¹ Converse, Philip E., ‘Popular Representation and the distribution of information’, in John A. Ferejohn and James H. Kuklinski, eds, Information and Democratic Processes (Urbana and Chicago: University of Illinois Press, 1990)Google Scholar.

² Budge, Ian, Klingemann, Hans Dieter, Volkens, Andrea, Bara, Judith and Tanenbaum, Eric, Mapping Policy Preferences: Estimates for Parties, Electors, and Governments 1945–1998 (Oxford: Oxford University Press, 2001)Google Scholar; Klingemann, Hans Dieter, Volkens, Andrea, Bara, Judith, Budge, Ian and MacDonald, Michael P., Mapping Policy Preferences: Estimates for Parties, Electors, and Governments in Eastern Europe, European Union, and OECD 1990–2003: Vol. 2 (Oxford: Oxford University Press, 2006)Google Scholar.

³ McGann, Anthony, The Logic of Democracy: Reconciling equality, Deliberation and Minority Protection (Ann Arbor: University of Michigan Press, 2006)CrossRefGoogle Scholar, chap. 2.

⁴ Downs, Anthony, An Economic Theory of Democracy (New York: Harper & Row, 1957)Google Scholar; Merrill, Samuel III and Grofman, Bernard, A Unified Theory of Voting: Directional and Proximity Spatial Models (Cambridge: Cambridge University Press, 1999)CrossRefGoogle Scholar.

⁵ Adams, James, Merrill, Samuel and Grofman, Bernard, A Unified Theory of Party Competition: A Cross-National Analysis Integrating Spatial and Behavioural Factors (Cambridge: Cambridge University Press, 2005)CrossRefGoogle Scholar.

⁶ See Pascoe-Watson, Gordon, ‘No left turns for deputy PM’, Sun, 31 May 2007Google Scholar; and David Cameron speech ‘We Stand for Social Responsibility’, 1 October 2006, http://www.conservatives.com/tile.do?def=news.story.page&obj_id=132485.

⁷ May, John D., ‘Defining Democracy: A Bid for Coherence and Consensus’, Political Studies, 26 (1978), 1–14CrossRefGoogle Scholar.

⁸ Wlezien, Christopher, ‘The Public as Thermostat: Dynamics of Preferences for Spending’, American Journal of Political Science, 39 (1995), 981–1000CrossRefGoogle Scholar; Wlezien, Christopher, ‘Patterns of Representation: Dynamics of Public Preferences and Policy’, Journal of Politics, 66 (2004), 1–24CrossRefGoogle Scholar.

⁹ Soroka, Stuart N. and Wlezien, Christopher, ‘Opinion–Policy Dynamics: Public Preferences and Public Expenditure in the United Kingdom’, British Journal of Political Science, 35 (2005), 665–689CrossRefGoogle Scholar.

¹⁰ Erikson, Robert S., Mackuen, Michael B. and Stimson, James A., The Macro Polity (Cambridge: Cambridge University Press, 2002)Google Scholar.

¹¹ See Teer, Frank and Spence, James D., Political Opinion Polls (London: Hutchinson, 1973)Google Scholar; Worcester, Robert M., British Public Opinion: A Guide to the History and Methodology of Political Opinion Polling (Oxford: Basil Blackwell, 1991)Google Scholar; Worcester, Robert M., ‘Pollsters, the Press, and Political Polling in Britain’, Public Opinion Quarterly, 44 (1980), 548–566CrossRefGoogle Scholar; Worcester, Robert M., ‘The Polls: Britain at the Polls, 1945–1983’, Public Opinion Quarterly, 48 (1984), 824–833CrossRefGoogle Scholar.

¹² Sanders, David, ‘The Impact of Left–Right Ideology’, in Geoffrey Evans and Pippa Norris, eds, Critical Elections: British Parties and Voters in Long-Term Perspective (London: Sage, 1999), pp.181–206Google Scholar; Ward, Hugh, ‘Preference Shaping and Party Competition: Some Empirical and Theoretical Arguments’, in Judith Bara and Albert Weale, eds, Democratic Politics and Party Competition (London: Routledge, 2006), pp. 245–270Google Scholar.

