ASMR
Automated analysis of media suicide reporting
Source: This example was kindly contributed by Markus Schäfer, Communication Science Research Group, Johannes Gutenberg-Universität, Mainz, Germany
The project described here aims to develop and test automated methods for identifying, analysing and predicting risk factors of media suicide reporting. The project is conducted in the context of the Centre for the Digital Foundation of Research in the Humanities, Social, and Educational Sciences (CEDIFOR).
Effective measures to prevent suicide are of enormous importance for health policy in most societies around the world. In Germany, more than 10.000 people die by suicide every year. The global number of suicides is estimated at more than 800.000 every year. The goal of responsible media coverage of suicides plays a crucial role in national and international prevention concepts. The way in which mass media report on suicides can have an impact on the development of the number actual suicides. The existence of the so-called “Werther effect” - an increase in actual suicide after a certain type of media reporting on suicides - is internationally well confirmed and there are several recommendations on how the media should (not) report on suicide to minimize the risk of copycat behaviour. Unfortunately, very little is known about how suicide actually is reported, also because manual analysis of coverage is quite complex and expensive. In the long term, the project aims to provide suicide prevention with a useful automated tool that allows continuous monitoring of media suicide content, facilitating fast reactions of health organisations to (an accumulation of) problematic media content.
The project works with a large corpus of suicide-related newspaper data in German language. INCEpTION is used as the annotation environment for the first explorative annotation campaign of the project. We use it for annotating suicide-related information in our data, such as presentation of (certain) suicide methods, framing of suicide and suicidal persons.
Features of INCEpTION that we consider of particular importance to our use-case are:
- linking with KB: helpful e.g. to identify and analyse celebrity suicide reporting
- active learning, will be used to speed up the annotation process
- monitoring of annotators’ progress
- monitoring of inter-annotator agreement
References
- Markus Schäfer & Oliver Quiring (2015) The Press Coverage of Celebrity Suicide and the Development of Suicide Frequencies in Germany, Health Communication, 30:11, 1149-1158, DOI: 10.1080/10410236.2014.923273 [Publisher]
