Power-Sharing and Voting: Conflict, Accountability and Electoral Behaviour at the 2015 Northern Ireland Assembly Election

Lead Research Organisation: Queen's University of Belfast
Department Name: Sch of Hist, Anthrop, Philos & Politics

Abstract

When citizens in Northern Ireland cast their vote in the 2015 Northern Ireland Assembly election, almost two decades will have passed since the Good Friday/ Belfast Agreement of 1998 established devolution and power-sharing. Thus, a rigorous analysis of the 2015 Assembly election will allow us to carefully consider the impact of the 'peace process' and the associated implementation of power-sharing governing structures. We investigate whether power-sharing has actually led to good quality electoral democracy for the people of Northern Ireland or has instead merely cemented and strengthened ethnic divisions. In our investigation, we address a number of specific questions:

1/ Have the power-sharing institutions maintained or reduced the importance of ethno-national (Protestant British versus Catholic Irish) campaigning and voting at election time? In our study we compare the strength of ethno-national voting over the 1998-2015 period by linking our proposed 2015 study to earlier ESRC funded studies in 1998 and 2003. We assess whether there has been an increase over time in 'cross bloc' voting (Protestants giving lower preference votes to nationalist parties, or Catholics giving lower preference votes to unionist parties). Also, we assess whether the importance of ethno-national ideology (voting on the basis of nationalist versus unionist beliefs) has become more or less important over time. We, furthermore, compare ethno-national ideology to non-conflict ideologies (such as economic left-right views or moral liberal-conservative views).

2/ Is electoral accountability possible? If there is only one party in government, it is easy for voters to identify who to hold responsible for government performance. In coalition governments it is harder for voters to clearly see which party to blame if things are going badly (or reward if things are going well). This difficulty is particularly acute when all of the parties are in Government and none are in Opposition, as is the case in Northern Ireland. We assess whether it is actually possible for voters to hold decision-makers to account in such a context.

3/ Is there an emotional basis to voting? Following on from questions 1 and 2 we assess the conditions under which voters rely on either ethno-national voting or peformance-based voting. We investigate the role of emotions and assess whether anger is associated with voting on the basis of the ethno-national conflict while fear is associated with performance-based voting.

4/ Does power-sharing discourage participation? Does the absence of a clear Opposition lead to reluctance to engage in politics? Also, does the sense that powersharing benefits one community over the other lead to alienation from politics, or non-electoral forms of political protest?

5/ What are Northern Ireland citizens' attitudes to political reform (such as developing a strong Opposition that would hold the Government to account, or the regulation of sensitive cultural matters such as flag flying and parades)? How do citizens form such attitudes? Why is it that disagreement with suggested reform can spiral into violent street protest?

Answering these questions is important for understanding how democracy operates in the sensitive setting of Northern Ireland. The answers are also important for our broader understanding of how to respond to violent ethnic conflict around the world. The 'Northern Ireland Model' is often held up as an impressive example of successful power-sharing and peace building. Our analysis will provide a detailed critique of the quality of electoral democracy in the Northern Ireland example that will be important for peace builders around the world (including organisations such as the United Nations). We will also maximise the relevance of our work for ordinary citizens in Northern Ireland, second level school teachers and students, Northern Ireland newspapers, documentary makers, and the Northern Ireland Assembly.

Planned Impact

Our study will influence non-academics in a range of ways.

BBC Radio 4 documentaries: We plan that our findings will form the basis of two Radio 4 documentaries on the subjects of 1/ Emotions and Politics and 2/ Is Accountability Possible in Politics?

Northern Ireland Newspapers: We have agreed with the leading newspaper political correspondents that we would have their input into the questionnaire design, that we would provide them with data relating to questions they were particularly intersted in and that we would write newspaper articles.

Electoral Commission: Our findings will inform the Electoral Commission (Northern Ireland) in terms of providing a fuller understanding of the attitudinal as well as demographic precursors to non-participation, and providing a fuller understanding of the ability of citizens to hold the power-sharing government to account at election time.

Community Relations Council: We have discussed the project with the Research Director (Paul Nolan) of the Community Relations Council and our data will be a valuable contribution to the annual Peace Monitoring Reports produced by the Council.

Northern Ireland Assembly: We plan that our findings will be presented and filmed at a special seminar held in Stormont (part of the QUB/UU/Assembly Seminar Series). This will inform political decision makers about the nature and quality of the citizen-politician link at election time.

British-Irish Parliamentary Assembly: We will write a report for this Inter-Parliamentary body which brings together representatives from Westminster, Dáil Éireann and the Northern Ireland Assembly. Specifically, we will focus on Committee A which addresses 'Sovereign Matters'.

European Union: We will provide a report for the External Action Service of the European Union. Specifically, we focus on the 'Conflict Prevention, Peace Building and Mediation' sub-division of the 'Security Policy and Conflict Prevention' Directorate.

