Hostname: page-component-7c8c6479df-995ml Total loading time: 0 Render date: 2024-03-29T12:34:22.741Z Has data issue: false hasContentIssue false

One-Party Politics and the Voter*

Published online by Cambridge University Press:  02 September 2013

Warren E. Miller
Affiliation:
University of California(Berkeley)

Extract

The existence of one-party systems of politics within the larger framework of the democratic society merits an increased share of the attention of students of American politics. For too long the common conception has been that one-party politics is a problem unique to the Southern states. This misconception persists even though systematic studies such as the monumental work of V. O. Key and Alexander Heard have illustrated that the politics of the so-called Solid South is in many respects the politics of the nation.

As a case in point, an examination of the composition of contemporary legislative delegations indicates that monopolistic control of elective offices is clearly not the exclusive province of the Southern Democrats. In 1955, three out of every four state legislative bodies or congressional delegations were so completely dominated by a single political party that that party controlled more than 66 per cent of the members of the group. Excluding the 15 Border and Southern states, fully half of the remaining 33 state legislatures were controlled by one party holding at least two out of every three seats; in only six states was the controlling margin below 55 per cent. Within the same group of non-Border, non-Southern states, 25 of the 33 congressional delegations were dominated by one party controlling two-thirds or more of the delegation members; only four delegations were so evenly divided as to give the majority party less than 55 per cent of the members.

Type
Research Article
Copyright
Copyright © American Political Science Association 1956

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

References

1 Key, V. O. Jr., and Heard, Alexander, Southern Politics in State and Nation (New York 1949)Google Scholar. See also Key, V. O. Jr., Politics, Parties and Pressure Groups, 3rd ed. (New York, 1953), pp. 224–31Google Scholar, and American State Politics (New York, 1956)Google Scholar.

2 Moos, Malcolm, Politics, Presidents, and Coattails (Baltimore, 1952), pp. 2429Google Scholar.

3 The time base selected for study is also more or less arbitrarily chosen. The entire analysis reported here was also carried out using the average division in votes cast for President over the years 1936–1948. The pervasiveness of the swing to the Republican presidential candidate in 1952 resulted in a relatively even shift of from 10 to 15 percentage points, from the 1936–1948 average, in most of the areas with which we shall be concerned. Consequently, the two analyses (one using the period 1936–1948 as a base, the other using the one year, 1952) are completely parallel with respect to most major findings. The final selection of 1952 as the base year followed from two major considerations. (1) The analysis is not centrally concerned with problems of historical trend or development. The assumption that significant trends did—or, indeed, did not—precede the 1952 election constitutes an assumption that lies outside the framework of the analysis. Although the historical or developmental context of the analysis is, for the most part, ignored in this discussion, it clearly calls for careful examination as a separate enterprise. (2) All of the survey data used in the analysis pertain to the election of 1952 and have been validated only for that election.

4 With the exception of the county election data, all of the data presented in this discussion were produced by the SRC study. A full description of that study, and of the data reported here, may be found in the volume The Voter Decides, by Campbell, A., Gurin, G., and Miller, W. E. (Evanston, 1954)Google Scholar. It should be noted that this investigation of one-party dominance was not anticipated in the design of the SRC study. Many of the apparent inadequacies in the data should be interpreted with this in mind.

5 Preliminary investigation of the Southern and Border state data indicated substantial discontinuities between them and the data representing the rest of the nation.

6 This is an important consideration which governs the categorization of counties. As can be seen in Table I, the “most Republican” counties were in 1952 more extremely pro-Republican than the “most Democratic” counties were comparably pro-Democratic. If the 1936–1948 average were used, instead of the data for 1952, the situation would be reversed (without seriously rearranging the counties), with few or no counties being as extremely pro-Republican as their most pro-Democratic opposite numbers. The difficulty of attempting to decide what absolute standards are sufficient to define “extremely Republican” or “extremely Democratic” under two such different situations as represented by the 1936–1948 period and the one year 1952 is avoided, without adjusting or otherwise violating the data, by the use of the relative “most Republican” and “most Democratic” designations.

Tables II, III, and IV, and the exploratory analyses on which they are based, provide an empirical justification for talking about counties as being distributed on a continuum of greater-to-lesser Democratic (or Republican) dominance. Adoption of a fixed breaking point to define one-party control would seem to imply a discontinuity, or at least curvilinearity, in relationships which, in fact, has not been observed in these data. One might assume, for example, that a partisan division as extreme as 70–30 would indicate a qualitatively different situation from that indicated by any more evenly balanced division, such as 65–35, 58–42, 51–49. One might assume further that once the 70–30 point was reached, more extreme divisions would not alter the essentially “one-party” nature of the situation. If this set of assumptions were to be verified, one would then expect a general case of relationships involving party dominance and any other related factor to be something like that in Figure 1. In Figure 1, of course, co-variation between party dominance and Factor “X” is found only in the middle of the two-party vote range and disappears when either party gains complete control. In all relevant instances in the analysis with which this discussion is concerned, the observed relationships were of the general type where co-variation extended in an apparently linear fashion from one extreme of partisan dominance to the other as in Figure 2. There was no indication that a real breaking point did in fact exist near either end of the dominance continuum. Consequently, the use of the concept of a continuum of relative dominance not only provides a needed solution to the methodological problem involved, but also appears to provide the solution most compatible with the relevant data. This is not, of course, intended to constitute an argument that all relationships involving party dominance are linear, nor that they are necessarily continuous.

7 Inasmuch as the individual vote-motivation relationships have been established elsewhere (see The Voter Decides, Chs. 7–9), Table II indicates that these relationships are so high that they persist even when one variable (voting behavior) is only indirectly reflected by the two-party division of all voters in the counties in which the possessors of these motivations being measured reside.

