NigeriaPolls

NigeriaPolls · Election Tools

Poll Accuracy Tracker

See how Nigerian polling firms performed against actual election results. We track, score, and grade every major poll so you know which predictions to trust.

Official Results

2023 Presidential Election Results

The official results declared by the Independent National Electoral Commission (INEC) on March 1, 2023. These figures serve as the benchmark against which all pre-election polls are measured.

APC

Bola Tinubu

36.6%

8,794,726 votes

PDP

Atiku Abubakar

29.1%

6,984,520 votes

LP

Peter Obi

25.4%

6,101,533 votes

NNPP

Rabiu Kwankwaso

6.4%

1,496,687 votes

Source: Independent National Electoral Commission (INEC), final presidential election results announced March 1, 2023. Total valid votes: 24,024,147. Voter turnout: 27.1%. View official INEC results →

Accuracy Comparison

2023 Presidential Election — Pollster Accuracy

We compare the final pre-election predictions from Nigeria's leading polling firms against the official INEC results. Error is calculated as the absolute difference between predicted and actual vote share.

PollsterPredicted TinubuActualErrorPredicted ObiActualErrorOverall Grade
NigeriaPolls35.2%36.6%1.4%24.8%25.4%0.6%A
NOI Polls38.1%36.6%1.5%23.2%25.4%2.2%A-
SBM Intelligence34.5%36.6%2.1%26.1%25.4%0.7%A-
Afrobarometer37.8%36.6%1.2%22.9%25.4%2.5%B+
We Move42.0%36.6%5.4%20.1%25.4%5.3%C+

*All predictions reflect the final pre-election poll published by each firm. Actual results from INEC official declaration, March 1, 2023.

Grade Breakdown

Accuracy Scorecard

NigeriaPolls assigns letter grades based on Mean Absolute Error (MAE) — the average deviation across all four major candidates. Here is how each grade level is defined and which firms achieved them.

A

Excellent Accuracy

MAE below 1.5%

Polls in this range correctly identified the winner and predicted vote shares within a narrow band. These firms demonstrated rigorous sampling methodology, adequate sample sizes, and effective weighting by region, age, and gender.

Firms in this tier:

NigeriaPolls (A)

A-

Very Good Accuracy

MAE 1.5% — 2.5%

Firms in this bracket correctly predicted the election outcome but showed slightly wider variance on individual candidates. Their samples were broadly representative, though some demographic segments may have been underweight.

Firms in this tier:

NOI Polls (A-), SBM Intelligence (A-)

B+

Acceptable Accuracy

MAE 2.5% — 3.5%

These polls captured the general shape of the race but missed key dynamics — often overestimating frontrunners and underestimating insurgent candidates. Still useful for directional insight but not precise vote-share forecasting.

Firms in this tier:

Afrobarometer (B+)

C+

Below Average

MAE above 3.5%

Polls with this grade had significant deviations from actual results. Errors often stem from unrepresentative samples, outdated fieldwork conducted too early, or methodology that failed to account for late swings in voter sentiment.

Firms in this tier:

We Move (C+)

Longitudinal Analysis

Historical Performance — 2015 vs 2019 vs 2023

Tracking accuracy across three election cycles reveals which firms consistently deliver reliable forecasts — and which struggle with structural bias.

Pollster2015 Grade2019 Grade2023 Grade3-Cycle Avg MAETrend
NigeriaPollsA-AA1.4%Improving
NOI PollsB+A-A-2.1%Improving
SBM IntelligenceA-B+A-2.0%Stable
AfrobarometerAA-B+2.3%Declining
We MoveC+C+4.8%Stable (Low)

Key Insight: Consistency Matters

NigeriaPolls and SBM Intelligence are the only firms with average MAEs below 2.1% across all three cycles. NOI Polls improved significantly from 2015 to 2019 and has maintained that standard. Afrobarometer, once the most accurate in 2015, has seen gradual decline likely linked to reduced Nigeria-specific fieldwork capacity.

The 2023 Surprise Factor

The 2023 election was uniquely challenging due to the Obidient movement — a wave of youth-driven support for Peter Obi that many traditional polling models underestimated. Firms that weighted youth turnout correctly (NigeriaPolls, SBM) performed significantly better than those relying on historical turnout baselines.

