"UBC POLI 385 Public Opinion Elections and Representation What is"
UBC POLI 385 Public Opinion, Elections, and Representation Models of Voting and Party Identification 1 What is a model of voting? Why do we need one? Takes all the potential influences on voting and specifies theoretical relationships among them Makes predictions about the relative strength of influences Tells us the order of social and psychological influences Tells us what influences might depend on others – e.g. do evaluations of leaders matter more for young voters? Tells us how to look for change versus stability, long‐term change versus short‐term change All this is partly to explain, partly to predict Models also guide us in looking for measurements: What variables are we trying to measure? So what questions do we ask survey respondents? 2 What are the determinants of voting? Write down as many influences as you can think of Number them in the (temporal) order that they would influence a person Then draw a diagram with arrows connecting them, indicating the causal relationships – arrows can flow both ways (you should be able to say why) Don’t worry about what you call these things, I’m interested to know what you call them 3 Three Models of Voting Sociological – just identifies that different groups behave differently in politics – doesn’t really illuminate the reasons for group differences Socio‐Psychological – all possible influences, with timing specified • room for long‐standing social influences • and more ‘proximate’ short‐term influences • allows rational cognition and emotion Rational Choice – Downs’ model of voting – Acknowledges information is a costly resource for decision‐making – But it should be entirely future‐oriented, using the past only as information about likely future performance – Mainly policy‐based: therefore a complicated function of ‘policy distance’ from parties and the importance of each policy area 4 Models of Voting: The “Funnel of Causality” 5 UBC POLI 385 Public Opinion, Elections, and Representation Party Identification http://www.politics.ubc.ca/fcutler/teaching/POLI385 6 Don’t hesitate to make your papers historical For example, you can discuss the CA/PC = Conservative Party merger. But don’t get carried away with the news; instead, use it to discuss party identification. – in particular, the contrast between old and new parties! Read 165‐ in Green, Palmquist, and Schickler in particular – but you should read the foregoing stuff to understand it! 7 “Unmoved?” Progressive Conservative / Conservative Party Identification 40 37 35 30 29 26 26 25 24 22 21 20 19 19 16 15 14 10 7 5 0 5 8 4 9 4 8 2 3 7 0 4 6 196 196 197 197 198 198 199 199 199 200 200 200 8 What is Party Identification (PID)? An ‘affective’, emotional orientation to political parties – learned/formed early in life – fairly stable (80%+ in USA 1956‐58‐60, 1972‐74‐76) A ‘habit of the heart’ – like other components of identity – ‘minimal group theory’ would predict it A cognitive shortcut? (Not in Canada, because we vote for parties) A direct predictor of vote choice – a ‘standing decision’, ‘brand loyalty’, or ‘shortcut’ useful if there is • unreliable or conflicting new political information • or a lack of new political information 9 Cognitive Effects of Party Identification (PID)? It mediates perception of most political information – Call this: filtering, projection, screening, selective perception, etc. – Identifiers adopt the party’s position on new issues • so NDP‐ers would be against an Iraq war… but British Labour party suppoters would not (this would be stronger the more obscure the policy) – e.g. Strong Conservative partisans didn’t see as much recession in the early 1990s as non‐partisans or other party identifiers – VERY POWERFUL MEDIATING EFFECTS – Note Green, Palmquist and Schickler’s perspective on this – They say it’s not too strong 10 The Classic American Theory of Partisanship Original formulation was social or socio‐psychological – Campbell, Converse, Miller, Stokes in The American Voter Learned early in life, based in emotion/symbols Because it is part of identity, political attitudes are made to be consistent with it (reducing cognitive dissonance) Strengthens with age Influences attitudes, but attitudes don’t influence it “the unmoved mover?” Such a good predictor of voting that other short‐term forces could only get to second place in terms of influence When added up in a country, it provides a baseline, “normal vote” – short‐term forces act on this baseline But in every other democracy, PID seemed to change along with voting much more than in the USA 11 The Revisionist American Theory of Partisanship Since PID does change, what changes it? If it changes on the basis of performance evaluations, leader evaluations, or issue positions, then it isn’t what we thought Some of it is real, long‐standing, affective PID Some of it is a ‘running tally’ of what parties have done PIDcurrent = PIDlast election + β (Retrospective Evaluations) + error