¹ The 1968 national sample survey(N = 1559) was made possible by a grant from the Ford Foundation, whose support we gratefully acknowledge. A total of 1559 citizens of voting age were interviewed, most of them both before and after Election Day. The preliminary nature of this report is to be emphasized, since the data on which it is based had not been fully cleaned at the time of writing. When the study is released through the Inter-University Consortium for Political Research, interested analysts may discover small discrepancies from the statistics reported here. Readers should also remember that all sample statistics are subject to varying amounts of sampling error in relation to the number of cases on which they are based.

² A deviating election is one in which the party commanding the identifications of a majority of the electorate is nonetheless voted out of power temporarily. See Campbell, A., Converse, P., Miller, W., and Stokes, D., The American Voter (New York: John Wiley, 1960), Chapter 19Google Scholar.

³ Congressional Quarterly, November 22, 1968, p. 3177 Google Scholar.

⁴ The percentage difference of 62% in candidate preference between blacks and whites is substantially larger than class differentiation or other social cleavages and partisanship within the United States in recent history or for democracies of Western Europe.

⁵ Such segregation is indicated simply because of the fact that within the black vote in 1968 there is next to no meaningful “variance” to be “accounted for.” When categories of “Nixon voters” and “Wallace voters” are presented, they are necessarily “lily-white” in composition. Therefore when “Humphrey voters” are contrasted with them, it is confusing if differences may be totally a function of the large admixtures of blacks in the Humphrey support, as opposed to differences which would stand up even with comparisons limited to whites.

⁶ See “Voting and Foreign Policy,” by Miller, Warren E., Chapter 7 in Rosenau, James N. (ed.), Domestic Sources of Foreign Policy (New York: The Free Press, 1967)Google Scholar.

⁷ A separate analysis, carried out by a colleague in the Survey Research Center Political Behavior Program and using the same body of data from the SRC 1968 election study, suggests, moreover, that many voters who thought the police used too little force deserted Humphrey in the course of the campaign while the minority who objected that too much force was used voted more heavily for the Democratic nominee. See Robinson, John P., “Voter Reaction to Chicago 1968,” Survey Research Center (1969), mimeoGoogle Scholar.

⁸ The decline was only on the order of ½ percent nationally, but the overall figures are somewhat misleading. Enormous efforts devoted to voter registration projects among Southern blacks between 1964 and 1968 appear to have paid off by increasing voter participation in that sector from 44% to 51%. Perhaps in counterpoint, Southern whites increased their turnout by 2%, thereby inching ever closer to the national norm. Thus the decline in turnout was concentrated outside the South, and there approached the more substantial drop of 4%. Even this figure paris misleading, since whites outside the South showed a 3% loss in percentage points of turnout, while nonwhites declined by almost 11 percentage points! See Current Population Reports, “Voter Participation in November 1968,” Series P-20, No. 177, December 27, 1968. Although such turnout figures, apart from the more general mobilizing of Southern blacks, are consistent with a proposition that whites were more eager to “throw the rascals out” than blacks, and that among whites, Southerners had the fiercest grievances of all, there is no hiding the fact of anemic turnout in most of the country in 1968. Interestingly enough, the decline from 1964 was uniformly distributed across the entire spectrum of party allegiances from loyal Democrats to strong Republicans.

⁹ The reader should also keep in mind several other things about Table 2. The “South” here refers, as it will throughout this paper, to the Census Bureau definition of the region that includes 15 states and the District of Columbia. Hence such border states as Maryland or West Virginia are included along with the deeper southern states of the old confederacy. Presumably, for example, George Wallace's rating among whites of a more hard-core South would be correspondingly higher. Secondly, it should be remembered for some of the lesser candidates that respondents knowing so little about a candidate as to be indifferent to him would end up rating him “50°.” Thus it would be questionable to conclude from Table 2 that LeMay was more popular than George Wallace, except in a very limited sense. Actually, three times as many respondents (nearly one-third) left LeMay at the indifference point as did so for Wallace. Thus lack of visibility helped to make him less unpopular. But among those who reacted to both men, LeMay was less popular than Wallace. Similarly, Wallace's low rating must be understood as a compound of an admiring minority and a hostile majority. The variance of Wallace ratings is much greater than those for other candidates, even in the South.

¹⁰ Just after the decision of Robert Kennedy to run and before Lyndon Johnson's withdrawal, the Gallup poll showed Democrats favoring Kennedy as the party's nominee by a 44–41 margin.

¹¹ Interestingly enough, the same generational cleavages among Southern white Democrats occur at an earlier age than those elsewhere. In that region, Humphrey-Johnson preferences hold a plurality in all age cohorts over 30, despite the fact that Kennedy support has an edge of better than three to one among those under 30 (N of 34), perhaps because the latter group has less of a memory of the fury in the deep South at the Kennedy family prior to the assassination of President John Kennedy in Dallas in 1963.

