Back To Career News

Can Social Media Status Updates Predict Unemployment Spikes?


Can your social media status updates predict spikes in the unemployment rate? They can, according to new research by SAS and United Nations Global Pulse.

The researchers first tracked and analyzed hundreds of thousands of Web-surfing citizens in the U.S. and Ireland between June 2009 and June 2011 to quantify the mood in their updates on blogs, news sites, forums and social networks. After comparing this data to actual unemployment figures, they discovered that just before a rise in unemployment rates, users increasingly talked about tightening their budgets; as the rise in unemployment continued, they discussed topics like foreclosures, evictions and canceled vacations.

Researchers classified updates based on conversation topics, such as "Definitely won't be buying new clothes this season," and sentiment, such as "I've felt so scattered since the layoffs." When such status updates occurred in the context of unemployment or a similar job-related term, they were logged for analysis.

"Social media and Internet content is like the letters and phone calls that have always informed organizations," said I-sah Hsieh, the global manager of International Development at SAS, to MainStreet. "Only now, it’s digital, public and massive in scale. This untapped treasure can provide real-time feedback on policies, improve public safety, enhance citizen relations and support important sociological research."

Do You Know What You're Worth?

You can read the SAS white paper here. Did you notice a change in your network's status updates during the recession?

More From Payscale

Employee Resignation Up in 2012: Should Employers Be Worried?
No Bloody Monday Job Loss, But Minority Unemployment Filing Steady
Should a 21-Hour Workweek Be the New US Standard?


(Photo credit: Robert S. Donovan/Flickr)

Marissa Brassfield
Read more from Marissa

Leave a Reply

Notify of
What Am I Worth?

What your skills are worth in the job market is constantly changing.