Log in Page Discussion History Go to the site toolbox

501 Readings Week 6

From BluWiki

Controversies in Voting Behavior Edited by Richard Niemi and Herbert Weisberg

Chapter 17 Chapter 17 is mainly an overview of the debate about the supposed stability or instability of party identification and outlines the debates presented in subsequent chapters.

Concept of Party Identification Original developments concerned with social psychology and defined party identification as an attitude that is associated with group. Out of the rational choice framework, party identification became a “running tally” of evaluations that reflected reactions to current politics.

Measurement 7 point classification of party identification used most frequently. Niemi and Weisberg’s two largest concerns with this measurement are its intransitivity and its dimensionality.

Arguments that Partisanship as routinely affected by politics Political identifications are to some degree tested by different political variables that allow for some change in the orientation of partisanship. This includes issue or leader evaluations, short term issues, and large cataclysmic events.

Argument that Partisanship is fairly immune to politics Began with Campbell’s American Voter which argued that childhood socialization crystallized party identification so as to render it a stable concept throughout a citizens life. Excluding realignments, party identification remains highly stable when it comes to a person’s ability to assert whether they are Democrat, Republican, or Independent. It has been argued that generational perspectives and measurement error are the main reasons for why it appears in some research that party identification is volatile.

Niemi and Weisberg argue that the debate continues as demonstrated in Chapter 19 and 20. Overall, both authors argue that some degree of party identification change exists but the real question that we continue to debate is whether that change is meaningful or not.

Chapter 18: Generational Changes and Party Identification by Warren Miller

This chapter seeks to identify “a set of circumstances under which the basic measure of party identification, applied to a significant body of voters, has produced results fully in line with the original concept of an enduring predisposition largely impervious to changing election specific events” (338).

1960-1988 considered a turbulent period in American politics which shaped “the basic party dispositions of the young cohorts of voters who have come of age since 1968 while not affecting the older generations in a similar way”(341). These young voters have altered the parameters of subsequent American national politics. Focuses specifically on the post-New Deal generation. Attempts to “compare citizens who differ primarily in the nature of the political epochs that influenced their initial awareness of the world politics” (339-40).

Limits of their study include the fact that they did not give any consideration to the party identification of the black citizenry or the non-voting citizen electorate.

Generational differences in party identification and voter turnout Voter turnout declines are transformed by the application of the cohort/generational analysis. The New Deal generation saw steady figures when it came to voter turnout. It was rather increasing size of the nonparticipant post- New Deal generation into the electorate that caused such large decline in turnout. This generation has a lower incidence of party identification and the lower voter turnout between identifiers and non-identifiesrs (342).

Generational Differences in Partisan Balance of Party Identification In the North there is relative stability among the stability of party identification. There does exist large generational differences but “there are not significant period or life cycle effects altering the party balance” between 1950 –early 1970’s (343). The post New Deal generation benefited the Democratic Party in 1970’s. After 1970’s the post new deal generation led to a rise in Republican strength and subsequent democratic decline. In the South, the erosion of democratic strength and republican growth begins after Kennedy election. Rate of change from one party to another was a direct function of education. This transition hits a peak in the late 1960’s and early 1970’s.

There was a general rise in the support of the Republican Party that was predicted to continue with the rise of the post-New Deal generation. This has not actually happened as the growth of the New Deal generation has led to a republican constituency that is less republican that its New Deal cohorts. “The post New Deal generation is making a continuing contribution to a realignment of the social foundations of Democratic and Republican politics, and it is leading the shift away from Democratic dominance and toward parity between the parties” (352). Overall there exists a “great persistence and stability of party identifications even in the presence of events that have an impact on younger voters” (353).

Chapter 19 Partisan Stability: Evidence of Aggregate Data by Donald Green, Bradley Palmquist and Eric Shickler

G.P.S argue that looking at aggregate data cannot explain information about individual change in partisan identification but that aggregate data about the overall proportions of Democrats and Republicans can help understand electoral changes over large periods of time. It can help to understand the stability of the Democrat to Republican ratio and can identify sources of change.

In this particular study, G.P.S. use Gallup Poll data in a regression model to understand how consumer optimism and president performance affect changes in macropartisanship.

