Abstract
The Northern Ireland peace process saw remarkable change in the policies of Sinn Féin. Having pledged to end Northern Ireland as a political entity, Sinn Féin accepted the 1998 Good Friday Agreement and has since been a strong supporter of the political institutions created under that deal. The party's backing for political institutions has amounted to the zeal of converts, the one-time ‘unequivocal support’ for the IRA's ‘armed struggle’ and repudiation of ‘partitionist institutions’ long displaced by participatory politics. This article shows how, in agreeing to work within political institutions, Sinn Féin was following the desires of the nationalist electorate, assisting the party in capturing majority support within that community. Drawing upon electoral survey data from 1998 and beyond, the article illustrates how, controlling for other demographic and socio-economic variables, it is Sinn Féin supporters who are now the most enthusiastic backers, if only marginally more than those of the Social Democratic and Labour Party, of the devolved power-sharing government in Northern Ireland.
Acknowledgement
The authors thank Professor Kai Arzheimer for suggestions on the paper.
Notes
The 2011 PSAI/PSA Northern Ireland Assembly Election Survey (Principal Investigator: Dr J. Garry) was based on telephone interviews with a weighted representative sample of 1,200, conducted by MRNI within 4 weeks of the Assembly election in May 2011.
Because of some sparse cells across party choice by satisfaction with power-sharing, we reran the model using each party category in turn as reference. Estimates remained stable. We also ran jackknife estimates of the model to confirm stability of standard errors with each case excluded in turn. The jackknife model saw more conservative significance levels for SDLP and Sinn Féin effects, but notably the Sinn Féin effect in contrast 3 remained significant at the 90 per cent level. The alternating reference and jackknife estimates are available from the authors on request.
We would usually provide two nests, beginning with sociodemographics and party support, and then introducing ideology. However, given the complexity of multinomial models, introducing an additional layer of nesting for this one parameter would lead to an excessively large table. Readers wishing to see the nested version can obtain this from the authors directly.