Latest News

latest

Social Media Data Not As Foolproof As Once Thought?

Writing's On The Wall

Since the inception (and subsequent boom) of social media, its ability to aggregate and forecast public opinion has become greater and greater. Instead of being the also-ran to traditional polling and public surveys it once was, social media has risen through the ranks to become, pretty much, the go-to tool in figuring out whatever the hell it is that the public love, hate or find moderately humorous at any given moment.

Facebook, Twitter, Instagram and their less established chums are now used by sociologists as an easy way to discover how we all tick. They're so simple to use that the process of data farming has boiled down to using third party software to automatically see how many hundreds of thousands of people have said what, and what they've tended to say it about. This means that corporations and scientists alike now use social media as their first port of call when working out their science-y and market-y stuff.

social media studies
source: sheknows.com

Social media, in other words, is awesome. But beware: the data we so often see quoted in scientific journals, newspapers and smarmy tech blogs isn't as infallible as your friendly social media guru would have you believe. In fact, Derek Ruths - a professor at McGill University and an active researcher of social media - has suggested that a large chunk of the data he is employed to collect is inherently flawed, and the entire system is in need of across-the-board refinement. In a column in the popular and creatively named scientific journal Science, he detailed many potential reasons for this, most of which make perfect sense when you take a step back and actually think about it.

The first of many problems is that our online posts tend to be far more exaggerated than what we say in real life. If you're unsure what I mean by this, consider how many times you've actually been audibly guffawing and straining your eyes through tears of laughter when typing out the acronym 'LOL'?

Ok, fine, so let's just disregard acronyms when accruing data. But consider the recently prominent debates about the extent that our online comments can get us in trouble with the law (if you have an aversion to discussions of rape threats, do not click this link). The entire case rests on the notion that the comments a Facebook user made on his wall were horrifically exaggerated for comic/shock effect. In other words, there was an arguably massive gap between the writing on his wall and the thoughts that were in his mind.

Many studies don't take this into account.

There's more. Some fail to consider the respective demographics of different social media platforms. Trawling through Instagram will bring you the voice of the younger generation, while Facebook's users are becoming increasingly middle-aged. Google+ is mainly inhibited by no-one, and Pinterest is populated by an army of wealthy women. It's why I love using it to pick up chicks.

social media studies
source: lucidcrew.com

Like political parties, different social networks vary by demographic, status and social outlook. If a study fails to take this disparity into account, it is irrecoverably flawed from the beginning and its results are tainted almost beyond use.

Another problem is that the source of some media research is strangely ambiguous (who conducted the analysis, their motivations, where they looked, etc). This raises the question of whether corporate interest is at the centre of even the most seemingly innocent of data announcements, which makes it difficult to completely confirm that certain social media based studies are entirely without corporate or personal bias.

The way in which these factors influence studies is so subtle that it is difficult for peer reviewers to even spot irregularities when they do occur. According to Professor Ruths, though, they're definitely an issue.

“Mounting evidence suggests that many of the forecasts and analyses being produced misrepresent the real world.
“These papers are no longer just circulating among academics...[perhaps] we’re hurting the field by trying to move it forward too quickly.”

Another key problem mentioned is that the platforms themselves can skew data in unpredictable ways. The National Post, which provided the quotes for this article in a thorough and interesting article, gave the example of Google using the auto-corrected versions of Google searches in its data instead of what we actually type, subtly skewing their statistics. Other social media platforms have different quirks, and unless we learn to take each of these into account, data garnered through social media will always be at least somewhat tainted.

Researchers seem to be ignoring these issues, which is turning Professor Ruths' warning that we are killing the field by trying to advance it before we are ready into a self-fulfilling prophecy.

source: dodho.com

Neither this research nor I are suggesting that social media trends should be completely ignored from this point onwards. In fact, I've created many articles which were based on research conducted through social media, and I stand by each of them. Right now, though, social media is a relatively new public speaking platform that the guys who collect data haven’t yet completely learnt to utilise.

In the words of Professor Ruths, there is “incontrovertible evidence that there is a strong social signature in social media data that we can use to gain insight. But there’s a big difference between saying it’s in there, and we are able to get it out.”

In other words, it is without doubt that social media has the potential to be an incredibly accurate and powerful polling tool. Techniques for farming its data are always improving, and pollers' tactics are getting more refined by the day. For now, though, we should take the social media studies we do see with a whole nugget of salt.

To learn about this issue in more detail, check out the original source here. It’s an interesting, eye-opening read, and well worth ten minutes of your time.


Emile is a postgrad from the University of Saint Mark and Saint John. He’s hoping to break into journalism or publishing, and won’t stop blogging until he’s managed it! Follow him @EmileAtSMF.

Contact us on Twitter, on Facebook, or leave your comments below. To find out about social media training or management why not take a look at our website for more info http://socialmediacambridge.co.uk/.
Social Media Data Not As Foolproof As Once Thought? Reviewed by Emile Cole on Tuesday, December 02, 2014 Rating: 5
All Rights Reserved by Social Songbird © 2010 - 2016
Powered By Blogger

Contact Form

Name

Email *

Message *

Powered by Blogger.