The results found that large economic shocks sustained over time alter partisan balance but once the president performance variable was added in, the effects of the economy were negated. The presidential performance variable only showed a less than 1% change in macropartisanship. G.P.S. even believe that this number may be inflated due to the measurements.

Overall there is a limited extent to which partisanship identification responds to the political environment. This study reaffirms the idea of stability in identification of the American electorate at the aggregate level.

Bibliographical info: Niemi, Richard and H Weisberg, eds. (2001). Controversies in Voting Behavior. Washington, D.C: CQ Press.

Chapter 20: (Erikson, MacKuen, Stimson) Premise: If partisanship is frozen in place by early adulthood, how can it be responsive to routine currents of economic and political fortune represented by consumer sentiment and presidential approval (364)? The answer is that a macro-level model of partisanship is needed that allows for partisanship as a function of new info, but with no forgetting of the past (364).

Content: Macropartisanship essentially contains a permanent memory of political and economic inputs…these inputs are identified as the same economic and political shocks that cause changes in consumer sentiment and presidential approval. Through their model (see pp 365, 366), the authors demonstrate that indeed the fundamentals of macropartisanship are comprised of the permanent political and economic shocks from the past and present (370). Although the (macropartisan) response to new economic and political inputs can be imperceptible at the time in which they happen, it is long lasting and becomes more visible over time.

Chapter 21: (Niemi and Weisberg) Premise: Is the party system changing, and if it is, in what ways? Yes, there has been meaningful change, but the exact nature and extent of that change has been hotly debated, as a survey of contemporary research shows.

Content: The editors identify three major veins of thought in contemporary research. The first of which is the body of work which suggests that the party system has changed in the form of a Post-New Deal Realignment. Generally, it is observed that some sort of durable changes are or have occurred, whether in the form of long lasting policy direction changes, the issue basis of the party system, or alignment of social groups with the parties. The editors note that none of these changes constitute a realignment as drastic as that of the New Deal in the 1930s, if at all. Proponents of post-new deal realignment contend that evidence for this school of thought can be found by closely examining realignment at the issue and social group levels (373). Specifically, analysis of social groups reveals realignment most clearly amongst white southern males (374). Changes, however, are also detectable among many other groups, including voters who identify with religious groups as well as between married and unmarried voters (376). Although there have clearly been observed changes especially among social groups, the editors contend that most observers probably do not view the Post-New Deal Realignment as the main recent party change in American politics.

More common in recent scholarly research is the contention that party system change is currently manifested in the dealigning shift toward candidate centered politics. Evidence for this ‘dealignment’ is cited in the increasing incidence of Independence and Independent and Third party candidate voting (379). Divided government and split ticket voting are also cited as further evidence of dealignment(380-81).

Finally, the third vein of thought regarding party system change is that rather than dealignment, we are observing a reinvigoration of partisanship (382). Evidence for this includes the contention that Independent identification is not actually as strong as certain studies and surveys indicate (Keith, et al 1992). As well, it is cited that parties still play an important and powerful role in the voting process, albeit different from the ‘golden age’ of the late nineteenth century (382-83). Ideological reorientation is also mentioned as evidence of realignment.

Chapter 23: (Aldrich and Niemi) Premise: There have been five agreed upon American party systems and following a ‘critical era’ in the 1960s there actually emerged a sixth American party system (405).

Content: The analysis of a variety of attitudinal and behavioral changes (micropatterns) allows one to collect them all into an identifiable macropattern from which can be drawn generalizations about changes during transitions between party systems (406). Among the micropatterns examined in relation to party identification were: partisanship of blacks and whites, support coalitions, and party images. Micropatterns examined in relation to issues included: issue-party and candidate linkages, basis of issue conflict, and the consideration of ‘most important problems’ (410). Analysis of these patterns in a variable-by-variable fashion shows consistency across a wide variety of measures (417). This analysis supposedly supports the contention that the party system was in equilibrium (until the 1950s); the equilibrium was disrupted during the critical era of the 1960s; a new equilibrium emerged by 1972 (see figure 23-10, p418). The empirical analysis of the micropatterns, along with the macropattern leads the authors to conclude that the new equilibrium reflects strongly a sixth American party system which is signified by its candidate orientation (419). Since the critical era, they contend that what has happened involves a new form of political parties: the candidate centered ones, and that this form has been sufficiently institutionalized since 1972 to conclude that the emergence of a sixth system has emerged rather than instability and dealignment as some contend (420).


