¹ Downs, Anthony, An Economic Theory of Democracy (New York: Harper and Row, 1957)Google Scholar; Davis, Otto A., Hinich, Melvin J. and Ordeshook, Peter C., ‘An Expository Development of a Mathematical Model of the Electoral Process’, American Political Science Review, 64 (1970), 426–48CrossRefGoogle Scholar; Enelow, James and Hinich, Melvin J., The Spatial Theory of Voting: An Introduction (New York: Cambridge University Press, 1984).Google Scholar

² Rabinowitz, George and Macdonald, Stuart Elaine, ‘A Directional Theory of Issue Voting’, American Political Science Review, 83 (1989), 93–121CrossRefGoogle Scholar; Macdonald, Stuart Elaine, Listhaug, Ola and Rabinowitz, George, ‘Issues and Party Support in Multiparty Systems’, American Political Science Review, 85 (1991), 1107–31CrossRefGoogle Scholar; Macdonald, Stuart Elaine and Rabinowitz, George, ‘Direction and Uncertainty in a Model of Issue Voting’, Journal of Theoretical Politics, 5 (1993), 61–87.CrossRefGoogle Scholar

³ See Rabinowitz, and Macdonald, , ‘A Directional Theory of Issue Voting’Google Scholar; Macdonald, , Listhaug, and Rabinowitz, , ‘Issues and Party Support in Multiparty Systems’.Google Scholar

⁴ Converse, Philip E., ‘The Nature of Belief Systems in Mass Publics’, in Apter, David E., ed., Ideology and Discontent (New York: The Free Press of Glencoe, 1964), pp. 206–61.Google Scholar See, for example, Stimson, James A., ‘Belief Systems: Constraint, Complexity, and the 1972 Election’, American Journal of Political Science, 19 (1975), 393–417CrossRefGoogle Scholar; Knight, Kathleen, ‘Ideology in the 1980 Election: Ideological Sophistication Does Matter’, Journal of Politics, 47 (1985), 828–53CrossRefGoogle Scholar; Jacoby, William, ‘Levels of Conceptualization and Reliance on the Liberal-Conservative Continuum’, Journal of Politics, 48 (1986), 223–32CrossRefGoogle Scholar; Lodge, Milton and Hamill, Ruth, ‘A Partisan Schema for Political Information Processing’, American Political Science Review, 80(1986), 505–19CrossRefGoogle Scholar; McGraw, Kathleen, Lodge, Milton and Stroh, Patrick, ‘On-Line Processing in Candidate Evaluation: The Effects of Issue Order, Issue Importance, and Sophistication’, Political Behavior, 12 (1990), 41–58CrossRefGoogle Scholar; Sniderman, Paul, Brody, Richard A. and Tetlock, Philip E., Reasoning and Choice: Explorations in Political Psychology (Cambridge: Cambridge University Press, 1991).CrossRefGoogle Scholar For a dissenting view, see Rahn, Wendy, Aldrich, John H., Borgida, Eugene and Sullivan, John L., ‘A Social-Cognitive Model of Candidate Evaluation’, in Ferejohn, John and Kuklinski, James, eds, Information and Democratic Processes (Urbana: University of Illinois Press, 1990), pp. 136–59.Google Scholar

⁵ See Davis, , Hinich, and Ordeshook, , The Spatial Theory of VotingGoogle Scholar; Rabinowitz, and Macdonald, , ‘A Directional Theory of Issue Voting’.Google Scholar

⁶ The development of intensity as an explicit mathematical function of certainty and concern appears in Macdonald and Rabinowitz, ‘Direction and Uncertainty in a Model of Issue Voting’. The view that voters are often uncertain about their own issue preferences is a novel feature of directional theory. Recent work by Zaller and Feldman offers a psychological theory and some empirical evidence in support of this view. See Zaller, John and Feldman, Stanley, ‘A Simple Theory of the Survey Response: Answering Questions versus Revealing Preferences’, American Journal of Political Science, 36 (1992), 579–616.CrossRefGoogle Scholar

⁷ The minus sign means that utility is negatively related to distance. The farther a party is from a voter, the less the voter will like the party.

