Much of the scholarly interest in critical realignments results from the pivotal role that ordinary citizens play during these periods. By altering their voting behavior, citizens hold political elites accountable and forge non-incremental change in policy outputs. A central question regarding realignments is thus how are citizens changing their behavior to hold elites accountable? Are citizens producing realignments by converting from one party to the opposition? Are previous non-voters becoming mobilized in response to emerging issues or crises? Or are one party’s supporters disproportionately abstaining from voting and altering the partisan balance in the process? This article makes four central contributions to our understanding of these realignment processes, or dynamics. We present a theoretical framework for the analysis of realignment dynamics, based upon the Michigan model of voting and its conception of the normal vote. Where previous dynamics studies have collectively only examined two realignments, we examine the dynamics of all presidential realignments in American electoral history. Where previous studies have often focused on national, sectional, or state levels of analysis, we focus on city- and county-level realignments, a critical advancement for an inherently local-level phenomenon such as critical realignments. Finally, unlike previous studies, we identify the factors that promote particular realignment dynamics. We find that the conversion of active partisans has produced most of the enduring change in voting behavior in the United States, with the relative contribution of different dynamics varying both across time and space. Political factors such as the strength of state and local parties and demographic factors such as changes in the size of local immigrant populations have each favored particular realignment dynamics in American electoral history.
This is a preview of subscription content, access via your institution.
Buy single article
Instant access to the full article PDF.
Tax calculation will be finalised during checkout.
Subscribe to journal
Immediate online access to all issues from 2019. Subscription will auto renew annually.
Tax calculation will be finalised during checkout.
As we discuss later, conversion is also an efficient realignment dynamic that is capable of producing marked changes in the partisan balance, as during conversion a voter is simultaneously subtracted from one’s party ledger and added to the opposing party’s ranks.
The data for this analysis are part of a demographic, electoral, and political archive collected by Peter F. Nardulli and a team of researchers at the University of Illinois at Urbana-Champaign. The data set includes observations for all counties and most major cities in the continental United States for each presidential election from 1828 to the present.
As Key (1959) recognized long ago, enduring electoral change can occur either through critical realignments (immediate structural shifts in voting patterns) or through secular change (gradual changes in voting patterns). The focus of this article is on critical realignments.
The success of the approach used here can be appreciated by comparing the average change in the normal vote across pairs of presidential elections for cities and counties that experienced critical realignments and those that did not. Local electorates which experienced a critical realignment according to our criteria had an average absolute change in the normal vote of 41 percentage points. Local electorates coded as not experiencing a critical realignment had an average absolute change in the normal vote of only 2 percentage points. The critical elections analyzed here thus capture important and dramatically distinctive events in American electoral history.
These values are weighted by the absolute value of an electoral unit’s change in the normal partisan balance and total votes, which varies dramatically across urban and rural units.
Separate analyses show that demobilization has played a larger role in producing secular electoral change. Thirty-seven percent of all secular change has been produced by demobilization, 44% by conversion, and only 19% by mobilization.
Because demobilization plays such a minor role in generating critical change we do not examine it in this analysis.
Andersen, K. (1979). The creation of a democratic majority, 1928–1936. Chicago, IL: University of Chicago Press.
Beck, P. A. (1974). A socialization theory of partisan realignment. In R. G. Niemi, et al. (Eds.), The politics of future citizens. San Francisco: Jossey-Bass.
Beck, P. A. (1982). Realignment begins? The Republican surge in the south. American Politics Quarterly, 10, 421–438.
Binkley, W. E. (1943). American political parties: Their natural history. New York, NY: Alfred A. Knopf.
Brown, C. (1991). Ballots of tumult: A portrait of volatility in American voting. Ann Arbor, MI: University of Michigan Press.
Burnham, W. D. (1970). Critical elections and the mainsprings of American politics. New York, NY: W. W. Norton.
Campbell, J. E. (1985). Sources of the new deal realignment: The contributions of conversion and mobilization to partisan change. Western Political Quarterly, 38, 357–376.
Campbell, A., Angus, P. E., Miller, W. E., & Stokes, D. E. (1960). The American voter. New York, NY: Wiley.
Clubb, J., Flanigan, W. H., & Zingale, N. (1980). Partisan realignment: Voters, parties, and government in American history. Beverly Hills, CA: Sage Publications.
