There has been an ongoing debate in the research literature on violations of trustworthiness and the effectiveness of various forms of repair. In this article, three studies compare the effectiveness of several variables hypothesized to repair perceived trustworthiness in a negotiation context: (a) the use of words (accounts or apologies) versus deeds (compensation for the costs of violation), separately and in combination; (b) whether words or deeds pointed to the past or to the future; (c) the effect of an active or passive third-party monitor of the negotiation, and (d) the type of trustworthiness (competence versus integrity) that was violated. Results show support for the repair effects for deeds over words, and that a focus on the past violation (through either words or deeds) is more effective than looking toward the future. Deeds, rather than words, also accounts for the strong impact of the combination of the two courses of action. In addition, results show that an active third-party intervention has a stronger impact on repair than a passive intervention. Implications for theory and practice are discussed along with suggestions for future research.