Image Repair Campaign Strategies Addressing Race: Paula Deen, Social Media, and Defiance

Abstract

This multi-method study examines how the use of social media in a crisis campaign involving race-related issues may affect a public figure’s credibility and perceived response appropriateness. First, image repair theory is used to analyze Paula Deen’s image repair campaign in the wake of the National Enquirer’s revelation that she admitted to using the “N-word” during a lawsuit deposition. Our analysis shows her response strategies were unsuccessful because her apology did not center on the allegations, and she was contradictory in her bolstering, minimization, and mortification strategies. We build on the Deen case study results by exploring the effectiveness of tweeted message strategies in a race-related crisis via Twitter. We use a mixed-design experiment examining how public figure type (politician v. TV celebrity) and response strategy (moral defense, performance defense, defiance defense, no defense) affect perceptions of a female public figure’s credibility and perceptions of the appropriateness of the response. Results show that any of the three responses are better than no response when addressing charges of racial insensitivity. A defiance defense, as newly tested strategy, and moral defense worked better for the TV celebrity condition than the politician condition. Implications are discussed.

Notes

¹ The distinction in changing “restoration” to “repair” was that Benoit felt that many times rhetors were unable to rebuild their reputations to what they had “owned” before. In going through a crisis, one’s image does not stay the same.

² There are other theories of crisis communication, such as Coombs’ Situational Crisis Communication Theory (SCCT) (Coombs & Holladay, 1996), which is grounded in much of the same literature, but typically relies on an experimental approach to support its tenants.

³ Mauchly’s test of sphericity was significant at p < .05, indicating a departure from homogeneity in the equality of the variances among treatment conditions. Sphericity violations often occur in large samples. The Greenhouse–Geisser correction ϵ was close to 1 indicating the departure from homogeneity was small. We took the conservative approach of using the Greenhouse–Geisser correction for all repeated measures analyses to reduce the probability of a Type I error (Field, 2009).