Converse, P. E. (1966). The concept of a normal vote. In A. Campbell, P. E. Converse, W. E. Miller, & D. E. Stokes (Eds.), Elections and the political order. New York: Wiley.
Darmofal, D. (2008). The political geography of the new deal realignment. American Politics Research, 36(6), 934–961.
Erikson, R. S., & Tedin, K. L. (1981). The 1928–1936 partisan realignment: The case for the conversion hypothesis. American Political Science Review, 75, 951–962.
Key, V. O., Jr. (1955). A theory of critical elections. Journal of Politics, 17, 3–18.
Key, V. O., Jr. (1959). Secular realignment and the party system. Journal of Politics, 21, 198–210.
Kleppner, P. (1987). Continuity and change in electoral politics, 1893–1928. New York, NY: Greenwood Press.
Ladd, E. C., Jr., & Hadley, C. D. (1978). Transformations of the American party system. New York, NY: W.W. Norton.
MacRae, D., Jr., & Meldrum, J. A. (1960). Critical elections in Illinois: 1888–1958. American Political Science Review, 54(3), 669–683.
Nardulli, P. F. (1995). The concept of a critical realignment, electoral behavior, and political change. American Political Science Review, 89(1), 10–22.
Nardulli, P. F. (2005). Popular efficacy in the democratic era: A reexamination of electoral accountability in the United States, 1828–2000. Princeton: Princeton University Press.
Niemi, R. G., Katz, R. S., & Newman, D. (1980). Reconstructing past partisanship: The failure of the party identification recall questions. American Journal of Political Science, 24(4), 633–651.
Shively, W. P. (1992). From differential abstention to conversion: A change in electoral change, 1864–1988. American Journal of Political Science, 36, 309–330.
Sundquist, J. L. (1983). Dynamics of the party system: Alignment and realignment of political parties in the United States. Washington, DC: The Brookings Institution.
Wanat, J., & Burke, K. (1982). Estimating the degree of mobilization and conversion in the 1890s: An inquiry into the nature of electoral change. American Political Science Review, 76, 360–370.
We thank the editors, the anonymous reviewers, Brian Gaines, and James Kuklinski for helpful comments and suggestions.
Assumptions Underlying the Logic of Dynamic Processes
The key assumption underlying the logic used to define the realignment dynamics variables is that the electoral shifts reflected in the difference patterns discussed above are the result of the simplest conceivable movement involving the smallest number of voters: the simplest movement of the smallest number assumption. For a simple mobilization pattern, (+,0,+,0), the “simplest movement of the smallest number” assumption would mean that the increases in core voters and core Republican voters are due simply to an influx of new core voters who aligned with the Republicans.
However plausible the simplest movement of the smallest number assumption is, it is not the only possible interpretation of a difference pattern such as (+,0,+,0). Consider the case in which the size of the core electorate and the size of the Republican constituency for a given county in a particular election both increased by 1,000 voters between t 0 and t 1. The simplest movement of the smallest number assumption attributes these changes to a mobilization of 1,000 non-core voters who became core Republican partisans. However, those increases could also have been generated by the following pattern of movements: (1) a demobilization of 1,000 core voters who were core Democratic partisans at t 0, perhaps a cohesive group that became disaffected over a series of policy changes or political events (e.g., a scandal, a recession, etc.); and (2) a mobilization of 2,000 non-core voters who divide their loyalties equally between the Democrats and Republicans.
While these electoral movements would result in the generation of a (+,0,+,0) difference pattern, this interpretation is less credible than that provided by the simplest movement of the smallest number assumption, for two reasons. First, the more elaborate movements described above require changes in the habitual behavior of three times as many voters (3,000 as opposed to 1,000). This, of course, makes this account of the electoral shifts less consistent with what we know about individual-level voting behavior. Second, if the increases in core voters and core Republican voters are equal, the more elaborate interpretation of the electoral movements requires that the political disturbances generating the changes in habitual behavior affect an equivalent number of core Democrats, non-core Democratic sympathizers, and non-core Republican sympathizers. This strains credibility on two counts. The most obvious is the numerical equivalency. Less obvious is that events of sufficient magnitude to mobilize non-core voters are not likely to activate non-core sympathizers of both parties in the same place at the same moment in time.
