Dr. Edzia Carvalho, Postdoctoral Researcher, Department of Political Science, University of Amsterdam, O.Z. Achterburgwal 237, 1012 DL Amsterdam, The Netherlands edziacarvalho@gmail.com Co-author: Dr. Kristi Winters, Datenarchiv für Sozialwissenschaften, GESIS Leibniz Institutes for the Social Sciences, Unter Sachsenhausen 6-8, 50667 Köln, Germany Kristi.Winters@gesis.org This research was funded by the British Academy, grant number SG090860.
Where the Lib Dems started and the first debate On 6 April 2010, 17 to 21 percent of respondents polled expressed a Liberal Democrat vote preference (Populus 2010; YouGov 2010). The first debate was held one week into the campaign and drew an audience of approximately 9.7 million people. Clegg stole the first leaders debate: his performance was fresh and he was viewed as a serious player (Wintour and Curtis, 2010; Sergeant, 2010, Stringer and Allman, 2010). All post-debate opinion polls reported that Clegg had won the first round (Lawes and Hawkins 2011, 62).
Observing Cleggmania live See the paper at Academia.edu for exciting real time reactions to Cleggs performance in the First Leaders Debate!
Cleggmania emerges Following the debate Liberal Democrat support in the polls rose from 17 to 31 percent (Lawes and Hawkins 2011, 68). In the media comparisons were made between Clegg and Che Guevara, Sir Winston Churchill, and Barack Obama (Burkeman 2010; Oliver and Smith 2010). There was speculation the Liberal Democrats might overtake Labour to become the main opposition party (Deacon and Wring 2011, 287).
The election results The results of the 2010 general election were disappointing for the Liberal Democrats. The party won 23 percent of the vote, just one point better than its share in the previous general election (Atkinson and Mortimore 2011, 78). The six BPC polls had overestimated the party s vote share by an average of 3.6 percentage points (Kellner et al 2011, 95). Bartle et al (2011, 148) observed the three parties ended up where they began the campaign.
The puzzle and possible answers Why did Cleggmania fail to live up to its expected results? Two accounts have been provided to explain: 1) The polls got it wrong, either because they measured support that did not exist or they failed to measure a decline in support (Atkinson and Mortimore 2011, 78). 2) Over-emphasis by the media on the effect of the Leaders Debates (Allen et al 2011) while ignoring the constraining factors, such as the limited number of viable seats Liberal Democrats could win, that prevented any bounce from translating into seats (Lawes and Hawkins 2011; Johnston and Pattie 2011).
Our analysis Our approach to the Cleggmania puzzle uses unique data: the perceptions and self-reported actions of voters themselves. Qualitative data allows researchers to examine the British general election as self-reflexively understood by the participants. The Qualitative Election Study of Britain (QESB) provides highquality textual data to analyze how people use language in their everyday interactions, their discourses with each other, and how they put their linguistic skills to use in building specific accounts of events, (Burr 2003, 17).
Qualitative Election Study of Britain Data We analysed transcripts from post-election focus groups conducted between May 18 and 24, 2010 in England (Essex and London), Scotland (Glasgow) and Wales (Aberystwyth). 40 participated, 30 of whom voted. Participants were invited to recount what happened leading up to Election Day and if, and for whom they voted. Vote choice narratives (and relevant side mentions) were identified in the transcripts, extracted, and prepared for analysis.
Qualitative analysis Our working assumption was that since participants knew how they had voted they would construct their vote choice story both to recount and to explain (or justify) their actions. We employed two methods of qualitative analysis, narrative and discourse, respectively and iteratively.
Narrative analysis Narrative analysis is a holistic approach that preserves the context and particularity of the data (Riessman 1993). The temporal ordering of events in a story is a necessary but not sufficient condition for the emergence of a story. The events in the sequence must be bound together by some principles of logical coherence (Franzosi 1998, 521). Stories draw upon subjective experiences and provide insights into concepts of identity and self (Smith 2000). Narrative analysis examines how the story-teller interprets things (Bruner 1990).
