BLANTYRE-(MaraviPost)-Statistician Tsilizani Mwalimu Kaombe has commended the use of graphics in Shyley Kondowe’s recent report on voter preferences but has raised concerns over the survey’s methodology.
In an interview with Maravi Post,Kaombe who is also a senior lecturer in statistics at the University of Malawi (UNIMA) said the graphical presentation made the findings accessible to a wide audience, but he questioned the scientific rigor behind the sampling process.
Said Kaombe, “While the sample size is provided in the introduction, there is little clarity on how it was determined.
“The report cites the researcher’s ability to reach the required sample size in each district as justification, but this approach not only invites potential bias in estimates, it also lacks a basis for scientific replication.”
He further noted that the report does not indicate how the sample was distributed across rural and urban populations, gender groups, or regions key factors in ensuring representativeness.
“The absence of such details raises questions about the accuracy of the estimates and may introduce additional sources of error,” he said.
Kaombe also highlighted that the methodology failed to mention the level of probability of Type I Error applied to the estimates, which he said is a standard requirement in statistical reporting.
“These issues are crucial in giving both investigators and the public confidence in the results. Without addressing them, such a survey risks being regarded as anecdotal evidence,” he said.
The new survey report indicates that President Lazarus Chakwera of the ruling Malawi Congress Party (MCP ) will win the next Tuesday’s presidential election with 55 percent against his close competitor Arthur Peter Mutharika of the former ruling Democratic Progressive Party (DPP) who is placed at 34.8 percent.
