Energy Analyst Kwadwo Poku has rejected claims that the support base of the New Patriotic Party (NPP) is dwindling, describing recent findings by pollster Mussa Dankwa as questionable and difficult to accept.

His comments come after findings by Global InfoAnalytics suggested that support for the New Patriotic Party is declining, while the National Democratic Congress is gaining ground among voters, according to early results from an ongoing March 2026 tracking poll.

Speaking on JoyNews AM Show on Tuesday, March 24, Mr Poku said while he acknowledges Mr Dankwa’s reputation as a pollster, he has concerns about the methodology and conclusions of the survey that is suggesting a decline in NPP support.

“I have no doubt he is a good pollster,” Mr Poku stated. “But I think he is trying to do too many things that for me, I have a problem with. In polls, you might get it right, or you might get it wrong. No pollster is 100 percent right.”

He cited global examples to support his argument, referencing United States presidential election, where several polls predicted a victory for Hillary Clinton, yet Donald Trump ultimately won.

Mr Poku further criticised what he described as inconsistencies in Mr Dankwa’s previous work, particularly a survey on cocoa farmers. According to him, the methodology used in that poll raised concerns about the credibility of its conclusions.

He explained that although Mr Dankwa claimed to have sampled respondents across all 16 regions, many of those surveyed were not cocoa farmers themselves but friends of cocoa farmers, and the report then equated their satisfaction with that of actual cocoa farmers.

Turning to the latest poll, Mr Poku questioned the credibility of the data collection process. He explained that Mr Dankwa had earlier indicated via TikTok that a nationwide poll involving 7,500 respondents across 83 constituencies would be conducted through face-to-face interviews.

However, Mr Poku expressed skepticism about the timeline.

“Four days later, on March 21, he comes back and says he is bringing findings from people he has met on the streets. You are going to talk to 7,500 people across 83 constituencies, and you have done that in four days? He must have over 600 people working day and night,” he argued.

Breaking down the figures, Mr Poku said that applying a 30 percent sampling benchmark to regions such as Ashanti would require coverage of at least 14 constituencies, involving approximately 1,260 respondents.

“You cannot stand in Kumasi and interview people and say that represents the entire Ashanti Region. You have to go to the length and breadth of the region to get a proper sample,” he noted.

Mr Poku maintained that it would be highly improbable to conduct such extensive face-to-face interviews within the stated timeframe, particularly across multiple constituencies.

“In a nutshell, Mussa Dankwa could not have interviewed 1,260 people in 14 constituencies in Ashanti Region in four days to arrive at that conclusion,” he stressed.

Kwadwo Poku stated that he has consistently analysed the work of pollster Mussa Dankwa within its proper context, noting that where the exercise involves calling people across the country to make predictions through opinion polls, the outcomes may or may not be accurate.

He raised concerns about the interpretation of data, arguing that party affiliation does not necessarily reflect voting intention.

“One of my biggest problems is that party affiliation is not the same as who people will vote for. How many people in Ghana will you meet on the streets, and they will truly tell you their party affiliation?” he questioned.

Mr Poku called for greater scrutiny of polling data and urged analysts and the public to interrogate such findings more critically.

“The polls have to have a bit more bite. Let’s interrogate the data. When you draw certain conclusions, some of us will question them. I do not believe he was able to achieve that sample size within that time,” he said.