Social media researchers have been warned against using unreliable data from social media platforms to represent definite human behavior in studies. Social media scientists from McGill University in Montreal and Carnegie Mellon University in Pittsburgh caution online researchers to be careful when citing studies derived from social media analytics, because social media firms have a way of presenting untrue information to the public.
It is noted that social media platforms cannot be relied upon to present very factual information to the public, and in fact, the statistics they present to the public are biased; and cannot be solely relied upon as representative of human behaviors and trends in any given way. Any major research study that relies on publicly displayed statistics from social media is bound to flawed and unreliable, because it is based on flawed and unreliable facts.
For instance, Pinterest is dominated by female users aged 25-34 years, a researcher that is not mindful of this fact may base his studies on general factors derived from the site without bias to the fact that it is dominated by 25-34-year old females.
Also, Facebook has a “Like” button but does not provide a “Dislike” button, and while it is easy to measure the number of approvals for piece of media item, it will be very impossible to measure disapprovals and negative responses that people have toward a piece of media item. This essentially means that the way social media platforms are structured to function affect the way people use them or behave in relation to items featured on such platforms.
Another thing is that spammers and bots behave like real humans when crawling social media sites and this gives a false impression of human activity. Researchers should therefore by mindful of the activities of these bots and spammers when utilizing social media information.
Derek Ruths of the McGill’s School of Computer Science advises researchers to classify social media information according to users, topics, events, posts, movements, activity, and other factors to be able to get a real and accurate picture of what they desire to derive for their research studies as representative of human behavior and trends.