¹ The major original voting studies are of course those of the Columbia and Michigan groups whose theories and data analyses are presented for the most part in Lazarsfeld and his associates' The People's Choice (New York: Columbia University Press, 1944)Google Scholar and Voting (Chicago: University of Chicago Press, 1954)Google Scholar and in Campbell, et al., The Voter Decides (Evanston, Ill.: Row Peterson, 1954)Google Scholar, The American Voter (New York: John Wiley and Sons, 1960)Google Scholar, and most recently Elections and the Political Order (New York: John Wiley and Sons, 1966)Google Scholar. Much has been said with respect to both the important contributions and theoretical shortcomings of these studies. What is perhaps the most significant commentary on the state of our knowledge about voting behavior is the structure of a recent book, Milbrath's, Lester Political Participation (Chicago: Rand McNally, 1965)Google Scholar, which relies heavily on the findings of all but the most recent work cited above. Rather than presenting theories of participation, Milbrath is forced to offer a propositional inventory whose only unifying theme is a conceptual model of his own with which he attempts to bring together the various discrete propositions belonging to a variety of conceptions.

² Inferences with respect to the irrationality of the voter have been based upon the relative stability of partisan affiliation, the low informartion level with respect to national policy questions and the apparent influence of primary groups on partisan attitudes. A major attack on the irrationality hypothesis is presented in Key's, V. O. The Responsible Electorate (Cambridge: Harvard Univ. Press, 1966)CrossRefGoogle Scholar, which is based upon an analysis of voters who switch parties from one election to the next, a switch which Key's analysis indicates to be issue oriented.

³ Riker, William H. and Ordeshook, Peter C., “A Theory of the Calculus of Voting,” this Review, 62 (03, 1968), 25–42 Google Scholar.

⁴ This distinction between images of man and models of man has been recently discussed by Searing, Donald D., “Models and Images of Man and Society in Leadership Theory,” Journal of Politics, 31 (02, 1969), 3–31 CrossRefGoogle Scholar. Generally models of man can be thought of as tentative conceptual frameworks which are recognized as being empirically relevant while images are more unself-conscious, reality-filtering mechanisms.

⁵ Kircheimer, Otto, “Private Man and Society,” Political Science Quarterly, 81 (03, 1966), 124 CrossRefGoogle Scholar. For a similar critique of consensus approaches which argues explicitly for rational choice conceptions to avoid biasing an analysis in favor of desired political structures, see Harsanyi, John, “Rational-Choice Models vs. Functionalistic and Conformistic Models of Political Behavior,” World Politics, 21 (07, 1969), 513–538 CrossRefGoogle Scholar.

⁶ Defining a belief system in terms of the degree of constraint among idea elements, Converse found that with respect to national, domestic and foreign policy, the average voter demonstrated very little constraint. See Converse, Philip, “The Nature of Belief Systems in Mass Publics,” in Apter, David (ed.) Ideology and Discontent (New York: Free Press, 1964), pp. 206–261 Google Scholar. For a similar approach emerging with the finding that American attitudes on national policy lack a clear pattern, see Axelrod, Robert, “The Structure of Public Opinion on Policy Issues,” Public Opinion Quarterly, 31 (Spring, 1967), 51–60 CrossRefGoogle Scholar.

⁷ Rogin, Michael P., The Intellectuals and McCarthy: The Radical Spector (Cambridge, Mass.: MIT Press, 1967), p. 228 Google Scholar.

⁸ The inspiration for this quasi open-ended interview technique was derived from the self-anchoring striving scale developed by Kilpatrick and Cantril. See Kilpatrick, F. P. and Cantril, Hadley, “Self-Anchoring Scaling: A Measure of Individuals' Unique Reality Worlds,” Journal of Individual Psychology, 16 (11, 1960), 158–173 Google Scholar. The scale's most notable application is in Cantril's, Hadley The Pattern of Human Concerns (New Brunswick, N. J.: Rutgers University Press, 1965)Google Scholar.

⁹ The usefulness of distinguishing between deductive and inductive rationality has been discussed by Gibson, Quentin, The Logic of Social Inquiry (London: Routledge & Kegan Paul, 1960)Google Scholar, chapter XIV, “The Assumption of Rationality.”

¹⁰ The original experiment is reported in Mosteller, Frederick and Nogee, Philip, “An Experimental Measurement of Utility,” Journal of Political Economy, 59 (1951), 371–404 CrossRefGoogle Scholar. In addition to the Riker-Ordeshook study cited above, applications of utility-type models include such diverse substantive concerns as foreign policy decision-making, e.g., Blechman, Barry M., “The Quantitative Evaluation of Foreign Policy Alternatives,” Journal of Conflict Resolution, 10 (12, 1966), 408–426 CrossRefGoogle Scholar, and child rearing practices, reported in Brim, O. G., et al., Personality and Decision Processes (Stanford, Calif.: Stanford University Press, 1962)Google Scholar.

