How NaijaElects measures public opinion

Good polling is only as trustworthy as the method behind it. This article explains, in plain language, how NaijaElects collects opinion, how we protect the data from manipulation, and — just as importantly — the limits of what our numbers can tell you.

What we ask

Every respondent answers a short, anonymous set of questions: which office they are voting on (presidential, governorship or House of Assembly), their candidate or party preference, their state, their age band, and the single issue that matters most to them. That's it. We deliberately keep the form under a minute to complete.

Crucially, we do not collect names, phone numbers, email addresses, PVC numbers, NIN, BVN or exact home addresses. Demographic questions use broad categories — a state and an age band — so that no individual can be identified from a response.

How results are aggregated

Individual responses are never published. Everything you see on the dashboard is an aggregate: the percentage of respondents choosing each candidate, the ranking of issues, the breakdown by state. This protects the privacy of participants and keeps the focus where it belongs — on the overall picture rather than any one person's vote.

Guarding against manipulation

Any open online poll faces a basic challenge: how do you stop a small group from voting many times to distort the result? We use several layers of protection:

• One response per office, per device. The system limits repeat submissions so a single device cannot flood the poll.
• Rate limiting. Rapid, repeated submissions from the same source are throttled automatically.
• Anonymised hashing. We store a one-way hash of network information for duplicate detection — never the raw IP address.
• Anomaly monitoring. Unusual spikes and suspicious patterns are flagged for review before they can skew public figures.

Sample quality

Not all responses are equal. Alongside the headline numbers we publish a sample-quality indicator that reflects how complete and consistent the incoming data is. A higher score means the current pool of responses is more balanced and reliable; a lower score is a signal to read the figures with extra caution.

The honest limitations

We believe transparency about weaknesses is what separates trustworthy polling from noise. Two limitations are worth stating plainly:

• Online samples skew. People who take part online tend to be younger, more urban and more connected than the electorate as a whole. That means open-poll figures should be read as directional sentiment, not a perfect mirror of the country.
• Self-selection. Respondents choose to participate, which is different from a randomly drawn scientific sample. Where we publish a scientifically sampled and weighted release, we say so explicitly and document the method.

A poll should tell you how confident to be in its own numbers. That's the standard we hold ourselves to.

How to read the dashboard

Use the segment filters to compare groups — for example, how younger voters differ from the overall pool. Switch the state view between map and list to see regional patterns. And treat sudden movements with healthy scepticism until they persist over time; genuine shifts in opinion usually build gradually.

Our commitment

NaijaElects is independent and non-partisan. We are not affiliated with INEC or any political party, we do not endorse candidates, and we do not present our figures as official results or predictions. Our only goal is to measure and share public sentiment as honestly as the method allows.