Zoodsma and Schaafsma’s (2022) research revealed an upward trend in the instances of political apologies rendered by countries in the 20 years leading up to their study. On an organisational level, Wang et al. (2023) speak of the contemporary prevalence of ‘apology culture’ (p. 2).

Corporate apologies rendered in the recent past include PricewaterhouseCoopers (PwC) Australia and Volkswagen. PricewaterhouseCoopers Australia communicated confidential information as to upcoming taxation changes to clients (Foley, 2023). Volkswagen’s vehicle emissions figures proved much higher than those advertised by the company (Associated Press, 2016; Sillars, 2022). Bain and Company and the South African Revenue Services published apologies for their involvement in what has been termed ‘State Capture’ in South Africa (SA), as revealed in the findings of the Zondo Commission (Laurence, 2022; SARS, 2022).

Research within the field of corporate crisis communication is growing and considers the impact of the formulation and rendering of corporate apologies on company stakeholders to manage public image and relations (for example, see Bentley & Ma, 2020). Studies such as Shao et al. (2022) have considered factors influencing the perception of a good corporate apology and the impact of a good apology on organisational performance and other indicators.

On an interpersonal level, what has been coined ‘cancel culture’ has regained prominence in contemporary times, with individuals and organisations reportedly being cancelled for offensive actions that have been publicised (Bies et al., 2021). Leaders’ apologies to their followers constitute an instance of an apology on an interpersonal level. Researchers such as Epitropaki et al. (2020), Kahkonen et al. (2021) and Sharma et al. (2023) have conducted comprehensive literature reviews in the field of leaders’ apologies. While their syntheses (and Epitropaki et al.’s [2020] proposed framework) remain untested in their entirety, empirical studies have considered the relationships among variables encompassed by these models. Wang et al. (2023), Grover et al. (2019) and Hetrick et al. (2021) have investigated the relationship between leaders’ apologies and follower trust, transgression severity and intentionality and follower attributions for unintentional transgressions.

Consideration of what constitutes an effective apology and the factors influencing the perception of the same potentially assists leaders in empathising with those receiving their apologies and considering relevant factors in apology formulation, thereby better equipping leaders to render effective apologies.

In sum, the objective of the present study is to inform an empirically tested model to explain the factors affecting followers’ perceptions of the effectiveness of their leaders’ apologies. The boundaries of the study need to be defined. Firstly, the focus of the present study is on one group of stakeholders, namely followers (employees). Secondly, the study focusses on an interpersonal level, on leader (manager) apologies – or apologies rendered by a leader to one or more followers internal to the organisation. Thirdly, the efficacy of leaders’ apologies is considered from the perspective of followers as one particular group of stakeholders.

The purpose of the study is to investigate factors influencing followers’ perception of the effectiveness of leaders’ apologies. The quality of the leader’s apology, which includes the leader’s (cognitive, affective and conative) apology content, the promptness of the apology and the delivery channel through which it is delivered, serves as an independent variable. Followers’ perception is operationalised in two dependent variables, namely firstly, their perception of an apology’s authenticity and secondly, their relationship with their leader post-wrongdoing and apology. A final dependent variable to be considered is followers’ perception of the fairness of the apology, which is defined as the extent to which the apology is viewed as evidencing distributive, procedural and interactional justice. The moderating variables that are considered include, firstly, the leader-follower relationship (LFR) prior to the wrongdoing, and secondly, the type of transgression. The latter is defined as either intentional or preventable.

Sverdrup and Stensaker (2018) highlighted that there are two main strands of apology research. In the first, researchers focus on apologies between individuals. In the second apology research strand, namely corporate apologies, researchers focus on apologies on an organisational level, considering groups of stakeholders. The present article falls within the first of these two strands.

Based on a literature review, Epitropaki et al. (2020) have proposed a leader transgression and subsequent relationship repair framework on an individual level. Their framework is comprehensive, including firstly, pre-transgression aspects such as the characteristics of the leader, of the follower and of the relationship between the two. Secondly, characteristics of the transgression itself are considered, as well as the individual’s cognitive and emotional evaluations of the same and of the impact on the LFR. Finally, post-transgression factors come into play, including relationship repair strategies. Epitropaki et al.’s (2020) study has not been empirically tested and does not clearly specify the elements of a leader’s apology.

In the realm of corporate apologies, Shao et al. (2022) developed a holistic corporate apology framework, by qualitatively analysing 40 corporate apology statements in conjunction with an analysis of existing corporate apology literature. The framework these authors developed is presented (exactly as it is presented in the original article) in Figure 1.

Shao et al.’s (2022) framework has not been empirically tested, and the latter framework focussed on corporate apologies rather than on leaders’ apologies. These authors’ framework included various stakeholders and was not specific to the impact of apologies on followers.