Weisberg, Herbert F. and Steven H. Greene. 2003. “The Political Psychology of Party Identification.” In Electoral Democracy. Eds. Michael MacKuen and George Rabinowitz. Ann Arbor, MI: University of Michigan Press.

What: An argument for the continued utility of social psychology in aiding our understanding of party identification and the implications of such for American electoral democracy.

The Argument: The Reference Group Theory of Party Identification: proved useful in explaining party identification. It did so by explaining that an individual’s social identity and preferences are influenced by attachment to groups and people attach themselves to groups that they are objective members of, that they otherwise identify with, or to those towards which they have a positive affect. Groups are real because they “psychologically” affect the ways that we behave. Groups serve to influence individual preferences by relaying group preferences to members. That is, perceptions, evaluations, and preferences are significantly influenced by the values, preferences, and evaluations of the groups with which an individual identifies (85 - 86).

Implications: This allows us to understand political identification. Simply, political parties can be considered groups. ‘Independent’ can also be considered a group. This allows us to understand that people identify themselves as belonging to a particular political group on the basis of identifying with and sharing the preferences of these groups. Political parties serve as reference groups because they serve to influence the development of attitudes and preferences towards the political world (86).

The Intergroup Relations Field: examines the impacts of in-group and out-group sentiments on behavior. Focus here has shifted towards the individual ‘internalization’ of group preferences. People tend to be biased towards and have more positive attitudes about in-groups, and this bias is related to one’s social identity as distinct from one’s personal identity (87 -88).

Implications: To the extent that one party is defined as ‘us’ and the other as ‘them’, identification with one party (or neither) can be influenced (87). Because attachment to in-groups may be associated with self-esteem, making parties salient as groups may be important for attracting identifiers (90). In-group bias and out-group hostility may be independent: identifying as a Democrat doesn’t necessarily suggest hostility towards Republicans or Independents (92). Individuals may emulate the behavior of trusted or liked others, the discovery that these others identify with a particular party may lead to identification with that party (92). ‘Independent’ becomes a group through either in- or out-group sentiments: if independence is associated with positive political values, this may help explain associating such a categorization with an in-group; while negative evaluations of the parties creates a bias against the parties as out-groups (93). Intergroup Relations Theory suggests testable hypotheses for the study of political identification (93-94).

Attitude Studies: Social psychological theories of group identification also rely on attitudes to explain. Attitudes are related to evaluations of certain objects, not directly observable, they are indirectly observable through behaviors and/or responses (94). There are three interrelated aspects of attitudes: cognitive, affective, and behavioral (94 – 95). Evaluation and affect are distinct; attitudes need not contain an affective aspect (95). Research has focused on two different structures of attitudes: internal composition and relationships between attitudes. Internal attitude structures may be bipolar or unipolar (97). Research has also examined attitude strength (99 – 104). Attitudes may create biases in information processing. Specifically, attitudes may bias exposure and attention to information, perception and judgment, or an individual’s memory (104).

Implications: Which aspects of attitudes are most relevant to party identification? Some studies have found cognitive partisanship to be most frequent but that affective partisanship is most stable (95). The understanding of the internal structure of attitudes may be most important for understanding party identification. There has been some research demonstrating that measuring the bipolar structure of a partisan attitude, by measuring likes and dislikes about parties, allows vote prediction. Attitudes towards the parties may be unipolar, indifference towards one may lead to identification with the other, while indifference towards both may lead to identification as independent (97). Areas in need of further research: are the relationships between attitudes towards parties and ideologies horizontal or hierarchical? Extremity and strength of an attitude need to be kept distinct: the strength of one’s party identification is more than a simple measure of one’s distance from ‘strong independent’ (99). Research on attitude strength has found stronger attitudes to be better predictors of behavior but has also raised concerns that self-reports do not accurately capture attitude strength (100). Because social psychology has developed many measures for attitude strength, studies of partisan attachment might benefit from their use (102). A further area in need of more research is the degree of evaluative-cognitive and/or evaluative-affective consistency in party identification (102 – 103). Political research might benefit from further attention to the information processing biases associated with party identification and the strength of this attachment (Zaller’s study is an example of such) (104 – 105). Because attitudes serve different functions, examining the function(s) that party identification plays might enhance understanding (105 – 106). The attitude – behavior question rears its head here. In determining the role of party identification in voting behavior, we need to address causation-correlation and direction of causation if applicable (107).

