Plus Ca Change … Redux? The New 2016 – 2020 ANES Election Study Panel

• In Event: The Role of Ideology in Public Opinion

Abstract

In 1979, Converse and Markus published a paper comparing the individual-level stability of party identification, on the one hand, and policy issues, on the other, in the original 1956 – 1960 CPS panel study, and the second major 4-year panel survey conducted between 1972 and 1976. In one of the foundational articles in all of political science, Converse (1964) had reported high levels of over-time stability of party identification in the original 1956-1960 panel, compared to strikingly lower levels of stability of policy attitudes over that same time period. The dramatic political events of the 1960s, including the end of legal segregation with the Civil Rights and Voting Rights Acts, the women’s liberation movement, growing disillusionment with the war in Vietnam, large scale urban riots in many American cities in the second half of the decade, the assassinations of John and then Robert Kennedy, the striking success of George Wallace and his American Independent third party in the 1968 election, and the end of the Democrat-dominated 5th American party system, all converged to lead to a noticeable decline of party identification (or a rise in the proportion of the electorate saying they are “independents”), a rise in the cross-sectional correlation among attitudes or “attitude structure,” and clearer links between parties and policies – all of which would lead us to expect some noticeable revisions of the story Converse (1964) had told about the relative stability of party identification and policy views during the 1950s – an expectation that was definitively not born out when Converse and Markus analyzed the 1972 – 1976 panel and found stability coefficients from the 1970s that essentially replicated those that Converse had reported earlier.
Forty years later, the ANES has released a new 4-year panel study of over 2000 respondents who completed both the 2016 and 2020 ANES surveys. The intervening 40 years of American politics have produced at least as many significant changes as those occurring during the 1960s (albeit, spread over a longer time period), including the rise of the right wing of the Republican party, the globalization of the U.S. economy, the collapse of the Soviet Union, repeated instances of ethnic “cleansing,” 9/11 and a general increase in international terrorism around the world, noticeable increases in the degree of polarization between the Democratic and Republican parties and of course Donald Trump, anti-democratic populism, and insurrectionist storming the capital. We will examine the individual level over-time stability of general political orientations like party identification, specific policy preferences, affect toward political figures and social groups, and most interestingly, indicators of social identification that are a big part of the current ANES surveys but were largely absent in the 1970s. We expect party identification to be every bit as stable between 2016 and 2020 as it was during the two earlier periods. But we also expect policy preferences and political affects to be more stable over the recent panel than they were during the 1970s. Of course Converse and Markus felt the same way about the 1972 – 1976 panel. Although there are no comparable items in the 1972 – 1976 panel to compare, we further predict that the stability of social identification items available in the most recent ANES surveys, will approach that of party identification.

Authors

• Richard R. Lau, Rutgers University, New Brunswick

• Michael Strawbridge, Rutgers University