Elections: Quantitative research method may result in bias – researcher

As South Africa heads for watershed elections this year, a leading research methodologist is calling for an end to the dichotomy between qualitative and quantitative research, arguing that opinion polls could lack credibility and be biased.

In an interview on the sidelines of the three-day World Conference on Qualitative Research in Johannesburg, Prof King Costa, managing director of the Global Centre for Academic Research, said credibility, trustworthiness and believability should be at the core of any study.

ALSO READ: ANC implements Motlanthe report, faces potential by-election losses

“As we head towards elections, people should bear in mind that opinion polls are predicated within the quantitative dimension – about numbers – while qualitative research brings out people’s insights and behaviour.

“Credibility is part of the rigour in qualitative research. A qualitative study has to have 100%, or close to that level, of believability, trustworthiness, credibility. There has to be peer debriefing.

“Qualitative research is neither subjective nor biased, with conventions used to validate outcomes,” said Costa.

Maintaining that bias and propaganda were possible in quantitative research, Costa said there were “many variables that could inform the outcome, with bias being number one”.

He said: “Most opinion polls use quantitative methods – theoretically or software-based, calibrating data and outcomes. “If the opinion poll instrument has not been tested for reliability and consistency – like an online questionnaire sent out to people – it is likely to give problems.

ALSO READ: Joseph Boakai sworn in as Liberian president

“Relying on people clicking when giving responses in online research may not give a true reflection of people’s opinions.

“That clicking can be done by people who probably do not have time to thoroughly go through questions on a five-point scale.

“A follow-up verification is crucial – failing which there is a risk of propaganda in these things.”

Calling for a dual approach in carrying out a study, Costa, who is also chair of the World Conference on Qualitative Research local organising committee, he said: “If you have a dual approach to research, with an opinion poll also being triangulated with qualitative interviews, then there would be a strong validation of results – making the study believable.

“On its own, it is difficult to rely on quantitative research.

“Predictive analysis with historical data is key in looking at the future outlook. While the dichotomy is historical, there is need to marry the two approaches in research. They should be together to produce credible research.”

ALSO READ: Mandla Msibi’s future hangs in the balance as disciplinary process gets going

Underscoring the importance of qualitative research was “very crucial in solving complex societal issues”.

“If you are a policymaker or practitioner – having to develop policy in healthcare, education or other related aspects – it could be difficult because human beings are complex.

“South Africa – like elsewhere in the world – needs research or the science to debunk or decode – bringing down walls by making us understand ourselves and allowing for multiple perspectives.

“It is important to find means of solving the divide by giving answers to complex problems. “We needed some form of qualitative dimension to understand human behaviour.

“Qualitative is not an opponent of quantitative research. Qualitative gives us an in-depth analysis as opposed to quantitative which generalises in terms of how many people share a perspective.”

The World Conference on Qualitative Research has delegates from Zambia, Angola, the Philippines, US, Swaziland and Lesotho, Brazil, UK and Portugal.