United Nations: We will provide the United Nations with a report which highlights the particular findings of our study that have most import for the UN's international peace building work. Specifically, we will focus on the Mediation Support Unit in the UN's Department of Political Affairs.

Second level schools and teachers in Northern Ireland: Our findings will be used as a web-based resource for Politics teachers and students in Northern Ireland. Our findings will also inform the annual QUB 6th form lectures for Politics teachers and students.

Blogs: The applicants have experience contributing to debates in the premier Northern Ireland politics blog, Slugger O'Toole. The findings of our project would be of great interest to the users of the blog. We will also utilise two LSE blogs, particularly for aspects of our study of relevance to a Britain-based audience and an international audience.
 
Description 10 Key Findings

There are 10 key findings of this research project to date.
Each key finding is the basis of a distinct research paper from the project. In relation to each finding, we offer below a summary of the finding in simple terms as well as specifying the status of the research paper the finding is associated with.

Key Finding Number 1: Socio-moral attitudes are also important as well as green-orange attitudes
We examine the structure of citizens' political attitudes in Northern Ireland. It is well known that attitudes to the underlying unionist-nationalist divide are of high importance. But is there also evidence that citizens think systematically in terms of economic left-right issues and in terms of moral liberal-conservative issues? In assessing the structure of political beliefs we ask whether citizens' views on a particular theme (unionism/nationalism, left/right, and liberal/conservative) are strongly related to each other. We know that the main set of issues relating to unionism/nationalism are strongly structured in the sense that if a voter has a strongly unionist position on the issue of, say, parades they are also highly likely to hold a strongly unionist position on the issues of flags, union with Britain and language rights and other such issues. Our main aim is to examine whether other clear 'dimensions' of politics are also evident. When we ask voters a wide range of questions about economic matters is it the case that a voter who adopts a left wing attitude on one issue (corporation tax) also adopts a left wing positions on others (such as nationalisation of railways and policies to reduce inequality). Similarly, do citizens who have a conservative position on one 'moral' issue (such as drug use) also have a consistently conservative position on other 'moral' issues (such as same sex marriage)? Knowing the answer to these questions will enable us to understand whether there is another dimension of politics that can potentially play a strong role in Northern Ireland politics, alongside the currently dominant unionism-nationalism dimension. If it turns out that voters only have 'coherent' or 'consistent' views on the unionism-nationalism dimension then the potential for a different attitude dimension to potentially shape Northern Ireland politics is low.
We use data from a Voter Advice Application (VAA) that we created for use by voters in the 2015 Westminster Election. In a VAA, an online user is asked a wide range of policy questions and is provided with an 'advice' score which tells then how close they are to each party taking all policy matters into consideration. We use the data generated from users answering the wide range of policy questions. We examine the data using methods designed to identity the underlying themes or structure in the data (Principal Components Analysis). Our main finding is that there is a strong liberal-conservative attitude dimension. Northern Ireland voters hold clear and consistent views on issues relating to drug use, prison sentencing, same sex marriage, abortion and similar issues on the 'liberal-conservative' dimension. Alongside a very strong and clear unionist-nationalist dimension, this strong liberal-conservative dimension suggests that if a 'second' dimension of politics is to emerge in Northern Ireland in addition to the unionist-nationalist dimension, it is likely to relate to the liberal-conservative dimension. A manifestation of this dimension in Northern Ireland politics in recent years relates to the issue of same-sex marriage and the controversy over the 'gay cake' issue.
Status
Paper: Dimensionality of Policy Space in Consociational Northern Ireland
Authors: John Garry, Neil Matthews and Jonathan Wheatley
Presentation: This paper has been presented in the Centre for Direct Democracy at the University of Zurich
Published: This paper has been published in Political Studies, one of the top two politics journals in the UK
65(2): 2017 - Published here:
http://journals.sagepub.com/doi/abs/10.1177/0032321716658917





