Africa-Press – Gambia. In recent days, a wave of commentary has emerged surrounding the opinion polls conducted by CEPRASS under the auspices of The Gambia Participates. While many commentators have received the report with optimism and recognition of its value, a small but vocal group has voiced discontent with its findings. Their focal point of contention lies in the sampling process, with repeated criticisms directed at the sample size, statistical rigour, design, demographic representation, research tools, and gender balance. In this article, I intend to demonstrate why the sample size is both correct and defensible, drawing on established academic standards and authoritative scholarly perspectives in political opinion research to substantiate my argument.
Sample size
Sampling is a critical determinant of the reliability and validity of any research, including opinion polls. In political science and survey research, the adequacy of a sample size greatly influences the accuracy of estimates, the generalisation of findings, and the credibility of conclusions. Renowned scholars such as Fowler (2014) and Kish (1965) emphasised that sample size should be sufficiently large to ensure demographic representativeness and statistical power. While Fowler posits that sample size should be determined by the intended confidence level, error margin, and population heterogeneity, Kish stressed the need for stratified random sampling to achieve precision and accuracy without necessarily increasing the sample size.
In The Gambia, with an estimated voting-age population of around a million, a sample size of 1,556 yields a margin of error of approximately ±2.5% at a 95% confidence level, assuming a simple random sample. This value falls well within the acceptable thresholds for national opinion polling, as noted by Groves et al. (2009), who argued that margins of error below ±5% are considered reliable for public opinion estimates. These values were modelled by global pollsters in the last US election, as Emerson College Polling showed a margin of error of ±3.4% for Florida and ±3.3% for Texas. Reuters and NPR both reported a margin of error of 3.5% in their national polls. Therefore, in both theory and practice, CEPRASS’s margin of error of ±2.5% was within the acceptable threshold. Questioning their sample size would suggest that one has not done their homework in terms of literature and accepted practice.
Methodological strength
As per the report, the survey employed stratified random sampling, a technique endorsed by Cochran (1977) for its ability to reduce sampling error and improve subgroup analysis. Stratification by Local Government Areas (LGAs) and urban/rural divisions ensured proportional representation, a key criterion for sample correctness. In fact, the inclusion of all 53 constituencies in the country enhances geographic coverage and minimises regional bias. The researchers went further to break down the demographics in a manner that included key variables like gender, age, employment status, and education to support their representativeness. For instance, they captured a diverse electorate, with over 60% of respondents having secondary or tertiary education, and a balanced distribution across age groups. This aligns with the recommendations of Salant and Dillman (1994), who argue that demographic diversity within a sample strengthens the validity of inferences. Therefore, any argument against representativeness, as far as the data is concerned, would be a deliberate insult to research and academic excellence.
Scholarly Endorsement and Comparative Perspective
Robert Samohyl (2021), writing in Survey Practice, critiques overreliance on confidence intervals and advocates for probabilistic modelling, especially in contexts with limited polling infrastructure. The complexity of research infrastructure in The Gambia is an open secret even to those who have never written an academic paper. Therefore, the use of STATA for disaggregated analysis and its stratified design reflects an awareness of such advanced methodological concerns, evidencing a high level of technical competence. Meanwhile, Patrick Fischer (2023) emphasises the need for sample sizes to be context-sensitive, balancing statistical rigor with logistical feasibility. This research overwhelmingly strikes that balance effectively, offering national-level insights while enabling regional and demographic disaggregation.
To be fair to the critics, even though they may not have found any interesting or sound sentence in the CEPRASS report, I found their critique of the gender imbalance genuine and on point. The imbalance in gender representation may affect the generalisability of gender-related findings, as male voters might not necessarily vote the same way as female voters. Even though there is no evidence to show that male and female voters behave significantly differently in The Gambia in terms of voting, it is only fair to ensure equality for fair representation. Therefore, this is an area where CepRass could have done better, or perhaps acknowledged the imbalance as a limitation of the research, which would have helped readers reach a more informed conclusion. However, since there is no recent empirical study confirming differing voting attitudes between genders, this cannot invalidate the findings. Also, since gender was one of many tested variables, an oversight in this area cannot cast doubt on the overall results.
Conclusion
The sample size of 1,556 respondents in The Gambia’s 2026 Presidential Election Opinion Poll is both methodologically robust and academically defensible. It adheres to well-established principles of survey design, including stratified sampling, proportional allocation, and demographic representativeness. Although minor limitations exist, particularly in gender balance, but the overall framework aligns with international best practices in political opinion research. Consequently, the sample size is not only correct but also appropriate for the survey’s objectives, providing credible insights into voter sentiment and electoral dynamics in The Gambia.
And now, for a little political spice: I would strongly urge the critics to, for once, recognise genuine hard work and resist the temptation of nitpicking for sport. If their objections are indeed rooted in academic rigor rather than political bias, the least they could do is cite actual scholarly references, which would be a basic gesture of respect for academic integrity. Better yet, they might consider joining forces with the People’s Mayor, Ahmed Bensouda, the most probable UDP flagbearer and the natural choice to lead any credible coalition. After all, The Mayor will certainly need their brilliance when salvaging The Gambia, and rest assured, his administration will happily provide them with the research infrastructure to keep them busy. Perhaps then, instead of chasing shadows, they can finally put their “beautiful work” to real use.
Source: The Standard Newspaper | Gambia
For More News And Analysis About Gambia Follow Africa-Press
