Desire or Disease? Framing Obesity to Influence Attributions of Responsibility and Policy Support

Abstract

The way we describe health threats affects perceptions of severity and preferred solutions to reduce risk. Most people agree obesity is a problem, but differ in how they attribute responsibility for development and decline of the disease. We explored effects of message framing on attributions of responsibility and support for public obesity policies using a 3 × 2 factorial design. Participants read one of six versions of a health message describing the negative effects of obesity. Message frames influenced respondent attributions and their support for policies to reduce obesity. Those who read a message that assigned agency to the disease (e.g., Obesity causes health problems) endorsed genetics as the cause to a greater degree than those who read a semantically equivalent message that instead assigned agency to people (e.g., Obese people develop health problems). In contrast, assigning agency to people rather than to the disease prompted higher attributions of individual responsibility and support for public policies. Explicit message frames that directly connected responsibility for obesity to either individual or societal factors had no effect on respondent perceptions. Findings suggest explicit arguments may be less effective in shifting perceptions of health threats than arguments embedded in agentic message frames. The results demonstrate specific message features that influence how people attribute responsibility for the onset and solution of obesity.

Efficacy of interventions to improve public support for evidence based public health and climate change policy measures: a systematic review.

“…The PRISMA diagram in Figure 1 outlines the data collection steps leading to the inclusion of 16 papers in the review (12)(13)(14)(15)(16)(17)(18)(19)(20)(21)(22)(23)(24)(25)(26)(27). Table 1 provides an overview of the intervention and design characteristics of each included study.…”

“…Two meta-analyses were conducted to provide a pooled estimate of intervention effect estimates for studies with continuous data ( Figure 2) and binary data ( Figure 3) respectively. Three studies were not suitable for inclusion in either meta analysis due to insu cient data (14,16,19). Barry and colleagues (14) reported public support for 10 policy measures as outcomes for their intervention, however this was reported in the text as a signi cant improvement, with only a p value reported.…”

“…Frederick and colleagues (16) used 16 policy items as their outcome measures, which they collapsed into three types and reported on a graph as cohen's d, with con dence intervals only available as lines on a graph, and the p values reported in the text were not exact. McGlynn and McGlone (19) reported outcomes for public support for 7 policy measures, which they reported as a summary measure for all seven using cohen's d. Con dence intervals and p values were unavailable. The ndings of the risk of bias assessment are presented in Table 3.…”

“…5/16 studies did not explicitly identify a theoretical basis for the intervention evaluated (12-14, 18, 27). Framing theory (not including inoculation theory, see below) was used in developing 6/12 promising interventions, 5 of which focussed on obesity policies (15,16,19,20,24) and one of which focussed on opioid pain reliever addiction policies (17). Skurka (26) also conducted an obesity framing study which showed no effect.…”

“…Skurka (26) also conducted an obesity framing study which showed no effect. Framing in this context involves communicating a particular meaning in a particular way in any medium, identifying some aspect of the information presented as more important than others (19). Frames may be essentially different versions or formulations of the same message, and the in uence of the frame is determined by which information is highlighted or emphasised (16,19), so that "any alteration to a message that in uences perspective can be considered a frame" (28).…”