Key Finding Number 2: Providing people with advice on parties' policies reduces 'tribal' voting
Voting in deeply divided societies such as Northern Ireland is very largely based on issues relating to 'the divide'. Thus, 'unionist' parties (Democratic Unionist Party and Ulster Unionist Party) are supported by Protestant/unionist voters and 'nationalist' parties are supported by Catholic/nationalist voters. This is largely explained by the durability and electoral relevance of the underlying identity-based divide between Irish nationalists and British unionists. However, it may also be explained in part by the fact that there is very little coverage of non-divide issues in the media, with the result that voters are uninformed about exactly what the parties positions are on economic and social-moral matters.
We examine what happens when voters are given advice on party policy positions on a wide range of policy issues. If voters had a clear idea about the policies of all parties in issues relating to the economy (the parties' views on a range of economic left-right issues) and socio-morality (the parties' policies on the liberal-conservative dimension of politics relating for example to same sex marriage, abortion, crime and so on) would this have an influence on their support for each of the parties. In other words, if a voter realises that they are actually close to a party from 'across the divide' once you take economics and social issues into account would their support for that party rise? Or will Northern Ireland politics, and party support, always be dominated by the unionist-nationalist divide, irrespective of what information is given to voters? To put it bluntly, if a Catholic voter learns that, taking all policies into consideration they are quite close to the UUP, would that lead to an increased likelihood of voting for the UUP? And similarly for a Protestant voters and Sinn Fein/SDLP.
We test this question with an experiment. We design a 'Voter Advice Application (VAA)' for the 2016 Northern Ireland Assembly Election (similar in design but distinct from the 2015 VAA discussed above). Any member of the public can use the VAA online. The user if asked to answer a range of policy questions. For this range of policy question we had a group of researchers identify the actual policy positions of all the main parties. Using an algorithm the VAA calculates how close in policy terms each user is to each party and provides an 'advice' score (e.g. you are very close to the DUP, not close at all to the SDLP, fairly close to Alliance, etc). The advice score is in the form of a highly user friendly bar chart: the higher the bar for a party the closer you are to them.
We disseminated our VAA widely during the election campaign and generated over 25,000 users of the app. We embedded an experiment in the VAA in order to help us understand whether the VAA was actually having any effect on the users. One randomly selected group of users received the 'advice' before giving a final verdict on how likely they were to vote for each party. A second randomly selected group of users were asked to indicate how likely they were to vote for each party prior to receiving the advice. Hence, we are able to compare the two groups in order to identify whether people who get advice behave differently from those who do not. Crucially, our aim is to see whether Catholic users, who receive advice are more likely to indicate a higher level of support for the DUP and the UUP than Catholics who do not. And similarly, are Protestants who receive advice more likely to indicate a higher level of support for Sinn Fein and the SDLP than those who do not.
We find that this is indeed the case when we examine the data. First of all we find that users did tend to follow the advice provided to a statistically significant extent. Second, we did find that providing advice on the full range of policy matters did lead to an increase in support for 'outgroup' parties (i.e. it led to Catholics having a higher level of support for the DUP/UUP and Protestants having a higher level of support for Sinn Fein/SDLP). The advice also led to higher levels of support for the cross community party (Alliance) and lower levels of support for 'in-group' parties (i.e. Catholics who received advice had lower levels of support for Sinn Fein and the SDLP, and Protestants who received advice had lower levels of support for the DUP/UUP).
The general point of these findings is that we demonstrate that the provision of objective and systematic advice to voters on all policy matters relating to all parties leads to a lessening of 'divide based' party support. The advice 'melts' the underlying divide somewhat: it moves the issue of voting somewhat away from merely being about unionism-nationalism and towards being about a wide range of policy matters. This movement opens up the possibility of people supporting parties that they never thought they would.

Status
Paper title: Voter Advice Applications Cause a Decline in Divide-Based Voting in Deeply Divided Places: Evidence from a Field Experiment in Northern Ireland
Authors: John Garry, James Tilley, Fernando Mendes, Neil Matthews, Jonathan Wheatley
Conference Presentation: This paper was presented at the Annual conference of the Political Studies Association of Ireland Conference, 2016
https://psai2016.files.wordpress.com/2016/05/psai-programme-final-version.pdf
This paper has been written up as an academic paper and is being finalised for submission the journal Party Politics
This paper has now been published in Party Politics (Garry et al. 2018)


