Devil in the detail: Designing referendum processes to support informed voter decision making

Ireland has been at the vanguard of countries that have integrated deliberative democratic innovations, most especially citizens’ assemblies, with direct democratic procedures. The 2018 referendum in Ireland which notably liberalised abortion provision typifies the success of the deliberative-direct model for many. The provision of abortion services was first discussed at a national citizens’ assembly which produced a comprehensive report favouring liberalisation. This was subsequently discussed by a committee of parliament which recommended to government that a referendum should be initiated to give effect to the recommendation of the citizens’ assembly. The government proceeded with the referendum and produced draft legislation and participated actively in informing voters. Importantly post-referendum research reported that voters had high levels of objective and subjective knowledge and there was a strong alignment between values and vote. But the deliberative turn is not a fail-safe for high quality and informed referendum votes as the Irish case also indicates. In 2023, another set of referendums were held that ostensibly followed the same route onto the ballot (citizens’ assembly, parliamentary committee, government decision) but these two referendum proposals were defeated by huge margins. Voters had low levels of knowledge of the proposals and although knowledge of the citizens’ assembly made a voter more likely to support the proposals, their overall numbers were so small that they had minimal effect. Directly comparing the two procedures, small but significant differences emerge. The citizens’ assembly took place during covid, media interest was dwarfed by the pandemic and the specific recommendation of the assembly was not adopted. Furthermore, no accompanying legislation was provided that explained what the legal effect of the changes would be. The voters smelled uncertainty and the old adage, don’t know-vote no prevailed. This paper will unpack the deliberative-direct fusion model and argue that the devil is in the detail when it comes to the deliberative turn in referendums.