Warranted Inferences of Mediation and Possibility of a Threshold Effect: Verifying the Mediating Effects of Anger in Situational Crisis Communication Theory

Rongting Niu (PhD student) and Xin Ma. (Forthcoming). “Warranted Inferences of Mediation and Possibility of a Threshold Effect: Verifying the Mediating Effects of Anger in Situational Crisis Communication Theory.” Public Relations Review.

Abstract: Scholars have called for communication research to verify the causal claims of mediation models from a research design perspective, instead of only proving mediation statistically. This study validates whether and how anger mediates the causal effects of crisis types on publics’ responses in Situational Crisis Communication Theory (SCCT), including reputation, negative word-of-mouth intention, and purchase intention in China. Two experiments were conducted based on the experimental-causal-chain design. Results in Study 1 demonstrate that the causal relationships between three crisis types and publics’ emotional and other responses in China are consistent with findings in Western contexts. In Study 2, the results of a 2 (anger: low, high) x 3 (crisis types: victim crisis, accidental crisis, preventable crisis) factorial experiment reveal significant mediating effects of anger on publics’ responses in the victim and accidental crisis conditions, but not for preventable crises. This novel finding suggests the possibility of a threshold effect of anger in the mediating process. Specifically, anger has a mediating effect on publics’ responses when it is below the high-anger threshold. However, anger may no longer be the mediator when it exceeds this threshold. This finding empirically challenges the common assumption that emotions have a linear relationship with publics’ responses, thus offering a new research avenue and deeper understanding of how emotions function in crises. Therefore, this research serves as a pioneer, calling for future studies to validate other theories involving mediation to yield fruitful insights.