top of page

Press release: methodology statement


This statement has been necessitated by several requests for clarification on methodology regarding the recent 2023 pre-election polls conducted by NOIPolls commissioned by ANAP Foundation.

NOIPolls Limited is an indigenous research and opinion polling organization founded in 2006. We have a long-standing history of conducting opinion polls on various socio-economic and political issues affecting Nigerians including pre-election polls since 2011. More information is available on the NOIPolls website.

With regards to pre-elections polls, we do not conduct polls for political parties or candidates but conduct polls for non-governmental organizations and independent organizations with goals to fostering accountability and encouraging civic responsibilities. We have been conducting pre-elections polls for ANAP Foundation since 2011.

The more recent pre-election polls we have conducted for ANAP Foundation include the 2015 and 2019 pre-presidential elections polls which showed President Muhammad Buhari of APC leading the then President Goodluck Jonathan of PDP in 2015 and also leading his main challenger HE Alhaji Atiku Abubakar of PDP in 2019. See links to the publications below:

Our election polls have not been limited to presidential elections but include gubernatorial elections with the following sample publications:

Including the most recent gubernatorial elections in Ekiti State.

Independently, we also have conducted pre-election polls including the Anambra State gubernatorial elections in 2021

In summary we have been carrying out credible pre-election polls since 2011 and our methodology have not changed.

For the presidential pre-election polls conducted in September 2022, which is the first in a series of pre-election polls commissioned by ANAP Foundation, a total sample size of 1,000[1], phone-owning Nigerians (aged 18+) proportionately representing the six geo-political zones in the country were interviewed. This was the same sample size and distribution used in 2015 and 2019. With this sample size, and the selection method, the results obtained represents the opinions of adult Nigerians with 95% confidence level and plus or minus 4.65%[2] margin of error. This includes a design effect of 1.5. The interviews were conducted in the language of preference of the respondent viz; English, Pidgin, Hausa, Igbo, and Yoruba languages. Interviews were done over the telephone. Nigeria’s tele-density at more than 97.45%[3] makes telephone polling a viable methodology.

NOIPolls adopts a proportionate, stratified random sampling design for polling. Respondents 18+ years are selected form our database of over 70 million phone owning Nigerians. Stratification is set on several key demographics such as age (18+), gender, geo-political zones, states, and Local Government Areas (LGAs) to make sure the final sample is representative of the actual population. The gender and the geo-political zone demographic distribution of the respondents are in the same proportion with the 2006 National Population Census.

In conclusion, while NOIPOLLS stands by its findings given our rigorous method of obtaining the opinion of Nigerians, we would like to point readers to the significant proportion of undecided voters (32%) and respondents that refused to disclose their choice (15%) making a total of 47% of respondents who said they are registered to vote and who can swing the outcome of the election in any direction. NOIPolls remains resolute to providing timely and relevant information on public opinion on various social and economic issues, thus give voice to ordinary citizens, while empowering decision makers in the private and public sectors for better performance and improved governance.


Dr. Chike Nwangwu

Chief Executive Officer,

NOIPolls Limited

[1] A minimum of a total of 1,000 respondents, drawn from the population of 18 years and above Nigerians is recommended using a stratified random sampling of respondent from the population. The sample size (minimum of 1,000) is calculated on the assumption that 50% of the population have a given characteristic or indicator of interest. The required sample size (n) for the general public is calculated with the assumption to estimate the expected proportion (p) of the indicator of interest with 3.1% margin of error. Using the formula for calculating 95% confidence interval, the equation for calculating the sample size is given below: where n is the minimum sample size required, is the margin of error (level of precision), is the estimate of the current outcome measure (estimated to be 0.50, which gives the widest confidence interval, is the level of significance (= 5%), = Standard normal deviate or Z score corresponding to the two-sided null hypothesis (= 1.96).

[2] Margin of error of 3.1% (assuming simple random sampling) multiplied by a design effect of 1.5 (accounting for variance in estimates accruing from the use of proportionate stratified random sampling in selecting respondents) gives plus or minus 4.65%.



bottom of page