How to read Nigerian election polls without getting misled

Every election season, screens fill with percentages. Some come from careful surveys. Others come from viral graphics with no method behind them. This guide explains how to read opinion polls in Nigeria — including open online trackers like NaijaElects — so you can tell useful signal from noise.

Start with what the poll claims to measure

A useful poll states its question clearly. “If the election were held today, who would you vote for?” measures current voting intention. That is different from approval of a sitting president, different from “who do you like personally,” and different from a forecast of who will win months later.

Official election results are something else entirely. Only INEC produces those after votes are cast and counted. Opinion polls — including ours — measure what respondents say at a point in time.

Ask who was invited to answer

There are three broad families of political polls:

• Probability / scientific surveys that start from a designed sample (often face-to-face or carefully recruited) and weight results to match the population.
• Open online polls where anyone who finds the link can respond. These are good for engagement and directional mood, but they overrepresent people who are online and motivated.
• Unlabelled graphics that show percentages with no sample size, no dates, and no method. Treat these as entertainment until proven otherwise.

NaijaElects runs an open online poll with quality controls. We describe the method on our methodology page and warn against reading every snapshot as a perfect mirror of Nigeria.

Check the base, not only the percentage

“Candidate A leads with 48%” means little if you do not know how many people answered. A lead based on a few dozen responses in one city is fragile. A lead based on thousands of responses spread across many states is more interesting — still not official, but harder to dismiss as a fluke.

Also check state coverage. Nigeria is large and diverse. A national number built mostly from two megacities can miss rural and regional patterns that decide real elections.

Watch the undecided and the persuadable

Head-to-head charts among “decided” voters can look dramatic while hiding a large undecided group. Late deciders, first-time voters, and people who care more about issues than parties can still move. When a dashboard offers a “persuadable” or undecided segment, use it — it often tells a more honest story than a two-candidate horse race graphic.

Prefer trends over one viral day

Genuine shifts in public mood usually build over weeks. One-day spikes often follow a major news event, a coordinated sharing campaign, or a temporary surge of partisans finding the poll link. Before you rewrite your mental map of 2027, ask: did this pattern hold for several updates, or did it appear once and vanish?

Separate issues from candidates

Voters can dislike a party’s candidate and still rank the economy or security as their top issue. Issue rankings help explain why preferences move. If cost of living dominates the issue list while candidate shares stay volatile, the campaign conversation and the private household conversation may not be the same thing.

Online polls have known biases

People who take polls online tend to be younger, more urban, and more connected than the full electorate. That does not make online data useless — it makes it incomplete. The responsible habit is to say: “Among people who responded online this week, preference looks like X,” not “Nigeria has decided.”

Weighting and scientific field work can reduce some of those biases. Until a release is explicitly labelled as a weighted scientific sample, assume the open stream is directional.

A trustworthy poll tells you how confident to be in its own numbers. If a graphic cannot explain its method, it is asking for your trust without earning it.

A simple checklist before you share a poll

1. Who ran it, and are they independent of the parties on the ballot?
2. What exact question was asked?
3. When were responses collected?
4. How many people answered, and from which states?
5. Is this an open online poll, a scientific sample, or unclear?
6. Does the source show limitations, or only a victory headline?
7. Am I about to call this an “INEC result” by mistake?

How NaijaElects fits

We publish live aggregates, segment filters, state views, and issue rankings so readers can inspect the data rather than accept a single slogan. We also publish explainers on how Nigerian elections work, voter registration, and current 2027 developments. Our job is measurement and education — not crowning winners early.