Executive Director of Global InfoAnalytics, Mussa Dankwah, has cautioned analysts, communicators, and self-styled political researchers against misinterpreting polling data, following the outcome of New Patriotic Party’s (NPP) 2026 presidential primary.
His comments come in the wake of predictions by researcher Dr. Evans Duah, who, ahead of the primaries, projected that none of the five candidates would cross the 50-plus-one vote threshold required for an outright victory—except former Assin Central MP, Kennedy Ohene Agyapong.
Speaking on Channel One TV’s The Point of View on Wednesday, January 28, Dr. Duah had argued, “A lot of people are missing the strategy and, if you ask me confidently, none of the five candidates will cross that 50+1 on Saturday. And if anyone does, it will be Hon. Kennedy Ohene Agyapong. I stand by this, and it is based on data.”
But when the votes were counted, former Vice President Dr. Mahamudu Bawumia emerged victorious in the January 31 presidential primary, securing 56.48 percent of valid votes and decisively defeating his closest contender, Kennedy Ohene Agyapong, who polled 23.76 percent.
In a Facebook post on Sunday, February 1, Mussa Dankwah analysed the outcome, providing a postmortem of the primary while reinforcing the limits of polling interpretation. He said Bawumia’s performance closely aligned with Global InfoAnalytics’ forecasts, missing their projection by just 0.5 percent—well within the margin of error. In contrast, Agyapong underperformed by 4.5 percent, putting him outside the forecasted range.
“Such actions gave undecided delegates an opportunity to rethink their choices, and Dr Bawumia ultimately benefited from that gaffe,” Dankwah observed, referencing controversial campaign conduct that included an insult-laden video circulating in the final weeks of the race.
He also highlighted the performance of Abetifi MP, Dr. Bryan Acheampong, who outperformed polling predictions by 5.5 percent, particularly among undecided and previously undisclosed delegates.
“The weekly tracking data also shows that DBA’s rise in the favourability rating was phenomenal, increasing by 5.4 in the last four weeks of the campaign compared to DMB, 2.1 and KOA, 0.1,” Dankwah said.
Dankwah used the post-primary analysis to underscore a key lesson: the difference between pollsters and researchers.
“There is a difference between pollsters and researchers,” he said. “Pollsters provide the lead and researchers go in to probe why. Pollsters can only triangulate. Small advice to wannabe party researchers, don’t turn polls into in-depth research. The objectives are not the same. Not everyone has the eyes to see what pollsters can see.