⁸ See Tversky, Amos, ‘Features of Similarity’, Psychological Review, 84 (1977), 327–52, at pp. 327–9.CrossRefGoogle Scholar

⁹ The symbol I in Equation 2 and the symbol 0 in Equation 1 both represent issue position. Different symbols are used because the interpretation of issue position is quite different in the two theories. In the analysis section of the article we will use the neutral symbol x to indicate position on an issue.

¹⁰ Because P_ij is subscripted by both individual and party, the theory allows the extent of penalty to vary across voters based on the level of agreement between the voter and the party. While varying penalties are intuitively appealing, empirical applications to date have assumed a uniform penalty across voters.

¹¹ The data are publicly archived. The 1988 American National Election Study is available from the ICPSR, PO Box 1248, Ann Arbor, MI 48106. The 1989 Norwegian Election Study is available from the Norwegian Social Science Data Service, Hans Holmboes gate 22, N-5007 Bergen, Norway.

¹² The questions dealt with liberal-conservative ideology, level of government services, defence spending, government-sponsored health insurance, government guarantee of jobs and standard of living, relations with Russia, women's appropriate role, government responsibility to aid blacks and government responsibility to aid minorities. The last two questions were asked to split-halves of the sample. The exact wording is available in the NES codebook.

¹³ The questions dealt with left-right ideology, agricultural support, environmental protection, immigration, health care, availability of alcohol and treatment of criminals. A translated version of each question appears in Appendix I.

¹⁴ See Shaffer, William, ‘A Congruence Model of Issue Voting’ (paper presented at the Annual Meeting of the Midwest Political Science Association. Chicago, 1993)Google Scholar; Pierce, Roy, ‘Directional Versus Proximity Models: A Second Opinion’ (paper presented at the Annual Meeting of the American Political Science Association. Washington, DC, 1993)Google Scholar; Merrill, Samuel, ‘Discriminating between the Direction and Proximity Spatial Models of Electoral Competition’, Electoral Studies (forthcoming)Google Scholar; Gilljam, Mikeal, ‘The Directional Theory under the Magnifying Glass: A Reappraisal’ (unpublished, Göteborg University).Google Scholar

¹⁵ See, for example, Brody, Richard A. and Page, Benjamin I., ‘The Assessment of Policy Voting’, American Political Science Review, 64 (1972), 450–8CrossRefGoogle Scholar; Kinder, Donald R., ‘Political Person Perception: The Asymmetrical Influence of Sentiment and Choice on Perceptions of Presidential Candidates’, Journal of Personality and Social Psychology, 36 (1978), 859–71CrossRefGoogle Scholar; Markus, Gregory B. and Converse, Philip E., ‘A Dynamic Simultaneous Equation Model of Electoral Choice’, American Political Science Review, 73 (1979), 1055–70CrossRefGoogle Scholar; Granberg, Donald and Holmberg, Soren, The Political System Matters (Cambridge: Cambridge University Press, 1988).Google Scholar

¹⁶ See, for example, Markus, Gregory B., ‘Political Attitudes during an Election Year: A Report on the 1980 NES Panel Study’, American Political Science Review, 76 (1982), 538–60CrossRefGoogle Scholar; Budge, Ian and Farlie, Dennis J., Explaining and Predicting Elections (London: Allen & Unwin, 1983)Google Scholar; Shanks, J. Merrill and Miller, Warren E., ‘Policy Direction and Performance Evaluation: Complementary Interpretations of the Reagan Elections’, British Journal of Political Science, 20 (1990), 143–235CrossRefGoogle Scholar, and ‘Partisanship, Policy and Performance: The Reagan Legacy in the 1988 Election’, British Journal of Political Science, 21 (1991), 129–97Google Scholar; Abramson, Paul R., Aldrich, John H. and Rohde, David W., Change and Continuity in the 1992 Elections (Washington, DC: Congressional Quarterly Press, 1994).Google Scholar

¹⁷ A third reason could also be given. Recent experimental work in political psychology finds that individual perceptions of candidate positions recalled from memory are often distorted and do not predict as well as the actual stands the candidates take. See Lodge, Milton, McGraw, Kathleen and Stroh, Patrick, ‘An Impression-Driven Model of Candidate Evaluation’, American Political Science Review, 83 (1989), 399–419CrossRefGoogle Scholar; Lodge, Milton, Steenbergen, Marco R. and Brau, Shawn, ‘The Responsive Voter: Campaign Information and the Dynamics of Candidate Evaluation’, American Political Science Review, 89 (1995), forthcoming.CrossRefGoogle Scholar

¹⁸ The mean party placements are consistent with the expert ratings shown for Norway in Laver, Michael and Schofield, Norman, Multiparty Government: The Politics of Coalition in Europe (Oxford: Oxford University Press, 1990), p. 263Google Scholar; and for Norway and the United States in Laver, Michael and Hunt, W. Ben, Policy and Party Competition (London: Routledge, 1992), pp. 276–84, 312–16.Google Scholar The mean, however, is not without criticism; we address this issue in Appendix III.