The following aspects of Shao et al.’s (2022) framework featured in the development of the questionnaire used for the current study: leader apology elements (LAE), authenticity, fairness (both seen from the perspective of the follower) and transgression type. The prior firm reputation was replaced by the prior LFR. Apology tradeoff, firm performance and spokesperson components were deemed not to be of relevance to leaders’ apologies.

Given that there is no empirically tested, comprehensive model to explain the impact of leaders’ apologies on followers, the question is posed, within the proposed research: can elements of Shao et al.’s (2022) framework be adapted and applied to leaders’ apologies in organisations?

The initial intention of the research was to test a model including 14 variables within Shao et al.’s (2022) framework, in the context of followers’ perceptions of leaders’ apologies in organisations. However, as will be unpacked with regard to methodology, the resultant measurement model limited the extent to which this was possible and determined which variables could be studied. These include LAEs, fairness, authenticity, prior and post-LFRs and transgression type.

A cross-sectional survey design was utilised as a change in followers’ perceptions of their leaders’ apologies over time was not being assessed. The research design was quantitative, consistent with the objective of examining causal relationships among variables.

The research population was comprised of all individuals who are working in an organisation. Non-probability sampling was used. The questionnaire was distributed via LinkedIn and WhatsApp, via SurveyCircle™ (2023) in particular and also SurveySwap™ (Ringle et al., 2022). SurveyCircleTM is a moderated platform with built-in checks to ensure participants do not complete too quickly. The questionnaire was also distributed to the researchers’ network, and via fellow students and colleagues to request volunteers to be part of the study. Volunteers were also requested to forward the same to their contacts.

A total of 486 responses, the final sample comprised 311 respondents who responded ‘yes’ to the filtering question, ‘My leader (past or present) committed a wrongdoing or caused harm (could be major, or relatively minor), towards me’. South Africans made up 50.20% of the sample, with the remainder hailing from over 50 other nationalities. Women comprised most of the sample (61.10%). The sample consisted of 49.50% white people, 22.50% black people, 19.60% Asian people, 4.80% mixed race people and 3.50% Other. Most of the respondents were between 35 years and 44 years of age (34.40%), while 28.60% were aged 25 years–34, 17.00% between 45 years and 54 years of age, 15.10% between 18 years and 24 years of age and 4.80% were 55 years of age or older. Most (37.00%) were in middle or senior management, 21.90% were Specialists, 14.80% in Executive positions, 13.20% in Administrative positions and 6.40% in Other positions. Out of the total number, 73.60% held postgraduate qualifications, while 1.30% did not have a matriculation or school-leaving qualification. For at least half of the sample that is South African, this qualification demographic is very high and therefore potentially a skewed sub-sample.

To assess the independent variable, LAEs, as well as two of the dependent variables (fairness and authenticity), new items were constructed by the authors. The same five items from Radulovic et al.’s (2019) seven-item LMX scale were used to assess the remaining dependent variable, post LFR and the moderating variable, prior LFR, Radulovic et al.’s (2019) original scale spoke to a hypothetical scenario; therefore, items were edited such that they were no longer scenario specific. Items assessing prior LFRs were prefaced, ‘[before] the wrongdoing or harm I experienced’, while post LFRs were prefaced, ‘[following] my leader’s apology (or, in the absence of one, following the process post-wrongdoing)’. Again, for the second moderating variable, Transgression type, new items were constructed by the authors. Questions were edited to take into account feedback received from pilot study participants. The final questionnaire is provided in Table 6 at the end of this article. In addition to the filtering question mentioned under the ‘Research participants’ section, two more filtering questions were added. The first (Q2) checked whether or not participants were fulltime employees. However, this requirement was removed early on in the data collection phase to broaden the sample base, and while the question was included, a ‘yes’ was not required to continue through the questionnaire. The second (Q3) filtering question ascertained how long ago the wrongdoing occurred in order to account for possible effects of the passage of time on memory. Six questions (Questions a–f) were included to collect demographic information concerning the sample gathered (sex, nationality, qualification level, job level, age and ethnicity). This was done in order to provide a detailed description of the eventual sample while still maintaining respondent anonymity. The questionnaire was pretested by three people, followed by slight edits to enhance user-friendliness and logical flow.

Responses were exported to Microsoft^® Excel^® Version 16.77.1 and imported into both IBM^® SPSS^® Version 28.0.1.0 and SmartPLS™ 4 (Ringle et al., 2022). Descriptive statistical techniques and exploratory factor analysis (EFA) were conducted, along with partial least squares structural equation modelling (PLS-SEM).

Written ethical approval was obtained from the Ethics Committee of the University of Pretoria’s Gordon Institute of Business Science (GIBS), prior to questionnaire distribution. Questionnaire responses were gathered anonymously. The first page of the questionnaire consisted of an informed consent clause, and respondents were informed that their completion of the questionnaire would be taken to indicate that they granted their informed consent for their responses to be utilised.

