Brunel University Research Archive(BURA) preserves and enables easy and open access to all
types of digital content. It showcases Brunel's research outputs.
Research contained within BURA is open access, although some publications may be subject
to publisher imposed embargoes. All awarded PhD theses are also archived on BURA.
Please use this identifier to cite or link to this item:
http://bura.brunel.ac.uk/handle/2438/29519| Title: | The role of economic news in predicting suicides |
| Authors: | Moscone, F Tosetti, E Vittadini, G |
| Keywords: | suicide;health outcomes;text analysis;emotions extraction;forecasting |
| Issue Date: | 7-Aug-2024 |
| Publisher: | Elsevier |
| Citation: | Moscone, F., Tosetti, E. and Vittadini, G. (2024) 'The role of economic news in predicting suicides', Economics and Human Biology, 0 (in press, pre-proof), 101413, pp. 1 - 19. doi: 10.1016/j.ehb.2024.101413. |
| Abstract: | In this paper we explore the role of media and language used to comment on economic news in nowcasting and forecasting suicides in England and Wales. This is an interesting question, given the large delay in the release of official statistics on suicides. We use a large data set of over 200,000 news articles published in six major UK newspapers from 2001 to 2015 and carry sentiment analysis of the language used to comment on economic news. We extract daily indicators measuring a set of negative emotions that are often associated with poor mental health and use them to explain and forecast national daily suicide figures. We find that highly negative comments on the economic situation in newspaper articles are predictors of higher suicide numbers, especially when using words conveying stronger emotions of fear and despair. Our results suggest that media language carrying very strong, negative feelings is an early signal of a deterioration in a population’s mental health. |
| Description: | JEL classification: I14; I15. Data availability: The authors do not have permission to share data. Appendix. Brief overview of the WordNet-Affect: A synset is a group of data elements that are considered semantically equivalent for the purposes of information retrieval. WordNet-Affect is an extension of WordNet Domains (see Magnini and Cavaglià, 2000), that includes a set of synsets suitable to represent affective concepts representing moods, situations eliciting emotions, or emotional responses. The authors specifically initially identified a set of words that directly refer to emotional states (e.g. fear, cheerful, sad). Then, they expanded this initial set by implementing an unsupervised algorithm that exploited a mechanism of semantic similarity to automatically acquire from a large corpus of texts (100 millions of words). The final data set includes 1641 terms characterising 28 different emotions. Further information on the approach are available at the web link https://wndomains.fbk.eu/wnaffect.html . |
| URI: | https://bura.brunel.ac.uk/handle/2438/29519 |
| DOI: | https://doi.org/10.1016/j.ehb.2024.101413 |
| ISSN: | 1570-677X |
| Other Identifiers: | ORCiD: Francesco Moscone https://orcid.org/0000-0001-5378-680X ORCiD: Elisa Tosetti https://orcid.org/0000-0003-0979-2828 101413 |
| Appears in Collections: | Dept of Economics and Finance Research Papers |
Files in This Item:
| File | Description | Size | Format | |
|---|---|---|---|---|
| FullText.pdf | Crown Copyright / The Author(s) © 2024. Published by Elsevier B.V. This is an open access article under the CC BY license (https://creativecommons.org/licenses/by/4.0/). | 583.16 kB | Adobe PDF | View/Open |
This item is licensed under a Creative Commons License
