We use cookies to distinguish you from other users and to provide you with a better experience on our websites. Close this message to accept cookies or find out how to manage your cookie settings.
Published online by Cambridge University Press: 01 August 2014
The middle-class character of the leadership of American reform movements has been well established. While leaders of the progressive movement in early twentieth-century America also conform to this pattern, the nature of the voting base of support for progressivism has not yet been established. The purpose of this paper is twofold: first, to examine whether or not middle-class voters supported progressive candidates at the polls; and secondarily, to test the relative strength of cultural factors (i.e., ethnicity and religion) versus class considerations as determinants of voting behavior in the early twentieth century. The results demonstrate that, at least in the key progressive state of Wisconsin, middle-class voters failed to support progressive candidates in either general or primary elections; to the contrary, they provided the bulwark of support for conservative opponents of reform. Using bi-variate and partial correlational analysis, the paper also shows that ethnocultural factors remained as the most powerful determinant of voter choice among urban voters in general elections, but that class considerations often proved more influential in motivating voters in primary election contests.
Research for this study was supported in part by grants from the American Philosophical Society and the Research Council of Rutgers University. The author is indebted to Michael Ebner of City College, CUNY, to James Wright of Dartmouth College, and especially to Warren F. Kimball and John C. Blydenburgh of Rutgers University for their critical comments on an earlier draft. The author would also like to acknowledge the assistance of David McLaury in collecting and manipulating the data, and of Thomas Hayton and John Dickinson in analyzing data and for suggestions regarding the construction of tables.
1 Mowry, George E., The California Progressives (Berkeley: University of California Press, 1951), chap. 4, particularly 87–89 CrossRefGoogle Scholar; Chandler, Alfred D. Jr., “The Origins of Progressive Leadership,” in Morison, Elting E., ed., The Letters of Theodore Roosevelt (8 vols., Cambridge: Harvard University Press, 1954), VIII, 1062–1065 Google Scholar. Mowry later extended his research to include more than 400 progressive leaders, arriving at the same conclusions. Mowry, , The Era of Theodore Roosevelt, 1900–1912 (New York: Harper, 1958), chap. 5Google Scholar.
2 Hofstadter's argument is too well-known to require any more extensive elaboration here; see The Age of Reform (New York: Knopf, 1955), chap. 4Google Scholar. Soon afterward, “status displacement” theories were applied with somewhat reckless abandon to several aspects of American history. A good summary of this development is McFarland, Gerald W., “Inside Reform: Status and Other Evil Motives,” Soundings, 54 (Summer 1971), 164–165 Google Scholar.
3 The Age of Reform was initially assailed as leaning towards a psychological interpretation of history. See, for example, Pollack, Norman, “Hofstadter on Populism: A Critique of ‘The Age of Reform,’” Journal of Southern History, 26 (11 1960), 478–500 CrossRefGoogle Scholar, and Hesseltine's, William B. review of the book in Wisconsin Magazine of History, 39 (Summer 1956), 280 Google Scholar. The best criticism along this line is Thelen, David P., “Social Tensions and the Origins of Progressivism,” Journal of American History, 66 (09 1969), 323–341 CrossRefGoogle Scholar, which uses recent findings in the behavioral sciences to refute Hofstadter.
Other scholars have emphasized the coalitional nature of progressivism, especially the contribution made by non-middle-class groups. See, for example, Huthmacher, J. Joseph, “Urban Liberalism and the Age of Reform,” Mississippi Valley Historical Review, 49 (09 1962), 31–41 CrossRefGoogle Scholar; and Buenker, John D., “Urban, New-Stock Liberalism and Progressive Reform in New Jersey,” New Jersey History, 87 (Summer 1969), 79–104 Google Scholar, and “The Progressive Era: A Search for a Synthesis,” Mid-America, 51 (07 1969), 175–193 Google Scholar.
