Differential Effects of Message Framing on Obesity Policy Support Between Democrats and Republicans

ABSTRACT

This study tests whether gain- and loss-framed messages about establishing obesity-reducing policies have different persuasive effects on Republicans and Democrats. In a randomized between-subject experiment, participants (N = 384) read a message emphasizing either benefits to a society by establishing policies aimed to reduce obesity (i.e., gain-framed message) or costs to a society that fails to establish those policies (i.e., loss-framed message). Results indicated that Democrats perceived the gain-framed message as more persuasive than the loss-framed message and the perceived argument strength fully mediated the framing effect on Democrats’ policy support; however, there was no framing effect on perceived argument strength among Republicans. On the other hand, the gain-framed message led Republicans to attribute the cause of obesity less to the individual level compared to the loss-framed message and the no-message condition. We observed no framing difference among Democrats on causal attributions. Theoretical and practical implications of these findings are discussed.

Notes

¹ The messages are available upon request.

² We chose five items from the original nine-item perceived argument scale (Zhao et al., 2009) to limit the length of online questionnaire. We excluded semantically similar items such as “The statement is a reason for __ that is convincing”; “The statement put thoughts in my mind about not wanting to ____”; “Overall, how much do you agree or disagree with the statement?”; “Is the reason for the statement gave for ____ a strong or weak reason?”

³ We ran parallel analyses with 3 (message: gain- vs. loss-framed message vs. no message) × 2 (political parties: Republicans vs. Democrats) conditions on all DVs, excluding Independents. The results were consistent with the 3 × 3 interactions, except on perceived argument strength: The interaction became insignificant when Independents were not included, F(1, 149) = 1.46, p = .229. For simplicity, we report only 3 × 3 interactions in text.