Researchers attempt to predict depression from Facebook posts


In October 2018, researcher Johannes Eichstaedt led a project to study how the words people use on social media reflect their underlying psychological state. Working with 1,200 patients at a Philadelphia emergency department, 114 of whom had a depression diagnosis, Eichstaedt's group studied their EMRs and up to seven years of their Facebook posts. Matching every person with a depressive diagnosis with five who did not, to mimic the distribution of depression in the population at large, from the final pool of 684 patients the researchers identified the most frequently used words and phrases and developed and algorithm to flag depression-associated language markers such as greater use of first-person singular pronouns and words reflecting hostility and loneliness. Using these the researchers were able to predict a diagnosis of depression up to three months before it would appear in their medical records.
writer: Robbie Gonzalez
publication: Wired

Related learning resources
Target Profile