Key Finding Number 3: Only a minority describe parties in unionist-nationalist terms
When people in Northern Ireland are asked what they think political parties in Northern Ireland stand for only a minority (42%) mention issues related to the unionist-nationalist divide, almost nobody mentions issues related to economic (left-right issues) or morality (liberal-conservative issues), and almost half either say they 'don't know' or give a very vague response. Of those who do characterise the parties in terms of unionism-nationalism we break down responses in terms of ideas (e.g. the party's views on a united Ireland versus maintaining the union with Britain) or groups (the party's ability to represent Catholics or Protestants) and we find that parties are three times more likely to be described in terms of ideas than groups. The way the parties are described is largely in line with expectations. The Democratic Unionist Party and the Ulster Unionist Party are almost exclusively described in unionist terms, Sinn Fein is almost exclusively described in nationalist terms and Alliance is almost exclusively seen in 'cross-community' terms. The surprise is the extent to which one of the main 'nationalist' parties, the Social Democratic and Labour Party (SDLP,) is seen as 'cross-community' by many people: one third of people described the party in these terms and two thirds described the party in nationalist terms.
When people in Northern Ireland are asked what they think the main political parties in Britain stand for the most typical response (42%) describes the parties in terms of economics: with the Conservatives almost exclusively seen as having economically right wing views and being pro-middle class and Labour regarded as economically leftist and pro-working class.
When people in Northern Ireland were asked what they thought the three main parties in the Republic of Ireland stand for: Fianna Fail, Fine Gael and the Irish Labour party, the predominant response was 'I don't know'. Hardly any respondents described the parties in either left-right, liberal-conservative or unionist-nationalist terms.
These findings are based on a survey we conducted of a representative sample of 1000 Northern Ireland respondents. In relation to each party (five Northern Ireland parties, two British parties and three Irish parties) respondents were asked an open ended question: 'What do you think this party stands for?' We coded the responses and analysed them in order to generate the above findings.
The reason we wished to examines responses to these questions was to increase our understanding of how people in Northern Ireland view their political parties. Is it the case that they perceive them simply in green-orange, or unionist-nationalist, terms and do not use economics and social class as a prism through which to characterise the parties? Our findings clearly show that Northern Ireland people are well able to describe parties in class and economic terms when that is relevant (i.e. this is how they describe the British parties) but they do not do so for Northern Ireland parties. Insofar as they regard their own Northern Ireland parties in ideological terms, it is very predominantly in ethno-national (i.e. unionist-nationalist) terms, but this it must be said is only the case for 42%. Thus: Northern Ireland people do not tend to describe their parties in ideological terms, but for those who do there is only one relevant ideological dimension: unionism-nationalism.
Status
Paper title: Party Images in Northern Ireland: What do citizens think their parties stand for?
Paper authors: John Coakley, John Garry, Neil Matthews and Brendan O'Leary
Conference presentation: Elections, Public Opinion and Parties (EPOP) Annual Conference, 2016, University of Kent, 9-11 September
https://blogs.kent.ac.uk/epop/files/2015/11/Programme-.pdf
This paper has been submitted to an academic journal and is currently under review with Irish Political Studies
This paper has been published in Irish Political Studies (Coakley et al. 2018)

Key Finding number 4: Protestants and Catholics differ in how they vote 'across the divide'
We have studied in depth how voters in Northern Ireland use the Single Transferable Vote (STV) electoral system. We found that voting 'across the divide' is rare. In other words, when voting very few Catholics give any preference (either a first preference or lower preference) for Unionist parties, and very few Protestants give any preference vote for Nationalist parties. However, we find that voters use the system very similarly to voters in other STV polities: they typically rank the same number of candidates (3 on average).
Most interestingly, we find that lower preference voting varies across community. Protestants were much more likely to transfer to the SDLP than to Sinn Fein whereas Catholics were almost equally likely to transfer to the UUP and the DUP. What this means is that how citizens view the parties of the 'other community' varies a lot: Catholics see the 'other' parties as very similar (and equally unpalatable) whereas Protestants see the 'other' parties as very different from each other, with the SDLP significantly more palatable than Sinn Fein.
We used our post-election survey to conduct this analysis. The survey is uniquely well placed to enable us to examine in detail these issues given its unusually large number of respondents (over 4000). We wrote up our detailed results in a report and published it in collaboration with the Electoral Reform Society.

Status
Report Title: The Northern Ireland 2016 Assembly Election - How Voters Use STV
Authors: John Coakley, John Garry, Neil Matthews and James Tilley
Published Report here:
http://archive.electoral-reform.org.uk/sites/default/files/2016%20NI%20Election%20Report.pdf
Our analysis received significant press coverage, including front page headlines of the Belfast Telegraph.


Key Finding number 5: Brexit is an ethno-national issue
We have used our data to investigate the basis of voting behaviour in the EU referendum in Northern Ireland. In our analysis, we found a very strong ethno-national structure to voting. Protestants were strongly in favour of leaving the EU and Catholics were very strongly in favour of remaining in the EU. Further we found some evidence of a class basis to voting: working class and lower educated citizens were more likely to want to leave the EU and middle class and more highly educated citizens were more likely to want to remain. Importantly, however, this class distinction only works strongly within Protestants. Almost all Catholics, irrespective of class and education level wish to remain in the EU.
We further examined how Northern Ireland people felt about Brexit one year later. When asked how they would vote if there was a second referendum on remaining/leaving the EU almost all (90%) of people who voted to Remain in the 2016 referendum indicated that they would continue to vote Remain if there was a second referendum. Somewhat in contrast, we found some slippage in the Leave vote: 75% of those who voted to Leave in the 2016 Referendum indicated that they would continue to vote Leave if there was a second referendum. Hence, there has been attitude change since the referendum and this has been in a Remain direction.
When asked about a range of issues associated with 'Brexit' there was much difference between Leavers and Remainers in terms of how they thought the economy in Northern Ireland would fare post-exit, with Leavers very considerably more optimistic than Remainers. However, there was one issue on which there was a consensus across Leavers and Remainers: the desire to avoid having a 'hard border' between North and South.
We used our unique 2-wave panel study to address these questions, which provided high quality and (for Northern Ireland) rare data from the same individuals over two time points (2016 and 2017).
Professor Garry presented the analysis of Brexit at a presentation at Stormont and provided Stormont with an associated policy briefing paper.
Description:
http://www.niassembly.gov.uk/assembly-business/research-and-information-service-raise/knowledge-exchange/knowledge-exchange-seminars-series-6/#outcome
Policy Paper:
http://www.niassembly.gov.uk/globalassets/documents/raise/knowledge_exchange/briefing_papers/series6/garry121016.pdf
Filmed Talk by Professor Garry at Stormont:
https://niassembly.tv/eu-referendum-vote-northern-ireland-implications-understanding-citizens-political-views-behaviour/