4 The best of these is Sherman, Richard B., “The Status Revolution and Massachusetts Progressive Leadership,” Political Science Quarterly, 78 (03 1963), 61–65 CrossRefGoogle Scholar. See also Kerr, William T. Jr., “The Progressives of Washington, 1910–1912,” Pacific Northwest Quarterly, 55 (01 1964), 16–27 Google Scholar; Tager, Jack, “Progressives, Conservatives and the Theory of the Status Revolution,” Mid-America, 48 (07 1966), 162–175 Google Scholar; and Potts, Eli Daniel, “A Comparative Study of the Leadership of Republican Factions in Iowa, 1900–1914” (Master's thesis, State University of Iowa, 1954)Google Scholar, part of which was published as “The Progressive Profile in Iowa,” Mid-America, 47 (10 1965), 257–268 Google Scholar.
5 Kerr included a brief analysis of the type of electoral districts which sent either progressives or stalwarts to the Washington legislature. “Progressives of Washington,” pp. 22ff.
6 Saxton, Alexander, “San Francisco Labor and the Populist and Progressive Insurgencies,” Pacific Historical Review, 34 (11 1965), 421–438 CrossRefGoogle Scholar; Rogin, Michael P., “Progressivism and the California Electorate,” in Rogin, and Shover, John L., Political Change in California: Critical Elections and Social Movements, 1890–1966 (Westport, Conn.: Greenwood Publishing Co., 1970), pp. 35–61 Google Scholar; Shover, , “The Progressives and the Working-Class Vote in California,” in Rogin, and Shover, , Political Change in California, pp. 62–89 Google Scholar.
7 The best account of Wisconsin politics during this era is Margulies, Herbert F., The Decline of the Progressive Movement in Wisconsin, 1890–1920 (Madison: State Historical Society of Wisconsin Press, 1968)Google Scholar. See also Maxwell, Robert S., La Follette and the Rise of the Progressives in Wisconsin (Madison: State Historical Society of Wisconsin Press, 1956)Google Scholar.
8 The superiority of the Wisconsin data helps to avoid some of the deficiencies of the California studies. For example, the only ethnic variable used by either Rogin or Shover is “per cent foreign born.” Since this is not broken down into smaller ethnic group components which might not vote alike, it tends to be largely meaningless: in most states, there never was any unified “foreign vote” because of basic religious and other cultural antagonisms among immigrant groups. The data for Wisconsin, however, permits the use of individual ethnic variables at the minor civil division level; it also permits the division of “working class” into skilled and unskilled categories.
Another problem with the work of Rogin and Shover is that they deal statistically with only one candidate in each election, usually Johnson. Thus votes for Johnson are equated with “progressive” votes; there is no comparison with other candidates running on the same ticket to serve as a control for Johnson's personal ap-peal or lack of it. As shown below, a similar analysis for Wisconsin using only the vote for La Follette would produce distorted results.
9 Wisconsin State Census, 1905 (Madison, 1906), I, 276–601 Google Scholar.
10 Middle-class percentages ranged from 6.2 in Oshkosh's Ward 6 to 55.8 in La Crosse's Ward 17. The average percentages of middle-class residents for Milwaukee and La Crosse were 23.8 and 24.4, respectively.
11 The greater range of the upper class plus middle class variable (from 7.6 per cent in Oshkosh, Ward 6 to 60.4 per cent in La Crosse, Ward 17) plus its higher mean (28.9 per cent in Milwaukee and 29.6 per cent in La Crosse) give it greater statistical reliability than either upper class or middle class alone. Only four wards—two in La Crosse and one each in Milwaukee and Madison—had upper-plus middle-class populations in excess of 50 per cent.
12 Kleppner, Paul J., The Cross of Culture: A Social Analysis of Midwestern Politics, 1850–1900 (New York: The Free Press, 1970)Google Scholar; Jensen, Richard, The Winning of the Midwest: Social and Political Conflict, 1888–1896 (Chicago: University of Chicago Press, 1971)Google Scholar. See also Swierenga, Robert P., “Ethnocultural Political Analysis: A New Approach to American Ethnic Studies,” Journal of American Studies, 5 (04 1971), 66–79 CrossRefGoogle Scholar.
13 Wyman, Roger E., “Voting Behavior in the Progressive Era: Wisconsin as a Case Study” (Ph. D. dissertation, University of Wisconsin, 1970), chaps. 8–15Google Scholar.
