| dc.description.abstract | This thesis studies whether and how information about the popularity of political
candidates and parties, such as the results of opinion polls, influences what people
think is going to happen at elections and how they intend to vote. It focuses on
the ‘bandwagon effect’ – the idea that people will vote for candidates or parties
because they are popular. This effect is consequential for debates about the regulation
of opinion polls, as well as campaign strategy and the discipline of election
forecasting. The thesis comes at this question theoretically, by rigorously defining
and classifying the bandwagon effect and sketching a causal model of how it comes
about. It also addresses methodological points related to the bandwagon effect,
making the case for major improvements in how it is measured. Finally, it does
significant empirical work, providing substantial new evidence on how popularity
information affects people’s attitudes and behaviours. Specifically, the key findings
of this research suggest that voters who saw polls suggesting Biden was in the
lead in their state in the 2020 US presidential election became less likely to vote
for him, that people think parties that are growing in the polls are more likely
to win elections even when they are not in the lead, and that voters in the 2017
UK general election who saw polls suggesting the Labour Party’s vote share was
3
growing became more likely to vote Labour. In order to make these findings, I
contribute new theoretical arguments and methodological approaches that show
how to rigorously study the bandwagon effect. I conclude by clearly stating the
scientific and political implications of my research and setting out fruitful avenues
available to researchers for future study. | en_US |
