Verb Voice and Crisis Responsibility: Key Findings from Fannes & Claeys (2022)

Overview

This article meticulously investigates how the specific content of an organizational crisis response, whether a denial or an apology, and the grammatical verb voice employed, either active or passive, cumulatively shape public attributions of crisis responsibility. In turn, these attributions are analyzed for their impact on organizational reputation. Through two distinct experiments, the research consistently demonstrates that an apology significantly increases responsibility attributions when compared to a denial, a factor that subsequently harms the organization's reputation. Conversely, the passive voice is found to reduce perceptions of responsibility more effectively than the active voice. However, this effect is not universal; it is contingent on the crisis response content, proving beneficial only when the chosen strategy is a denial, and not when the organization opts for an apology.

Theoretical framework and key concepts

The theoretical foundation of this study is largely rooted in the Situational Crisis Communication Theory (SCCT), which establishes a direct link between responsibility attributions and the potential for reputational damage. SCCT advocates for defending an organization's reputation by aligning the crisis response strategy with the public's perceived responsibility. Responsibility attributions themselves are dynamic, influenced by both the explicit message content, such as a denial versus an apology, and subtle linguistic cues like verb voice. A denial, for instance, can effectively lower responsibility attributions if the public accepts the organization's no-responsibility frame. Conversely, an apology, while seemingly conciliatory, can inadvertently elevate responsibility attributions by explicitly or implicitly accepting blame. Verb voice, differentiating between an explicit agent (active voice) and a de-emphasized or omitted agent (passive voice), plays a crucial role. The passive voice has the potential to distance the sender from the action and reduce perceived agency, thereby possibly lowering responsibility attributions. A central premise of the study is to test whether verb voice acts as a moderator, influencing the effect of crisis content on both responsibility and reputation, and whether this moderation is dependent on the specific nature of the response, i.e., whether it is a denial or an apology.

Research design and hypotheses (summary)

Study 1 employed a 2 \times 2 factorial, between-subjects design, manipulating the crisis response strategy (denial vs. apology) and the verb voice (active vs. passive). The primary measures included responsibility attributions and organizational reputation. The core hypothesis posited that an apology would lead to higher responsibility attributions than a denial, subsequently damaging reputation. Furthermore, it was hypothesized that a passive voice would reduce responsibility more than an active voice, with an anticipated interaction effect specifically for a denial, but not for an apology. Building upon this, Study 2 expanded the design to a 2 \times 3 factorial, varying the crisis response strategy (denial vs. apology) and incorporating three levels of verb voice (active, agentive passive, agentless passive). This extension aimed to test the generalizability of findings by introducing a new crisis scenario (cadmium in toys) and to provide a more nuanced differentiation between various passive constructions.

Study 1: design, measures, and main results

Study 1 engaged N = 207 Dutch-speaking Belgians, distributed across four conditions in a 2 \times 2 layout. The experimental stimuli carefully manipulated the crisis response between a denial and an apology, alongside the verb voice, using an agentless passive form in the passive conditions. Measures included crisis responsibility, assessed with two items on a 7-point scale, and organizational reputation, measured with nine items on a 7-point scale. Mediation and moderation analyses were conducted using the PROCESS macro. Key findings revealed a significant indirect effect of the crisis response strategy on reputation, mediated through responsibility attributions, with a coefficient of b = -0.45, SE = 0.11, 95\%CI[-0.68, -0.25]. This indicates that an apology led to higher responsibility attributions, which in turn resulted in a worse reputation. Notably, the direct effect of the crisis response strategy on reputation was not statistically significant, thereby supporting the mediation hypothesis through responsibility attributions. The indirect effect of verb voice on reputation via responsibility was not significant overall. However, a crucial moderated mediation was observed: the crisis response strategy was found to moderate the indirect effect of verb voice on reputation through responsibility. Conditional effects demonstrated that a passive voice yielded a more substantial reputation repair than an active voice specifically when the strategy was a denial (b = 0.21, SE = 0.09, 95\%CI[0.05, 0.42]), while no such benefit was observed for an apology (b = -0.01, SE = 0.06, 95\%CI[-0.11, 0.11]). The primary implication from Study 1 is that verb voice effectively helps reduce responsibility and protect reputation predominantly when the organization chooses to deny its involvement in the crisis.