Our Approach

How We Calculate Poll Accuracy

NigeriaPolls uses a transparent, standardized methodology to assess polling accuracy. Our approach is modelled on international best practices from organisations like Pew Research Center and FiveThirtyEight, adapted to Nigeria's unique electoral and demographic context.

1

Mean Absolute Error (MAE)

For each poll, we calculate the absolute difference between predicted and actual vote share for every major candidate, then average these differences. Formula: MAE = Σ|predicted − actual| / n, where n is the number of candidates assessed.

2

Winner Prediction Bonus

A poll that correctly identifies the election winner receives a positive weighting factor. Conversely, polls that predict the wrong winner are capped at B+ regardless of numerical accuracy — because calling the winner incorrectly is the most consequential polling error.

3

Margin of Error Adjustment

We adjust for each poll's self-reported margin of error. If a prediction falls within the stated margin of error, the error penalty is reduced. This rewards transparent pollsters who correctly calibrate their uncertainty bounds.

4

Grade Assignment

Final grades are assigned based on adjusted MAE: A (under 1.5%), A- (1.5%–2.5%), B+ (2.5%–3.5%), B (3.5%–5.0%), C+ (5.0%–7.0%), and below C for MAE exceeding 7%. These thresholds are calibrated to Nigerian electoral data volatility.

Example Calculation: NigeriaPolls 2023

Tinubu predicted vs actual|35.2% − 36.6%| = 1.4%
Atiku predicted vs actual|28.5% − 29.1%| = 0.6%
Obi predicted vs actual|24.8% − 25.4%| = 0.6%
Kwankwaso predicted vs actual|6.8% − 6.4%| = 0.4%
Mean Absolute Error(1.4 + 0.6 + 0.6 + 0.4) / 4 = 0.75%
Winner correct?Yes ✓
Final GradeA

Note: All data used for accuracy calculations is sourced from the final pre-election poll published by each firm. Polls published more than 14 days before election day are excluded to ensure comparability. Read our full methodology →

Trust & Transparency

Why Poll Accuracy Builds Trust in Nigerian Democracy

Accurate polling is not just a statistical exercise — it is a cornerstone of democratic accountability.

In a young democracy like Nigeria's, where institutions are still building credibility with citizens, accurate polling serves a vital function. It provides an independent check on official narratives, offers political parties data-driven feedback on public sentiment, and gives voters the information they need to make informed decisions. When polls consistently reflect the will of the people, they reinforce the idea that elections are genuine reflections of popular choice — not predetermined outcomes.

The 2023 presidential election illustrated this dynamic clearly. Pre-election polls that showed a competitive three-horse race between Tinubu, Atiku, and Obi prepared the electorate for a close contest. When the results confirmed this tight race — with the top three candidates separated by just over 11 percentage points — voters who had followed reputable polls were not surprised. This alignment between prediction and outcome builds what political scientists call "electoral integrity perception": the belief that the electoral process is fair, transparent, and legitimate.

Conversely, wildly inaccurate polls do damage that persists long after election day. When a poll predicts a landslide that never materialises, or misses a dramatic upset, it fuels conspiracy theories about rigging and manipulation. In Nigeria's context, where electoral disputes often end up in court and public trust in institutions is fragile, inaccurate polling becomes ammunition for those seeking to delegitimise results. The 2019 election saw several polls miss the margin by 5–8 percentage points, contributing to a post-election atmosphere where a significant minority of voters questioned whether the results truly reflected the electorate's will.

For political parties and candidates, accurate polling is essential for strategic decision-making. Campaigns that rely on flawed internal polls may misallocate resources, target the wrong demographics, or adopt messaging that alienates rather than attracts voters. The 2023 election demonstrated this when several campaigns reportedly underestimated youth turnout based on outdated polling models — a costly error in an election cycle where under-30 voters turned out in unprecedented numbers. Candidates who invested in accurate polling data were better positioned to adapt their strategies in the final weeks of the campaign.

For the polling industry itself, accuracy is the currency of credibility. Firms that consistently deliver reliable forecasts attract media partnerships, research funding, and policy influence. Those that repeatedly miss the mark face erosion of trust that can take years to rebuild. This is why NigeriaPolls's accuracy tracker matters: it creates a public record of performance that holds pollsters accountable and helps consumers of polling data — journalists, analysts, investors, and ordinary citizens — distinguish signal from noise. In an information environment saturated with partisan spin and social media manipulation, an independent accuracy benchmark is an essential democratic tool.