¹⁹ Ironically this implies that prior work using the proximity model to explain voting behaviour may owe its effectiveness to the directional model.

²⁰ See Weisberg, Herbert F. and Rusk, Jerrold G., ‘Dimensions of Candidate Evaluation’, American Political Science Review, 64 (1970), 1167–85.CrossRefGoogle Scholar

²¹ The demographic model for the United States includes the following variables: south, black, Hispanic, female, youngest, oldest, education, low income, high income, subjective working-class, union membership, public sector employment, Catholic, Jewish, other religion and ‘born again’. The model for Norway includes: region, periphery, female, education, class, subjective working-class, union, public sector, farm, language, religiosity and alcohol use.

²² The penalty concept in directional theory is not explicitly operationalized. This means that the penalty is incorporated in the intercept, implicitly making it a constant across all voters. There is some evidence that the Socialist Left and the Progressive parties were penalized in the 1989 election in Norway; these parties had distinctly more extreme issue positions on average than the other parties. See Macdonald, , Listhaug, and Rabinowitz, , ‘Issues and Party Support in Multiparty Systems’. pp. 1119–20.Google Scholar

²³ We use 0.01 as the level of significance because we have a large number of cases and are conducting a large number of tests. See Mock, Carol and Weisberg, Herbert F., ‘Political Innumeracy: Encounters with Coincidence, Improbability, and Chance’, American Journal of Political Science, 36 (1992), 1023–46.CrossRefGoogle Scholar

²⁴ See, for example, Campbell, Angus and Valen, Henry, ‘Party Identification in Norway and the United States’, in Campbell, Angus, Converse, Philip E., Miller, Warren E. and Stokes, Donald E., Elections and the Political Order (New York: Wiley, 1966)Google Scholar; Granberg, and Holmberg, , The Political System Matters.Google Scholar

²⁵ We consider an alternate measure of political sophistication based on accuracy of party placements in Appendix III.

²⁶ The demographic model explains an unusually large amount of variance for the Christian People's party, due primarily to the impact of religiosity and personal alcohol use. Personal abstinence is commonly included in the set of demographic predictors in Norway because teetotallers have been a traditional source of support for the Christian party. We follow this practice while noting that it greatly dilutes the impact of the issue question about government regulation of the sale and use of alcohol. When personal abstinence is not included as a demographic variable, the directional model is discernibly superior to the proximity model for the low and medium sophistication groups; the two models remain indistinguishable for the high sophistication group.

²⁷ The small number of high sophisticates on the extreme left give Labour a lower rating than the Liberal party or the Centre party. This does not correspond with either the directional or proximity perspective, but is consistent with a special antagonism towards Labour caused by displeasure with their conduct in government or a desire to separate them from the Socialist Left.

The results for the Labour party among the high sophisticates resemble the results Listhaug, Macdonald and Rabinowitz report for the Social Democrats in Denmark in 1979. The Danish analysis was based on less direct evidence of decision mechanisms and was not performed by level of sophistication; nevertheless, the parallels between the cases are worth noting. Both are social democratic parties with long histories of participation in government and both have competitors on their left seriously challenging them for votes. See Listhaug, Ola, Macdonald, Stuart Elaine and Rabinowitz, George, ‘Ideology and Party Support in Comparative Perspective’, European Journal of Political Research, 25 (1994), 111–49.CrossRefGoogle Scholar

