"Interactions between humans and machines in text-as-data approaches: Challenges of reliability, validity and interpretation"

12.04.2019 12:30 - 13:30

Hajo Boomgaarden, Professor of Methods in the Social Sciences at the Department of Communication

Automated text analysis has become a common tool across the social sciences. However, the different techniques being applied – of course – cannot operate without human intervention. Users of such techniques either provide information on the input side or interfere when it comes to the interpretation of results. Both types of interventions are potentially problematic, given that researchers have to juggle the validity of computer-generated data and the reliability and validity of human judgments in the different processes. The presentation draws on four studies to show the challenge of interactions between humans and machines in such text-as-data approaches.

Hajo Boomgaarden

Data Science @ Uni Vienna

SR Geschichte 1

Main Building, 1st floor, Universitätsring 1, 1010 Wien