Researchers at the University of Pennsylvania, Purdue University, Stanford University, and the National Institute on Drug Abuse analyzed Facebook posts to differentiate language associated with depression and with loneliness.
The researchers obtained consent to collect 3.4 million posts from nearly 3,000 individuals, then administered depression and loneliness surveys to measure psychological states and feelings of social isolation before linguistically screening the posts for loneliness- and depression-related language.
One approach used a University of Texas database to categorize words according to meaning and grammatical function.
A machine learning-based method extracted common words, phrases, and topics often found in posts by lonely or depressed participants.
The researchers found depression-associated language refers mainly to emotions, while loneliness-associated language focuses more on cognition.
View Full Article
Abstracts Copyright © 2022 SmithBucklin, Washington, DC, USA