Project report

In a randomised field experiment undertaken at four council supplementary elections in 2021 groups of electors were exposed to three different messaging strategies. Their participation was then assessed to calculate the effectiveness of the treatments in raising turnout. The field experiment was a success: it was well-received by the four councils that agreed to allow it to take place at their elections; there was absolutely minimal negative feedback to the conduct of the experiment, and some electorally and statistically significant results were generated.

The results of the experiment provide a number of valuable lessons about what works and what does not work in terms of stimulating South Australian electors to vote at council elections. These lessons include the following: conventional messaging reminding and encouraging electors to vote is not effective; messaging that exerts social pressure on electors can lead to significant increases in participation; electors are more likely to vote when told that participation is being monitored; social pressure messages are far more cost-effective at generating additional votes than traditional messages used in South Australia; electors from different areas and demographics react somewhat differently to messaging, which means that a one-size-fits-all approach is unlikely to be successful or cost-effective at raising turnout at council elections.

Project evaluation

A post-project evaluation was undertaken by the author and overseen by the Electoral Commission of South Australia’s Research Advisory Committee. Feedback from councils and experiment participants, budget, and achievement of goals were all assessed as part of the evaluation.

The project was completed slightly under budget. The funding provided through this grant was used efficiently and effectively. Of the $32,870 in funding provided, total spending amounted to $31,821 (both figures ex. GST).

While still successful, the experimental treatments were perhaps slightly less effective than anticipated. In retrospect, this may partly have been a consequence of moderating some of the language in the second and third treatment letters before they were approved for sending (as discussed in the Final Report). This was the first time that Electoral Commission SA had undertaken an experiment of this type, so a cautious approach was adopted. Concerns about potential complaints and criticism of the treatments ultimately proved unfounded, however, as these were absolutely minimal. As discussed in the Research Report, the stronger the social pressure applied, the greater the increase in voter turnout.