Can Twitter messaging help corporations mitigate the impact of ethical scandals? We topic-model pre-scandal tweets of 92 ‘offenders’ to investigate
Can Twitter messaging help corporations mitigate the impact of ethical scandals? We topic-model pre-scandal tweets of 92 ‘offenders’ to investigate
Purpose: in this exploratory study, we examine whether Twitter messaging can help mitigate the harm corporations suffer in the aftermath of ethical scandals.
Design/methodology/approach – We apply web Application Programming Interfaces (API) on the Guardian and New York Times news archives to find corporations that suffered scandals between 2014 and 2019, revealing 92 publicly listed companies in the United Kingdom. Using Twitter API and the Python library, Getoldtweets, we extract historical, pre-scandal – i.e. pre-2014 – tweets of the 92 firms. We topic-model the tweets data using Latent Dirichlet Allocation (LDA). We then subject the topics to Multidimensional Scaling (MDS) to examine commonalities among them.
Findings: LDA reveals 10 topics which group under five themes; these are Product Marketing, Urgent Signalling of ‘Greenness’, Customer Relationship Management, Corporate Strategy, and News Feeds. MDS suggests that the topics further congregate into two meta-themes of Future-oriented versus Immediate, and Individual versus Global.
Implications: provided they are sincere and legitimate, corporations’ tweets on global issues with a green agenda should help cushion the impact of ethical scandals. Overall, however, our findings suggest that Twitter messaging could be a double-edged sword, and underscore the importance of strategy.
Originality/value: trhe paper offers a first exploration of the relevance of corporate Twitter messaging in mitigating ethical scandals.
Ethical scandal, Ethical reputation, Reputation continuity, Web Crawling, Twitter Messaging, Topic modelling, Latent Dirichlet allocation; Multidimensional scaling
Raheja, Shivani
15917745-e740-4e6a-b264-a5e3e77b7629
Chipulu, Maxwell
12545803-0d1f-4a37-b2d2-f0d21165205e
Raheja, Shivani
15917745-e740-4e6a-b264-a5e3e77b7629
Chipulu, Maxwell
12545803-0d1f-4a37-b2d2-f0d21165205e
Raheja, Shivani and Chipulu, Maxwell
(2020)
Can Twitter messaging help corporations mitigate the impact of ethical scandals? We topic-model pre-scandal tweets of 92 ‘offenders’ to investigate.
Society and Business Review.
(In Press)
Abstract
Purpose: in this exploratory study, we examine whether Twitter messaging can help mitigate the harm corporations suffer in the aftermath of ethical scandals.
Design/methodology/approach – We apply web Application Programming Interfaces (API) on the Guardian and New York Times news archives to find corporations that suffered scandals between 2014 and 2019, revealing 92 publicly listed companies in the United Kingdom. Using Twitter API and the Python library, Getoldtweets, we extract historical, pre-scandal – i.e. pre-2014 – tweets of the 92 firms. We topic-model the tweets data using Latent Dirichlet Allocation (LDA). We then subject the topics to Multidimensional Scaling (MDS) to examine commonalities among them.
Findings: LDA reveals 10 topics which group under five themes; these are Product Marketing, Urgent Signalling of ‘Greenness’, Customer Relationship Management, Corporate Strategy, and News Feeds. MDS suggests that the topics further congregate into two meta-themes of Future-oriented versus Immediate, and Individual versus Global.
Implications: provided they are sincere and legitimate, corporations’ tweets on global issues with a green agenda should help cushion the impact of ethical scandals. Overall, however, our findings suggest that Twitter messaging could be a double-edged sword, and underscore the importance of strategy.
Originality/value: trhe paper offers a first exploration of the relevance of corporate Twitter messaging in mitigating ethical scandals.
Text
Final Edited Version Twitter Topic Modelling Corporate Messaging
- Accepted Manuscript
More information
Accepted/In Press date: 12 October 2020
Keywords:
Ethical scandal, Ethical reputation, Reputation continuity, Web Crawling, Twitter Messaging, Topic modelling, Latent Dirichlet allocation; Multidimensional scaling
Identifiers
Local EPrints ID: 444474
URI: http://eprints.soton.ac.uk/id/eprint/444474
ISSN: 1746-5680
PURE UUID: 9dfb64d9-77b2-4567-96c5-58f74e63ab7b
Catalogue record
Date deposited: 20 Oct 2020 16:33
Last modified: 17 Mar 2024 02:54
Export record
Contributors
Author:
Shivani Raheja
Author:
Maxwell Chipulu
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics