¹ Kuhn, Thomas S., The Structure of Scientific Revolutions, 2nd ed. (Chicago: University of Chicago Press, 1970)Google Scholar.

² Ibid., pp. 52–65.

³ American Political Science Review, 59 (March, 1965), 7–28CrossRefGoogle Scholar.

⁴ It is becoming increasingly apparent that the survey-research profile of individual voters laid down by the Michigan, team in The American Voter (New York: Wiley, 1960)Google Scholar, and in their Elections and the Political Order (New York: Wiley, 1966)Google Scholar, must be significantly modified for the post-1964 period—and on the basis of survey analysis itself. See, for example, Pomper's, Gerald two articles, “Toward a More Responsible Two-Party System? What, Again?”, Journal of Politics, 33 (November, 1971), 916–940CrossRefGoogle Scholar, and “From Confusion to Clarity: Issues and American Voters, 1956–1968,” American Political Science Review, 66 (June 1972), 415–428CrossRefGoogle Scholar; also, Nie, Norman, “Mass Belief Systems Revisited: Political Change and Attitude Structure” (mimeo draft, August 1972)Google Scholar; and especially Miller, Arthur H., Miller, Warren E., Raine, Alden S. and Brown, Thad A., “A Majority Party in Disarray: Social and Political Conflict in the 1972 Election” (paper presented at the 1973 meeting of the American Political Science Association)Google Scholar. In view of these findings, which are of course congenial to the arguments made in this paper, it seems about time that we had a complete retesting or replication of The American Voter, and one which is relevant to the contemporary electoral scene.

⁵ Converse, Philip E., “Change in the American Electorate,” in The Human Meaning of Social Change, ed. Campbell, Angus and Converse, Philip E. (New York: Russell Sage Foundation, 1972), chapter 8 (pp. 263–337)Google Scholar. The controversy at issue here is centered in the first half of Converse's essay, pp. 263–301.

⁶ Rusk, Jerrold G., “The Effect of the Australian Ballot Reform on Split-Ticket Voting: 1876–1908,” American Political Science Review, 64 (December, 1970), 1220–1238CrossRefGoogle Scholar. Also see the exchange of views between Rusk, and Burnham, in this Review, 65 (December, 1971), 1149–1157Google Scholar.

⁷ Converse, , “Change in the American Electorate,” pp. 287–288Google Scholar.

⁸ To take but two contemporary examples of this problem: As of the 1970 election, there were no personal-registration requirements at all in 37 (or 42.0 per cent) of Ohio's 88 counties, partial registration for small cities in another ten (11.4 per cent), and countywide registration in the remaining 41 (46.6 per cent). In such a case, of course, no uniform standard exists even at the level of enumeration of electors provided by registration lists. Moreover, while Ohio gives the total number of ballots cast for its cities, it provides no such information for its county returns. In Massachusetts, on the other hand, complete information concerning the numbers of registered voters is published for each town and precinct; and the total number of votes cast, including blanks, is reported by town along with the vote for each candidate. Even in this case, of course, there is no indication of the number of potential voters–those who met legal requirements for the suffrage, but who did not register in the first place. In most European elections and in Canadian elections as well, it is possible to identify: (a) the total eligible electorate; (b) the total vote cast; (c) the number of invalid ballots cast; (d) the total valid vote cast; (e) the partisan distribution of that vote. For a more detailed discussion of heterogeneity and estimation problems in American Electoral Reporting, see the author's “Estimates of Potential Electorate,” Introduction for Section Y, 1974 edition of U.S. Census Bureau, Historical Statistics of the United States (forthcoming).

⁹ This urban/nonurban stratification has remained quite widespread down to the present day, particularly in Midwestern states with many low-density rural counties. In addition to the Ohio case mentioned in Footnote 7, it is worth noting that as late as 1960, 87 of Missouri's 115 counties had no personal-registration requirements. As a group, these counties had turnout rates about 12 per cent higher than those of the state's two major metropolitan centers. In New York, as of 1949, 10 counties had no personal-registration coverage at all, while a majority of election districts in 35 additional counties also had no such coverage. In 1967 a permanent personal-registration statute was adopted for the entire state.

¹⁰ Pennsylvania is a typical example of the personal-registration transition which is occurring gradually in Ohio and has been completed in New York. The 1906 personal-registration statute applied these requirements to cities of the third class and larger, and was replaced by a statewide registration requirement in 1937. For text of the 1906 law, see any volume of The Pennsylvania (Smull's) Manual between 1907 and 1935. For text of the subsequent statute, see the 1937 or succeeding volumes of this publication.

¹¹ The effectiveness of machine manipulation of the vote in Philadelphia can best be appreciated by examining turnout and voting in two wards: the 12th, a river ward populated heavily by immigrants and solidly controlled by the city Republican organization, and the 21st, a middle-to-upper class and politically independent suburban ward. Statewide turnout rates during this period were 32.5 per cent in 1922, 45.8 per cent in 1924, 30.9 per cent in 1926, 62.6 per cent in 1928 and 40.3 per cent in 1930. The inversion of usual criteria of turnout—especially extreme nationally in the 1920s—is as remarkable in the contrast between lower-class Ward 12 and upper-middle class Ward 21 as in the extraordinary control of partisan outcomes in the former. Without further data presentation, it is also possible to make one more assertion: the decline in turnout among middle-class “independent” wards such as 21 from 1890 through 1930 was vastly heavier than in the lower-class wards which the machine controlled, in the former case approximating the rate of turnout decay in the state's 37 nonregistration counties. If the 1906 statute was designed to protect middle-class and native-stock interests against the big-city machines and their ethnic clienteles, it quite failed to achieve its purpose. In any event, there is a striking parallelism between post-1890 declines in most of those wards of Philadelphia where anti-organization sentiment was strongest and similar behavior in the nonregistration territories of the state at large. Once again—and forcefully—it suggests the intervention of a behavioral variable which can plausibly be linked to political alienation.

¹² One notes the relevant parts of the Keystone (Progressive-Republican) party platform of 1910: “A political trust, managed by cunning politicians, threatens the Commonwealth. Some of the conspirators are labeled Republicans and some Democrats, but they are all in league against the people and act in harmony with one treasonable purpose and under the orders on one head …. Both of the tickets nominated and both of the platforms adopted, it is well understood, were dictated by the same authority and were intended for the delusion of the voters and the further confirmation of the power of the political machine.” The Pennsylvania (Smull's) Manual, 1911, p. 520Google Scholar. It is perhaps also worth noting that the Socialist party reached its off-year apogee of 5.3 per cent of the total vote in this election, with the Keystone candidate winning 38.3 per cent, the Democrat 13.0 per cent and the Republican Organization's candidate 41.6 per cent.

¹³ It is obvious that this metric is based upon an assumption of intercategory homogeneity in every respect other than that of the personal-registration variable. Such an assumption may or may not be partially or wholly valid. More elegant techniques of decomposition are undoubtedly called for. Even so, this is a possibly useful effort at first approximation.

¹⁴ Burnham, , “Changing Shape,” pp. 23–26Google Scholar. See also Burnham, Walter Dean, Critical Elections and the Mainsprings of American Politics (New York: W. W. Norton, 1970), pp. 71–90Google Scholar; and Schattschneider, E. E., The Semisovereign People (New York: Holt, Rinehart & Winston, 1960), pp. 78–96Google Scholar.

¹⁵ The best discussion still remains that of Key, V. O. Jr.,, Southern Politics (New York: Knopf, 1949)Google Scholar. See also Lewinson, Paul, Race, Class and Party (New York: Grosset & Dunlap, 1965 ed.)Google Scholar.

¹⁶ It would be closer still if the partisan means included the Democratic percentages of the total vote cast for all statewide offices in the two periods.

¹⁷ Flinn, Thomas A., “Continuity and Change in Ohio Politics,” Journal of Politics, vol. 24 (August, 1962), 521–544CrossRefGoogle Scholar; Fenton, John H., Midwest Politics (New York: Holt, Rinehart & Winston, 1966), pp. 117–154Google Scholar.

¹⁸ Burnham, , Critical Elections, pp. 91–134Google Scholar, and Appendix Tables I-V.

¹⁹ For a classic presentation of this point for mostly European data before 1937, see Tingsten, Herbert, Political Behavior (London: P. S. King, 1937), pp. 10–78Google Scholar; see also Campbell, Angus, Converse, Philip E., Miller, Warren E. and Stokes, Donald E., The American Voter (New York: Wiley, 1960), pp. 483–493Google Scholar; and Rokkan, Stein, Citizens, Elections, Parties (New York: David McKay, 1970), pp. 385–394Google Scholar.

²⁰ Converse, , “Change in the American Electorate,” pp. 269–270Google Scholar.

²¹ Kentucky's 1920 turnout was 71.8 per cent, more than 20 points above the national average, and falling midway between a mean of 81.2 per cent for the 1900–1916 period and a mean of 63.1 per cent for the period 1924–1940. It is perhaps worth noting that the 1920 presidential election was closer in Kentucky than was any subsequent election except for 1952.

²² The sex-stratified data—available down to the precinct level in the city of Chicago—are found in the 1916, 1917, 1920, and 1921 volumes of the Chicago Daily News Almanac.

²³ Rusk, “The Effect of the Australian Ballot Reform.”

²⁴ Except with regard to “rolloff,” or ballot fatigue, which in Massachusetts underwent a sudden and substantial increase with the introduction of ballot reform. The failure of the partisan indicators to undergo any major change in the same reformera period suggests the probability that ballot fatigue affected the followings of both parties about equally.

²⁵ Burnham, , Critical Elections, p. 115Google Scholar.

²⁶ Ibid., especially pp. 52 and 109–110.

²⁷ Jensen, Richard, The Winning of the Midwest 1888–1896 (Chicago: University of Chicago Press, 1971), especially pp. xii–xivGoogle Scholar. In the present author's view, this seminal work is “must” reading for anyone interested in diachronic analysis of American electoral politics.

