The majority of young blacks believe the Democratic Alliance would bring back apartheid if it won the next election. A substantial 91% of South Africans are highly satisfied with online shopping. At least 42% of young South Africans believe foreigners are bad for SA.
Social media and online survey results have made for some eye-catching headlines over the past few months, but commentators say this cannot be taken as a depiction of the overall population.
According to director of the Wits journalism programme, Anton Harber, social media only addresses a narrow sector of the population, so it can be used as a dipstick of that sector and not as a reflection of a whole country's view.
"Research via social media cannot claim to represent anyone except those who use that particular social media. It would be like doing interviews in Hyde Park Centre and then claiming it represents the whole country," says Harber.
He notes that people can often get taken in with controversial results of social media surveys, which leads to results being overstated. "Journalists need to take a close and sceptical look at the data behind the headlines, so as not to be taken in by those who hype up the results for self-serving purposes."
World Wide Worx MD Arthur Goldstuck concurs that an online survey is reliable to the extent that it is representative of the audience being reflected. "If we do an online survey of a community of 2 000 people, eg, a school, and 500 people whose demographics we know precisely participate, that is highly reliable. If we do an online survey that is open to the whole population and 500 random people choose to fill it in, that is a highly unreliable measure of the population," says Goldstuck.
Honest response?
Mike Sharman, owner of digital communications agency Retroviral, points out research done via social media often offers researchers the opportunity to see people's view on certain subjects, without them being aware they are being surveyed. "I think sometimes with traditional research, the interviewee might feel like the interviewer is expecting a certain response from them, making them less honest. When you use Twitter, for example, you can search a certain topic and see what people really think without them knowing their opinions are being used for research," he explains.
Fears that online survey results could be unreliable because people can use a fake identity when taking part is not only limited to online research, says Sharman. He notes it is possible to provide fake personal details when taking part in a one-on-one survey or other traditional methods.
Jodene Shaer, social media publicist and owner at Lifeology, points out that the anonymity of online research can actually make results more reliable, as participants are likely to be more honest if they know their identity is being protected.
"One has to remember, however, that every country has its own online psyche," she says. "SA's overall energy online is much more reserved than the international standard. We tend to be less outspoken than others."
Harber feels anonymity is one of the many limitations on social media research. "You are going to get a higher lie, exaggeration and frivolity factor."
Quick turnaround
Research company, Pondering Panda, churns out survey results, having conducted over four million interviews in two years, and claims it typically gets 2 000 responses in a 24-hour period. The company only uses social media platform Mxit as its sample group, but says its survey results are still highly reliable. As an example, the company Web site shows a comparison between the results of a Pondering Panda survey and that of SA's national census, which shows the results are similar.
Harber remains convinced that good and trustworthy research takes time and resources. "If you want research you can rely on because it is done with care, detail and accuracy, then it takes time and money. There is no easy and cheap way to do something solid and trustworthy. Quick, cheap research will get you quick, cheap results. They may be accurate, but you will probably never know if it is."
Goldstuck notes that academic research methods are not a guarantee of sure-fire results. He says while these research methods tend to be highly scientific, they often rely on samples that are too small, or qualitative methodologies that do not reflect a population. "The keys to reliability and validity are sample frame, sample size, and random sampling within that frame and within the context of matching sample to targeted population."
"Online research has definitely changed the [research] industry," says Sharman. "It has given us more tools and platforms [to conduct research] and allows people to be a lot more vocal on their opinion of different things."
Shaer concurs, saying social media allows researchers to monitor public opinion much more and broadens the scope of who can be researched and how. "It has just given a whole new dynamic to the industry."
Social media research definitely holds benefits for the industry, adds Harber. "These communication channels have brought change and of course we must embrace them. We must just do it with caution and not pretend they provide an alternative to the hard slog of good research."
Share