¹³ Stimson, James, Public Opinion in America: Moods, Cycles and Swings. (Boulder, Colorado: Westview, 1999)Google Scholar; Stimson, James, Tides of Consent: How Public Opinion Shapes American Politics. (Cambridge: Cambridge University Press, 2004)CrossRefGoogle Scholar.

¹⁴ Stimson, James A., Mackuen, Michael B. and Erikson, Robert S.. ‘Dynamic Representation’, American Political Science Review, 89 (1995), 543–565CrossRefGoogle Scholar; Erikson et al. The Macro Polity.

¹⁵ Lippmann, Walter, Public Opinion (New York: Free Press, 1997), p. 127Google Scholar.

¹⁶ Kingdon, John W., Agendas, Alternatives and Public Policy (Boston, Mass.: Little Brown, 1984), p. 153Google Scholar. See also Orwell, George, Orwell: The Observer Years (London: Atlantic Books, 2004), pp. 4–5Google Scholar.

¹⁷ Zaller, John R., The Nature and Origins of Mass Opinion (Cambridge: Cambridge University Press, 1992)CrossRefGoogle Scholar. See also Stimson, James A., ‘The Micro Foundations of Mood’, in James H. Kuklinski, ed., Thinking about Political Psychology (Cambridge: Cambridge University Press, 2002), pp. 253–280CrossRefGoogle Scholar.

¹⁸ Budge, Ian, ‘Identifying Dimensions and Locating Parties: Methodological and Conceptual Problems’, in Richard S. Katz and William Crotty, eds, Handbook of Party Politics (London: Sage, 2006), pp. 422–434CrossRefGoogle Scholar.

¹⁹ Erikson et al., The Macro Polity.

²⁰ See Stimson, James A., ‘Public Policy Mood: A Personal Narrative’, Political Methodologist, 12 (2004), pp. 9–14Google Scholar, at p. 12.

²¹ Converse, Philip E., ‘Belief Systems in Mass Publics’ in David Apter, ed., Ideology and Discontent (New York: Free Press, 1964), pp. 209–264Google Scholar, at p. 209.

²² Further questions and administrations have been added to the dataset. As of 28 March 2007, it contained 481 questions. This database, together with the estimated preferences for government activity, will be deposited with the Data Archive at the University of Essex.

²³ It has been suggested by some reviewers that the meaning of the ‘middle’ or supposed ‘neutral’ response might vary over time and that this might cast some doubt on our coding case. The classic case that such reviewers would seem to have in mind would be the ‘keep things as they are’ response option that is available to respondents on various tax and spending questions. When a right-wing party is in power, for example, it might be supposed that those who select this option might be revealing conservative attitudes towards change. This might well be the case for some questions and some respondents. It may well be that we could improve our estimates were we able to incorporate such refinements. Identifying the sort of questions for which this might be a problem would however be subjective and identifying the time at which the meaning of the neutral category changes would inject a degree of subjectivity. In any event, our method relies on covariation across a range of items, many of which do not conceivably suffer from this problem. For the moment – just as with the European preference series – we prefer to maintain a degree of consistency and focus on those responses that clearly take ‘sides’.

²⁴ This excludes very few questions. British pollsters and political scientists have focused largely on domestic affairs in their questionnaires.

²⁵ Schumann, Howard and Presser, Stanley, Questions and Answers in Attitude Surveys: Experiments on Question Form, Wording and Context (Thousand Oaks, Calif.: Sage, 1996)Google Scholar; Zaller, The Nature and Origins of Mass Opinion.

²⁶ See Gallup, George H., The Gallup International Public Opinion Polls, Great Britain, 1937–1975 (London: Random House, 1976)Google Scholar, 2 volumes.

²⁷ It should be noted, however, that we are relying here on important questions about trade unions and nationalization, probably the key issues defining the political battleground of the time. These sorts of items tend to load very heavily on our estimated preference series, something that boosts our confidence in our measure for these earlier years.