14 The term “ethnocultural factors” here refers to the role of both ethnicity and religion, even though only ethnic variables are used in the analysis. During this era, religion was at least as important as, if not more important than, ethnicity as a determinant of party identification and voting behavior. Unfortunately, the lack of any published religious data at the minor civil division level makes it impossible to use religion as a quantifiable variable. With the exception of the Germans and a few other groups who were divided internally along religious lines, ethnicity and religion tended to be indivisible. Two good examples are the Poles and the Scandinavians. For a general discussion of the role of religion in Wisconsin politics during the Progressive Era, see Wyman, “Voting Behavior,” chaps. 8, 9.
15 The variable “percentage Anglo-American” used in this study is composed of native-born of native parents plus British stock (English plus Scots plus Welsh).
16 In 1900, McKinley polled 60.06 per cent of the total vote to 59.83 per cent for La Follette, a minuscule difference of .23 per cent. In 1908, Taft won 54.01 per cent to 54.74 per cent for Davidson, a difference of .73 per cent. In striking contrast, La Follette ran 12.7 per cent behind Roosevelt in 1904.
For a discussion of Wisconsin election campaigns, see Margulies, Progressive Movement in Wisconsin, chaps. 2, 3; and Wyman, “Voting Behavior,” chap. 7. On La Follette in 1900, see Twombly, Robert C., “The Reformer as Politician: Robert M. La Follette in the Election of 1900” (Master's thesis, University of Wisconsin, 1964), pp. 58, 61–100 Google Scholar.
17 In Table 1 and other tables, some ethnic and occupational variables (usually “middle class” and either “German” or “Democratic ethnic groups”) do not ap-pear for some cities. As explained earlier, there is no need to use both “middle class” and “upper plus middle class” as distinct variables. Similarly, in a few cities per cent German and per cent Democratic ethnic groups are virtually identical in their distribution, and one variable is omitted to prevent needless duplication. In Sheboygan, for example, the two upper- plus middle-class wards and the Republican ethnic group wards were the same.
18 The source for all ward level ethnic data is “Cultural-Ethnic Backgrounds in Wisconsin” (11-volume typewritten manuscript in possession of Rural Sociology Department, University of Wisconsin), a retabulation of the manuscript returns for the Wisconsin State Census, 1905. All ethnic percentages are the percentages of total family heads (the base) who are of foreign stock (foreign-born plus native-born of foreign parents). The “Republican” and “Democratic” ethnic variables were produced by adding the percentages of total family heads for each of the groups mentioned above who fell into either category. All election data comes either from the Wisconsin Blue Book for the year following the election or from the reports of the county boards of canvassers, in the Archives and Manuscripts Division, State Historical Society of Wisconsin. For a more comprehensive explanation of the methodology involved, see the methodological appendix in Wyman, “Voting Behavior.”
19 The Social-Democrats grew at the expense of both the Republican and Democratic parties. For a comprehensive account of Milwaukee politics in this era, see Still, Bayrd, Milwaukee: The History of a City (Madison: State Historical Society of Wisconsin Press, 1948), chaps. 11, 12Google Scholar. On the Socialists, see Wachman, Marvin, History of the Social-Democratic Party of Milwaukee, 1897–1910 (Urbana: University of Illinois Press, 1945)Google Scholar.
20 Although the Social-Democrats were strongest among German skilled workers, religious factors were also present. Few Catholic Germans voted Socialist, and most German Socialists came from the freethinker and Protestant elements. In a recrudescence of anti-Catholicism in 1914 and thereafter, Social-Democratic candidates invariably won the approval from militant anti-Catholic organizations.
21 In Milwaukee, per cent upper plus middle class correlated .498 with per cent Anglo-American and .411 with the Republican ethnic variable. In La Crosse, the correlation coefficients were .507 and .383, respectively.
22 For each dependent variable, the standard regression equation [Y (vote) = a + b 1 x l (ethnicity) + b 2 x 2 (class)] was calculated, using the stepwise method.
