Tweeting About Mental Health

Big Data Text Analysis of Twitter for Public Policy

by Mikhail Zaydman

Download eBook for Free

FormatFile SizeNotes
PDF file 1.4 MB

Use Adobe Acrobat Reader version 10 or higher for the best experience.

This dissertation examines conversations and attitudes about mental health in Twitter discourse. The research uses big data collection, machine learning classification, and social network analysis to answer the following questions 1) what mental health topics do people discuss on Twitter? 2) Have patterns of conversation changed over time? Have public messaging campaigns been able to change the conversation? 3) Does Twitter data provide insights that match the results obtained from survey and experimental data? This dissertation finds that Twitter covers a wide range of topics, largely in line with the impact that these conditions have on the population. There is evidence that stigma about mental illness and the appropriation of mental health language is declining in Twitter discourse. Additionally the conversation is heterogeneous across various self-forming communities. Finally, I find that public messaging campaigns are small in scale and difficult to evaluate. The findings suggest that policy makers have a broad audience on Twitter, that there are communities engaged with specific topics, and that more campaign activity on Twitter may generate greater awareness and engagement from populations of interest. Ultimately, Twitter data appears to be an effective tool for analysis of mental health attitudes and can be a replacement or a complement for the traditional survey methods depending on the specifics of the research question.

Table of Contents

  • Chapter One


  • Chapter Two

    Background and Motivation

  • Chapter Three


  • Chapter Four

    Characterizing the Mental Health Conversation

  • Chapter Five

    Changes over Time

  • Chapter Six

    Comparing Dissertation Findings with the Literature

  • Chapter Seven

    Conclusions, Policy Implications, Limitations, and Future Work

Research conducted by

This document was submitted as a dissertation in January 2017 in partial fulfillment of the requirements of the doctoral degree in public policy analysis at the Pardee RAND Graduate School. The faculty committee that supervised and approved the dissertation consisted of Douglas Yeung (Chair), Luke Matthews, and Joie Acosta.

This report is part of the RAND Corporation Dissertation series. Pardee RAND dissertations are produced by graduate fellows of the Pardee RAND Graduate School, the world's leading producer of Ph.D.'s in policy analysis. The dissertations are supervised, reviewed, and approved by a Pardee RAND faculty committee overseeing each dissertation.

Permission is given to duplicate this electronic document for personal use only, as long as it is unaltered and complete. Copies may not be duplicated for commercial purposes. Unauthorized posting of RAND PDFs to a non-RAND Web site is prohibited. RAND PDFs are protected under copyright law. For information on reprint and linking permissions, please visit the RAND Permissions page.

The RAND Corporation is a nonprofit institution that helps improve policy and decisionmaking through research and analysis. RAND's publications do not necessarily reflect the opinions of its research clients and sponsors.