There is limited research on the process of conducting such referendums and how to communicate the contents of peace agreements and constitutions to the electorate. The 1998 referendum in Northern Ireland has received the greatest attention in the literature. Available research focuses on factors that can influence whether a referendum is likely to be won, including discussion of campaign strategies.
These factors include:
- allocating sufficient time for planning, informing and implementation;
- undertaking broad-based consultation and inclusive participation;
- implementing a variety of voter education strategies targeting all levels of society;
- ensuring media coverage, which is likely to inform what people think;
- communicating a ‘no alternative’ narrative that draws on prospect theory; and
- reaching marginalised groups.
This report includes four country examples:
- Northern Ireland (1998): an inclusive community strategy with a non-partisan message, designed to appeal to all. Civil society and the media enabled regular communication at the local level. The referendum campaign focused on messages of a better future, reconciliation and equality, rather than on the details of the agreement. This focus on ‘no alternative’ is considered to be a key factor in the referendum outcome.
- South Africa (1992): question framing and wording made it difficult for the referendum question to garner opposition. There was much cooperation on the Yes side across political parties, civil society, and other sectors such as big business, the media, universities etc. Similarly to Northern Ireland, the Yes campaign relied heavily on the language of ‘no alternative’.
- Cyprus (2004): Simultaneous twin referendums with Greek Cypriots and Turkish Cypriots, with little public consultation or polling of preferences during the negotiation process. Campaigns were aimed exclusively at each constituency.
- Guatemala (1999): Little effort to include indigenous groups in the process and lack of investment in voter education services resulted in a very low turnout and a rejected proposal to ratify the Indigenous Rights Accord. This case demonstrates the importance of preparing a favourable context in order for a referendum to succeed in cases of protracted conflict.