Article Summary
Introduction
Partisanship appears to be an increasingly important cleavage in elite political behavior, structuring voting in legislative chambers and dividing Supreme Court justice opinion. However, these increasingly entrenched coalitions at the elite level exist alongside increasing voter identification as independents, and high-levels of straight-ticket voting. How might these two trends coexist?
In their paper, Abramowitz and Webster contend that decreasing voter alignment with one’s own party is related to a trend they call negative partisanship, wherein people dislike the out-party more than they like their own party. As negative partisanship increases, so too does the likelihood of voting for (against) your favored (disfavored) party up and down the ballot. This leads to electoral results across all partisan contests appearing increasingly aligned with those of the president (sometimes called the “nationalization” of politics). Such trends have important implications for democratic representation, competition, and governance in an increasingly divided country.
Analytical Approach
The authors use survey data from the American National Election Studies cumulative file between 1972 and 2012. They conduct a number of descriptive analyses using these data, including tracking the number of independents (including leaners), straight-ticket voting, and in- and out-party affect over time.
To determine the connection between affect and voting behavior, the authors focus on the 2008 and 2012 elections. Using a binary operationalization of party loyalty (self-reported intent to vote for the presidential, senate, and house candidate of a respondent’s own party) in a logistic regression, the authors measure the effects of negative partisanship (opposing party feeling thermometer), own party feeling thermometer, strength of partisanship, and party of incumbents. To better understand the causal direction of the relationship, the authors run a second regression using the same dependent variable but a pre-election measures for vote intention and feeling thermometers.
The authors also conduct several analyses using aggregated data on electoral outcomes. Specifically, the authors regress constituency-level outcomes in house (2-year increments) and senate (10-year increments) races on the corresponding constituency-level presidential outcome and the party of the incumbent, noting both the effect sizes/significance of the amount of variance explained. They conduct a similar analysis with state legislative elections, but with the share of seats a party holds in every state as the corresponding measure.
Main Findings
Negative partisanship seems to significantly structure political behavior in contemporary U.S. politics. The share of consistently loyal “leaning” independents has risen sharply since 1972, from roughly 55% in 1972 to 81% 2012. This parallels a similar trend in strong and weak partisans. Additionally, while feelings toward one’s own party have remained largely stable, feelings toward the opposing party have dropped 15 points (on a 101 point scale) over the same time period.
These two trends appear linked. In their regression analysis, out-party affect is the strongest, significant predictor of consistent party loyalty in the 2008 and 2012 elections, with almost double the coefficient size as in-party affect. The finding holds in the panel analysis, with opposing party FT significantly predicting party loyalty in t+1.
The results using observational data give a similar picture. From 1972 to 2014, the coefficient of presidential partisanship predicting house outcomes has nearly doubled, the coefficient for the party of the incumbent is less than half, and the R squared value has risen from .77 to .92. Similar trends exist for Senate elections and state legislative delegations. More generally, the percentage of contest winners consistent with the party advantage of the president is now above 90% in House elections and above 80% in Senate elections.
Implications
While identification with parties has declined, loyalty to such parties has increased, potentially increasing the influence of presidential politics at the expense of personal candidate advantages. These results have numerous implications for competition, representation, and governance.
One specific implication is the growing structural advantage of Republican candidates. Because Democratic candidates are inefficiently distributed in a spatial manner (concentrated in more urban areas), more loyal partisan behavior across districts will likely lead to more Republicans being elected, despite similar overall vote counts.
Additionally, the content of representation may change from character and specific policy considerations to purely partisan ones, potentially leading candidates to vote in lock-step with other partisans. This is likely to mean there will rarely be mismatches in how a district votes for president versus other races.
Finally, it is unlikely this is a short-term trend, as a polarizing electorate begets polarized elected officials, and vice-versa. It is unclear to the authors what might stop these mutually-reinforcing trends.
Questions left unanswered
The question the authors ask at the beginning of the paper is what mechanism might be responsible for increasing loyalty to a party even while identification with parties is declining. Their answer is negative partisanship, but the content of that negative partisanship is ambiguous, and has very different implications for political representation. For example, if rising animosity is due to policy disagreement, it is entirely understandable why voters would prefer to vote more loyally for one party. Indeed, this might maximize quality representation, since they are voting for candidates closer to their own ideals. Understanding the content of animosity, then, is important for evaluating the democratic and representational consequences.
Methods and Analysis
Was the study and its analyses pre-registered?: No
Did the study rely on proxy variables to measure polarization?: No
Were standard p-value thresholds used (p<.05 or 95% Confidence Intervals that don’t overlap zero)?: Yes
- Largest p-value presented as significant: 0.05
Were correlational results interpreted with causal language?: No
Limitations / Weaknesses
One problem the authors confront in the paper is the causal ordering of their mechanism; i.e., do negative evaluations of the opposing party come prior to or after decisions to vote for certain candidates. Their solution is to regress party loyalty in t+1 on pre-election (t) party feeling thermometers and a control for pre-election candidate preference. However, this doesn’t solve the endogeneity problem, as pre-election candidate preference might causally have an effect on pre-election opposing-party warmth. A cross-lagged panel design may be more applicable in this situation, but is unfortunately not possible because the pre-election only asks about presidential vote intention, whereas the DV in the post-election wave is made up of self-reported voting in presidential, house, and senate elections.
Additionally, a lot of emphasis is placed on the primacy of the presidential contest as the unifying agent behind partisan behavior: “the influence of presidential partisanship has risen.” This conclusion seems premature and cuts against the other main argument of the paper that partisan consideration comes prior to vote choice. While the results suggest down-ballot contests finally are coming in with presidential contests, the opposite could be true, or a third, confounding variable could be structuring both to become more correlated.
Open Data & Analyses
Does the article make the replication data publicly available?: No
Does the article make the replication analysis scripts publicly available?: No
Article Citation
Abramowitz, A. I., & Webster, S. (2016). The rise of negative partisanship and the nationalization of U.S. elections in the 21st Century. Electoral Studies, 41, 12–22. https://doi.org/10.1016/j.electstud.2015.11.001
Bibtex
1
2
3
4
5
6
7
8
9
10
11
12
@article{ABRAMOWITZ201612,
title = {The rise of negative partisanship and the nationalization of U.S. elections in the 21st century},
journal = {Electoral Studies},
volume = {41},
pages = {12-22},
year = {2016},
issn = {0261-3794},
doi = {https://doi.org/10.1016/j.electstud.2015.11.001},
url = {https://www.sciencedirect.com/science/article/pii/S0261379415001857},
author = {Alan I. Abramowitz and Steven Webster},
keywords = {U.S. elections, Partisanship}
}