Our analysis has also been written up as an academic paper and has been presented at the annual conference of the Political Studies Association of Ireland, 2017, at Dublin City University, 13-15 October:
https://psai2017.files.wordpress.com/2017/05/programme-psai172.pdf

Paper: Explaining Vote Choice in the "Brexit" Referendum and the Stability of Voter Preferences
Authors: John Coakley and John Garry

An academic paper is being finalised and will be submitted to the Journal of European Public Policy

Professor Garry will be presenting the findings also at a number of event:

Oct 30 2017: Presentation at Westminster, Portcullis House, organised by the ESRC's UK in a Changing Europe

November 15: Invited presentation at the Foreign Office in London

November 15: Public presentation at the Irish Centre in London



Key Finding number 6: 'Northern Irish' identity has a range of meanings to people and politicians

National identity in Northern Ireland is typically regarded as being either British or Irish. However, surveys have typically recorded approximately a quarter of respondents opting for the category of 'Northern Irish' rather than either 'British' or 'Irish'. According to the results of the 2011 census 21 percent of people described themselves as Northern Irish. The question arises as to what exactly this Northern Irish national identity actually means, if anything? Is it a 'middle way' identity, different from and half-way between Irish and British? Or is it simply another way of being Irish or British: a Catholic can be Irish and also 'Northern Irish' because they are geographically from the northern part of the island, and a Protestant can be British and also Northern Irish because they are from the particular part of the United Kingdom that is Northern Ireland? The answer to this question is important as it will provide insight into whether 'Northern Irish' as an identity can potentially act as a 'shared identity' that will overcome the stark British-Irish binary.
We empirically investigate the question using data from qualitative focus groups and qualitative interviews with elected politicians (MLAs). We find that a range of distinct ways of thinking and talking about this identity choice emerge and we generate a 'typology' or systematic categorisation of the different meanings.
Status
Paper: Taxonomy of a nation: mass and elite evidence on the meaning of Northern Irish
Authors: Kevin McNicholl, Clifford Stevenson and John Garry
Presentation at the annual conference of the Political Studies Association of Ireland, 2017, at Dublin City University, 13-15 October:
https://psai2017.files.wordpress.com/2017/05/programme-psai172.pdf

This academic paper has been completed and submitted to Political Psychology

This paper has been published in Political Psychology (McNicholl et al. 2018)

Key Finding number 7: A 'Northern Irish' politician would attract votes from the 'other' community
We examine how voters react when they are faced with evaluating a political candidate who is described as being from the 'other community' and is also described as 'Northern Irish'. Specifically, imagine a Catholic voter who is faced with a Protestant candidate. What difference does it make to that Catholic voter if the candidate is described as 'Northern Irish' rather than what would typically be expected (British). Does it make the Catholic voter somewhat more likely to support the candidate on the grounds that 'Northern Irish' signals a more moderate political attitude than 'British' does? Similarly, is a Protestant voter likely to be more favourably disposed to a Catholic candidate if that Catholic candidate is 'Northern Irish' rather than 'Irish'?
We address these questions in order to shed light on the electoral implications of identity description of candidates in divided places. More specifically, we seek to understand whether, in the minds of voters, the 'Northern Irish' identity is a signal of political moderation and centrism.
We generate data from a representative sample of the Northern Ireland population (1000 respondents). Our study involves an experiment in which the respondents are presented with the description of a candidate. All Catholic respondents are presented with a candidate who is Protestant but one random half of Protestant respondents are told the candidate is Northern Irish and the other random half are told the candidate is British. A similar design is put in place for Catholic respondents. We find that it is indeed the case that support for the outgroup politician (i.e. a Catholic candidate if the respondent is Protestant; and a Protestant candidate if the respondent is Catholic) is higher if the candidate is described as 'Northern Irish'. This demonstrates that in the minds of voters a candidate who describes themselves as Northern Irish will be perceived as being more moderate.
We also use qualitative interview data from MLAs to investigate what they feel about calling themselves Northern Irish. Some MLAs do enthusiastically do so in a way that is consistent with Northern Irish indicating a compromise centrist position between Irish and British, but it is not very common. Also, some enthusiastically engage with the Northern Irish identity but in a way that simply complements the crucial identity they already have (Irish or British), and many do not engage at all with the identity, particularly Sinn Fein MLAs.
Status
Paper: Shared Identities' and Voting Across the Divide in Deeply Divided Places
Authors: John Garry, Kevin McNicholl and Clifford Stevenson
This paper has been presented at the annual conference of the Political Studies Association of Ireland, 2017, DCU, 13-15 Oct
https://psai2017.files.wordpress.com/2017/05/programme-psai172.pdf

