Professor Smart Sarpong defends his poll forecasting Dr. Mahamudu Bawumia as the favored presidential candidate with 38.9 percent in the upcoming 2024 general elections.
As the Director of Research and Innovation at Kumasi Technical University, Professor Sarpong asserts that his survey indicates the New Patriotic Party’s (NPP) flagbearer leads John Dramani Mahama and the National Democratic Congress (NDC) in Ahafo, Ashanti, Bono, Central, Eastern, North East, Western, and Western North regions.
However, the survey has faced criticism, particularly from members of the NDC, who question its validity due to Prof. Sarpong’s perceived bias towards the NPP and Dr. Bawumia.
In response, Prof. Sarpong defended his credibility and track record for conducting impartial polls, citing his accurate predictions of the 2020 general elections.
“In the report that I have shared, I have included links to about two of my previous polls. The first one is the election 2020 and the other is the regional election in the Ashanti Region. In 2020, I presented my elections research findings and in there, the NPP was supposed to have a first-round victory with 51.4% and that was a major one.
“One characteristic about the findings that I put out there is that I meet the press and give them enough time to seek clarifications because they are the source of determination.
“On four major occasions, I have met the press on varying research on national interests, and I have heard people talk about track records, but there are fine brains in the universities who can get involved in election research and can help shape discussions, but sometimes our primary work is to do the university teaching and research collation. But on occasions where I get the time, I can do these things and I have always allowed the press to interrogate my findings and also seek clarifications.”
Prof. Sarpong argued that, his survey met all the needed sampling methods to make a poll statistically accurate and credible.
“There are requirements for selecting samples, and in our instance, we were supposed to do a minimum of 15,500 because the population of registered voters in Ghana as of 2020 was about 17 million and about 13.4 million came out to vote so there is that range between 13.4 million and the 17 million. Anybody who seeks to do a sample will have to do a minimum sample size of about 15,500 and that is what we used in our case.
“Even in selecting that sample, you need to make sure that sub-samples are represented, and that is very key to sampling. The concept of representation. I agree that you don’t need to do 275 constituencies to demonstrate representation, but the law of statistical regularity and the law of central limit theorem suggest that the more sub-samples you include, the more accurate your work.
“Any study that reaches all the characteristics of the population is statistically stable and statistically reliable than those that are unable.”
ALSO READ: