Do Traditional Party Campaigns Matter Anymore? Experimental and Survey Evidence from Britain

'If you believe in something, write it on a piece of paper and stick it through a letterbox'

This quote, attributed to former Member of Parliament David Penhaligon (1944-1986), epitomises traditional electioneering in Britain. But local campaigns that had long been characterised by labour-intensive activities such as canvass visits and leaflets are changing. Dwindling numbers of local activists means that parties are struggling to maintain widespread contact with voters through such labour-intensive means. While developments in communication technology have introduced new, digital forms of campaigning, traditional activities remain resilient. It is against this backdrop that the motivation behind this thesis is presented - should traditional campaigning be allowed to whither? I explore the role and impact of campaigning in Britain today, focusing on the most common modes - leaflets and canvass visits.

First, I examine the role these activities play when it comes to increasing participation at election time. I make an empirical contribution to a largely US-based, non-partisan Get Out The Vote (GOTV) literature by conducting new voter mobilisation field experiments in the May 2017 round of English local elections. The study represents one of the few partisan GOTV experiments to be carried out in Britain, and the first individual-level experiment anywhere to include postal voters. Embedded in a real party campaign, the experiments show that leaflets, and a combination of canvass visits and leaflets, both have positive effects on voter turnout at a local election. Meanwhile, there is evidence of a mobilisation 'ceiling effect' among higher turnout postal voters.

I then use the marked electoral registers from the subsequent general election to examine the downstream effects the treatments had on turnout. I find that the effects of campaign contact detected in the May local election were quickly subsumed at the general election in June, when underlying turnout rose substantially. Together, the experiment and the downstream analysis provide new empirical evidence that traditional campaigns mobilise voters at lower saliency elections.

Finally, I investigate the role that voters' preferences for different types of campaign contact play in conditioning campaign effects. While research into campaign activities is extensive, this represents the first attempt to examine voters' preferences when it comes to how they are contacted. Using survey data of voters in Wales, I show that there is considerable heterogeneity in preferences, driven largely by levels of political interest and previous exposure to different activities. I also find evidence that wanting to be contacted conditions campaign effects, contributing to ongoing debates about which voters are most influenced by campaigns.

Overall, I provide evidence that while under threat, traditional party campaign activities still play a vital role in increasing turnout at lower saliency elections - though the temporal durability of effects are susceptible to being consumed at subsequent higher turnout elections. Given that low turnout elections are becoming increasingly common in Britain, however, I argue that the importance of traditional campaign activities is unlikely to diminish for some time to come.