Narrative structure We identified the following elements that could be applied to all vote choice narratives: Scene setting: Introduction and/or background for the narrative; Dilemma: a dilemma or complicating factor; Turning point: event(s) important to heighten or resolve the dilemma(s); Resolution: A resolution to the dilemma; Justification: explanations or rationalization; Outcome: the outcome of their narrative; Reflections: further thoughts on the outcome Coda: a short re-statement, may include a verbal link to the beginning or to the present.
Qualitative data analysis Original text Diane: I didn t know who I was going to vote for. I liked the Lib-Dems and thought -is it going to be a wasted vote? Then I saw the debate, which was here, and that made up my mind really who I did want to vote for which was Lib-Dem and I don t think they could have asked for a better result really because they- there was no way they were going to get voted in anyway so for them that s the best outcome. Narrative analysis applied to text Diane: [Dilemma] I didn t know who I was going to vote for. [Dilemma] I liked the Lib-Dems and thought -is it going to be a wasted vote? [Turning point] Then I saw the debate, which was here, [Resolution] and that made up my mind really who I did want to vote for which was Lib-Dem [Reflection] and I don t think they could have asked for a better result really because theythere was no way they were going to get voted in anyway so for them that s the best outcome.
Discourse analysis Once organized, the content was analysed employing a form of discourse analysis. We applied Gee s (2008) principles to identify the common and unique content. We analysed the ways participants used language to make things significant (or not), how they used language to enact identities, the social goods they communicated in their stories, and how they connected or disconnected things in their descriptions to make them relevant or irrelevant to each another (Gee 2008, 9-13).
Qualitative data analysis Narrative and Discourse analysis applied to text Diane: [Dilemma: Undecided] I didn t know who I was going to vote for. [Dilemma: A wasted vote?] I liked the Lib-Dems and thought -is it going to be a wasted vote? [Turning point: Saw debate] Then I saw the debate, which was here, [Resolution: Will vote Lib Dem] and that made up my mind really who I did want to vote for which was Lib-Dem [Reflections: Happy with result] and I don t think they could have asked for a better result really because they- there was no way they were going to get voted in anyway so for them that s the best outcome.
Higher order analysis Narratives were examined for patterns in structures or language. Differentiated by the elements of the story, how they were ordered, the values cited, and how dilemmas were resolved. We create criteria of organization for vote choice typologies based upon Dilemma or not, and Content of the justifications.
Four narrative types Through this process we identified four typologies: 1) Started loyal, stayed loyal voters, 2) Dated Nick but it didn t stick * voters, 3) Won-over voters, and 4) Strategic voters. The remaining 5 narratives were collected under the category Other. *The dated Nick category pays homage to the 2004 Democratic primary s Dated Dean, Married Kerry.
How they hang together Three typologies are characterised by two factors: where participants fell on partisanship (partisans, not partisans) and vote choice (decided or undecided). Started loyal voters generally had preferred party or other values that underpinned their vote choice and did not waver in their vote choice. Dated Nick voters expressed a prior preferred party but described their vote choice as a dilemma. Won over voters did not express a preferred party or value and described their vote choice as a dilemma. (Limited to LD voters for this analysis)
Strategic voters The fourth typology has unique dimensions that distinguish it from the other three. Strategic voter is characterised by a common vote for the Liberal Democrats but their justifications are based in maximising personal value. Our analysis identified three types of strategic voters: Tactical- Stop another party Satisficing - A viable party that is politically proximal Principle/policy - Expressive voting for a principle or policy
Typologies of QESB vote choice narratives and their characteristics Typology Characteristics 1. Started loyal Partisanship or value, Decided 2. Dated Nick Partisanship, Undecided 3. Won Over No partisanship, Undecided 4. Strategic Tactical Stop another party Satisficing A viable party that is politically proximal Principle/policy Expressive voting for a principle or policy
Started Loyal, Stayed Loyal Name (Vote) Structure John (Lab) O: Knew who to vote for, Ss: Personal history, J: Trade unionist family, J: Thatcher, J: Been Labour ever since, C: Walked in knowing Geoff (Lab) O: Voted Labour, J: Labour from the start Jody (Con) J: Brown s failings, J: Waste in public sector, Ss: No way vote Labour, O: Voted Con, J: Always going to vote Con Anthony (Con) O: Voted Tory, J: Tory supporter, Ss: No problem to put the X, J: Campaigned for them George (PC) J: Support my country, J: Policies I agree with, J: Looking after vulnerable, O: It didn t go our way Rita (Green) Ss: Green involvement, O: Vote Green, J: Shift in politics, J: Seems more positive Helen (LD) Ss: Was easy for me, J: Lib Dem majority, J: MP did good work, C: Easy for me Andrea (LD) O: Voted Lib Dem, Ss: Never wavered, D: Tired after work, seat so safe, a wasted vote? C: But I voted
Analysis of SLSL voters These narratives lack dilemmas or any sense of drama. The narratives simply report voters doing what they intended to do. Key lies in the justifications (J) given. Non-Lib Dem voters: cite own values and linked to their chosen party influenced through a trade unionist father, support my country, a more positive vote. Lib Dem voters: do not mention values or identity, both mention constituency dynamics in their narratives.