8 Because of the complexity of the data involved, the relative margin of support (or opposition) will be used throughout the analysis as a summary index of the distributions being examined.

9 An extension of the analysis indicates that the uniformity of the three motivation-vote relationships is not merely the result of the fact that each motivational factor is correlated with each of the other two as well as with the vote. First, the voting behavior of persons with different motivational patterns was examined. Persons possessing strong pro-Democratic motivations (pro-Democratic on all three factors or pro-Democratic on two and neutral on one) voted more heavily Democratic in extreme Democratic counties (with a Voting Partisanship Index of +70) than in extreme Republican counties (where their Vote Index dropped to +36). Similarly, persons exhibiting mildly pro-Democratic motivations (pro-Democratic on one factor and neutral on two, or pro-Democratic on two and pro-Republican on one) varied consistently from a +43 Vote Index in Democratic territory to +24 in Republican country. The comparable data for persons with Republican motivations showed the same pattern of consistent relationships.

In a second test, each segment of Table III was reproduced, but with the second and third motivational factors controlled. Although the number of comparisons was limited because of insufficient cases in many cells, this procedure again supported the thesis that each relationship observed in Table III exists independently of the other two factors. It should also be noted that the functional relationship between the county vote and the individual vote raises the possibility that the persistence of this relationship under the controlled conditions depicted in Table III is artifactual. Although this possibility cannot be denied with certainty, the preponderance of judgments obtained from statisticians on this point supported the presumption that the findings were, in fact, not spurious.

10 Schattschneider, Elmer E., Party Government (New York, 1942), pp. 7480Google Scholar.

11 This suggestion is made without benefit of insight into-the nature of the effects which county environment may have on the evolution of political motivations. Although this is a fit subject for further study, we are here concerned only with the subsequent problem: given motivating factors, however developed, what effect does county environment have on the relationship between them and individual voting behavior?

12 It should be noted that the presence of a relatively largo number of persons with no observed partisanship, relative to any one motivational factor, and the existence of a sizable group of non-voters allows these two ways of studying the vote-motivation relationship to produce results which are complementary but which have quite different implications. See note 9, above.

13 Hereafter, when intra-party comparisons of groups of voters are made, the terms minority and majority will be used. The discrimination will be made only on the base of lesser and greater support from county environment. Thus, Democratic voters in counties voting 41–50 per cent Democratic will be minority voters when compared to Democratic voters in the 51–60 per cent Democratic category, and they will be majority voters when compared to Democratic voters in counties voting 31–40 per cent Democratic.

14 For either Issue or Candidate Partisanship, the probability is less than one out of 100 that the observed relationship occurred by chance. For Party Identification, the combined data for both Democratic and Republican voters approach a completely random distribution. Out of 12 possible comparisons, six conform to the pattern observed for Issue and Candidate Partisanship and six run directly counter to that pattern.

15 Key, V. O. Jr., “The Direct Primary and Party Structure: A Study of State Legislative Nominations,” this Review, Vol. 48, pp. 126 (March, 1954)Google Scholar.

16 Another possible explanation might be thought to lie in the distribution of 1948 Truman supporters who voted Republican in 1952. Some 10 or 11 per cent of the residents in each group of counties are to be found in this category of 1948–1952 switchers. Removing them from the columns of Republican voters does not, however, greatly change the entries in Table IV and does not eliminate the aberrations.

17 Because our concern is with understanding the behavior of voters, we shall not examine in detail those data in Table IV pertaining to non-voters. Nevertheless, it should be noted that the non-voters appear to be extremely sensitive to their political environment; with but one deviation they reflect in all three motivations the relative impact of the postulated differences in environment.

18 If one were to assume that under most circumstances Party Identification, too, gives less support to minority voters than to majority voters, the problem of explaining the minority vote would, of course, only be accentuated.

19 For discussions of the use of these items as indicators of instability see Janowitz, Morris and Miller, Warren E., “The Index of Political Predisposition in the 1948 Election,” Journal of Politics Vol. 14, pp. 710–27 (June, 1952)CrossRefGoogle Scholar, and also The Voter Decides, pp. 17–26. It should be noted that the absence of supporting partisanship is not the same as absence of the motivational factor. Of the two dimensions measured for each factor, intensity or strength and direction or partisanship, we are concerned only with the latter.

20 It is, of course, conceivable that we are observing a rather unique threshold-contingency relationship between motivational factors and behavior. If an individual's motivation is sufficient to produce a vote, it is also sufficient to produce the entire range of behaviors open to citizens; the addition of motivational factors, as in the case of the majority voters, does not produce any additional variance in behavior. Such an explanation is less than satisfactory on theoretical grounds and runs counter to observed relationships involving the same data. See The Voter Decides, Chs. 10, 11.

21 Berelson, B., Lazarafeld, P. F., and McPhee, W. N., Voting (Chicago, 1954), p. 126 and note 2, pp. 126–27Google Scholar.

22 The authors of Voting were, of course, aware that external events do alter the equilibrium which they posit in their analysis.

23 Even if it were asaumed that all inter-party exchange was provided by the same relatively small group of changers who circulate, election-to-election, from one party to the other, still we should find at any one point in time that minor party members were relatively more heterogeneous in terms of previous political affiliation and experience, than their counterparts who were fortunate enough to reside in predominantly friendly territory.

24 It might be noted in passing that serious attention could not be given this admonition in the current analysis because of the restricted size of the sample. Early efforts to analyze the various patterns of motivation which might be associated with varying states of one-party dominance were abandoned when it became obvious that the number of cases in crucial cells too frequently diminished beyond the point of no return.