As controversy surrounds the latest opinion survey, the bigger question is whether critics can identify genuine methodological flaws.
“The credibility of an opinion poll must be judged by its methodology, not by whether its findings favor our preferred candidate.”
The publication of a new political survey by the Institute for Governance Reform (IGR) has placed Mohamed Omodu Kamara, popularly known as Jagaban, at the center of Sierra Leone’s emerging debate over the APC flagbearer race. The report gives him an aggregate preference score of 65.8% among APC supporters and delegates, well ahead of the other aspirants listed.
The result is striking—and, for some, likely unwelcome. But neither reaction settles the question that matters: is the survey credible? A poll is not invalid simply because it produces an unfavourable outcome for a political actor. Equally, a result that pleases a candidate or their supporters does not, by itself, prove that the research was scientifically sound. The proper test is the quality and transparency of the method.
IGR says the second edition of Sierra Leone’s Political Marathon draws on interviews with 2,400 likely voters and 1,025 randomly selected delegates from the two main parties. The survey was conducted in September and October 2026. Those details offer a starting point, not a complete audit. To assess the findings, readers need to understand how respondents were selected, how interviews were conducted, how the results were weighed, and how the different groups were combined.
That last point is particularly important. The report presents an aggregate APC preference score for Jagaban, while also showing substantial differences between ordinary supporters and delegates. Dr Ibrahim Bangura, for example, recorded 14.9% among supporters and 37.5% among delegates. Dr Samura Kamara’s figures moved in the opposite direction: 34.1% among supporters and 10.4% among delegates. These are not minor distinctions. They describe different constituencies with different roles in a flagbearer contest.
A combined score can be useful, but only if its construction is clear. How much weight was assigned to supporters and how much to delegates? Were the figures adjusted to reflect the parties’ actual selection rules or the relative size of each group in the sample? Were respondents asked to choose one candidate, or could they express support for more than one? Without clear answers, critics have a legitimate reason to ask for further explanation. But that is a call for methodological transparency—not proof that the poll is fraudulent or invalid.
The survey itself cautions that reported preferences are not votes cast at party conventions, and are not predictions of the 2028 Presidential Election. That distinction should be kept firmly in view. A survey can measure what respondents say at a particular time; it cannot guarantee what delegates will do when they enter a convention. Party hierarchies, endorsements, campaign resources, personal loyalties, changing political conditions and pressure on delegates can all affect the eventual outcome.
IGR’s report acknowledges some of these risks, warning that political pressure, influence networks and fears of repercussions could affect delegates’ willingness to disclose their preferences. This is a material qualification. If respondents fear consequences for speaking openly, their answers may not reflect their private choices. The question for readers is whether the report explains how researchers sought to reduce that risk—for example, through confidential interviews, careful sampling or other safeguards.
The findings also show why the headline figure should not be treated as a simple forecast. On a separate question about candidates’ ability to address national challenges, 71.9% of APC supporters selected Jagaban, while 70.9% of delegates did so. These figures concern perceived capacity, not necessarily first-choice support in the flagbearer race. Similarly, the report says nearly half of respondents considered the ability to finance campaigns and support party members the most important candidate-selection factor. That may help explain political preferences, but it does not make financial capacity equivalent to electability or public approval.
There is useful context in IGR’s earlier work. Its December 2025 report found that only 26% of surveyed party members and supporters said they had voted in party primaries, while large majorities in both the APC and SLPP favoured allowing all card-holding members to vote directly. The earlier report also found that no SLPP aspirant secured majority support on its electability measure, while Jagaban was the only one among 26 potential aspirants across both parties to exceed 50% endorsement. This history may help readers understand the institute’s approach and the political questions it has tracked. It does not, however, substitute for scrutiny of the latest survey’s methods.
The same care is needed when interpreting the SLPP results. Ordinary supporters were closely divided among Dr David Moinina Sengeh, Dr Kandeh Yumkella and First Lady Fatima Bio. Among delegates, the reported preferences were different, with Fatima Bio and Vice-President Mohamed Juldeh Jalloh also scoring strongly. Such divergence is not inherently evidence of a flawed survey. It may reflect real differences between the wider support base and the people who participate in party conventions. But it makes transparent reporting of the sample, questions and calculations all the more important.
Political criticism of polls is healthy when it is specific. Critics should ask whether the sample was representative, whether the questions were neutrally worded, whether respondents could answer privately, how non-responses were handled, and how the published percentages were calculated. They should identify an error that can be examined and, where possible, provide evidence. Researchers, in turn, should publish enough methodological detail for independent readers to assess the work.
What does not advance public understanding is dismissing a poll solely because its leading candidate is not one’s preferred choice. Nor should supporters present a favourable result as conclusive proof of broad public backing. A poll is a snapshot shaped by its questions, sample and timing—not a convention result, a mandate or a prophecy.
The IGR report has put numbers into a political conversation that will intensify as the 2028 elections approach. Those numbers deserve neither automatic acceptance nor reflexive rejection. The fair response is to examine the evidence, demand clarity where details are missing, and distinguish methodological criticism from political disappointment. In a democracy, disagreement with a finding is a reason to ask better questions—not, on its own, a reason to declare the poll invalid.
