¹ Campbell, Angus, Converse, Philip E., Miller, Warren E., and Stokes, Donald E., The American Voter (New York: John Wiley and Sons, 1960), Chapter 8Google Scholar.

² Converse, Philip E., “The Nature of Belief Systems in Mass Publics” in Apter, David E., editor, Ideology and Discontent (Glencoe: The Free Press, 1964), pp. 238–245 Google Scholar, and Converse, Philip E., “Attitudes and Non-Attitudes: Continuation of a Dialogue” (Survey Research Center, The University of Michigan, February, 1963), 27 ppGoogle Scholar. (Mimeographed.)

³ Key, V. O. Jr., Public Opinion and American Democracy (New York: Alfred A. Knopf, 1961), p. 445 Google Scholar; Sorauf, Frank J., Party Politics in America (Boston: Little, Brown and Company, 1968), pp. 163–165 Google Scholar.

₄ Campbell, Angus, Converse, Philip E., Miller, Warren E., and Stokes, Donald E., Elections and the Political Order (New York: John Wiley and Sons, 1966), p. 287 Google Scholar.

⁵ Campbell, et al., The American Voter, p. 186 Google Scholar.

⁶ Stokes, Donald E., “Some Dynamic Elements of Contests for the Presidency,” American Political Science Review, 60 (March, 1966), p. 27 CrossRefGoogle Scholar. It should be pointed out that the graphs in Stokes' article do not indicate the relative importance of each component for individual voting choice; rather the graphs represent the net effect of each component on the overall outcome of each election. Stokes multiplies the beta weights—that is, the importance of each component in predicting individual voting choice—by the distance that the average attitude of the entire electorate is displaced from the neutral position on the attitudinal component. His formula is:

Thus, the relative importance of issues and candidates in individual voting decisions cannot be discerned directly from Stokes[ data since his findings are a function of both b _i, and _i.

In his original component analysis, published in 1958, Stokes did present data showing the relative importance of each component for individual voting choice; bar graphs indicated the relative strengths of the beta weights in the 1952 and 1956 presidential elections. Stokes, Donald E. et al., “Components of Electoral Decision,” American Political Science Review, 52 (June, 1958), p. 381 CrossRefGoogle Scholar. However, these individual weights have not been included in the subsequent presentations of the component analysis published in The American Voter and in the 1966 article cited above.

⁷ Key, V. O. Jr., Public Opinion and American Democracy, p. 465 Google Scholar.

⁸ Campbell, et al., The American Voter, p. 143 Google Scholar.

⁹ There are some exceptions; a few studies have used open-ended questions to analyze issues. In an effort to measure the impact of the Goldwater campaign, Kessel examined several sets of Survey Research Center open-ended questions: those which measure attitudes toward the parties and the presidential candidates as well as the “most important problems” (which provides the basis for this present study). Kessel shows the distribution of responses to these questions for interviews taken during different stages of the campaign. See Kessel, John H., The Goldwater Coalition (Indianapolis and New York: The Bobbs-Merrill Company, 1968), pp. 271–280 Google Scholar. As we have seen, Stokes based his component analysis on the open-ended questions measuring likes and dislikes for the parties and the candidates. However, m his analysis, the impact of specific issues cannot be directly determined since all foreign policy items were placed in a single category, and all domestic issue responses were lumped together into three broad categories (parties as managers of government, domestic issues, and group-related attitudes). Stokes, “Some Dynamic Elements of Contests for the Presidency,” op. cit.

¹⁰ A “salient” issue may range in centrality from a readily verbalized phrase that is “on top” of the respondent's mind to an attitude which is well developed and of deep concern. When asked an openended question, the respondent must have some awareness of an issue and be able to verbalize this cognition. The problem with the Likert type scale is that it measures only the affective component of an attitude, if indeed it measures an attitude at all. Open-ended questions measure both affect and cognition. In using these questions, the researcher comes much closer to measuring an attitude which is on the respondent's mind (salient) at the time of the interview—an attitude which presumably was salient before the interview and will remain so afterwards.

¹¹ The Survey Research Center changed some code categories in 1964 from those used in 1960. Thus, these data are not exactly comparable. This is particularly true of the foreign aid and Supreme Court issues. Foreign aid responses were not distinguished from other foreign policy responses in 1960, but they were in 1964. Thus, foreign aid appears to be a new issue in 1964; it was not.

The 1964 interview question specifically asked about the Supreme Court; it had not been mentioned in the 1960 question. Thus, responses concerning the Supreme Court in 1964 may be a function of question wording rather than a sudden interest in the Court's activities. All other issues which appear for the first time in 1964 are a result of spontaneous responses, not question wording or coding.

¹² The concept of “issue public” was originally used by Converse in “The Nature of Belief Systems in Mass Publics,” op. cit.

¹³ The opposite finding regarding the importance of foreign affairs in 1964 is reported in Stokes, “Some Dynamic Elements of Contests for the Presidency,” op. cit., p. 21. His finding, however, is based on quite a different set of questions.

¹⁴ It should be noted that in 1964 a relatively large cluster of people did begin to voice concern about Vietnam.

¹⁵ See Kessel, , The Goldwater Coalition, Chapter 9, for a more thorough analysis of the impact of the 1964 campaignGoogle Scholar.

¹⁶ A study using 1958 data reported low salience of specific items of Congressional legislation. Perhaps this was in part a result of the lack of legislative activity under Eisenhower, and in part a result of incomplete questioning of respondents. Stokes, See Donald E. and Miller, Warren E., “Party Government and the Saliency of Congress” in Campbell et al., Elections and the Political Order, pp. 198–199 Google Scholar.