Final Implications: Party identification may not be linked to political stability at all. Electoral rules and incentives of leaders may provide a better explanation, as suggested by Rational Choice (111-112). We need to disentangle party identification, ideological identification, and electoral rules in order to understand the role played by each in political stability (113). Lumping all independents into one group may reduce theoretical clarity; understanding independence might be enhanced by use of social psychological theories (115).

The Impact of Party Identification (Chap. 6) The American Voter Campbell, Converse, Miller, Stokes

The jist of the chapter

How do party identifications affect voting behaviors? -The influence of party identification on attitudes toward political objects extends through time. -If the individual has developed attitudes not consistent with his party allegiance, that allegiance presumably will work to undo the contrary opinions. -Republicans are less likely to vote Democrat than Democrats are likely to vote Republican.

The Development of Party Identification (Chap. 7)

The jist of the chapter

This chapter is a brief inquiry into the development of party identifications. -A high degree of correspondence between partisan preferences of subjects and that of their parents. -Political affiliation corresponds with early voting preferences. -Cataclysmic national events have the power produce substantial realignment in long-standing divisions of political sentiment. -Distribution of partisan attachments still follows the same regional lines laid down at the time of the Civil War. The scope of reversal 1. Youth- great depression swung a heavy portion of young electorate to the Democratic Party and put a hold on those voters that has still not been relinquished. 2. Economic groups- Depression had major effect on changes among electorate concerning their economic situation. Legacy left by Republicans lingered in the minds of those who lived through the depression for decades. 3. Minority groups- Roosevelt’s egalitarian practices led to substantial increases in affiliation to the Democratic party from Catholics, Jewish people and African Americans.


Franklin, Charles H. and John E. Jackson. 1983. “The Dynamics of Party Identification.” The American Political Science Review. Vol. 77(4): 957 – 973.

What: A model of the evolution of party identification and the effects causing change to such.

Past understandings of party identification: were rooted in psychology of groups. Once a positive or negative evaluation was attached to a party, these attitudes served as organizing mechanisms for individual “political cognitions”. Put simply, identification serves as an information shortcut. Change in identification was discounted, or considered very rare. Emerging from socialization, identification was seen as a long-term, stable factor of individual political lives, this conception put party identification exogenous to the electoral process (957).

Rational Choice: views identification as malleable, identification is endogenous to the electoral process. Identification is the result of individual evaluations of their own issue positions and preferences and the fit of these with a party’s positions and preferences. Identification, because it is based on continuing evaluation, is considered responsive to new information. Stability of identification results from stability of individual or party preferences and platforms (958).

Recent work: questions both. Simply, we need to account for stability and change, not one or the other. The new model: Party identification at any given time is “a function of previous identification…some other influences [“any factors that might systematically effect” the measured identification]…and a random disturbance…” (958). This allows an empirical test of the two models described above (959).

Three hypotheses tested: 1. That party identification is based on current evaluations of the fit between the individual’s preferences and positions and those of the party. 2. That past votes influence subsequent party identification. 3. That attachment to party becomes more stable with age (960).

Results: The most significant impact on party identification comes from evaluations of fit between individual and party positions and preferences. Voting does not exert a significant force on subsequent identification. Past identification influences current identification and this influence increases with age (965).