This paper is being finalised for submission to Electoral Studies

In preparing the groundwork for this analysis McNicholl and Garry provided an analysis of the Northern Irish identity at Stormont
Description
http://www.niassembly.gov.uk/assembly-business/research-and-information-service-raise/knowledge-exchange/knowledge-exchange-seminar-series-4-2014-2015/#n
Policy Briefing Paper
http://www.niassembly.gov.uk/globalassets/documents/raise/knowledge_exchange/briefing_papers/series4/northern_ireland_identity_garry_mcnicholl_policy_document.pdf
Filmed Presentation
http://www.niassembly.gov.uk/assembly-business/research-and-information-service-raise/knowledge-exchange/knowledge-exchange-seminar-series-4-2014-2015/mr-john-garry-qub-and-mr-kevin-mcnicholl-qub/




Key Finding number 8: People would accept decision making by a Citizens' Assembly
The power-sharing institutions in Northern Ireland have suffered from gridlock and stalemate as the major political parties disagree with the each other. The latest example is the absence of a government in Northern Ireland since the 2017 Assembly election (and for a period prior to it also) due to disagreement between the DUP and Sinn Fein on the merits of an Irish Language Act (among other issues).
We examine the possible ways of responding to such gridlock. Specifically, we examine the extent to which there may be support among the public for a Citizens' Assembly to address the controversial issues that led to gridlock, and attempt to reach agreement on the issue and remove it as a political obstacle to political progress and stability.
A Citizens' Assembly is made up of a random sample of ordinary citizens who are representative of the Northern Ireland population terms of age, gender, religious background and so on. The citizens are asked to consider evidence and argument on the particular contentious issue at hand, reflect upon and consider the issue, and vote on the issue to reach a decision on it.
We generate data from a representative sample of 2000 respondents. We use the topical example of the Irish language as a contentious issue. We describe to respondents a situation in which there is gridlock on this issue and we present them with a range of ways of seeking to overcome it: interparty talks, a new election, a referendum, a Citizens Assembly. We assign respondents to one or other of a range of categories. Some are told that interparty talks have reached a decision, others are told a referendum was held to reach a decision, others that a Citizens' Assembly was established and reached a decision, and others were told that there was a new election held which led to a decision being reached. Crucially we conducted the experiment such that all respondents were told that the decision reached was one that they (the respondent) disagreed with. So, respondents who indicated that they were in favour of an Irish Language Act were presented with a decision that was opposed to an Act (and those opposed were told the resulting decision was in favour).
Thus, all respondents in our study got something that they did not want. We asked people how much they were prepared to accept this decision. Crucially, we compare acceptance levels across decision making categories in order to identify whether people in the Citizens' Assembly category were more (or less) likely to accept this decision compared to people in the other decision making categories.
We find that those in the Citizens' Assembly group are just as likely to accept the decision as those in any other group. Hence, this is strong evidence suggesting that there is a potential role for a Citizens' Assembly to be established to seek to address and overcome contentious issues that gridlock the power sharing arrangements.

Status
Paper: The perceived legitimacy of mini-publics in divided places: Experimental evidence from Northern Ireland
Authors: John Garry, Jamie Pow, John Coakley, David Farrell, Brendan O'Leary and James Tilley
Paper presented at the annual conference of the Political Studies Association of Ireland, 2017, DCU, 13-15 October
https://psai2017.files.wordpress.com/2017/05/programme-psai172.pdf
This paper is being finalised for submission to the American Journal of Political Science
This paper is currently under review at World Politics




Key Finding number 9: Parties have moved to the centre ground and this affects voting

Thirty years of ethno-national violent conflict drew to a close with the 1998 Belfast/Good Friday Agreement which implemented power sharing arrangements. These arrangements involve a coalition government made up of parties from both communities, with parties rewarded with ministerial positions in proportion to votes and seats won, and with powers for each community's representative to veto legislation. We suggest that there are strong incentives in the system which will lead to extreme parties becoming more moderate over time. We further suggest that parties in each community bloc are likely to become more distinct on new issues areas relating to economic left-right and moral liberal conservative.

We analyse all the parties' manifestos in all Assembly elections between 1998 and 2016 inclusive. Our main finding is that there was a major movement in the unionist bloc, as the DUP moved from an extreme position to a position where it occupied the same moderate space as the UUP. On the nationalist side there was a move in a moderate direction by Sinn Fein but not to the same extent.