The consequences of fundamental societal transformations (the challenges of settling the frontier or dealing with the industrial revolution), major economic events (the Great Depression), or major scandals (Watergate) are not likely to be experienced differently by previously uninvolved citizens in the same locale at a given point in time. For example, while it is conceivable that Hoover’s handling of the Depression could alienate core Republican voters, it is unlikely that it would also mobilize Republican sympathizers among the non-core electorate and motivate them to become electorally active Republicans. It is even more unlikely that an equivalent number of core and non-core Republicans would be affected differently by Hoover’s conduct.
Despite the fact that the simplest movement of the smallest number assumption is more plausible and more compatible with individual-level voting research, it is impossible to refute rival hypotheses definitively using only aggregate data. It is, nonetheless, possible to conduct an empirical test that will shed some light on the viability of the simplest movement of the smallest number assumption, as well as the plausibility of rival hypotheses. The “equivalency test” offered here examines the equivalency of the electoral shifts that the assumption assumes to be the only factors driving enduring electoral change. Consider, for example, a county that experienced a critical realignment that favored the Republicans and in which all of the critical change was attributed to mobilization. In such a case, all of the observed changes should be in the proportions of core voters and core Republican voters. If this county had an increase in the proportion of core voters of .10, an increase in the proportion of core Republican voters of .10, and no changes in the proportions of core Democratic or minor party voters, then it would qualify as evidencing equivalency under the equivalency test.
A finding that the shifts assumed to be determinative are equivalent would provide empirical support for the simplest movement of the smallest number assumption in two ways. First, it would demonstrate that there is not much electoral change that is unaccounted for by the analyses. Second, it would make rival hypotheses more implausible because of the tenuousness of equivalent shifts among multiple sets of voters. The empirical examination of the equivalency test provides solid support for the simplest movement of the smallest number assumption. Two versions of the equivalency test were constructed. The “near equivalency test” shows that 97% of the electoral shifts analyzed were within ±.01 of one another. The “identical equivalency test” found that 85% of the electoral shifts were identical.
Coalitional Invariance and the Simplest Movement of the Smallest Number Criterion
The second test that can be conducted to provide some empirical sense of the viability of the simplest movement of the smallest number criterion is the coalitional invariance test. When a single dynamic process is producing the electoral change in a city or county, the coalitional makeup of at least one component of the local electorate will necessarily remain constant as a consequence. For example, consider the mobilization thesis. If the net gain experienced by the Democratic Party in a county is due entirely to the mobilization of new voters, then the coalitional makeup of the core Republican voting population will be unchanged in the election. This invariance in the coalitional make-up of electoral constituencies is less likely to hold if electoral change is produced by a more complex set of movements, that is, if the simplest movement of the smallest number assumption does not hold.
The coalitional invariance test can be expressed very concretely in the form of six hypotheses:
As noted above, if all of the electoral change in a local electorate is due to the mobilization of new voters to one of the parties, then the coalitional makeup of the core voting population for the party not attracting new voters will be unchanged. This leads to two separate hypotheses, one for cases of Democratic mobilization and one for cases of Republican mobilization.
If the net gain experienced by the Democrats is due entirely to the mobilization of new voters, then the coalitional makeup of the local Republican core constituency will be unchanged from t 0 to t 1.
If the net gain experienced by the Republicans is due entirely to the mobilization of new voters, then the coalitional makeup of the local Democratic core constituency will be unchanged from t 0 to t 1.
If all of the electoral change in a local electorate is due to the demobilization of voters from one of the parties, then the coalitional makeup of the core voting population for the other party will be unchanged. This leads to two hypotheses, one for cases of demobilization favoring the Democratic Party and one for cases of demobilization favoring the Republican Party.
If the net gain experienced by the Democrats is due entirely to the demobilization of core voters, then the coalitional makeup of the local Democratic core constituency will be unchanged from t 0 to t 1.
If the net gain experienced by the Republicans is due entirely to the demobilization of core voters, then the coalitional makeup of the local Republican core constituency will be unchanged from t 0 to t 1.
If all of the electoral change in a local electorate is due to the conversion of core voters and not to the demobilization of core voters or the mobilization of new voters, then the coalitional makeup of the core voting population will be unchanged. This leads to two hypotheses, one for cases of conversion favoring the Democratic Party and one for cases of conversion favoring the Republican Party.
If the net gain experienced by the Democrats is due entirely to the conversion of core voters, then the demographic makeup of the local core voter constituency will be unchanged from t 0 to t 1.
If the net gain experienced by the Republicans is due entirely to the conversion of core voters, then the demographic makeup of the local core voter constituency will be unchanged from t 0 to t 1.
