Reputation repair is a paradigm within public relations and crisis communication. The reputation repair paradigm is currently focused on the symbolic strategies organizations use to repair their reputations in the aftermath of a crisis. This dissertation proposes significant revisions to the reputation repair paradigm and builds a series of scale measures and a revised model of reputation repair to achieve this goal.

Using moral foundations theory, situational crisis communication theory (SCCT), image repair theory, and input from 20 participants with expertise in public relations, this dissertation designs new measures for instructing information, adjusting information, reputation management messages, offensiveness of a crisis, and perceived virtuousness that buffers against reputational harms posed by crisis. This dissertation then refines and validates these measures with a pilot test with 797 participants recruited from mTurk. Finally, it concludes with an experiment testing these measures in a crisis situation operationalized as a potentially deadly fire in a building. The experiment used 1,000 participants recruited from mTurk in a 2 (crisis types: rumor or organizational misdeed) x 2 (offensiveness: high or low) x 2 (instructing information: yes or no) x 2 (adjusting information: yes or no) x 2 (crisis response: denial or rebuilding) factorial design to test the effect of SCCT’s matching construct of response strategies and the proposed revised model of reputation repair that explains how messages, offensiveness of a crisis, and perceived virtuousness impact post-crisis reputation.

This dissertation finds that matching strategies according to SCCT have a very small effect (ή2 = .005) on post crisis reputation while reputation management messages overall have a very strong structural effect on post crisis reputation (.814). Further, it finds that the revised model of reputation repair explains how messages, perceived offensiveness, virtuousness, and post-crisis reputation interrelate and that the revised model changes slightly under different situations. Implications for theory and practitioners are discussed.