Media personality, Nana Aba Anamoah, has cast doubt on the credibility of election polls conducted by Global InfoAnalytics, which favours former President John Dramani Mahama in the upcoming 2024 general election.
In a post shared on her X (formerly Twitter) on November 18, 2024, Nana Aba questioned the methodology used in the survey, particularly the representativeness of the sample size.
“Assuming the minimum sample size is what you’re using now, how do you have so much confidence that an analysis of 2,600 people can truly determine the outcome of a decision to be taken by (using the previous election results) over 6 million people?” she wrote.
Nana Aba acknowledged the role of statistical analysis but argued that relying on a small sample size, in her view, undermines the accuracy of the results.
Global InfoAnalytics founder, Mussa K. Dankwah, responded to Nana Aba’s concerns, defending the validity of the methodology.
He explained that for a voting population of 18.7 million, a sample size of approximately 2,600 is statistically sufficient to achieve a confidence level of 95% and a margin of error of 1.9%.
“It is not difficult to calculate a sample size required for a given population; it is available free on the internet. But let me help you here: for a voting population size of 18.7 million, if you want to conduct a poll in this population with a confidence level of 95% and a margin of error of 1.9%, the minimum sample you need to interview is 2,600, roughly,” he stated.
However, Nana Aba maintained her stance, emphasising that the issue goes beyond mathematical calculations.
“It’s not about calculator and margins of error. It’s about a representative sample size. Your latest poll used 2,623 people. I’m not questioning statistical analysis and accuracy, but how can you tell me that figure is representative?” she reiterated.