NigeriaPolls · Standards
Research standards and disclosure
What we publish with every result, and what you are entitled to ask for. Written against the disclosure standard the profession actually uses.
In short
A number you cannot interrogate is not evidence, it is an assertion. So we publish what sits behind every result: who paid for it, who was asked, how they were found, what they were asked in those exact words, when, through which channel, and what was excluded.
The list below follows Section A of the American Association for Public Opinion Research disclosure standards, the items a research organisation should make public at the moment results are released. Anything not published is available on request within 30 days, which is the standard AAPOR sets.
We are not currently a member of AAPOR's Transparency Initiative. An application is in progress and we will say so here when it is granted, not before.
The standard
Eleven things we disclose
For every published result. Where a finding rests on more than one sample or more than one mode, each is disclosed separately.
01Data collection strategy
What kind of research produced the result: a survey, a focus group, an in-depth interview programme, or a combination, and which part of the finding comes from which.
02Who sponsored it and who conducted it
The sponsor of the research and the party that carried it out. Where the original source of funding differs from the sponsor, both are named.
03Measurement instruments
The exact wording and presentation of every question a reported result rests on, the response options offered, and any preceding question or context that could reasonably be expected to influence the answer. For qualitative work, the discussion guide and any instructions given to moderators.
04Population under study
Who the research is about, with the decision rules that defined them: location, age, and any other characteristic used to include or exclude someone. Nigerian adults aged 18 and above is a population; registered voters in Kano State is a different one, and we say which.
05How the sample was generated and recruited
The sample design, and how participants were contacted or recruited. We state explicitly whether a sample is probability-based or not, because that governs what can honestly be said about the result.
06Use of quotas
Where quotas were applied, what they were set on and at what level, by state, age band, gender, or any other characteristic.
07Modes of data collection
Which of telephone, SMS, WhatsApp or in-person interviewing produced the data, and the breakdown between them where more than one was used.
08Dates of data collection
The exact field period. A figure without its field dates cannot be assessed, and in an election cycle a stale number quoted as current is how good data ends up doing harm.
09Sample size and precision
The achieved sample size for every figure reported, including each subgroup quoted separately, and the margin of sampling error at 95 per cent confidence where a probability sample makes that meaningful.
10Weighting
The variables used to weight, the source of the population benchmarks, and the trimming applied. We weight by iterative proportional fitting against age, gender, state and geopolitical zone, benchmarked to published National Population Commission demographics.
11Processing and quality assurance
How responses were checked and what was excluded. Every recorded interview is transcribed and verified against the approved questionnaire, the logged answers and the assigned disposition. Interviews that fail are excluded and the reason is recorded, so exclusions are auditable rather than invisible.
What else we commit to
Beyond disclosure
We publish limitations alongside findings. Every method misses somebody. A study fielded only on digital channels under-represents rural, older and lower-income Nigerians, which is why in-person interviewing is a standard channel for us rather than an exception, and why we say where a sample is still thin.
Polls are snapshots, not forecasts. A poll measures stated intention during a field period. It does not predict an outcome, and we report it as the former.
Corrections are published, not quietly edited. Where a published figure turns out to be wrong we correct it visibly and say what changed.
Respondents are told what is happening. Interviewers identify NigeriaPolls and the purpose of the call. Where a call is recorded the respondent is told before recording begins and may decline and continue without being recorded. Outbound calling is screened against the NCC do-not-disturb register, and we keep our own do-not-contact list honoured across every study including commissioned work.
Published data is openly licensed. Under CC BY 4.0, so anyone can check our arithmetic, reuse the figures commercially with attribution, and disagree with us using our own numbers.
The full method is at nigeriapolls.com/methodology, our data protection position at nigeriapolls.com/legal/privacy, and our capabilities at nigeriapolls.com/capabilities.
FAQ
Frequently asked questions
What are the AAPOR disclosure standards?
The American Association for Public Opinion Research publishes a disclosure standard setting out what a research organisation should make public when it releases findings. Section A covers the items to disclose immediately: the data collection strategy, who sponsored and conducted the work, the measurement instruments, the population under study, how the sample was generated and recruited, any use of quotas, the modes of data collection, the field dates, and how the data were processed and quality assured.
Is NigeriaPolls a member of the AAPOR Transparency Initiative?
Not yet. We hold ourselves to the disclosure standard now and publish against it, and an application to the Transparency Initiative is in progress. We will say so here when it is granted and not before. We would rather be judged on what we publish than on a badge.
What does NigeriaPolls disclose with every poll?
The sponsor and who conducted the work, the population under study and how it was defined, the sampling and recruitment method, any quotas applied, the collection modes used and the breakdown between them, the field dates, the achieved sample size, the weighting variables, the margin of sampling error at 95 per cent confidence, and the exact question wording.
Can I see the questionnaire for a published poll?
Yes. Exact question wording and response options are part of what we disclose, including any preceding context that could reasonably influence an answer. If something you need is not on the page, ask and we will provide it within 30 days, which is the standard AAPOR sets for material not immediately published.
Who funds NigeriaPolls research?
Our own published polls and indices are funded by us. Commissioned studies are funded by the client, and where we publish anything arising from commissioned work the sponsor is named. Where the original source of funding differs from the sponsor, both are disclosed.