²⁸ Converse is particularly concerned to create the impression of endemic and massive voter frauds in nineteenth-century rural areas—not surprisingly, in view of his basic hypotheses—though part of the discussion of the fraud problem is based on Progressive-era formulations by Joseph Harris and others whose focus was urban rather than rural. Converse, pp. 288–293, 300.

²⁹ Jensen, pp. 34–35.

³⁰ Converse, p. 287.

³¹ Hartz, Louis, The Liberal Tradition in America (New York: Harcourt, Brace, 1955)Google Scholar; see also Hartz, Louis, ed., The Founding of New Societies (New York: Harcourt, Brace and World, 1964), especially pp. 1–22Google Scholar.

³² Converse, p. 289.

³³ For an extremely valuable discussion of the intellectual background of American political science generally and of voting analysis in particular, see Jensen's, Richard two articles, “History and the Political Scientist” and “American Election Analysis: A Case History of Methodological Innovation and Diffusion,” in Politics and the Social Sciences, ed. Lipset, Seymour Martin (New York: Oxford University Press, 1969), pp. 1–28 and 226–243Google Scholar.

³⁴ One perhaps suggestive quantitative characteristic of American electoral history has been the close association between periods of critical realignment such as those of 1856–60, 1894–96, and 1928–36 and waves of electoral mobilization or participation increases. Analysis of extreme-polarization situations in current American electoral politics likewise suggests inversions of “conventional wisdom” when racial polarizations are involved: for example, in the Cleveland mayoral election of 1967 the mean turnout in wards more than two-thirds nonwhite was 69.2 per cent, compared with a citywide average of 54.8 per cent; and the Republican (white) candidate received about 80 per cent of a workingclass white vote which has typically been Democratic by margins of 2 or 3 to 1.

³⁵ In addition to Jensen, see also Kleppner, Paul J., The Cross of Culture: A Social Analysis of Midwestern Politics, 1850–1900 (New York: Free Press, 1970)Google Scholar. For a review of this and two other recent works in the “new history,” see Burnham, Walter Dean, “Quantitative History: Beyond the Correlation Coefficient,” Historical Methods Newsletter, 4 (March, 1971), 62–66CrossRefGoogle Scholar.

³⁶ The term was developed in detail by Wiebe, Robert, The Search for Order, 1877–1920 (New York: Hill & Wang, 1967)Google Scholar, a nonquantitative but extremely useful reference for this critical transition period.

³⁷ Jensen, , The Winning of the Midwest, pp. 59–60Google Scholar.

³⁸ Ibid., pp. 61–62, 310–314.

³⁹ Burnham, Walter Dean, “Political Confessionalism and Political Immunization,” Journal of Interdisciplinary History, 3 (Summer, 1972), 1–30CrossRefGoogle Scholar.

⁴⁰ For a lucid and illuminating discussion of this party contribution to Union victory, see McKitrick, Eric, “Party Politics and the Union and Confederate War Efforts,” in The American Party Systems, ed. Chambers, William N. and Burnham, Walter Dean (New York: Oxford University Press, 1967), pp. 117–151Google Scholar.

⁴¹ To a significant extent, this effort may require pursuit of exemplary “case studies” down to a very microscopic, “local-history” level. A good case in point is that of corruption in rural voting among the northern states. Here we may choose the case of Adams County, Ohio, which received extensive discussion in Jensen, The Winning of the Midwest (see also sources cited there). One may summarize this case concisely. (1) Between about 1870 and 1911 the county was the scene of “undoubtedly the worst example of the corruption of the ballot ever known in American history …” (Jensen, p. 38). (2) About 90 per cent of the county's voters came to be involved, with each receiving on the average of $8.75. (3) The parties evidently wanted to curb the practice, but the voters would not let them. (4) In 1911, 1690 men were finally disfranchised for five years by a judge for illegal vote selling (about one-fifth of the electorate). (5) At no time did this lead to turnout figures in excess of 100 per cent, or to landslides for either party. As to the former, it may be inferred that the local (corrupt) voters had no incentive to bring in outsiders to share the spoils; and there were very possibly normative constraints on “repeating” by local voters as well.

This is a curious case indeed; but systemwide inferences about “corruption” in earlier periods must be subjected to such micro examination. Between 1872 and 1908, the mean turnout rate in Adams County was 93.1 per cent of the estimated potential electorate, only 2.9 per cent higher than the state-wide mean turnout rate. Moreover, in 1916—after elections had supposedly been “purified” but the 1,690 disfranchised culprits of 1911 had been restored to their voting rights—the county's turnout rate was 87.3 per cent, now more than 10 per cent above the state's participation rate, but less than 6 per cent below its corrupt pre-1911 mean.

It should also not go without remark that—as is well known—notorious examples of electoral corruption were widely used by the Progressives to obtain enactment of personal-registration statutes. But this rural county, perhaps proportionately the most corrupt of all, was entirely uncovered by such requirements uatil 1962; and a very large part of it is uncovered still.