Studying the values of hard-to-reach populations: content analysis of tweets by the 21st century homeless

  • Authors:
  • Jes A. Koepfler;Kenneth R. Fleischmann

  • Affiliations:
  • University of Maryland, South Wing, College Park, MD;University of Maryland, South Wing, College Park, MD

  • Venue:
  • Proceedings of the 2012 iConference
  • Year:
  • 2012

Quantified Score

Hi-index 0.00

Visualization

Abstract

This paper describes a content analysis of a corpus of 5,313 tweets from 32 individuals collected during a three-week period in March/April 2011. The corpus comprised two study groups: Group H -- Twitter users who self-identified as homeless or formerly homeless in their Twitter profiles, and Group NH -- a random, stratified sample of Twitter users who did not self-identify as homeless and who shared similar Twitter characteristics with those in Group H. The study uses the Meta-Inventory of Human Values for Informal Communication (MIHV-IC) to study value expression in tweets. Two rounds of inter-coder reliability testing demonstrated the challenges of reliably detecting human values in tweets. Analysis of categories with substantial inter-coder agreement indicated significant differences between the two groups for helpfulness and wealth. This approach provides a promising opportunity for detecting the values of hard-to-reach populations such as the 21st century homeless.