Ongoing Projects
Behavioural effects of individual-level message tailoring: evidence from a field experiment
Political campaigns are adopting increasingly sophisticated software to automate highly tailored communication with supporters and potential supporters. This allows campaigns to adapt messages using a wider range of individual-level information, such as individuals’ preferences, previous communications, or issue priorities. Despite a growing literature on micro-targeting by campaigns, previous literature mostly focuses on messages tailored to demographic characteristics. Less is known about the effects of tailoring using individual-level information, which is an increasingly viable option for campaigns. Furthermore, existing literature is reliant upon survey experiments, meaning there is limited understanding of the behavioural effects of targeted political communication in real-world settings. This paper presents the results of a large-scale field experiment (N=11,743) conducted collaboratively with a British pressure group during a real petition campaign. The campaign organisation’s mailing list was randomly assigned to a pure control group or to one of three emails: (1) untailored; (2) tailored; and (3) tailored with explicit reference to the tailoring. The findings reveal that whilst both tailored emails mobilise petition signatures, the effect is substantially larger when emails are explicitly tailored. Notably, no significant difference was observed between plain tailored and untailored conditions. These results provide two theoretical contributions. Firstly, despite theoretical fears of backlash to over-tailored communication in previous literature, individuals respond positively when they are explicitly told that content is individually tailored to them. Additionally, in line with organisational theories of activism, the results provide evidence that tailoring at the individual level acts as a signal of a campaign organisation’s responsiveness.
Mobilising supporters’ influence: A survey experiment during the 2026 UK local elections
Recent research suggests that messages received from other voters may effectively persuade and mobilise voters during elections. Yet, despite increased efforts by campaign practitioners to leverage the social influence of their supporters in ‘relational campaigning’, relatively little is known about what motivates supporters to attempt influence their social connections, and how such influence can be mobilised. In particular, we know little about what motivates supporters to influence others, what they ask their peers to do, and which peers they select to influence. In this paper, I conceptualise attempting to influence ties as a form of political participation that can be mobilised by campaigns and test whether information about effectiveness and the risk of (social) disapproval from others increase supporters’ willingness to attempt to influence others. Additionally, I examine how voters choose discussion partners from their personal networks when seeking to influence others.
Replications
Computational Reproduction and Robustness Replication for Hollyer, Klašnja and Titiunik (2022): A Replication Report from the Nottingham Replication Games
✍️ with Edmund Kelly and Angela Odermatt
Part of the Institute for Replication’s mega-reproduction project. See Brodeur et al. (2026).
Hollyer, Klašnja, and Titiunik (2022) analyse the trade-off that political parties face between running programmatic campaigns and fielding charismatic candidates, whose electoral appeal may come at the cost of undermining the party brand. They argue that higher electoral volatility prompts parties to rely on charismatic candidates, even though they might not be as loyal to the party’s programmatic stance. They substantiate their argument with a cross-national dataset and a quantitative case study in Brazil. We computationally reproduced and conducted further robustness tests for their cross-national study by translating the Stata code to R. Next, we conducted a computational reproduction and some additional robustness tests for the quantitative case study. We find that their cross-national analysis is reproducible, albeit with some minor discrepancies. The quantitative case study is also largely reproducible and both are robust in several ways. We conclude by making some suggestions about data dissemination and robustness checks for authors of regression discontinuity designs.
Socially-mediated campaign contact: An RCT testing the spillover of SMS messages in the 2025 English local elections
The spillover of campaign contact within households is well established, but communication is necessary for effects to spillover to neighbours. Previous work treats spillover as a by-product rather than the primary purpose of contact. However, campaigns are increasingly employing ‘relational’ strategies where they ask supporters mediate contact by communicating with their personal networks. Additionally, existing causally-identified research has modelled spillover in single constituencies, limiting exploration of moderating contextual factors. One underexplored factor that could moderate neighbourhood communication - and thus spillover - is rurality. This paper presents national field experiment conducted with a British political party during the May 2025 local elections. Party members across all areas with elections (N=27,330) were randomly assigned at the neighbourhood level to receive SMS messages asking them to remind their housemates and neighbours to vote. By measuring effects on supporters and their social ties across urban and rural areas, this paper tests the effectiveness of mobilising supporters’ social influence and contributes to the literature on how networks and context moderate political behaviour.