¹² Although there is some slight tendency for pre-convention supporters of McCarthy to be relatively young, the distribution by age is more homogeneous than expected, and much more so than is the case for Kennedy. It is possible that young people supporting McCarthy as the only alternative to the Administration switched more heavily than the middle-aged to Kennedy when he announced his candidacy.

¹³ See also the account for New Hampshire by Harris, Louis, “How Voters See the Issues,” Newsweek, 03 25, 1968, p. 26 Google Scholar.

¹⁴ Key, V. O. Jr., The Responsible Electorate: Rationality in Presidential Voting, 1936–1960 (Cambridge, Mass.: Belknap Press, 1966), pp. 7–8 CrossRefGoogle Scholar.

¹⁵ Stokes, Donald E., “Some Dynamic Elements of Contests for the Presidency,” this Review, LX (03, 1966), 19–28 Google Scholar.

¹⁶ This is not to say that it would be inconceivable for identification with one of the two traditional parties to correlate with preference for some third-party candidate. For example, it is possible that most of the voters for Henry Wallace's Progressive Party in 1948 were identified with the Democratic Party. However, it is clear that in such an instance “party loyalty” would have been a rather spurious name for the motivating factor. In the case of George Wallace, even this kind of spurious correlation is absent, except insofar as his Democratic origins and the invisibility of his American Independent Party label made it easy for Democrats to support him. Indeed, in the context of this argument it will be fascinating to discover whether Republicans and Democrats invoked different images of Wallace's party location in order to satisfy their need for consonance while voting for a man who reflects their own issue commitments.

¹⁷ Another domain of issues surrounding the “cold war” as it confronted the nation in the 1950's with controversies over foreign aid and trade with communist countries shows only modest correlations with the candidate rankings, and Nixon and Humphrey ratings show more of a parity with the Wallace correlations, although in an absolute sense the latter continue to outrun the former sharply in the South and mildly elsewhere. See Table 5.

¹⁸ These include such considerations as that of the causal direction underlying the observed relationships; or known and systematic biases in recollection of a presidential vote four years later; or the superficiality of the issues that show such patterns, as opposed to issues thought basic by sophisticated observers; or blatant misinformation supporting the issue positions registered; or a tendency for the less informed to “shift” more quickly than the better informed, with position on any given issue held constant, etc.

¹⁹ We much prefer an interpretation which hinges on a general inattention which is endemic because information costs are relatively high where little information is already in hand, and the stakes are rarely seen as being very large. While such a “condition” is likely to persist in mass electorates, there is nothing about it which is immutable given the proper convergence of circumstances.

²⁰ This was not true across every issue domain. The most notable exception was in the area of social welfare issues such as medicare and full employment guarantees, on which issues Wallace voters were significantly more “liberal” than Nixon voters, and almost matched the liberalism of Humphrey voters. This admixture was of course familiar in Wallace's frequent appeals to the underdog and the working man, in the tradition of Southern populism.

²¹ For reasons discussed elsewhere, a rather large proportion of the American electorate—nearly half—is found at this point of ideological neutrality.

²² Converse, P., Clausen, A., and Miller, W., “Electoral Myth and Reality: The 1964 Election” this Review, 59 (06, 1965), 321–336 Google Scholar

²³ It is quite possible, however, that some of this support might have moved to Wallace had the Republican Party nominated anybody but Nixon or Reagan, among the main contenders.

²⁴ Campbell, A., Converse, P., Miller, W., and Stokes, D., Elections and the Political Order (New York: John Wiley, 1965), Chapter 12Google Scholar.

²⁵ Converse, Philip E., “Of Time and Partisan Stability,” Journal of Comparative Politics (issue to be announced)Google Scholar.

²⁵ The South shows somewhat diluted patterns here, compatible with the likelihood that for at least some Southern Democrats, a vote for Wallace was not conceived as a defection.

²⁷ The American public seems to have a very low tolerance for unusual or “showy” forms of political dissent. Responses to an extended set of items in the 1968 study on the subject are appalling from a civil libertarian point of view. At the most acceptable end of the continuum of “ways for people to show their disapproval or disagreement with govermental policies and actions” we asked about “taking part in protest meetings or marches that are permitted by the local authorities” (italics not in original question). Less than 20% of all respondents, and scarcely more than 20% of those giving an opinion, would approve of such subversive behavior, and more than half would disapprove (the remainder accepted the alternative presented that their reaction “would depend on the circumstances”). In view of such assumptions, the overwhelmingly negative reaction to the Chicago demonstrations despite sympathetic media treatment (cited earlier) is hardly surprising.