²⁸ In some cases, the inclusion of the phrase ‘the government’ causes problems because it is not always clear whether the question will be interpreted as meaning the present government or government in general. We have included such questions in those cases where it seems to refer to government in general.

²⁹ Young, Hugo, This Blessed Plot (London: Macmillan, 1998)Google Scholar; Forster, Anthony, Euroscepticism in Contemporary British Politics (London: Routledge, 2002)CrossRefGoogle Scholar.

³⁰ Nairn, Tom, The Left Against Europe? (London: Penguin, 1973)Google Scholar.

³¹ This represents a heroic generalization. For exceptions, see Young, This Blessed Plot, chap. 5.

³² See Box-Steffensmeier, Janet M., Knight, Kathleen and Sigelman, Lee, ‘The Interplay between Macro-Partisanship and Macro-Ideology: A Time Series Study’, Journal of Politics, 60 (1998), 1031–1049CrossRefGoogle Scholar; Box-Steffensmeier, Janet M. and De Boef, Suzanna, ‘Macropartisanship and Macroideology in the Sophisticated Electorate’, Journal of Politics, 63 (2001), 232–248CrossRefGoogle Scholar. This paper suggests that ‘for the politically sophisticated group, there was a connection between macropartisanship and macroideology’, p. 245.

³³ Scarbrough, Elinor, Political Ideology and Voting Behaviour: An Exploratory Study (Oxford: Clarendon Press, 1984)Google Scholar.

³⁴ One might reasonably ask how we can observe the ‘median’ when we have discarded the distribution of individual responses. The answer is that we cannot. What we actually do is solve for the central tendency of preferences and then interpret that result as the preferences of the median voter.

³⁵ The algorithm is developed in Stimson, Public Opinion in America. It has been used in various studies including Cohen, Jeffrey E., ‘The Polls: Public Favourability toward the First Lady’, Presidential Studies Quarterly, 30 (2000), 575–585CrossRefGoogle Scholar; Chanley, Virginia A., Rudolph, Thomas J. and Rahn, Wendy M., ‘The Origins and Consequences of Public Trust in Government: A Time Series Analysis’, Public Opinion Quarterly, 64 (2000), 239–256CrossRefGoogle Scholar; Erikson et al., The Macro Polity, chaps 6 and 7; Kellstedt, Paul M., The Mass Media and the Dynamics of American Racial Attitudes (Cambridge: Cambridge University Press, 2003)CrossRefGoogle Scholar; Voeten, Erik and Brewer, Paul R., ‘Public Opinion, the War in Iraq, and Presidential Accountability’, Journal of Conflict Resolution, 50 (2006), 809–830CrossRefGoogle Scholar; Baumgartner, Frank R., De Boef, Suzanna L. and Boydstun, Amber E., The Decline of the Death Penalty and the Discovery of Innocence (Cambridge: Cambridge University Press, 2008)CrossRefGoogle Scholar.

³⁶ That depends, as always, on the degree to which each series is a valid indicator of the concept in question, here left–right public opinion. This is a matter for which the empirical results will be informative.

³⁷ One might think that the problem is easily solved, that the mean of a series can serve as an expected value. But this is an infinite regress problem. Without either the full series or a random sample of it, we have no valid estimate of that mean.

³⁸ The extraction algorithm can also be used to estimate a second dimension. This accounts for far less of the variation in the longitudinal variation and is more difficult to interpret. It does not, however, appear to represent the sort of liberal–authoritarian dimension or ‘tough v. tender’ dimension of the sort anticipated in the textbooks. We explore the issue of dimensionality further in Bartle, John, Dellepiane, Sebastian and Stimson, James A., ‘The Dimensionality of British Political Preferences, 1950–2005’ (paper presented at the 5th ECPR General Conference, Potsdam, 2009 – available from the authors)Google Scholar.