Because H₂ was accepted, the quality of the content of the leader’s apology, its timing in relation to the wrongdoing, the channel via which it is communicated (public or private) and the degree to which the apology enhances the follower’s perceptions of distributive, procedural and interactional justice is likely to have a positive effect on to the extent to which the follower perceives the leader’s apology to be sincere and genuine in nature.

The quality of the contents, timing, channel of the leader’s apology and the extent to which it enhances perceptions of fairness are together likely to enhance the quality of the LFR after the apology has taken place, given that support was found for H₃ in the current study. Because no support was found for H₅, the quality of the LFR prior to both the leader’s wrongdoing and apology is unlikely to either dampen or enhance the relationship between LAEs including fairness, and the extent to which the apology is perceived to be authentic. Similarly, as H₆ was not supported, the relationship prior to both the leader’s wrongdoing and apology is not likely to impact on the relationship between followers’ perception of LAEs and their relationship with their leader post-wrongdoing and apology.

The extent to which the follower perceives the wrongdoing to have been accidental and not preventable on the part of the leader enhances the relationship between the quality of the apology elements and the follower’s perception of the authenticity of the apology (H₈). Also, the more leaders’ transgressions are viewed as accidental and not preventable by followers, the more positive the relationship between elements of leaders’ apologies and post LFR (H₉).

The findings contribute both to leader apology andto LMX research. The results confirm Sharma et al.’s (2023) finding that the nature of leaders’ wrongdoing can influence the extent to which apologies are effective. Basford et al.’s (2014) finding that effective leaders’ apologies to be associated with higher quality LMX relationships are also confirmed by the present study.

Epitropaki et al. (2020) asserted that the quality of the LMX relationship could serve as a boundary condition for relationship repair, and Radulovic et al. (2019) found that high-quality LMX relationships could enhance relationship repair post-wrongdoing. In support of these authors’ studies, the current study found that the quality of the LMX relationship prior to the wrongdoing moderated the impact of the quality of the LAEs on the LMX relationship post-wrongdoing (and, if relevant, apology). Epitropaki et al.’s (2020) and Radulovic et al.’s (2019) findings are not fully supported by the current study; however, the quality of the LMX relationship prior to the wrongdoing was not found to moderate the impact of the apology elements on its perceived authenticity. While high-quality LMX relationships may serve as a boundary condition to relationship repair, they may not serve as one to followers’ perceptions of leader apology authenticity.

Managers and supervisors, in their capacity as leaders, are therefore cautioned to work on the quality of their apology in terms of its content, its promptness and the channel via which it is delivered. Managers should not assume that the strength of their relationship with their followers will enhance the extent to which followers are convinced of leaders’ apologies’ authenticity. Extra care should be taken should their followers perceive the leaders’ wrongdoing to have been preventable.

Leaders’ proactively working on the quality of their relationships with their followers is likely to protect (and possibly enhance) those relationships, if a leader commits a wrongdoing and has to apologise, provided the quality of the elements of the apology is perceived to be comprehensive, prompt and fair in nature. Again, should followers perceive a transgression to have been accidental or as not preventable on the part of the leader, this is likely to enhance the impact of the quality of the elements of the leader’s apology on the LFR post-apology.

Leaders should take the time to prepare their apologies to ensure that the wrongdoing is acknowledged, defined and explained at length revealing a deep understanding of its magnitude and harm. The apology should include an acceptance of full responsibility for the wrongdoing and detail as to remediation as well as a plan of action to build and maintain relationships with those affected by the wrongdoing. Leaders’ expression of empathy and remorse during the apology will assist in its effectiveness (Shao et al., 2022).

The present study opens avenues for future research. Firstly, while a large and diverse sample was sourced, South Africans were best represented, as were women, white people, those 35 years–44 years of age, middle to senior managers and those with post-graduate degrees. Future studies should assess the generalisability of results. Secondly, the discriminant validity of the measurement model compromised the extent to which the model could confirm the adaptation of Shao et al.’s (2022) model as initially intended. Future studies could refine the model.

Thirdly, while the research assumed Shao et al.’s (2022) definition of a leader apology as that which included the contents stipulated in these authors’ adapted model, the extent to which respondents agreed with this definition was not assessed. Future research should establish what elements are essential for a leader’s apology to be perceived as such by followers.

Fourthly, the variables such as LAE and fairness were conflated as a result of the EFA in the present study. It is recommended that the relationship between these two variables be explored in more depth in future studies.

Finally, 176 of the original 486 respondents stated their leader had not committed a transgression against them; thus respondents may assume a threshold above which an act is considered a wrongdoing. Future research should consider possible demographic and contextual differences that could contribute to respondents falling above or below such a threshold. Also, the significant decline found in post LFRs suggests followers’ perceptions of the nature of transgression could play a role in the post-wrongdoing outcome. Future studies, establishing the nature of the transgression and the factors that influence it, will inform situations in which apologies are necessary and impactful.