So PID is last PID plus or minus current political factors This means that lots of real political events/changes are incorporated in it over time Some people have long memories, some have short ones This theory has some bite – e.g. btween ’72 & ’74, Nixon Pardon has more effect than Ford Performance ’74‐’76 Ford Performance more influential than Nixon Pardon 12 But Can We Measure it? In Canada we ask: “In FEDERAL politics, do you usually think of yourself as a Liberal,Conservative, N.D.P, or none of these?” (now Green, also Reform, Alliance, etc.) So is “usually think of yourself as a” the right question for getting at a long‐standing emotional orientation? It’s that question that generates: – 85% of those with a Party ID voted for the party in 2000 – Identifiers 40% more likely to vote for their party than non‐identifiers, all else equal 13 Importance of Party Identification for Voting 14 Party ID 1992 (no election) and Voting in 1993 (percentages to be read across rows – e.g. 71% of those who were Liberal in 1992 voted Liberal in 1993) PARTY VOTED FOR NOVEMBER 1993 lib pc ndp rfm bq other none d.k. refused total liberal 71 7 2 9 5 3 0 1 3 28 federal pc 19 39 2 25 9 2 0 1 2 19 party n.d.p. 32 4 36 14 7 5 0 0 2 11 affiliation| reform 9 4 4 80 0 2 0 0 0 4 Sept. 1992 bloc 4 4 1 0 85 1 1 0 4 9 none 32 12 3 15 26 5 1 0 6 24 d.k. 40 6 0 17 26 2 0 0 9 4 refused 25 0 6 0 25 0 0 0 44 1 Total 37 13 6 16 19 0 0 4 100 15 PID anchors the ideological / competitive landscape 16 Is it too good? Too ‘close’ to the voting decision? But is it too good ? – if people think the question is asking about the present – or if they are worried about appearing inconsisent – … then the question will simply reflect current voting intention and we shouldn’t use it for explaining vote choice! 15% of identifiers do vote against (defect from) that party In unusually volatile electoral situations, more defection: Like 1993: 26% PC: 19% NDP: 7% Reform: 5% BQ: 7% % of Electorate: Liberal 26% - PC: 19% - NDP: 7% - Reform: 5% - BQ: 7% PC: Loyalty Rate: Liberal 80% - PC: 40% - NDP: 52% - Reform: 86% - BQ: 92% NDP: Reform: BQ: The last two should indicate that we’re not measuring something immutable, “learned at mother’s knee” – People “usually think of themselves as” Reformers and BQ? NEW PARTIES? But it is still by far the most powerful determinant of voting 17 How Stable is PID in Canada? Periods of Stability vs. Periods of Instability (new parties) 18 Aggregate Stability in Canadian PID 1965‐1988 PID in Canada 1965-1988 PC Liberal NDP Other/None 50 45 40 35 30 25 20 15 10 5 0 1965 1968 1974 1979 1980 1984 1988 19 Party Identification and Socialization Jennings, Stoker, & Bowers Interviews of parents (1965) and children (1973, 1982, 1997) and those children’s children (1997) Partisanship & Vote Choice are the political attitudes most similar between parents and children at all time periods – stronger than church attendance and view of bible! “the highest concordance tends to be on objects that are more concrete, salient, long- lived, and affect laden. Accordingly, measures involving partisanship, religion, race, and cognition lead the way.” (p.9) What Increases Transmission? Politicization & Parent Attitude Strength (Consistency/Stability) “When the parent's attitudes are unstable, transmission is weak or nonexistent. But when they are strong and fixed, transmission rates are high, often dramatically so.” Those with strong transmission (high initial correspondence) were 20% more likely to show stability in PID (62% to 41%) ALL THIS APPLIES TO THE ‘PROTEST GENERATION’ & ‘GENERATION X’ 20 Party Identification in a Federal System Canadians have “Two Political Worlds” to deal with Do we think they should identify with the same party? Surely a good deal of inconsistency is a GOOD thing – it tells us that there is some separateness in the political ‘worlds’, and people are dealing with that adequately! Nearly half of identifiers do not identify with the same party Are parents transmitting split‐level identifications or is it evidence of the revisionist, ‘evaluative’ theory of PID? The research shows that Canadians do update PID at one level based on evaluations of parties at the other level. – Does this make sense in BC? For which voters? – Example of the Ontario NDP and NDP support in 1993! 21 Cross‐Level Consistency in Identification? 22 Party ID and Anti‐Partyism Originally thought that Party ID was natural… those without it had particular anti‐party feeling But might it be better to have fewer partisans – more people making up their minds each time with new evidence Pro: – Party ID gives an accurate standing decision – PID mobilizes people to vote – PID gives the system some stability – PID is a link to the country’s social landscape Con: – Party ID can generate ‘mistakes’ – people voting against self‐interest – It would be better to base evaluations on current information – Allows parties to lead public opinion 23 No real drop in PID in Canada 1965‐1997! 24