³⁹ Marr, Andrew, A History of Modern Britain (Basingstoke, Hants.: Macmillan, 2007), p. xiGoogle Scholar, p. 28 and p. 70.

⁴⁰ Marr, , A History of Modern Britain, p. 70Google Scholar.

⁴¹ Crewe, Ivor, Särlvik, B., Bo and Alt, James, ‘Partisan Dealignment in Britain ,1964–1974’, British Journal of Political Science, 7 (1977), 129–190CrossRefGoogle Scholar; Crewe, Ivor, ‘The Labour Party and the Electorate’, in Dennis Kavanagh, ed., The Politics of the Labour Party (London: George Allen & Unwin, 1982), pp. 9–49Google Scholar.

⁴² Donoghue, Bernard, Prime Minister: The Conduct of Policy under Wilson and Callaghan (London: Jonathan Cape, 1987), p. 191Google Scholar.

⁴³ Crewe, Ivor, ‘Has the Electorate Become Thatcherite?’ in Robert Skidelsky, ed., Thatcherism (London: Chatto & Windus, 1988), pp. 25–49Google Scholar; Crewe, Ivor, ‘Values: The Crusade that Failed’, in Dennis Kavanagh and Anthony Seldon, eds, The Thatcher Effect (Oxford: Oxford University Press, 1989), pp. 239–250Google Scholar; Crewe, Ivor and Searing, Donald D., ‘Ideological Change in the British Conservative Party’, American Political Science Review, 82 (1988), 361–384CrossRefGoogle Scholar; Heath, Anthony, Jowell, Roger and Curtice, John, How Britain Votes (Oxford: Oxford University Press, 1985)Google Scholar; Heath, Anthony, Curtice, John, Jowell, Roger, Evans, Geoffrey, Field, Julia and Witherspoon, Sharon, Understanding Political Change: The British Voter 1964–1987 (Oxford: Pergamon Press, 1991)Google Scholar.

⁴⁴ The dyad ratios algorithm has no analytic solution for standard errors of the annual estimates. We proceed instead by bootstrapping. We first introduce variation by re-sampling our raw data. Instead of the fixed percentage reported by a survey organization, ρ, we take a draw for each (non-missing) item and year from a distribution which is normal with mean ρ and a standard deviation given by the sampling error of ρ, σ = √[ρ(100 − ρ)/N]. This allows us to repeat the estimation of the latent series as many times as we wish, each sampling slightly different data and returning different results. The standard deviations for each period of all those independent replications, 1,000 in the case at hand, become the standard error estimates that we report. We have not yet been able to establish why the errors are larger towards the end of the series.

⁴⁵ Wlezien, ‘The Public as Thermostat’.

⁴⁶ The full factor loadings for the 349 items are available from the authors on request.

⁴⁷ The earlier British Election Studies also asked about the reported left–right position, but the filter question generated so much missing data that the resulting marginals were not included in our database.

⁴⁸ The latter, it should be remembered, is not included in the former.

⁴⁹ Some indication that this is the case is provided by a simple regression of current preferences and current self-reported left–right position on the same indicators lagged one year. In the model of self-reported positions, lagged self-reported left–right positions and lagged policy preferences are correctly signed and statistically significant. In the model of current preferences, both lagged preferences and lagged self-reported left–right positions are significant, but the latter is incorrectly signed, suggesting that if they were rising one year ago, preferences fall. Self-reported left–right positions: Constant = −3.23 (2.95)*, Lagged preferences = 0.82 (0.08), Lagged self-reported positions = 0.26 (0.06), Adjusted R ² = 0.93, N = 31; Public preferences: Constant = 3.73* (2.47), Lagged preferences = 1.09 (0.07), Lagged self-reported positions = −0.16 (0.05), Adjusted R ² = 0.93, N = 31 (*Not significant at p < 0.05, two-tailed).

⁵⁰ Webb, Paul, The Modern British Party System (London: Sage, 2000)Google Scholar, chap. 3; Evans, Geoffrey, ‘Europe: A New Cleavage?’ in Geoffrey Evans and Pippa Norris, eds, Critical Elections: British Parties and Voters in Long-term Perspective (London: Sage, 1999), pp. 207–222CrossRefGoogle Scholar.

