The Impact of Peer, Politician, and Celebrity Endorsements on Volunteering: A Field Experiment with English Students

ABSTRACT

Endorsement is used by charitable organizations to stimulate public support, including monetary donations. This article reports a field experiment that examined the effect of leader and peer endorsement on student volunteering. The experiment was conducted with over 100,000 students from five UK universities and compared the effect on volunteering rates of email endorsements by politicians, celebrities, and peers, to a control group that received an email but no endorsement. We examined outcomes seven weeks after the original e-mails, including click-throughs to volunteering unit websites, attendance at volunteering training, registration with volunteering units, and actual volunteering. Peer endorsements reduced click-throughs to volunteering unit websites. There were positive treatment effects for endorsement by politicians on subsequent training but no significant effects of any of the endorsements on our other outcome measures. Overall, we found little support for the provision of leader and celebrity endorsement, and confirm negative effects for peer endorsement.

Acknowledgments

+Output from the project, Citizen Contribution to Local Public Services: Field Experiments in Institutions Incorporating Social Information ES/J012424/1. The paper was previously presented to panel: Political Mobilization Experiments in the Digital Age: A Comparative Perspective, at the American Political Science Association annual meeting, Washington, 27–31 August 2014, and to the Ninth 9th Annual Randomised Controlled Trials in the Social Sciences Conference, at the University of York, 10–11 September 2015. We are grateful to the participants, in particular David Nickerson, for comments. We thank ESRC for financial support for the project. We also are thankful for the help of staff in the five universities from volunteering units and central data services, without whom we could not have done the research. We greatly appreciate the contribution of Esteban Diamani, who worked on the project in autumn 2013, and Mike Elliott for his help with managing the data. The data for this article are deposited with the UK Data Service. Reference: Citizen contribution to local public services - Part 1. Data catalogue. UK Data Service, 10.5255/UKDA-SN-852194.

Disclosure statement

No potential conflict of interest was reported by the authors.

Supplementary material

Supplemental material for this article can be accessed here.

Notes

1. Approved by University College London Research Ethics Committee on 18 July 2012, Reference: 3949/001. Amendments approved 27 August 2014.

2. In the sections that follow we adopt the CONSORT guidelines for reporting randomized controlled trials http://www.consort-statement.org/ Accessed 09/02/2014.

3. The e-mail footers did contain the name of the research project and also an unsubscribe link which recipients could click to remove themselves from the mailing list (those unsubscribing were removed from the list and therefore not sent the reminder e-mail).

4. We thank staff at the University of York trials unit, in particular Hannah Buckley, for their help.

5. As the supporting table in Appendix 1 shows, two of four mentioned politicians are women and two of mixed descent (British Italian and British Hungarian), with an average age of 44 years. As the supporting table in Appendix 2 shows, two of the four celebrities are women and two are from mixed or non-white backgrounds (Jamaican/English and Black African), average age of 43 years. In each of the endorsement e-mails of students, two out of four are women, one from an Asian background, and the students were drawn from a variety of undergraduate and postgraduate courses.

6. Backbench MPs are those in Westminster style parliamentary systems such as the United Kingdom who do not hold government office and if in opposition, are not spokespersons for that party.

7. See http://www.marketingprofs.com/charts/2013/10751/email-open-and-click-through-rates-benchmarks-by-industry.

8. Taking a more conservative test of multiple comparisons (Bonferroni), we find that the students differ from the control at p = .059 and from the celebrities at p = 0.003. Other comparisons are not statistically significant.

Additional information

Funding

This work was supported by the Economic and Social Research Council [ES/J012424/1].