A test of these hypotheses can be conducted using a subset of the cases involving enduring electoral change in conjunction with data on the ethnic, religious, and racial composition of local electorates. Four summary ethnic composition variables were used (Northern European, Southern European, Eastern European, Scandinavian) along with a Foreign Born variable. Each of these variables is a measure of the proportion of the city- or county-level population that were members of the relevant ethnic or immigrant population. Eight religious affiliation variables capturing the largest denominations were also employed (Baptist, Congregationalist, Episcopalian, Jewish, Lutheran, Methodist, Presbyterian, Roman Catholic), along with one racial composition variable, African American. As with the earlier variables, each of these variables measures the proportion of the city- or county-level population that were members of these populations.
Only a subset of the cases in which enduring partisan gains were generated by a “pure” dynamic process (i.e., a single operative process) could be used for the coalitional invariance test. The cases included in each of the following difference patterns were used in testing the six hypotheses (the set element in bold is the constituency being examined in the invariance tests):
H1: (+,+,0,0), (+,+,0,+);
H2: (+,0,+,0), (+,0,+,+);
H3: (−,0,−,0), (−,0,−,−);
H4: (−,−,0,0), (−,−,0,−);
H5: (0,+,−,+), (0,+,−,0), (0,0,−,+);
H6: (0,−,+,+), (0,−,+,0), (0,−,0,+).
The number of cases used in the coalitional invariance test was limited by the fact that complete data for the demographic composition variables does not exist until the election of 1872. Thus, all incidences of enduring net partisan gain before that election were unavailable for the analysis.
The procedure used to test the coalitional invariance hypotheses was as follows. The first step was to create six data files that included only cases which involved net partisan gains and fell within one of the 14 difference patterns listed above. For each case in each data set the data entry for the election that immediately preceded it was identified and added to the data file. Thus, if Kalamazoo County, Michigan in 1880 met the inclusion criteria to test H1 (i.e., there was a Democratic net gain that was generated solely by an increase in the size of the core electorate), then the data entry for Kalamazoo in 1876 was added to the file.
A Post variable was created to differentiate the election year preceding the net partisan gain from the year in which the net gain was realized (the Post variable was scored ‘1’ for the year in which the gain was experienced and ‘0’ for the year preceding the gain). The Post variable was then used to construct a set of interaction terms involving the fourteen social composition variables listed above (e.g., Northern European*Post, Baptist*Post, African American*Post, etc.). The fourteen social composition variables and their interaction terms were then used in a regression analysis to test the six hypotheses listed above. The interaction term will be insignificant if the coalitional make-up of the electoral constituency expected to remain constant is invariant across the two elections. The interaction term will be statistically significant if the social composition of the constituency changes appreciably. Interaction terms were only tested if the social composition variable itself had a statistically significant effect in the regression analysis.
The results of the six tests are reported in Table 4. Separate analyses were conducted for all instances of net partisan gains and just for cases involving critical realignments. The entries in Table 4 reflect only the significance of the interaction terms. If the social composition variable itself did not have a significant entry, the entry is marked with a NS designation. If the interaction is not statistically significant it is designated with an entry of ‘…’. Two different standards were used for the significance tests. Because the sample sizes were so large for those analyses that included all instances of net partisan gains (between 8,000 and 18,000 cases), a .001 level of statistical significance was used. The number of realignment cases analyzed was much smaller (between 40 and 140). When fewer than 50 cases existed no analyses were conducted. When more than 50 cases were available a significance level of .01 was used.
Interaction terms that were significant at the .001 level are designated in two ways, ‘+++’ for positive interactions and ‘−−−’ for negative interactions. Interaction terms that were significant at the .01 level are designated similarly: ‘++’ for positive interactions and ‘−−’ for negative interactions.
Out of the 100 interaction terms tested, there were only five instances in which the interaction term was statistically significant. All of these were in regression analyses involving all cases of net partisan gain. In no analyses examining only critical realignment cases were the interaction terms statistically significant. This lends further empirical support to the viability of the simplest movement of the smallest number assumption.
About this article
Cite this article
Darmofal, D., Nardulli, P.F. The Dynamics of Critical Realignments: An Analysis Across Time and Space. Polit Behav 32, 255–283 (2010). https://doi.org/10.1007/s11109-009-9103-3
- Voting behavior
- Democratic accountability