⁵¹ See Carmines, Edward G. and Stimson, James A., Issue Evolution: Race and the Transformation of American Politics (Princeton, N.J.: Princeton University Press, 1989)Google Scholar; Stimson, Tides of Consent, chap. 3.

⁵² See Davies, Evan, Public Spending (London: Penguin, 1998)Google Scholar.

⁵³ Since Britain has three major parties, not two, we see the Liberal Democrats complicating the picture by sometimes falling between the other two parties and sometimes emphasizing different issues. We ignore the complication.

⁵⁴ See Feldman, Stanley and Zaller, John R., ‘The Political Culture of Ambivalence: Ideological Responses to the Welfare State’, American Journal of Political Science, 36 (1992), 268–307CrossRefGoogle Scholar.

⁵⁵ Wlezien, , ‘The Public as Thermostat’, pp. 985–9Google Scholar.

⁵⁶ Wlezien, ‘The Public as Thermostat’.

⁵⁷ A Dickey–Fuller test indicates the presence of a unit root in the preferences series. So, henceforth, we will work with first differences in the series or analogously employ an error correction formulation which is appropriate for integrated time series.

⁵⁸ Kavanagh, Thatcherism and British Politics, chaps 3–4.

⁵⁹ Jenkins, Mrs Thatcher's Revolution, chap. 1.

⁶⁰ Ponting, Clive, Breach of Promise (London: Hamish Hamilton, 1989)Google Scholar.

⁶¹ We have also considered analysis of the longer period, 1965–2005. That produces a compromise result where the effects are also highly significant, but less sharp than the shorter period of Table 3.

⁶² Enders, Walter, Applied Econometric Time Series (Hoboken, N.J.: Wiley, 2004)Google Scholar, chap. 6; Wooldridge, Jeffrey M., Introductory Econometrics: A Modern Approach (Mason, Ohio: Thomson, South Western, 2006)Google Scholar, chap. 18; Kennedy, Peter, A Guide to Econometrics (Oxford: Blackwell, 2008)Google Scholar, chap. 19.

⁶³ Short-term relationships between Δx and Δy may also be observed, but are relatively uncommon with unit root dependent series.

⁶⁴ We were unable to find a single measure of government expenditure as a proportion of GDP for the whole period 1950–2005. Therefore, we collected a series of overlapping indicators from a variety of sources and used the extraction algorithm to estimate a single series. The longest single-source period was found in White, G. and Chapman, H., ‘Long-term Trends in Public Expenditure’, Economic Trends, 408 (1987), pp. 124–128Google Scholar; but we also used HM Treasury, OECD and Eurostat. These resulting series explained: 86.85 per cent of the variation in the data had a mean of 39.64 and a standard deviation of 3.42. The complete original dataset, together with the extraction output, is available from the authors on request.

⁶⁵ The measure of average income tax is based on analyses carried out by Frances Lynch and her colleagues at the University of Westminster. It measures average income tax levels for specified ‘typical’ individuals (see Johnson, Paul, Lynch, Frances and Geoffrey Walker, John, ‘Income Tax and Elections in Britain, 1950–2001’, Electoral Studies, 24 (2005), 393–408CrossRefGoogle Scholar). One of the most striking features of debates about tax in Britain is the extent to which debate resolves around income tax, even though it generates a relatively small proportion of government revenue. We have been provided with the data for male single industrial workers only. It is clear from Figure 2 in Johnson et al., ‘Income Tax and Elections in Britain’, that this correlates very well with average tax levels for married industrial workers too. These estimates are available only up to 2001, with the consequence that N falls to 51.