Study 1: interpretation and implications

The findings from Study 1 clearly indicate that the content of a crisis response directly drives perceptions of responsibility and subsequently impacts reputation. Specifically, apologies tend to increase attributed responsibility and, relative to denials, cause harm to an organization's reputation. Furthermore, the form of the language used also plays a significant role: a passive voice generally reduces responsibility perceptions. However, this effect is conditional, providing a benefit only when combined with a denial. When an organization issues an apology, the grammatical voice form appears to have little to no discernible impact on public perception of responsibility.

Study 2: design, measures, and main results

Study 2 involved N = 317 participants and employed a more complex design: 2 (crisis response: denial vs. apology) \times 3 (verb voice: active, agentive passive, agentless passive). The measures for crisis responsibility and organizational reputation remained consistent with Study 1, with the addition of a more nuanced differentiation between agentive and agentless passive forms. The key findings replicated the indirect effect of crisis response strategy on reputation through responsibility attributions, yielding b = -0.33, SE = 0.09, 95\%CI[-0.51, -0.17]. This again indicated that an apology led to higher responsibility, resulting in a worse reputation. Intriguingly, Study 2 also found a significant direct positive effect of an apology on reputation (b = 0.38, SE = 0.14, 95\%CI[0.10, 0.66]), a nuance not prominent in Study 1. The indirect effect of verb voice on reputation via responsibility was not significant when comparing active versus agentless passive (b = 0.04, SE = 0.05, 95\%CI[-0.05, 0.15]) or active versus agentive passive (b = -0.00, SE = 0.05, 95\%CI[-0.09, 0.09]). However, a significant moderated mediation was found for both comparisons: active versus agentless passive (Index = -0.19, SE = 0.10, 95\%CI[-0.40, -0.01]) and active versus agentive passive (Index = -0.19, SE = 0.09, 95\%CI[-0.37, -0.03]). The conditional effects further elucidated these interactions: in a denial scenario, an agentless passive voice tangibly reduced responsibility and improved reputation relative to an active voice (e.g., b = 0.14, SE = 0.08, 95\%CI[0.001, 0.31]). Conversely, when an apology was issued, there was no meaningful advantage of using a passive voice over an active voice (b = -0.05, SE = 0.06, 95\%CI[-0.16, 0.06]). This leads to the implication that the benefits of the passive voice in reducing responsibility primarily manifest when the overall crisis strategy is a denial and specifically when the passive construction is agentless. The passive form does not significantly mitigate the reputational fallout of an apology.

Study 2: interpretation and implications

Study 2 successfully replicated the fundamental pattern observed in Study 1, reaffirming that apologies tend to increase responsibility attribution which consequently harms reputation. Conversely, denials are shown to be effective in limiting responsibility and protecting reputation, particularly when the organization's guilt has not been definitively established. A critical insight from this study is the nuanced importance of language: the agentless passive form demonstrably reduces perceived responsibility when used in the context of a denial statement. However, an agentive passive form, which still identifies the agent, does not achieve the same beneficial effect. This underlines that the specific construction of the passive voice is crucial for its impact on public perception.

Practical implications for crisis communication

For effective crisis communication, practitioners should avoid relying on passive phrasing as a means to mask responsibility when issuing an apology. Public perception tends to infer fault from apologies irrespective of the grammatical voice used, and passive apologies might even be perceived as insincere, thus potentially backfiring. However, if an organization decides to issue a denial, strategically employing a passive, agentless construction can be highly effective in reducing public perceptions of responsibility and consequently protecting organizational reputation. This approach proves more effective than either an active denial or a passive agentive form. It is crucial to internalize that the benefit derived from the passive voice is highly context-dependent; it is not a universally applicable solution and warrants cautious application, especially when considering the ethical dimensions of acknowledging or denying guilt.