¹¹ The concept of inductive rationality, as it is employed here, refers primarily to the various cognitive consistency models which hypothesize that information processing in attitude formation and choice behavior is partly responsive to motivations for maintaining consistent cognitions about objects and situations. Rosenberg and Abelson, for example, speak of the “induction of new relations,” when an individual seeks to render new information compatible with other pre-existing impressions: see Rosenberg, Milton J. and Abelson, Robert P., “An Analysis of Cognitive Balancing,” in Rosenberg, Milton J. et al. Attitude Organization and Change (New Haven: Yale University Press, 1960, pp. 112–163)Google Scholar.

¹² Among the most significant theories of this type are Heider, Fritz, “Attitudes and Cognitive Organization,” Journal of Psychology, 21 (1946), 107–112 CrossRefGoogle ScholarPubMed, and The Psychology of Interpersonal Relations (New York: John Wiley and Sons, 1958)Google Scholar; Osgood, Charles E. and Tannenbaum, Percy H., “The Principle of Congruity in the Prediction of Attitude Change,” Psychological Review, 62 (1955), 42–55 CrossRefGoogle ScholarPubMed; Festinger, Leon, A Theory of Cognitive Dissonance (Stanford, Calif.: Stanford University Press, 1957)Google Scholar; Abelson, Robert P. and Rosenberg, Milton J., “Symbolic Psychologic: A Model of Attitudinal Cognition,” Behavioral Science, 3 (1958), 1–13 CrossRefGoogle Scholar; and Sherif, Muzafer and Hovland, Carl I., Social Judgment (New Haven, Conn.: Yale University Press, 1961)Google Scholar. A comprehensive review of the relevant literature can be found in McGuire, William J., “The Current Status of Cognitive Consistency Theories,” in Feldman, S. (ed.), Cognitive Consistency: Motivational Antecedents and Behavioral Consequences (New York: Academic Press, 1966)Google Scholar. Insightful critiques of a number of these theories can be found in Brown, Roger, Social Psychology (New York: Free Press, 1965)Google Scholar, Chapter 11, and Insko, Chester A., Theories of Attitude Change (New York: Appleton-Century-Crofts, 1967)Google Scholar. The most recent versions of the various consistency theories appear in Abelson, Robert P. et al., Theories of Cognitive Consistency: A Source Book (Chicago: Rand McNally, 1968)Google Scholar.

¹³ This similarity has been noted explicitly in Feather, Norman T., “Subjective Probability and Decision Under Uncertainty,” Psychological Review, 66 (1959), 150–164 CrossRefGoogle ScholarPubMed; and Edwards, Ward, “Behavorial Decision Theory,” Annual Review of Psychology, 12 (1961), 473–498 CrossRefGoogle Scholar. I am indebted to Rufus P. Browning for calling these references to my attention.

¹⁴ Fishbein, Martin, “A Behavior Theory Approach to the Relations Between Beliefs about an Object and the Attitude Toward the Object,” in Fishbein, Martin (ed.), Readings in Attitude Theory and Measurement (New York: John Wiley and Sons, 1967), p. 394 Google Scholar.

¹⁵ See Rosenberg, Milton J., “Cognitive Structure and Attitudinal Affect,” Journal of Abnormal and Social Psychology, 53 (1956), 367–372 CrossRefGoogle ScholarPubMed, and R. B. Zajonc, “Structure of the Cognitive Field,” unpublished doctoral dissertation, University of Michigan, 1954. Algebraic versions of both these theories are developed by Fishbein op. cit. pp. 394–395.

¹⁶ This diagram is taken from Bross, Irwin D. J., Design for Decision (New York: Free Press, 1958), p. 29 Google Scholar.

¹⁷ A theory which focuses explicitly on the way in which cognitive structures process information is presented along with some interesting empirical applications in Schroder, Harold M., Driver, Michael J. and Streufert, Siegfried, Human Information Processing (New York: Holt, Rinehart and Winston, 1967)Google Scholar.

¹⁸ See Davis, Otto A. and Hinich, Melvin, “A Mathematical Model of Policy Formation in a Democratic Society,” in Bernd, Joseph L. (ed.), Mathematical Applications in Political Science II (Dallas, Texas: Southern Methodist University Press, 1966), pp. 175–208 Google Scholar, and an elaboration by the same authors, “Some Results Related to a Mathematical Model of Policy Formation in a Democratic Society,” in Bernd, Joseph L. (ed.), Mathematical Application in Political Science III (Charlottesville, Va.: The University of Virginia Press, 1967), pp. 14–38 Google Scholar. For another application and extension of the same theoretical calculus see Ordeshook, Peter C., “A Theory of Electoral Behavior and Party Strategy,” Unpublished doctoral dissertation, University of Rochester, 1968 Google Scholar. I am indebted to Professor Ordeshook for bringing this theoretical formulation to my attention.

