Politics of opinion polls
On April 28, 2014 The Nation carried a headline which read: “Poll Says JB to Win Vote”. The news item, as expected, generated mixed reactions. The ruling party exclaimed that the poll had just rubberstamped what they knew all along while the opposition Democratic Progressive Party (DPP) and Malawi Congress Party (MCP) branded it as hogwash.
A definition of opinion poll at this juncture would be relevant. Wikipedia describes an opinion poll as an inquiry into public opinion conducted by interviewing a random sample of people. As per this definition, it is pretty enlightening to know that opinion polls are not always about elections; they can be on myriad range of topics that are in the public domain.
Getting back to our issue: why is it that the likes of associate professor Edge Kanyongolo and professor Chijere Chirwa, both from Chancellor College, and Augustine Magolowondo of Netherlands Institute for Multiparty Democracy, said they would love to know of the methodology that was used before inputting their comment?
Well, matters of research findings are delicate that is why the issue of methodology is so important. To understand this issue of methodology better, consider what doctors do when you go to the hospital. Any learned doctor will never give his/her patient a prescription before diagnosing them of what they are suffering from. They have various instruments they use to arrive at a conclusion on whether you have malaria, tuberculosis or pneumonia. There is a particular methodology they utilise and so is the researcher.
Researchers employ specific methodologies to come up with credible and generalised findings. One step towards the achievement of this is the use of random (probability) sampling. Malawi Electoral Commission (MEC) reported that 7.3 million Malawians registered to vote. However, the researcher says they only sought the opinion of 1790 people to come up with the findings. I have to state here that there is nothing wrong in what they did as you would agree with me that it was impractical and impossible for him to seek the opinion of each and every registered voter. Nonetheless our concern should be whether he used random sampling.
It is possible that he went to Domasi in Zomba or Kapoloma in Machinga and sought the views of 1790 people on their preferred candidate. This too is research but it is unscientific. This is where random sampling comes in handy. A researcher has to accord each and every respondent in his population of interest the chance to participate in the research. In this case our concern then should be whether this researcher accorded the 7 million plus voters an equal chance of participating in this research. How did he select the participants? Did he use the entire voters roll from MEC and conducted a draw. Or he selected every 200th individual from the voters roll? Did he select participants from the rural only or just the urban only? How did he take care of the issue that the North registered fewer voters compared to the Centre and South? This is the fundamental question the research gurus who were queried to input their comment on the findings raised.
When research is conducted properly, it provides important insights and helps to provide solutions to lingering issues. However, research is also a subject filled with delicate nuances that can easily lead to misguided conclusions. Research is so an important field in that it offers a way to infer meaning and insights from a small subset of data, which can be applied to a much broader context.
History tells us that The Chicago Tribune in 1948 conducted a research in the run up to the presidential elections and according to its findings Thomas Dewey would trounce Harry Truman to pieces. To the surprise of The Chicago Tribune, Truman triumphed. All this happened because the newspaper botched up its sampling methodology and consequently came up with false findings.
Research findings on election outcomes are a sensitive topic. This is why the opposition has denied them. Party enthusiasts have even questioned the credibility of the research institution and notable research gurus have questioned the sampling methodology that was used. I personally do not know the motives of the research institution. There are questions that need answers as regards who funded the research and whether they did it as part of their social responsibility and for what purpose. However, it should concern the opposition parties more if a representative sample was used to come with the findings.
The author is a Chancellor College graduate of economics currently pursuing an MA in South Korea
