By Alhaji Kemo Conteh,
Founder and senior partner, GDMS
Attention:
Independent Electoral Commission
CSOs
Media
Introduction:
This Memorandum conveys is a non-partisan, evidence based methodological critique of the CepRass Perception Survey on the 2026 Presidential Elections. It goes deeper into the inquiry of methodologies than the mostly surface level avalanche of analysis and commentary which has dominated Gambian media since the release of the report. It is framed to zoom on the methodological transparency of the exercise, contribute to the national, constitutional and legal objective of electoral integrity, and effective and instructive voter literacy in the on-going public discourse. It seeks not to challenge CepRass and any affected organisation or institution.
Background and purpose
Presently, public discourse in the Gambia is being significantly influence by the recently released perception survey conducted by the Centre for Policy, Research and Strategic Studies (CepRass) on the electability of the candidates in the 2026 presidential election.
Given the high sensitivity of the electoral processes and the potential impact of opinion polling on voter behaviour, political stability, and democratic legitimacy, it is essential that such survey meet clear methodological and transparency standards. Reflecting on this, this memorandum seeks to provide a technical, non-partisan critique of the publicly available methodological elements of the CepRass survey, with the objective of:
Supporting electoral integrity
Enhancing public understanding of polling limitation, and encouraging best practice standards in future-election-related research scope and sources of sssessment
This critique is based on:
Publicly released survey results and media reports; statements attributed to CepRass representatives and comparison with internationally accepted election polling standards.
It is important to note that no full technical methodology report has been publicly released for the survey under review. Accordingly, this memorandum assesses both what is known and what is not disclosed, as the absence of information itself is a material methodological issue
Key Methodological Observations:
Survey classifications
This CepRas exercise is best classified as a cross-sectional public perception survey, not a predictive election pool, while this distinction sometimes reflects in public commentary, it has not always been clearly communicated in media reporting, potentially increasing the risk of misinterpretation by the public comparison with electoral integrity best practice, internationally accepted election standards (e.g) Afrobarometer, Pew Research, IFES-supported survey) require; clear probabilistic sampling or exhibit acknowledgement of non-probability methods; full methodological disclosure; explicit margins of error and transparent funding and governance arrangement.
The CepRass survey align partially with these standards in scope and intent, but fall short in transparency and statistically rigor as required for high stakes electoral contexts.
Sampling design and representativeness
The survey is described as a “nationwide” and appears to use geographic stratification across regions or local government areas.
Sample size is reportedly comparable to prior CepRass survey (approximately 1,500 respondents).
Risks to electoral integrity and public discourse
If perception surveys are reported without adequate methodological context, there is a risk of:
Voter demobilisation or bandwagon effects; misrepresentation of electoral competitiveness;
Undermining public confidence in the electoral process; politicisation of research outputs.
These risks are not unique to CepRass and apply broadly to the election polling in emerging democracies.
Key methodological gaps:
No publicly disclosed sampling frame (e.g, IEC voter register, census enumeration areas).
No explanation of within-stratum respondent selection. No information on household selection procedures or respondent randomisation and no disclosure of weighing or post-stratification adjustments.
Without this information, it is impossible to verify independently, the statistical representation of the sample. Consequently, margins or error, and confidence intervals cannot be reliably inferred.
Data collection procedures:
Although CepRass has not given any public documentation of enumerator training supervision, or quality assurance, data collection appears to have been conducted trained enumerators through structured interviews, likely face-to-face with digital tools.
No reporting of response rate or non-response bias, no disclosure of fieldwork dates and duration.
The implication for this is that the absence of fieldwork documentation limits assessment of data reliability and potential interviewer of timing effects.
Questionnaire design and question framing:
The survey includes vote intention, candidate preference, and perceived likelihood of victory.
Results are presented as percentages of respondents and the concern for this is:
The full questionnaire has not been published.
There is no evidence of pretesting or neutrality checks.
External commentary suggests possible priming of framing effects, particularly in how candidates or alliances were referenced.
The implications for this that without access to the instrument, it is impossible to rule out measurement bias or question-order effects.
Data analysis and reporting:
Analysis relies primarily on descriptive statistics.
Undecided voters are explicitly reported, which is methodologically appropriate.
CepRass has stated that finding reflects perceptions at specific point in time.
No demographic or geographic weighing is disclosed.
No confidence intervals or error margins are reported.
No distinction made between likely voters and the general adult population
Implication: Findings should be interpreted strictly as indicative perceptions, not estimates of electoral outcomes.
Transparency and disclosure concerns
From an electoral integrity standpoint, the most significant issue is partial methodological transparency. Specifically absent from the public domain are:
A full methodology annexes
Sampling are weighing procedure
Questionnaire instrument
Fieldwork details
Funding and commissioning disclosure
The omissions do not imply misconduct, but they limit independence verification, and responsible interpretation and public trust.
Recommendations to:
Research institutions
Publish a full technical methodology annex with all election-related survey
Clearly label perception survey as non-predictive.
Disclose sampling frames, weighing, and margins of error.
Independent Electoral Commission-IEC
Issue public guidance on the interpretation of election opinion polls
Encourage methodological transparency as a condition for public credibility
Civil society and media
Report poll finding with mythological caveats
Avoid presenting perceptions data as forecasts or outcomes
Promote media literacy on opinion polling.
Conclusion
CepRass has a deserved reputation for electoral prediction in the Gambia but electoral perception surveys that do not meet international norms and standards such as using non-probability sampling, small sample sizes, or poor question framing can still successfully predict election results due to advanced statistical adjustments, the sheer simplicity of the electoral math, or, in some cases, luck. While reputable polls aim for a precise, representative cross-section of the population, substandard polls can sometimes hit the mark by correcting for skewed data or by tapping into the same underlying voter sentiment that determines the final outcome.
This particular CepRass perception survey provides useful insight into voter sentiment at a particular moment in time. However, due to limited methodological disclosure, its finding should be interpreted with caution and should not be treated as predictive of the 2026 election outcome. Strengthening transparency and adherence to international best practice will enhance public trust, protect electoral integrity, and contribute positively to democratic consolidation in The Gambia.