When we examine the voters we find that this movement by the DUP matters. In 1998 when the DUP and UUP were very different, voters' positions on the unionist-nationalist dimension were very strongly predictive of voting. However, in 2016 when the DUP and the UUP share the same moderate position, voters' positions on the unionist-nationalist dimension do not predict voting.

What these results show is that in the power sharing context there has been significant party movement and this movement has influenced citizens voting behaviour.

Status
Paper: Political Cleavages and the Evolution of the Northern Ireland Party System under Power-Sharing 1998-2016
Presented at the Paper presented at the annual conference of the Political Studies Association of Ireland, 2017, DCU, 13-15 October
https://psai2017.files.wordpress.com/2017/05/programme-psai172.pdf
Authors: James Tilley, Neil Matthews and John Garry
Paper is being finalised for submission to the British Journal of Political Science
This paper is currently under review at the journal Government and Opposition



Key Finding number 10: Religion matters
We find that the particular religious denomination that citizens belong to has an influence on their vote choice. In particular when we compare Free Presbyterian voters to other types of Protestant voters we find the former more likely to support the DUP than the UUP. This is not surprising given the historical development of the DUP, established by Iain Paisley who also established that particular church. We also compare non-religious voters to religious voters in both communities in terms of their political behaviour (voting versus non-voting and also party choice).
Status
Paper: Electoral Choice and Religion: Northern Ireland
Authors: Neil Matthews and John Garry
Forthcoming Publication in: Encyclopaedia of Politics and Religion, published by Oxford University Press, edited by Paul Djupe
Exploitation Route We have coded the manifestos of all the main five parties from 1998 to 2016. To do this we developed and tested a coding frame specifically designed for deeply divided places. We suggest that other scholars interested in systematically assessing political parties ideological positions in divided places can apply our coding frame.

Our analysis of our Voter Advice Applications paves the way for other scholars to develop our experimental work and further assess how information provision effects citizens. Our work was one of the first interventionist field experiments in this area and our methodology is of potentially widespread application in any democratic setting.

Our analysis of vote transfer behaviour may be used by other researchers in STV systems to systematically understand the details of how voters use the ballot paper.
Sectors Government, Democracy and Justice

URL https://www.qub.ac.uk/sites/NIAES2016/
 
Description Our project has contributed to debates about Northern Ireland politics in a range of media outlets such as Newsweek, The Conversation, the Belfast Telegraph, the Belfast News Letter. We have contributed to widely read blogs such as the LSE blog on British Politics and the main Northern Ireland current affairs blog 'Slugger O'Toole'. I have also appeared on Five Live radio and acted as advisor to the BBC/RTE on an all-island programme on public opinion. We have also presented work (presented by the project post-doc) on the gendered nature of Northern Ireland politics at Stormont as part of the Knowledge Exchange Seminar Series (KESS) run by the Northern Ireland Assembly in collaboration with Queen's University Belfast and Ulster University (presentation 9 March 2016) All of these impacts are detailed on our project website: https://www.qub.ac.uk/sites/NIAES2016/ Immediately after the May 2017 election we conducted, as promised, our survey of a large sample of Northern Ireland citizens (n over 4000). We asked questions in our survey relating to the EU referendum and we used this data to write a policy briefing paper for Stormont, and we presented this in person in a filmed presentation at Stormont as part of the Knowledge Exchange Seminar Series. http://www.niassembly.gov.uk/assembly-business/research-and-information-service-raise/knowledge-exchange/#outcome This presentation was on 12th October 2016 and was introduced by Mike Nesbitt, MLA. We conducted and further disseminated analysis of this issue in the Northern Ireland media: http://www.newsletter.co.uk/news/brexit-understanding-why-people-voted-as-they-did-in-the-choice-of-a-lifetime-1-7630272 We further disseminated our findings at: http://qpol.qub.ac.uk/public-opinion-challenge-ni/. Also, our project has had a direct impact on citizens in Northern Ireland in terms of us providing citizens with objective advice at election time. Our advice was offered in the form of Voter Advice Applications (at the 2015 and 2016 elections). Any member of the public can use the VAA online, and the user if asked to answer a range of policy questions. For this range of policy question we had a group of researchers identify the actual policy positions of all the main parties. Using an algorithm the VAA calculates how close in policy terms each user is to each party and provides an 'advice' score (e.g. you are very close to the DUP, not close at all to the SDLP, fairly close to Alliance, etc). The advice score is in the form of a highly user friendly bar chart: the higher the bar for a party the closer you are to them. In the 2016 elecion VAA we disseminated our VAA widely during the election campaign and generated over 25,000 users of the app. We embedded an experiment in the VAA in order to help us understand whether the VAA was actually having any effect on the users. One randomly selected group of users received the 'advice' before giving a final verdict on how likely they were to vote for each party. A second randomly selected group of users were asked to indicate how likely they were to vote for each party prior to receiving the advice. Hence, we are able to compare the two groups in order to identify whether people who get advice behave differently from those who do not. Crucially, our aim is to see whether Catholic users, who receive advice are more likely to indicate a higher level of support for the DUP and the UUP than Catholics who do not. And similarly, are Protestants who receive advice more likely to indicate a higher level of support for Sinn Fein and the SDLP than those who do not. We find that this is indeed the case when we examine the data. First of all we find that users did tend to follow the advice provided to a statistically significant extent. Second, we did find that providing advice on the full range of policy matters did lead to an increase in support for 'outgroup' parties (i.e. it led to Catholics having a higher level of support for the DUP/UUP and Protestants having a higher level of support for Sinn Fein/SDLP). The advice also led to higher levels of support for the cross community party (Alliance) and lower levels of support for 'in-group' parties (i.e. Catholics who received advice had lower levels of support for Sinn Fein and the SDLP, and Protestants who received advice had lower levels of support for the DUP/UUP). The general point of these findings is that we demonstrate that the provision of objective and systematic advice to voters on all policy matters relating to all parties leads to a lessening of 'divide based' party support. The advice 'melts' the underlying divide somewhat: it moves the issue of voting somewhat away from merely being about unionism-nationalism and towards being about a wide range of policy matters. This movement opens up the possibility of people supporting parties that they never thought they would. Hence, we argue that our research has directly influenced in a positive way the extent to which Northern Ireland citizens are informed about party policy at election time. Also, we analysed our post-election 2016 large scale survey (number of respondents of over 4000) to shed light on the way in which citizens cast their vote at election time using the Single Transferable Voting system.The survey is uniquely well placed to enable us to examine in detail these issues given its unusually large number of respondents (over 4000). We wrote up our detailed results in a report and published it in collaboration with the Electoral Reform Society. Published Report here: http://archive.electoral-reform.org.uk/sites/default/files/2016%20NI%20Election%20Report.pdf Our analysis received significant press coverage, including front page headlines of the Belfast Telegraph. Professor Garry will also be presenting the findings from the survey, particularly as they relate to the issue of Brexit, at the following: Oct 30 2017: Presentation at Westminster, Portcullis House, organised by the ESRC's UK in a Changing Europe November 15: Invited presentation at the Foreign Office in London November 15: Public presentation at the Irish Centre in London These presentations were conducted and enabled the researchers to build up links and relations with the Foreign Office which led to subsequent invitations to liaise.
Sector Government, Democracy and Justice
Impact Types Policy & public services

