One study raised the example of the swine flu outbreak in 2009 and the way panic quickly spread online, causing peculiar behaviors across Asia driven by the misinformation posted online. They suggest that the degree of connectivity within a network is a strong indicator of how rapidly misinformation may spread.
The accuracy of Twitter
In a similar vein, a recent study from researchers at Georgia Institute of Technology explored the accuracy of information posted on Twitter, and came to some rather damning conclusions, with roughly 75 percent of all tweets (once spam had been removed) not at all credible.
The researchers spidered around 1 percent of the tweets in the Twitter database, stripped out the spam and then sorted the tweets into categories.
They then recruited a team of volunteers on Mechanical Turk to go through each tweet to rate it for both accuracy and certainty.
Over a 96 day period this amounted to roughly 60 million tweets, with around about 1,000 news stories covered by them. Once analyzed, it emerged that just 1/4 of the tweets were what you might regard as trustworthy.
So even accounting for spam, that’s 45 million tweets that would politely be regarded as junk. What’s more, much of this misinformation was rather serious in nature, with misinformation particularly rife when it came to Ebola.
The authors have compiled their data into an open repository called CREDBANK in the hope that others can build on their work.
With much of social media happening instantly, determining credibility is often a challenge, and previous studies have highlighted how mainstream media are as guilty as anyone.
You might imagine that news organizations fit the bill as such an influencer, and that was the finding of a recent study that cast blame on the media for failing to do their part in ensuring the accuracy of the information they share.
“Rather than acting as a source of accurate information, online media frequently promote misinformation in an attempt to drive traffic and social engagement,” the authors say.
It’s probably not helped by the sheer number of people who will share content that they haven’t read, much less had verified. Whilst there are various attempts to provide such verification, they do nonetheless require the individual to be sufficiently motivated to check.
I suspect that until the social networks have a means by which they can rubber stamp content as credible, we will continue to share junk as it’s much easier to do that than check its authenticity.