Mental illness and bipolar disorder on Twitter: implications for stigma and social support

Alexandra Budenz, Ann Klassen, Jonathan Purtle, Elad Yom Tov, Michael Yudell, Philip Massey

Research output: Contribution to journalArticlepeer-review


Background: Mental illness (MI), and particularly, bipolar disorder (BD), are highly stigmatized. However, it is unknown if this stigma is also represented on social media. Aims: Characterize Twitter-based stigma and social support messaging (“tweets”) about mental health/illness (MH)/MI and BD and determine which tweets garnered retweets. Methods: We collected tweets about MH/MI and BD during a three-month period and analyzed tweets from dates with the most tweets (“spikes”), an indicator of topic interest. A sample was manually content analyzed, and the remainder were classified using machine learning (logistic regression) by topic, stigma, and social support messaging. We compared stigma and support toward MH/MI versus BD and used logistic regression to quantify tweet features associated with retweets, to assess tweet reach. Results: Of the 1,270,902 tweets analyzed, 94.7% discussed MH/MI and 5.3% discussed BD. Spikes coincided with a celebrity’s death and a MH awareness campaign. Although the sample contained more support than stigma messaging, BD tweets contained more stigma and less support than MH/MI tweets. However, stigma messaging was infrequently retweeted, and users often retweeted personal MH experiences. Conclusions: These findings demonstrate opportunities for social media advocacy to reduce stigma and increase displays of social support towards people living with BD.

Original languageEnglish (US)
Pages (from-to)191-199
Number of pages9
JournalJournal of Mental Health
Issue number2
StatePublished - Mar 3 2020


  • bipolar disorder
  • Mental health
  • social media
  • Twitter

ASJC Scopus subject areas

  • Psychiatry and Mental health


Dive into the research topics of 'Mental illness and bipolar disorder on Twitter: implications for stigma and social support'. Together they form a unique fingerprint.

Cite this