The limits of framing effects: citizen perceptions of councilor compensation

ABSTRACT

Existing research provides evidence that the framing of information can substantively alter how citizens perceive institutions of government and a range of policy issues. While prior work acknowledges that there are limits to the effects of framing, less attention has been given to examining contexts in which framing fails to shape perceptions. Using an exploratory survey experiment, we compare the effects of political knowledge, perceived ideological distance, and more deliberative thinking to framing regarding councilor pay, an issue for which citizens consistently express negative sentiment. When provided real-world information showing local councilors are compensated less than their counterparts in comparable cities, citizens are somewhat more likely to rate their councilors as underpaid. However, framing effects are not observed when respondents use more deliberative thinking. Further, we find that explained variance in perceptions is more strongly associated with political knowledge, ideological distance, trust, and, in the case of deliberative thinking, gender.

KEYWORDS:

• Information framing
• local government
• councilor compensation
• public opinion
• deliberative thinking

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/03003930.2023.2174979

Notes

1. Other differences certainly exist (e.g., at-large vs. district seats, staffing, and spending), but these characteristics were not introduced in the randomized conditions.

2. We also asked respondents to estimate the size of the council and whether they had ever attended a council meeting. Less than half of the respondents attempted to recall the size of the council (among those who did, 77 percent responded correctly). When we substitute attending a council meeting, coefficients are strong and in the hypothesized direction but not always significant. We used writing to the council in final models given stronger model fit and significant F tests.

3. It is possible for this variable to range from −4 to 4; no respondents in the analysis both self-reported an ideology in the most conservative category and perceive the council to be in the most liberal category.

Additional information

Notes on contributors

Amanda Rutherford

Amanda Rutherford is an associate professor and Director of the Undergraduate Honors Program at the O’Neill School of Public and Environmental Affairs at Indiana University-Bloomington. Her research focuses on representation, performance accountability policies, personnel management, and executive decision-making.

Cullen C. Merritt

Cullen C. Merritt is an associate professor in the O’Neill School of Public and Environmental Affairs at Indiana University–Purdue University Indianapolis. He conducts research on public management, organization theory, and publicness.