Narrative elements of DNDS voters Name (Vote) Ida (Lab) Structure D: Between Lab / Lib Dem, O: I voted Labour, J: My preferred party Roger (Lab) Ss: Between Lab / Lib Dems, D: Not used to decisions, O: Voted Labour, Tp: How will I feel? Cathy (Con) D: Undecided, Ss: Always voted Con, D: Could have been swayed, D: Liked Clegg, D: Strong Con constituency, D: Had to decide, O: Voted Con, J: Always voted that way Shirley (Con) Ss: Was going to vote LD, Ss: Family Tory, Ss: Always voted Tory, Ss: Give LD a chance, Tp: LD won t get in, Tp: Tory guys outside, O: Voted Con, Res: LD won t do it, I ll vote Tory Patricia (Con) Ss: Open-minded, Ss: Clegg good in debate, D: Lost confidence in Clegg, Res: Examine leaders, J: Gordon had his chance J: Went for best leader
Analysis of DNDS voters Recount and resolve the voter s dilemma. Start with setting scene (Ss) or dilemma (D) - undecided. Roger and Shirley report being undecided as they walked into the voting booth. Insights are found in the justifications (J), and the resolutions (Res). How will I feel when I see the results come in? just something in me was just like I ve always voted that way.
Identity is powerful For wavering QESB voters a sense of partisan identity overrode other considerations. We cannot transfer the rationales from participants to survey respondents. However this provides an additional account for the overestimation of Liberal Democrat support: People may have sincerely thought they would vote Liberal Democrats but when faced with the reality of the ballot box they could not bring themselves to vote against their party.
Elections are not surveys We feel that this key evidence from people s own experiences provides another piece to the puzzle as to why Cleggmania did not materialize. Shirley and Roger are examples of persons who could have reported a Liberal Democrat voting intention until the day before the election, but on the day they cast their votes for the same party they always had. The act of casting one s vote carries a unique weight in that moment that cannot be captured by survey questionnaires or lab experiments.
Won over to Lib Dem voters Name (Vote) Nicole (LD) Diane (LD) Stacey (LD) Ian (LD) Structure D: Undecided, O: Voted Lib Dem, Ss: Safe Tory seat, D: Vote won t mean much, J: Agreed with LD principles, J: Liked Clegg, J: Show of support D: Undecided, D: Wasted vote? Tp: Leaders debate, Res: That really made up my mind, O: Voted Lib Dem D: Undecided, Ss: Not the Tories, Tp: Leader debate, Tp: Impressed with Clegg, Ss: Considered Labour, J: Constituency dynamics, O: Voted Lib Dem, Ref: Didn t make a difference O: Voted Lib Dem, J: I agreed with them, J: Not Brown, J: Dad s influence, J: Not Cameron Nathan (LD) O: Voted Lib Dem, Ss: Happy about debates, J: Disillusioned, J: Not Labour or Tories, J: Lib Dem hype, Tp: Could vote and not waste it
Analysis of Won over voters Concerns about casting a wasted vote, or their vote not meaning very much comes up in three of the four narratives. These stories provide insights into the decisions of people without strong partisanship who used many pieces of information to arrive at a decision. Cleggmania was one piece of information that helped shore up support with wavering Lib Dem voters.