For a comprehensive explanation of both correlation and regression, see Blalock, Hubert M., Social Statistics (New York: McGraw-Hill, 1960), chaps. 17–19Google Scholar, and Dollar, Charles M. and Jensen, Richard J., Historian's Guide to Statistics: Quantitative Analysis and Historical Research (New York: Holt, Rinehart, and Winston, 1971), chap. 3.Google Scholar
23 Milwaukee and La Crosse are the only two of the six cities with enough wards to make statistical analysis reliable. The normal N in all tables involving correlation coefficients is 23 for Milwaukee (19 in 1900) and 21 for La Crosse (20 in 1900). In all calculations, one or a few cases in some variables were removed after scattergram analysis revealed that they might be exerting undue influence upon the correlations. The most serious of these was the removal of three wards from the middle-class and upper- plus middle-class variables in Milwaukee—those wards with the highest percentages; thus all calculations involving those variables have an N of 20 rather than 23. This was done to increase the reliability of the analysis, despite the fact that removal of the three wards reduced the correlations (e. g., the correlation between GOP vote, 1908, and upper plus middle class is .842 with an N of 23 compared to .706 with an N of 20). For an explanation of the effect of extreme cases and the rationale for their removal, see Blalock, , Social Statistics, pp. 290–292 Google Scholar.
24 The beta is the standardized regression coefficient. The author feels that in the case at hand it is superior to other measures as an indicator of the relative strength of each independent variable. See Blalock, , Social Statistics, pp. 344–346 Google Scholar; and Dollar, and Jensen, , Historian's Guide to Statistics, pp. 88–89 Google Scholar. Other results, such as the b's, standard error of b, and partial r's, were eliminated for the sake of clarity and simplicity in the tables.
25 Significance levels state the probability that results obtained could be the product of random variation. A P level of .05 means that there is a five per cent chance that the result could be accidental, provided there was no true relationship; a level of .001 reduces that chance to one in a thousand.
26 For the politics of the La Follette administrations, see Margulies, Progressive Movement in Wisconsin, chap. 2, and Barton, Albert O., La Follette's Winning of Wisconsin (Madison, 1922)Google Scholar.
27 Kenosha, Sheboygan, and Oshkosh are omitted from Table 3 only because of space limitations. The voting patterns were similar to those of La Crosse and Milwaukee.
28 Correlation coefficients between the same ethnic and occupational variables and the rest of the Republican ticket—represented by the vote for secretary of state—fell roughly midway between those with Roosevelt and La Follette. For example, the correlation between upper plus middle class and vote for secretary of state was .346. These results indicate that some stalwarts cut the progressive state ticket as well as La Follette, but that many other stalwarts voted the entire GOP ticket except for governor.
29 These results are not included in the tables.
30 The strong reaction along both ethnic and class lines to the Social-Democratic party produces the rare phenomenon of both Republican and Democratic votes correlating positively (or negatively) with the same social variable. The difference in the Social-Democratic percentages for the two offices represents a defection of about 800 normally Socialist votes that went to La Follette.
31 The vote for governor is the only Social-Democratic variable presented because there was no discernible difference in the correlation coefficients involving the Social-Democratic vote for president, governor, or secretary of state. The Social-Democratic vote was not included in Table 4 because it was of little consequence in La Crosse; in 1904, Social-Democratic candidates for president and governor received more than three per cent of the vote in only one ward.
32 The La Follette faction of the party had assiduously courted Scandinavian voters since 1894. For the voting patterns of Wisconsin Scandinavians, see Wyman, “Voting Behavior,” chap. 12.
33 A possible explanation of why ethnicity was more influential in Milwaukee is that while Anglo-American middle-class voters rejected La Follette, middle-class Germans supported him.
34 In order to obtain relatively reliable indicators of stalwart bolting, the vote for Roosevelt and Taft was added together to get an approximation of the true Republican vote in 1912, and McGovern's vote was then subtracted from it; a similar comparison was made by subtracting Wilson's vote from Karel's.
35 This theme is developed more fully in Wyman, “Voting Behavior,” chap. 7.
36 In Kenosha, for example, in 1910 Lewis, from nearby Racine, won three of the city's eight wards, severely undercutting McGovern in that city. In Oshkosh, the presence of local candidates in both 1910 and 1914 complicates any analysis, but nothing occurred to contradict the general pattern of weak middle-class support for progressive candidates.
37 McGovern won easily in La Crosse, capturing 20 of 21 wards. There were no distinguishable ethnic or class patterns in his vote, and analysis of it is omitted from Table 8.
38 Buenker, , “The Progressive Era,” pp. 184–192 Google Scholar.