¹⁷ Stokes, “Some Dynamic Elements of Contests for the Presidency,” op cit., Figures 1 and 2.

¹⁸ The exact wording of the 1964 interview questions was:

After each response, people were asked:

¹⁹ The data presented in Figure 1 and in the remainder of this analysis are based on issues that the respondents felt were of high salience. That is, responses were not included unless the respondent said he was extremely worried about the issue and/or the problem was on his mind a lot. When this measure of degree of salience is not used, perception of party differences is less sharply defined. There is an increase in “no difference” and “don't know” answers or a tendency to match the inappropriate party with the issue. This mismatching or lack of knowledge is to be expected since those respondents who have not thought about an issue or who are not very worried about it will be unlikely to seek out, or pay attention to, cues indicating party positions on the issue.

²⁰ In analyzing the Survey Research Center's open-ended questions about parties and candidates in 1964, Kessel found that attitudes about candidates were “far more salient than the attitudes about parties … questions about candidates engendered half again as many comments as the questions about parties.” See Kessel, John H., The Goldwater Coalition, p. 258 Google Scholar. Also, Stokes found the candidate component to be an unusually strong factor in 1964. See “Some Dynamic Elements of Contests for the Presidency,” op. cit.

²¹ Ellsworth performed a content analysis on a sample of campaign speeches in 1964, and concluded that “the ideological spread between Johnson and Goldwater did little or nothing to produce an increase in emphasis upon issues by the candidates themselves in their [speeches].” Ellsworth, John W., “Policy and Ideology in the Campaigns of 1960 and 1964” (unpublished paper), p. 8 Google Scholar. Polsby and Wildavsky make a similar observation about the 1964 campaign. They state that “in a campaign in which the parties were further apart on issues than they have been in our time … there was relatively little discussion of issues. Certainly, the existence of wide and deep policy differences did not, as has sometimes been thought, lead directly to a campaign focusing on specific issues.” Polsby, Nelson W. and Wildavsky, Aaron B., Presidential Elections, second edition (New York: Charles Scribner's Sons, 1968), p. 205 Google Scholar.

²² The issues on which less than 60 percent perceived party differences were: 1) war-peace, 2) corruption in government, 3) farm problems, 4) labor relations, 5) isolation, and 6) the Supreme Court. Mentions of “war” or “peace” as well as “corruption” in 1964 had little in the way of specific referents; such comments were often unaccompanied by any further substantive remarks. These vague problems were probably mentioned by respondents who were hard pressed to say something, and it is thus not surprising that these people were unable to recognize party differences. However, it is difficult to understand why only 52 percent of those who were concerned with farm problems and only 58 percent of those who mentioned labor relations perceived party differences on these questions. The parties have taken clear stands in these matters over the years. As for the isolationists, they have not been able to find a party which would do what they want since the late 1940's. The Supreme Court, as would be expected, was not considered in partisan political terms, at least not in 1964.

²³ When asked during the 1964 interview whether they had heard if “Congress did anything this year in the way of civil rights,” 77 percent mentioned the Civil Rights Bill. Of these, 95 percent were aware that Johnson favored it and 84 percent correctly perceived Goldwater as being against it.

²⁴ On some items, the SRC codes do not indicate whether the respondent favored the issue or was against it. When both pro and con respondents are lumped into the same category, it is not surprising that both parties are mentioned as best able to handle the issue. Where provision was made for coding issue position, clearer party differences emerged. A good example of this is the strong differences perceived by those who were favorable to civil rights or integration and those who were anti-civil rights.

²⁵ See Stokes', critique of Downs entitled “Spatial Models of Party Competition” for further elaboration of this point This article is found in Campbell, et al., Elections and the Political Order, pp. 161–179 Google Scholar.

²⁶ Further analysis of these data indicate that one of the ways in which party identifiers seem to resolve the psychological conflict created by being concerned about an issue on which the opposite party has the more appropriate position is to claim that there is no difference between the parties on that issue. These respondents might “know” the correct party positions on these issues, but prefer not to undergo the discomfort of saying the opposition party is better in this area.

²⁷ The author found that there is a strong tendency for those with no concept of issues to remain home on election day.

²⁸ Respondents were not encouraged by interviewers to mention more than 3 issues. Had all issues salient to the respondents been recorded, the average would have been higher.

²⁹ This is not to say that Independents are “good citizens” in all respects; the author confirmed previous findings which show that Independents are indeed less involved and participate less regularly. But their concern over substantive issues of the day is none-the-less relatively strong.

³⁰ When I analyzed the Survey Research Center questions which measured likes and dislikes for the two parties in 1964, my findings agreed with the conclusion in The American Voter that Independents have fewer attitudes toward parties than do party identifiers.

³¹ Only 32 of these 106—about 8 more than would be expected by chance—were Independents.

³² It might be noted in Table 5 that about one fourth of the Independents did not perceive party differences on any issue; however, an equal proportion of weak Democrats and weak Republicans also failed to do so.

³³ Where conflict occurred between traditional party ties and evaluation of which party would handle the issues best, the conflict did not cause lower turnout. Apparently, the strong force of issue concerns led a high percentage of these people to the polls—many of them to vote for the opposite party.

³⁴ Of course, this correlation is based on a measure which includes all issues an individual had in mind and is thus not directly comparable to previous studies which have used agree-disagree answers to single issue statements.

³⁵ The attitude toward candidate measure was based on open-ended questions which asked what the respondent liked and disliked about each candidate. The net attitude was determined by the algebraic sum of comments for and against the respective candidates.

³⁶ A similar conclusion was reached by Key, V. O. Jr., in The Responsible Electorate (Cambridge: Harvard University Press, 1966)CrossRefGoogle Scholar. However, his data and methods were completely different from those employed in this study.