Theoretical contributions and takeaways

This research makes significant theoretical contributions by clearly demonstrating how linguistic cues, specifically verb voice, intricately interact with the substantive content of a crisis response (denial vs. apology) to shape public attributions of responsibility and, by extension, post-crisis organizational reputation. The study enriches Situational Crisis Communication Theory (SCCT) by illustrating that the effects of crisis response strategies on reputation are not merely direct but operate through mediating responsibility attributions. Furthermore, it establishes that linguistic form can moderate these effects, introducing critical boundary conditions that depend on the specific defense strategy employed. A crucial takeaway for academics and researchers in the field is the imperative to consider both what is communicated (the message content) and how it is communicated (the linguistic formulation) when endeavoring to predict the complex outcomes of crisis communication efforts.

Limitations and directions for future research

The present research, while insightful, is subject to certain limitations. The two experiments were conducted with Dutch-speaking Belgian samples, implying that the generalizability of these findings to diverse cultures, languages, and broader media contexts necessitates further investigation. Furthermore, real-world crisis scenarios are often influenced by a multiplicity of mediators beyond mere responsibility attributions, such as public perceptions of credibility, empathy, or organizational past transgressions, which were not comprehensively explored. Future research avenues could fruitfully explore additional crisis response strategies, including the use of excuses, justifications, expressions of empathy, or bolstering, and examine their potential interactions with verb voice. Investigations into how oral versus written delivery modalities might alter the effects, or cross-cultural differences in the interpretation and impact of passive constructions, would also be valuable. Finally, assessing additional mediators and tracking long-term reputational trajectories could provide a more holistic understanding of crisis communication dynamics.

Takeaways for quick recall

When considering crisis response strategies, the choice between an apology and a denial is critical: apologies generally escalate responsibility attributions and can harm reputation, particularly when guilt is not definitively established, compared to denials. The verb voice employed also wields influence, with the passive voice effectively lowering responsibility perceptions more than the active voice. However, this benefit is strictly contingent on the response being a denial; an apology's impact on public perception remains largely unaffected by the verb voice. Furthermore, the efficacy of the passive voice in reducing responsibility is predominantly realized when the agent of the action is entirely omitted (agentless passive); an agentive passive, which still identifies an agent, fails to produce the same advantageous effect. Practically, it is advisable to reserve passive phrasing for denials that aim to distance the organization by removing the agent, and to avoid passive apologies if the objective is to minimize public perception of responsibility.

References ( core ideas cited )

The study draws extensively on existing literature, citing core ideas that underpin its theoretical framework and hypotheses. Key references include the foundational work on Situational Crisis Communication Theory (SCCT) and its articulation of the link between responsibility attributions and resultant reputational damage, as expounded by Coombs (2007) and Coombs & Holladay (2002). Psycholinguistic findings concerning how the active versus passive voice influences perceptions of responsibility are also central, with studies by Henley et al. (1995), Knobloch-Westerwick & Taylor (2008), and Platow & Brodie (1999) being crucial. Additionally, prior research on the comparative effects of apologies versus denials on public trust and credibility, including works by Schoofs et al. (2019), Kim & Sung (2014), and Coombs & Holladay (2008), provides further context and support.

Appendices (stimuli at a glance)

For Study 1, the stimuli involved a crisis scenario concerning \"Softeck,\" with varying wordings for denials and apologies. The manipulation of active versus passive voice was specifically applied to the CEO's statements within these scenarios. Study 2 utilized a new crisis scenario, \"TimberToys,\" involving cadmium in toys. This scenario also featured manipulations of denial versus apology responses, but expanded the verb voice manipulation to include three distinct forms: active, agentive passive, and agentless passive.