Implications: Party identification is neither fixed nor exogenous. Party identification can change; it is responsive to the shifting preferences of the individual or of the shifting positions of the parties. Rather than a result of socializing influences, party identification is the result of accumulated evaluations of individual and party preferences and positions, such evaluations occur during campaigns and terms in office. But, previous identification puts a “brake” on change, especially as an individual ages (968). We can explain realignment by understanding that salient issues might create a shift in individual or party preferences and positions, and/or individual evaluations of leaders and parties (968 – 969). Because younger people’s identifications less stable, a younger population may be less politically stable. A dynamic structure to party identification: Party identifications, individual preferences, and perceived party issue positions seem to be reinforcing. For example, an individual perceives a party position to be in line with his or her own preferences, this strengthens identification or creates identification. With identification, an individual may shift his or her preferences to be more in line with the perceived party position, which leads to better evaluations and a stronger party identification. This individual eventually becomes a “consistent-appearing”, stable voter. This dynamic model also allows that these interactions could create a feedback loop leading to negative evaluations and eventual change in party identification (969). Party leader actions influence partisan identification, strength of the party is not outside of the control of leaders (970).


Abramowitz, Alan I and Kyle L. Saunders. 2006. “Exploring the Bases of Partisanship in the American Electorate: Social Identity vs. Ideology.” Political Research Quarterly. Vol. 59(2): 175 – 187.

Conclusion: Party identification is more strongly associated with ideological preferences than with social identity created by group membership.

Responding to: Green, Palmquist, and Schickler (GPS) challenge the notion that identification in the U.S. is based on evaluations of positions, policies, or performance, and argue that instead, identification is more closely associated with social identity as constructed from group membership. GPS challenge the Downsian model (outlined in Franklin and Jackson, 1983) and argue that party identification is a more emotional than rational attachment. While downplaying the role of party ideology and issue stances, they do allow that realignment can occur when parties’ radically shift ideological or issue stances, but claim that dramatic shifts are rare.

Reaction: GPS’s social identity theory stands opposed to ideological realignment theory that holds that because the parties have become increasingly ideologically polarized, Americans have also increasingly selected a party identification aligned with their own ideological preferences. We have seen a realignment of loyalties along these ideological lines (175).

The Argument: Shifts in identification are associated with ideology. Group identification is only weakly related to party identification. Evidence shows a more rational basis to party identification than that presented by GPS. Increasing ideological and partisan consistency among voters has had political consequences (176).

The Evidence: Trends in Party Identification: GPS claim that partisanship outside of the south has seen little change, but empirical evidence doesn’t support this (176). Identification outside of the south has not been stable, as claimed by GPS. Whites outside of the south have also increasingly identified with the Republicans, the Democratic advantage has been lost (177).

Ideology in the American Electorate: Realignment has been significantly influenced by ideology. Ideological conceptualization within the American electorate depends, to some extent on the clarity and availability of ideological cues provided by elites (178). The Reagan and post-Reagan eras have seen the ideological differences of the parties become more salient and clear to the American electorate. This has made it easier for the electorate to base party identification on ideological preferences (177). This is enhanced by the increased ability of the American electorate to conceptualize and use ideology as demonstrated by: the increased ability of NES respondents to place themselves and the parties on a liberal-conservative continuum and the increased consistency between ideology and policy preferences among respondents (178).

Group Membership, Ideology, and Partisan Change: Evidence supports the ideological realignment hypothesis. For example, increases in Republican identification were largest among those identifying themselves as conservative (179).

Ideological Realignment vs. Partisan Persuasion: GPS hold that party identification will influence ideology (partisan persuasion) (180 – 181). The Partisan Persuasion hypothesis suggests that we will find increasing conservativism as groups realign. This is not supported by the data. Rather, conservativism is relatively stable amongst those groups. Simply, conservative groups became more republican, which supports the Ideological Realignment hypothesis (181). The 1992-1996 NES Panel Survey also supports the notion that ideological orientation influences party identification (182).

Social Identity, Ideology, and Party Identification: The Ideological Differences hypothesis suggests that rather than group identification influencing party identification, it may be that between group differences in party identification are related to ideological and policy preference differences between groups (183). The 2004 exit polls demonstrate that within groups, conservatives identify as Republicans and liberals as Democrats. The differences between conservatives and liberals within social groups are larger than differences between groups suggesting that ideology trumps social identity in party identification (184-185).

Implications: Voters with consistent ideological and partisan identifications are more loyal than those who are not. Increasing consistency between party and ideological identifications is likely to lead to continued partisan voting (186).

Site Toolbox:

Personal tools
GNU Free Documentation License 1.2
This page was last modified on 2 March 2010, at 16:25.
Disclaimers - About BluWiki