Article Summary
Introduction
American partisans are quick to distinguish between “us” and “them” along party lines, with the attributions of positive and negative traits also falling largely along those lines. In their paper, Schwalbe, Cohen, and Ross attempt to determine the existence of a deeper psychological divide between; that opposing partisans believe the other side sees the world in a fundamentally unobjective manner. This “objectivity illusion,” where one side believes they see the world free from normative biases but the other side’s perceptions are heavily influenced by such biases, could have deleterious effects on normal democratic functioning in the country.
The authors also consider potential asymmetries in objectivity illusion and downstream attitudinal effects. For example, do Republicans believe themselves to be more objective than Democrats than Democrats do the opposite? Additionally, do differences in views on objectivity influence consumption of political media, evaluations of candidates, and preference for candidates? Such findings may speak to the formation of attitudinal polarization.
Analytical Approach
The authors field two studies over the course of the 2016 U.S. presidential election; one longitudinal and descriptive, the other one-time only with a randomized component. Both studies are fielded via MTurk.
In study one, an original sample of 870 participants was fielded in the hours prior to the first presidential debate. Respondents were asked to evaluate the normative and nonnormative bases of support for Trump and Clinton; e.g. do Clinton supporters voice support because of Clinton’s knowledge of facts and history or because of misinformation. From this, the authors create an index of the objectivity illusion: “the mean of three perceived normative influences minus the mean of four nonnormative influences on political allies minus political adversaries.” They also ask respondents to evaluate the candidates on a number of traits (e.g. compassion, patriotism). In stage 2 (n = 696), completed the day after the debate, respondents were asked to evaluate the performance of the presidential candidates in the debate and an additional question on interest in receiving a book gift either critical or supportive of their preferred candidate (randomized). Finally, in stage 3 (n = 509), occurring after the second presidential debate, respondents were again asked to evaluate candidate performance and to respond to a question on social distance (preference for child marrying member of out-party).
Study two (n = 1,715), completed two days before election day, repeated the same normative/non-normative question batteries as study one, but with the focus on a randomized component where respondents were presented with a blog post written either in support of Trump or Clinton. Respondents were asked to evaluate the normative/non-normative factors influencing the blog author, as well as personal traits they might exhibit. The book lottery experiment was also repeated.
Main Findings
Members of both parties show strong patterns of the objectivity illusion, attributing normative bases of beliefs to supporters of their preferred candidate while attributing non-normative bases of beliefs to supporters of the rival candidate. While strong Trump supporters are slightly more intense in this belief then strong Clinton supporters, the results are largely symmetric. The stronger the support for one’s own candidate, the more objective (unobjective) they believe allies (adversaries) to be. This pattern also extends to evaluations of candidate traits.
How does the objectivity illusion influence perceptions of debate performance? Those higher on the illusory scale were much more likely to believe their own candidate won each subsequent debate. Objectivity illusion was also strongly correlated with subsequent changes in support for their own candidates; strength of support grew for those high on objectivity illusion, but not those lower on the scale, even conditional on pre-existing support. It also predicted willingness to receive a book favoring the opposite candidate and how upset a respondent would be if their son or daughter married someone of the rival party.
The results of study two were also consistent with this general trend. Trump and Clinton supporters consistently and symmetrically rated the blog writer as more objective when supportive of their own candidate and less objective when supportive of a rival candidate. They were also more likely to associate positive (negative) traits with the writer when writing favorably about their own (rival) candidate. These trends grew stronger as strength of support for their candidate grew. Finally, small but nontrivial amounts of both Clinton and Trump supporters (around 9%) were willing to describe the blog writer supporting the opposing candidate as “evil.”
Implications
These results expand our understanding of the depth of partisan animosity in the U.S. Not only do partisans hold negative views of each other, they believe the other side to see the world in a fundamentally incorrect way. Furthermore, these perceptions verge on the extreme for a small number of partisans who see the other side as evil or likely to commit acts of terrorism. Partisans also seem to do so symmetrically, however, an important finding given previous research focusing solely on Republican support.
The authors contend these results speak to how attitude polarization occurs. Namely, partisans process information in a biased way, and the internalization of biased information often acts to confirm the biases that previously existed. This feedback loop continues recursively, potentially furthering political polarization in the process as negative views of the opposing party are continually reinforced.
Questions left unanswered
The authors measure the objectivity illusion from a purely personalistic viewpoint; one viewing their own views as objective while seeing others as not. However, they do not fully answer the question of whether this objectivity is illusory from a more pragmatic perspective. For example, do stronger supporters of Clinton and Trump actually hold views that are more likely to be misinformed than weaker supporters, or supporters of the opposite side? We don’t fully know if their objectivity bias is illusory; it very much could be the case they are being more objective than others.
Furthermore, it is unclear whether the stated increase in polarization occurs on affective or ideological terms, or if it remains durable after the election has concluded. This process of polarization in candidate support may be part of a electoral cycle where voters become stronger in support as the election day draws nearer.
Methods and Analysis
Was the study and its analyses pre-registered?: No
Did the study rely on proxy variables to measure polarization?: N/A
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
Because the authors are not conducting a causal analysis of the effect of objectivity illusion on attitudes, it is very difficult to rule out endogenous relationships. Perhaps the most plausible of confounders is the generalized tendency of partisans to ascribe any and all negative things to members of the opposing party. In that sense, partisans may not truly believe they are more objective than the other side; they are just “cheerleading”, assigning positive attributes to their own side and negative ones to the other, regardless of content. The authors admit this is a possibility in their discussion, but point to the fact objectivity illusion predicted later partisan strengthening while the denigration of personal traits did not. This does not rule out the possibility of a more fundamental, latent trait predicting both trait denigration and objectivity illusion, however. More work should be done to rule out this possibility, especially in an MTurk sample used for such surveys.
Open Data & Analyses
Does the article make the replication data publicly available?: Yes
Does the article make the replication analysis scripts publicly available?: Yes
Article Citation
Schwalbe, M. C., Cohen, G. L., & Ross, L. D. (2020). The objectivity illusion and voter polarization in the 2016 presidential election. Proceedings of the National Academy of Sciences, 117(35), 21218–21229. https://doi.org/10.1073/pnas.1912301117
Bibtex
1
2
3
4
5
6
7
8
9
10
11
12
13
14
@article{
doi:10.1073/pnas.1912301117,
author = {Michael C. Schwalbe and Geoffrey L. Cohen and Lee D. Ross },
title = {The objectivity illusion and voter polarization in the 2016 presidential election},
journal = {Proceedings of the National Academy of Sciences},
volume = {117},
number = {35},
pages = {21218-21229},
year = {2020},
doi = {10.1073/pnas.1912301117},
URL = {https://www.pnas.org/doi/abs/10.1073/pnas.1912301117},
eprint = {https://www.pnas.org/doi/pdf/10.1073/pnas.1912301117}
}