Humans are very complex creatures. That's why we still remain the best types to study by scientists and other researchers. So far, though, many of these professionals have looked into social media to get the details they need.
Before, research especially in social or behavioral science worked this way: they identified their target respondents, generate the questions appropriate to their audience, obtain the ideal sample size that would represent the general population, and perform the study, which could include asking each of the sample size questions.
However, with the advent of social media such as Facebook, Twitter, and Instagram, researchers can get the data they think they need for the study quickly and cheaply as most of them come for free.
But then are these data reliable enough? Researchers from McGill University and Carnegie Mellon University don't think so. As much as those that use social media for study think the sample is representative in the general population, it doesn't. In other words, data from social media can be misleading and bias.
An article about it is currently published in Science on Friday, November 28.
This issue is not something that will be taken lightly since thousands of research papers now mentioned obtaining "big data" through these social media accounts.
The use of social media is understandable. As Juergen Pfeffer, one of the lead authors, pointed out, at the height of the Boston Marathon tragedy, he was able to obtain more than a million related tweets that show different types of human behavior in a short amount of time and without spending any money.
Social media makes people more assertive and opinionated as well, and it's highly possible to get some type of reaction for any issue, which ranges from the most mundane to the most controversial that may involve politics, war, and religion.
Moreover, because they can gather a lot of data, those who use social media for research think they can already reduce or eliminate the bias, but they don't since the demographics, for instance, of these websites already lean to certain types of people. The supposed respondents may not be the sample audience you're looking for, which already damages the credibility of the study results.
