Social Intelligence Stories that Caught Our Attention: Volume Sixteen

Here’s the social intelligence stories that caught our attention across the web the week of the 19^th of August.  This week we have a data scientist sceptical over data, racial bias in hate speech, Glassdoor data proving an important point, forcing people to read articles before they share, and how hate speech travels across social media.

A Data Scientist Who is Sceptical About data

Andrea Jones-Rooy, Professor of data science at NYU pens a great article about her scepticism of data. She warns:

“But in the frenzy, we’ve conflated data with truth. And this has dangerous implication for our ability to understand, explain, and improve the things we care about.”

Read the full article here.

Racial Bias Observed in Hate Speech Detection Algorithm from Google

This is an interesting article from Davin Coldeway at TechCruch.  I’m not the most trusting person of AI when it comes to social data analysis – it doesn’t get context properly and this research goes to prove the point (and is a little dangerous).

Researchers at the University of Washington believed that databases of hate speech currently available might have racial biases baked in.  When exploring Tweets that had been identified as being “hateful”, “offensive”, “abusive” etc, they analysed the language to find if the tweets were strongly associated with African American English or white-aligned English.

The result?  African American English is more likely to be labelled offensive.

Read the full article here.

Glassdoor Research Shows Employee and Customer Satisfaction Related

Do you use Glassdoor data in your research? This new study finds that employee and customer satisfaction is related.   Looking at a panel of 293 large employers across 13 industries the study found a strong statistical link between employee well-being reported on Glassdoor and customer satisfaction among a large sample of some of the largest companies today.

Read the full article here.

What if Social Media Companies Forced Us to Read Articles Before Sharing Them?

Can you remember the study that highlighted more than 60% of the links shared on social media were shared without the sharer ever reading the content? That’s huge. It’s misleading and plays right into the hands of vanity metrics – if the article is not read, where’s the value?

In this Forbes article, Kalev Leetaru questions whether we should be forced to read articles before sharing them.

Read the full article here.

How Hate Travels Across Social Media

A new study by researchers from George Washington University and the University of Miami mapped the interconnected network of hate across Facebook and VKontakte and regions in a new study published in Nature.

The study found that hate clusters often regenerate and spread across platforms and around the world, even when they are banned.  The research suggests platforms should remove smaller hate groups to weaken the larger ones.

Read more about the study here.

‍