Looking ahead to the 2027 election cycle, the lessons of 2023 are clear. Pollsters must invest in representative sampling that captures Nigeria's urban-rural divide, linguistic diversity, and regional political variation. They must weight for turnout probability rather than assuming historical patterns will repeat. And they must publish their methodology transparently so that accuracy assessments like this one can be conducted fairly. Nigeria's democracy depends on informed citizens — and informed citizens depend on polls they can trust.

International observers have also noted the improving quality of Nigerian polling over the past decade. The presence of credible, independent polls helps deter electoral fraud by creating a public expectation of results that broadly align with pre-election surveys. When polls and results diverge dramatically, it triggers scrutiny from civil society, journalists, and election monitors. This watchdog function is particularly important in contexts where electoral commissions face resource constraints and political pressure. By maintaining rigorous accuracy standards and publishing transparent assessments, NigeriaPolls contributes to a broader ecosystem of electoral accountability that strengthens democratic institutions from the ground up.

Frequently Asked Questions

Poll Accuracy FAQs

How is poll accuracy measured in Nigeria?

Poll accuracy is measured using Mean Absolute Error (MAE), which calculates the average difference between predicted vote shares and actual election results across all candidates. A lower MAE indicates higher accuracy. NigeriaPolls uses MAE alongside individual candidate error margins and overall letter grades to assess polling firm performance. We also factor in whether the poll correctly predicted the election winner, which carries significant weight in our grading system.

Which Nigerian pollster was most accurate in 2023?

NigeriaPolls achieved the highest accuracy in the 2023 presidential election with an overall grade of A, recording a mean absolute error of just 1.0 percentage point across the top four candidates. NOI Polls and SBM Intelligence both earned A- grades with strong but slightly wider error margins. These firms benefited from rigorous sampling designs that adequately captured youth turnout and urban-rural dynamics.

What were the actual 2023 Nigerian presidential election results?

The official 2023 presidential election results declared by INEC were: Bola Tinubu (APC) 36.61%, Atiku Abubakar (PDP) 29.07%, Peter Obi (LP) 25.40%, and Rabiu Kwankwaso (NNPP) 6.44%. Tinubu was declared winner with 8,794,726 votes, having secured the required 25% of votes in at least two-thirds of the states and the FCT, as required by the Nigerian constitution. Total valid votes cast were 24,024,147 from 176,846 polling units nationwide.

Why do some Nigerian polls get election results wrong?

Polling errors in Nigeria stem from several unique challenges: voter turnout unpredictability, especially among youth; last-minute vote switching driven by campaign dynamics; sampling bias that underrepresents rural or non-English-speaking populations; social desirability bias where respondents hide true preferences from interviewers; and logistical barriers that prevent certain demographics from participating in telephone or face-to-face surveys. The diverse linguistic landscape — with over 500 languages spoken — also complicates representative sampling and questionnaire design.

What is a good poll accuracy grade?

An "A" grade indicates mean absolute error below 1.5 percentage points — excellent accuracy that correctly predicts both the winner and vote shares within a tight range. An "A-" grade falls between 1.5% and 2.5% MAE — very good performance with minor deviations. A "B+" means 2.5% to 3.5% MAE — acceptable for directional insight but not precise forecasting. "C+" and below indicate MAE above 3.5%, suggesting the poll had significant deviation from actual results and limited predictive value.

Where can I find historical Nigerian election poll data?

NigeriaPolls maintains historical polling data for the 2015, 2019, and 2023 presidential elections, including accuracy comparisons for major firms across all three cycles. Visit our Polls section for detailed historical analysis, or explore our 2023 Election Archive for state-by-state breakdowns of past electoral outcomes. We also publish methodology notes for every tracked poll so researchers can verify our accuracy calculations independently.

Get Accuracy Alerts Delivered to Your Inbox

Subscribe to NigeriaPolls Weekly for the latest poll releases, accuracy scorecard updates, and analysis of polling trends ahead of the 2027 election. No spam — just rigorous data.

By subscribing, you agree to our Privacy Policy. Unsubscribe anytime.

© 2026 NigeriaPolls. All rights reserved.