability-based sample.
That makes headlines describing the result as though Tinubu has established a measurable national lead particularly questionable.
There is a huge difference between saying:
“Tinubu leads Obi by 1.54 percentage points in a representative national poll.”
and saying:
“Tinubu received 1.54 percentage points more of the responses submitted through an open online platform.”
The first statement makes a national inference.
The second describes what actually happened.
The “21 states” claim is also being overstated
Another headline says Tinubu has taken 21 states.
That sounds like an electoral map.
But the underlying data does not justify treating these online tallies as actual state-by-state electoral predictions.
The poll report says Tinubu had the highest tally in 20 states and the FCT, while Obi and Atiku led in eight states each.
But “highest tally” does not mean “majority of voters.”
A candidate can finish first with a minority of responses.
More importantly, some states had extremely small participation.
The reported Kogi contest, for example, had Tinubu on 39 responses and Atiku on 38—a difference of one response.
Calling that an indication that Tinubu has genuinely taken Kogi in 2027 would be absurd.
It is an online participation result.
Nothing more.
The regional figures actually undermine the victory narrative
If the purpose of the headline is to demonstrate overwhelming Tinubu dominance, the regional data does not support that interpretation.
The poll shows a deeply divided political landscape.
Tinubu leads the South-West, North-West and North-Central.
Obi leads the South-East and South-South.
Atiku leads the North-East.
Obi's South-East result is particularly strong, with 5,796 responses, representing 63.97% of that zone's responses.
Atiku's North-East result is 55.46%.
Tinubu's North-West figure is 43.10%, while his South-West figure is 43.53%.
That is not a picture of one politician sweeping the country.
It is a picture of three competing political blocs.
And Tinubu did not even win a majority of the poll
This point should not be ignored.
Tinubu received 35.41%.
Obi received 33.86%.
Atiku received 26.67%.
Together, the three major candidates accounted for almost the entire exercise.
But Tinubu's own share means that 64.59% of the recorded responses went to somebody other than Tinubu.
So how does a result in which nearly two-thirds of respondents did not select Tinubu become a headline about an overwhelming presidential victory?
It doesn't.
At most, it establishes that Tinubu was the largest single choice in this particular exercise.
That is very different from demonstrating majority national support.
Trust is not the same as voting intention
There is another conceptual problem.
This is a National Trust Poll.
Trust and voting intention are not identical.
A Nigerian may say that one politician is the person they trust most but still vote for another candidate when faced with an actual ballot.
Voting decisions can be influenced by political alliances, party structures, regional considerations, candidate selection, campaign promises, economic conditions, security concerns and the eventual configuration of the presidential race.
The 2027 election is still ahead.
A September 2026 online trust exercise cannot settle what voters will do when they eventually stand before a ballot box.
Oxford-Metro itself anticipated political misuse
Perhaps the strongest evidence against sensational interpretations comes from Oxford-Metro's own words.
In its explanation of Public Pulse, the organisation acknowledged that politicians could attempt to misuse its figures as evidence of electoral strength.
It said it would remind the public that its figures represent open participation on a website, not projected vote shares or official results.
That warning should be placed beside every headline claiming that the poll proves a candidate will win.
The pollster has already told the public what the numbers do—and do not—mean.
This does not prove the poll is fake
There is an important distinction here.
Discrediting the interpretation of the poll is not the same thing as accusing Oxford-Metro of fabricating the responses.
The organisation says it uses controls designed to detect duplicate voting, bots and suspicious activity. It also says suspicious submissions can be excluded.
Those safeguards may help ensure that the responses recorded are genuine platform participation.
But even perfectly genuine responses do not become a representative national sample simply because they are genuine.
Real respondents can still produce an unrepresentative sample.
That is the central statistical problem.
What the poll actually tells us
Strip away the political marketing and the result becomes much more interesting.
It tells us that among people who voluntarily participated:
- Tinubu was first with 35.41%.
- Obi was extremely close behind at 33.86%.
- Atiku remained a substantial third at 26.67%.
- No candidate reached 50%.
- Tinubu and Obi were separated by only 1.54 percentage points.
- Support was heavily divided geographically.
- Lagos and the FCT supplied more than half of all responses.
- The poll was unweighted.
- A conventional margin of error does not apply.
- Oxford-Metro says the result is not representative of Nigeria's electorate.
- Oxford-Metro says it is not a prediction of the 2027 election.
That is a very different story from “Panic as Tinubu beats Obi and Atiku.”
The real panic may be over the headline, not the election
There is nothing wrong with reporting that Tinubu finished first in the Oxford-Metro exercise.
There is something wrong with transforming that narrow result into evidence that he has already established a national electoral advantage.
The difference between Tinubu and Obi is just 1,067 responses.
More than half of all responses came from Lagos and the FCT.
The participation was self-selected.
The results were unweighted.
There is no conventional margin of error.
And the pollster itself says it is not a representative survey or a prediction of the 2027 election.
So Nigerians should be careful with the political propaganda now being built around the numbers.
Tinubu did not beat Obi and Atiku in the 2027 election.
There has been no 2027 election.
He did not even beat them in a scientifically representative national opinion poll.
He simply received the largest number of responses in an open online participation exercise.
That may be worth discussing.
It may even be an interesting indicator of digital political sentiment.
But turning it into evidence that Tinubu is already winning Nigeria is a leap the methodology simply cannot support.
The poll is real. The responses may be real. The political spin is where the problem begins.
Comments
Post a Comment