Abstract
This study examines two credibility-enhancing tactics when managers use Twitter as a platform to disclose a message about an employee training programme, which is an important dimension of corporate social responsibility (CSR). I conduct an experiment where I manipulate (1) whether a firm’s CEO retweets the company disclosure tweet in their personal Twitter account (hereafter, ‘retweeting’), and (2) whether the company’s tweet mentions the organisation that cooperates with the company’s CSR activities when the organisation, as an involved party, has a verified Twitter account (hereafter, ‘mentioning’). I document higher disclosure credibility for retweeting than non-retweeting, as it establishes a closer association between the CEO and the message. In addition, mentioning increases the social presence of the parties involved, as users can observe the interactions between them and the disclosing firm in real time. Furthermore, I find that the higher social presence of an involved party (e.g. customers, suppliers, and managers) due to mentioning results in higher disclosure credibility. Finally, investors are more likely to sceptically scrutinise signals from an internal party, which makes CEO retweeting less effective than mentioning. The effect of retweeting is subdued when both tactics are used.
Acknowledgements
I appreciate helpful comments from the joint editor, two anonymous reviewers, Jun Han, Xin Wang, Fei Du, Steve Wu, and Wei Chen. I appreciate helpful comments from workshop participants at The University of Hong Kong, University of International Business and Economics, Beijing Institute of Technology, and Dongbei University of Finance and Economics. I gratefully acknowledge the financial support from Faculty Development Fund of Faculty of Business and Economics, The University of Hong Kong. I gratefully acknowledge the financial support from Humanities and Social Sciences Youth Foundation, Ministry of Education of the People's Republic of China (22YJC790110).
Disclosure statement
No potential conflict of interest was reported by the author(s).
Supplemental data
Supplemental data for this article can be accessed here http://dx.doi.org/10.1080/00014788.2024.2314452.
Notes
1 Twitter (undergoing a rebrand as X in July 2023) requires a verified account to meet the following eligibility criteria: the account must have a display name and profile photo, be active in the past 30 days, and have a confirmed phone number. The account must have no recent changes to the profile photo, display name, or username; must have no signs of being misleading or deceptive; and must have no signs of engaging in platform manipulation and spam.
2 The approval rate is the ratio of the number of accepted assignments to the total number of assignments since a worker registered for MTurk. A higher approval rate implies higher work quality. Available at http://docs.aws.amazon.com/AWSMechTurk/latest/RequesterUI/mechanicalturkconcepts.html
3 The background information mentions that the firm was founded in 1991. Therefore, the tweet about the firm’s anniversary provides no new information, and participants in all conditions access the same set of information.
4 The participants must click on @FUTURE to proceed to the next page and see the profile summary.
5 The order of retweeting/non-retweeting manipulation and mentioning/non-mentioning manipulation is counterbalanced.
6 The order of the CSR disclosure, introduction to the firm, and introduction to FUTURE is counterbalanced. The CSR disclosure and introduction to the firm are always next to each other, as both these statements are posted on the ARMANO website.
7 The attention check question asked the participants to answer in a certain way (Oppenheimer et al. Citation2009, Steelman et al. Citation2014): ‘To show that you have been paying attention, please select “Neutral” as your answer’. The participants were provided with 11 options (as in other questions), and the right answer was the option ‘Neutral’.
8 Of the 127 participants in the four manipulated conditions, 30 failed at least one of the two manipulation check questions. The remaining analysis is based on the full sample. The results are similar to those when the participants who failed the manipulation check questions are included.
9 I also conduct a within-subjects test, with similar results (see Internet Appendix 4) to those of the between-subjects test. The design of the within-subjects test is presented in Internet Appendix 8, pages 49–78.