⁶⁶ We have two alternative paths by which preferences might be related to levels of government spending and taxing. In one, preferences at earlier times affected election results and altered the control of government. So current spending may be a positive function of preferences at various lags. The thermostatic hypothesis introduces a different path. Current electorates react to the current and previous size of government mix negatively. As government grows, they move to the right in response. As it declines, they move to the left. So when we observe a statistical connection between them, are we observing a web of endogeneity or can the two be sorted out? They differ in two important respects, which is helpful to sorting out causality claims. In timing, the logic requires the first path to associate changes in government only with lagged values of preferences. It is previous elections that put the current government in power. The thermostatic effect is the reverse; the electorate cannot react to changes in government until those changes have occurred, and therefore current preferences should reflect only previous policy. The second difference is the important matter of sign. One relationship, preferences → elections → policy, is positive; the other, policy → (−) preferences, is negative. Thus, the data can help sort out alternative interpretations. The sign and time order that we observe supports only the second, thermostatic, interpretation. We cannot rule out that the first path might exist. But if it does, it is too weak to leave behind evidence.

⁶⁷ Ivor Crewe, ‘Now relax, it's a dead cert’, New Statesman, 12 February 2001. See also King, Anthony, ‘Why Labour Won – At Last’, in Anthony King, ed., New Labour Triumphs: Britain at the Polls, 1997 (Chatham, N.J.: Chatham House, 1998), pp. 177–207Google Scholar.

⁶⁸ Stimson, , Public Opinion in America, p. 98Google Scholar.

⁶⁹ See Erikson et al., The Macro Polity, chap. 7.

⁷⁰ Butler, David, ‘Reflections on British Elections and Their Study’, Annual Review of Political Science, 1 (1998), 451–464CrossRefGoogle Scholar, p. 455.

⁷¹ Labour actually obtained a lower share of the vote in February 1974 but obtained 301 seats to the Conservatives’ 296. It obtained a tiny majority of just five seats in the subsequent election of October 1974.

⁷² Newton, Kenneth, ‘Caring and Competence’, in Anthony King, ed., Britain at the Polls, 1992 (Chatham, N.J.: Chatham House, 1992)Google Scholar.

⁷³ Sanders, David, ‘The New Electoral Battleground’, in Anthony King, ed., New Labour Triumphs: Britain at the Polls (Chatham, N.J.: Chatham House, 1998), p. 217Google Scholar.

⁷⁴ Toynbee, Polly and Walker, David, Did Things Get Better? An Audit of Labour's Successes and Failures (London: Penguin Books, 2001), p. 104Google Scholar.

⁷⁵ Information provided by Judith Bara (Queen Mary and Westfield College, University of London).

⁷⁶ See Bartle, John, The New British Politics: Election Update (Harlow: Longman, 2005), p. 22Google Scholar.

⁷⁷ King, Gary, ‘Replication Replication’, PS: Political Science and Politics, 3 (1995), 443–499Google Scholar.

⁷⁸ Erikson et al., The Macro Polity, chaps 6 and 7.

⁷⁹ Wlezien, ‘The Public as Thermostat’ and ‘Patterns of Representation’.

⁸⁰ Paldam, Martin, ‘The Distribution of Election Results and Two Explanations of the Costs of Ruling’, European Journal of Political Economy, 2 (1986), 5–24CrossRefGoogle Scholar; Paldam, Martin and Skott, Peter, ‘A Rational Voter Explanation of the Cost of Ruling’, Public Choice, 83 (1995),159–172CrossRefGoogle Scholar; Naanstead, Peter and Paldam, Martin, ‘The Costs of Ruling’, in Han Dorussen and Michael Taylor, eds, Economic Voting (London: Routledge, 2002), pp. 17–44Google Scholar.

⁸¹ Stimson, Public Opinion in America and Tides of Consent.

⁸² Erikson et al., The Macro Polity, p. 5.

⁸³ Converse, ‘Popular Representation and the Distribution of Information’.

⁸⁴ Bartle, John, Dellepiane, Sebastian and Stimson, James A., ‘The Dimensionality of British Political Preferences, 1950–2005’ (paper presented at the 5th ECPR General Conference, Potsdam, 2009)Google Scholar.