 
Title Survey used in the study of voting at the 2016 Northern Ireland Assembly election 
Description Survey of 4000 respondents directly after the Northern Ireland Assembly election in 206 
Type Of Material Database/Collection of data 
Year Produced 2018 
Provided To Others? Yes  
Impact We argue that post-conflict consociational arrangements in ethnically divided societies incentivize moderation by political parties, but not policy differentiation outside the main conflict. This results in little policy-driven voting. Analysing party manifestos and voter survey data, we examine the evolution of party policy and cleavage voting under power-sharing in Northern Ireland 1998-2016. We find a reduction in ethno-national policy differences between parties and that ethno-nationalism has become less important in predicting vote choice for Protestants, but not Catholics. We also find little party differentiation in other policy areas and show that vote choices are largely independent of people's policy stances on economic or social issues. Our findings are thus largely consistent with a 'top-down' interpretation of political dynamics. 
URL http://doi.org/10.5255/UKDA-SN-8293-1
 
Description Investigating the Single Transferable Vote in Partnership with the Electoral Reform Society 
Organisation Electoral Reform Society
Country United Kingdom 
Sector Learned Society 
PI Contribution Our team wrote an analysis of how the Single Transferable Vote is used by citizens to either transfer votes or not transfer votes across ethnic lines.
Collaborator Contribution The Electoral Reform Society is committed to a greater understanding of proportional electoral system. The ERS agreed to liaise with our team in terms of publishing and disseminating our report.
Impact "The 2016 Northern Ireland Assembly Election: How Voters Use STV" written by J. Coakley, J. Garry, N. Matthews, and J. Tilley. This report was published in soft copy and attracted much publicity in the media (see above url). It will be formally launched on 23 March 2017 in hard copy form at an event in The Mac, Belfast.
Start Year 2017
 
Description Our project has published a range of media articles in Newsweek, The Conversation, The Belfast Telegraph, The Belfast News Letter and various blogs. Our report published with the Electoral Reform Society (see collaborations) attracted publicity in the Belfast Telegraph. The PI engaged in many media interviews during the 2017 Assembly election also. 
Form Of Engagement Activity A press release, press conference or response to a media enquiry/interview
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Media (as a channel to the public)
Results and Impact A large range of articles published in Newsweek, The Conversation, The Belfast Telegraph, The Belfast News Letter and various blogs.
Year(s) Of Engagement Activity 2014,2015,2016,2017
URL https://www.qub.ac.uk/sites/NIAES2016/