²⁸ Extensive work has been done on the role of sophistication in the use of issues and ideology. Stimson demonstrated the tendency of more sophisticated respondents to make greater use of issues in their political judgements (see Stimson, , ‘Belief Systems’).Google Scholar This finding has been replicated in experimental work by Lodge, and Hamill, , ‘A Partisan Schema for Political Information Processing’Google Scholar, and McGraw, , Lodge, and Stroh, , ‘On-Line Processing in Candidate Evaluation’.Google Scholar Recently Sniderman, Brody and Tetlock found that low sophisticates showed virtually no reliance on issues in voting, while issues were important criteria for high sophisticates. See Sniderman, , Brody, and Tetlock, , Reasoning and Choice.Google Scholar The only exception to this general pattern that we are aware of is Rahn, , Aldrich, , Borgida, and Sullivan, , ‘A Social-Cognitive Model of Candidate Evaluation’.Google Scholar

²⁹ This pattern is largely replicated when mean squared errors are considered instead of variance explained.

³⁰ Stokes, Donald E., ‘Spatial Models of Party Competition’, American Political Science Review, 57 (1963), 368–77.CrossRefGoogle Scholar

³¹ Rabinowitz, George, Spatial Models of Electoral Choice: An Empirical Analysis (Chapel Hill, NC: Institute for Research in Social Science, 1973)Google Scholar; ‘On the Nature of Political Issues: Insights from a Spatial Analysis’, American Journal of Political Science, 22 (1978), 793–817.Google Scholar

³² Miller, Warren E. and Shanks, J. Merrill, ‘Policy Directions and Presidential Leadership: Alternative Interpretations of the 1980 Presidential Election’, British Journal of Political Science, 12 (1982), 299–356CrossRefGoogle Scholar; Shanks, and Miller, , ‘Policy Direction and Performance Evaluation’Google Scholar, and ‘Partisanship, Policy and Performance’.

³³ Budge, and Farlie, , Explaining and Predicting Elections.Google Scholar

³⁴ See Hastie, Reid and Park, Bernadette, ‘The Relationship between Memory and Judgment Depends on Whether the Task is Memory-Based or On-Line’, Psychological Review, 93 (1986), 258–68CrossRefGoogle Scholar; Lodge, , McGraw, and Stroh, , ‘An Impression-Driven Model of Candidate Evaluation’Google Scholar; McGraw, , Lodge, and Stroh, , ‘On-Line Processing in Candidate Evaluation’.Google Scholar

³⁵ Rabinowitz, and Macdonald, , ‘A Directional Theory of Issue Voting’, pp. 103–7, 115–18.Google Scholar

³⁶ The mixed model has the additional statistical advantage of avoiding the collinearity problems which arise if the distance and scalar product measures are included in the same regression. The distance and scalar product measures are, of necessity, highly correlated because the scalar product is a major component of distance. In contrast, the scalar product and length variables are only modestly correlated. This is because the value of the scalar product can be positive or negative while the length variable can have only positive values, and the value of the length variable is large when individuals are extreme on either side of an issue. Hence, large positive length values are associated with both large positive and large negative scalar product values, leading to a low correlation between the two variables.

³⁷ Powell, Lynda, ‘Analyzing Misinformation: Perceptions of Congressional Candidates' Ideologies’, American Journal of Political Science, 33 (1989), 272–93.CrossRefGoogle Scholar

³⁸ Listhaug, Ola, Macdonald, Stuart Elaine and Rabinowitz, George, ‘Issue Perceptions of Parties and Candidates: A Comparison of Norway and the United States’, Scandinavian Political Studies, 17 (1994), 273–87.CrossRefGoogle Scholar

³⁹ Accuracy of party and candidate ideological placements is one of the indicators Luskin uses to measure political sophistication. Education and interest are the two best predictors of sophistication available on both the US and Norwegian surveys. See Luskin, Robert C., ‘Measuring Political Sophistication’, American Journal of Political Science, 31 (1987), 856–99CrossRefGoogle Scholar, and ‘Explaining Political Sophistication’, Political Behavior, 12 (1990), 331–61.Google Scholar

⁴⁰ To construct the accuracy of placement measure, we compared a respondent's placement of the parties and candidates to the mean placement on each issue. Each time a respondent placed two objects in the same order as the mean placements, the accuracy score was increased by 1; each time a respondent placed two objects in the opposite order, the accuracy score was decreased by 1; if the respondent placed two objects at the same location, the score was left unchanged. In the United States the placement accuracy scores ranged from a low of – 43 to a high of 53 with a mean of 22.6. In Norway the scores ranged from – 38 to 123 with a mean of 73.7.