¹⁹ The linear algebra notation is preserved in order to maintain continuity with other applications of the theory. In normal mathematical notation, the calculus could be rendered as L_c

(Where L _c is the loss experienced by an individual voter with respect to candidate c, is the square of the distance between the individual's ith attribute preference and his perception of the candidates' position on that attribute, n is the number of attributes that the voter deems to be relevant to his voting choice, and a _i is the salience or importance of the ith attribute of the candidate with respect to his voting choice.)

²⁰ In Davis and Hinich's original version of the calculus, the A matrix was assumed to be constant for all voters. In their reformulation, they changed this assumption to allow for variation in the A matrix for the population of voters.

²¹ The sample was drawn on a random probability basis from the list of registered voters on the Island of Oahu which contains approximately seventy-five percent of the voting population of the State of Hawaii.

²² The questions asked to elicit voting criteria in the personal quality issue, and interpersonal cue categories are as follows: “In assessing the personal qualities of a presidential candidate, what things do you think about”; “What are the foreign and domestic policy questions with which you are, at present, most concerned”; “Can you think of any individuals or groups whose political opinions you pay attention to.”

²³ An analysis which attempts to differentiate policy orientations in terms of their complexity or sophistication can be found in Kessel, John, “Cognitive Dimensions of Political Activity,” Public Opinion Quarterly, 29 (Fall, 1965), 377–389 CrossRefGoogle Scholar.

²⁴ An explication and empirical assessment of these two competing hypotheses is presented by McGrath, Joseph E. and McGrath, Marion F., “Effects of Partisanship on Perceptions of Political Figures,” Public Opinion Quarterly, 26 (Summer, 1962), 236–248 CrossRefGoogle Scholar.

²⁵ The authors of Voting (see footnote 1) placed heavy emphasis on perceptual balance mechanisms in their explanation of voter decision-making. Since that time studies employing a variety of techniques have yielded contradictory results. Investigations which support the perceptual balance thesis include Froman, Lewis A. Jr., and Skipper, James K. Jr., “Factors Related to Misperceiving Party Stands on Issues,” Public Opinion QuaHerly, 26 (Summer, 1962), 265–272 CrossRefGoogle Scholar and Sigel, Roberta S., “Effect on Partisanship on the Perception of Political Candidates,” Public Opinion Quarterly, 28 (Fall, 1964), 483–496 CrossRefGoogle Scholar. Investigations which support a stimulus or non-perceptual bias thesis include Froman, Lewis A. Jr., and Skipper, James K. Jr., “An Approach to the Learning of Party Identification,” Public Opinion Quarterly, 27 (Fall, 1963), 473–480 CrossRefGoogle Scholar, and McGrath and McGrath op. cit.

²⁶ See, for example, Sherif, Carolyn, Sherif, Muzafer, and Nebergall, Roger E., Attitude and Attilude Change: The Social Judgment-Involvement Approach (Philadelphia: Saunders, 1965)Google Scholar. They found that the degree of ego involvement of respondents (with respect to their partisanship) was systematically related to their perceptions of campaign appeals in the 1900 presidential election.

²⁷ Because it is often asserted that various kinds of irrationality in voting are more in evidence among voters of lower socio-economic status, we have correlated income level with the index of the degree of perceptual distortion. The correlation coefficient in this case is a non-significant .04.

²⁸ The factor analysis model employed is a principle components model. All factors with an eigen value greater than one were preserved for rotation. For a comprehensive discussion of various factor models and their implications see Harmon, Harry H. Modern Factor Analysis, (2nd ed.; Chicago: The University of Chicago Press, 1967)Google Scholar.

²⁹ The overall prediction of 85.5 percent is based upon two different kinds of voting intentions. 39 of the 201 respondents included in the prediction did not indicate a definite voting choice but rather named two or more candidates among whom they said they were deciding. They were considered correctly predicted if the loss function equations predicted one of those that they named. A separate analysis of only those with definitive choices, however, indicates that the 85.5 percentage is appropriate with or without the 39 respondents.

³⁰ The predictions in terms of the loss function equations were examined for different income groups and it was found that there is very little variation in percentage prediction on the basis of income level.

³¹ It should be recognized that the sizes of the parameters in the loss function equations (a five point scale for salience and a seven point scale of attribute distances) are arbitrary. For this reason, several analyses of the overall predictions were conducted with the relative sizes of the saliences and distances varied. It was found that the scales as originally employed provided the optimum prediction.

³² For a discussion of the relative explanatory contributions of regression and correlation coefficients in multiple linear regression, see Ezekiel, Mordecai and Fox, Karl A. Methods of Correlation and Regression Analysis (New York: John Wiley and Sons, 1959), section III.Google Scholar