Operational dilemmas in safety-critical industries: The tension between organizational reputational concerns and the effective communication of risk

Abstract

Organizations involved in safety-critical operations often deal with operational tensions, especially when involved in safety-critical incidents that is likely to violate safety. In this paper, we set out to understand how the disclosures of safety-critical incidents take place in the face of reputational tension. Based on the case of the Nigerian National Petroleum Corporation (NNPC), we draw on image repair theory and information manipulation theory and adopt discourse analysis as a method of analyzing safety-critical incident press releases and reports from the NNPC. We found NNPC deploying image repair as part of incident disclosures to deflect attention, evade blame and avoid issuing apologies. This is supported by the violation of the conversational maxims. The paper provides a theoretical model for discursively assessing the practices of incident information disclosure by an organization in the face of reputational tension, and further assesses the risk communication implications of such practices.

Introduction

Safety-critical operations such as in the petroleum industry consist of complex sociotechnical systems of ‘people in multiple roles and their artifacts [e.g., assets, technology and organizations]’ (Lwears, 2012: 4561). These sociotechnical systems need to possess the highest levels of safety integrity as failures or malfunctions from the system(s) would lead to catastrophic incidents with the potential of violating public, environmental and asset safety (Ambituuni, Hopkins, Amezaga, Werner, & Wood, 2015). Because of this operational dynamic, there are a number of tensions and trade-offs within safety-critical industry organizations. These tensions are captured in literature and include the profit versus safety tension (Baron & Pate-Cornell, 1999; Kettunen, Reiman, & Wahlström, 2007; Perin, 2005; Reiman & Rollenhagen, 2012); the organization versus the regulator's divergent views on safety regulation and documentation (Ambituuni, Amezaga, & Werner, 2015; Kettunen, Reiman, & Wahlström, 2007); and the safety versus performance tension (Amalberti, 2001). These trade-offs and tensions need to be managed to maintain safe sociotechnical operations. But what further tension exists in the event that safety is violated by an incident involving such organizations, for example, a failure of petroleum pipeline? Indeed, incidents involving such safety-critical petroleum operations are likely to have devastating consequences (Ambituuni et al., 2015) with image damaging implications to the operating organization.

Thus, when an organization is involved in an image-damaging incident, such an organization is likely to deploy certain image repair or restoration practices (McCoy, 2014) amidst reputational tension. The contemporary organization research on image repair brings to light the practices of image repair discourse (Arpan & Roskos-Ewoldsen, 2005; Benoit, 1994; Muralidharan, Dillistone, & Shin, 2011). Furthermore, the seminal work of Benoit (1994) on image repair provides a theoretical conception of how organizations attempt to correct negative public perception of themselves following a specific incident or series of incidents. Recent research has opened up the domain of critically assessing the safety measures in organizations involved in safety-critical incidents, [e.g., British petroleum (BP) and the 2010 Deepwater Horizon disaster (Amernic & Craig, 2017)] in light of the image repair strategy deployed by the organization (Harlow, Brantley, & Harlow, 2011). Research also shows that effective disclosure of incidents by an organization ensures that adequate safety and emergency management information is available and lessons learned (Crichton, Ramsay, & Kelly, 2009; Enander, 2018). Whilst the aforementioned research provides justification for incident disclosure, the reputation of the organization remains at tension with the principles of effective incident and risk communication (Smillie & Blissett, 2010) especially when there is the need to communicate and disclose such information with the accurate veracity, manner, quality and quantity.

In this paper, we set out to develop an understanding of how the effective communication of safety-critical incidents takes place in the face of reputational tension. We focus on the disclosure of incidents by the Nigerian National Petroleum Company (NNPC) following incidents involving NNPC's safety-critical assets. We turn to image repair theory (IRT) (Benoit, 1994) and information manipulation theory (IMT) (McCornack, Levine, Solowczuk, Torres, & Campbell, 1992) as a twin theoretical lens through which we analyzed the archival data from the NNPC using a case study approach. Our choice of the NNPC is purposive because of the historic records of safety-critical incidents that the organization has been involved in (see e.g., Ambituuni et al., 2015, Ambituuni, Ochieng, & Amezaga, 2018; Anifowose, Lawler, van der Horst, & Chapman, 2012). As a petroleum industry organization, the NNPC operates in the context of societal dynamics of risk management, hence, and knowledge of incidents involving its operational assets builds both a social construct (Haukelid, 2008; Silbey, 2009) and ontological realism (Aven & Renn, 2010) of its performance. Therefore, we ask, when involved in a safety-critical incident, how does the NNPC communicate safety-critical incidents in the face of reputational tension? What image or reputational repair practices are deployed and how is incident information disseminated from the perspectives of veracity, manner, quality and quantity.

We found that during incident disclosure, and in the face of reputational tension, NNPC violates conversational maxims to deflect attention, evade blame and avoid issuing an apology whilst deploying image repair strategies. This approach has wider safety and risk communication implication. We found that prioritizing the image of the organization through deflection during incident disclosure falsely influences reception and acceptance of risk information by potential risk-receptors which has further safety implication on emergency response, especially where the seriousness and potential consequence of the incident is downplayed such that it further exposes people to greater danger. Our findings also contribute theoretically by providing a framework that will enable critical assessment of the veracity, manner, quality and quantity of information released by organizations whilst disclosing incidents.

Looking Through a Twin Theoretical Lens of IRT and IMT

The knowledge that an organization has been involved in an incident with safety and environmental implications threatens the reputation of the organization. Indeed, for such an organization, the incident, or the asset involved in the incident will be seen as violating the safety of people, the environment and properties, hence putting them at risk. This is especially true when defining safety as a condition where nothing goes wrong or, more cautiously, as a condition where the number of things that went wrong is acceptably small (Hollnagel, 2016). Boholm & Corvellec (2011) conceptualized risk as stemming from situated cognition that ascertains a relationship of risk linking two objects, a risk object (e.g., explosion, fire, petroleum pipeline leaks) and an object at risk (e.g., humans or natural environment at risk). The causal and contingent link is such that the risk object is considered, in some way and under certain circumstances, to threaten the valued object at risk.

The risk perspective of an object at risk can be constructed by the ‘object’ itself, for example, when people raise concern about their own safety in the event of safety-critical incidents, or on behalf of the object for example, when people raise concern about the safety of the environment. This perspective is shaped by the ontology of risk to its epistemology,that is, from the nature of risk to risk as a social construct (Hilgartner, 1992). The ontology of risk conceptualizes the extent to which risk exist independent of an assessor or communicator perspectives, whilst the epistemology of risk conceptualizes risk in the context of knowledge of risk and lack thereof, therefore, contributing to making issues of risk complications. Hence, when an organization is seen to have put an object at risk, such an organization engages discourse and communication as a way to socially construct its reputation and/or repair its image and shape the epistemology of risk (Coombs, Frandsen, Holladay, & Johansen, 2010; Coombs & Holladay, 2012). Indeed, the timing and the point at which an organization deploys image repair strategies following an incident is equally important (Coombs et al., 2010; Sturges, 1994). Image and reputation management communication should only happen after addressing public safety, especially in safety-critical situations. Risk receptors must know how to safeguard themselves based on factual incident information before the organization considers reputational concerns (Seeger, 2006; Ulmer, 2001).

Benoit (1994) asserted that image repair (also known as image restoration) is a discursive practice that provides persuasive message or group of messages in response or defence of negative threats to the image of an organization. Threats to image consist of two elements,that is, the ‘blame’ element which is linked to the beliefs of an audience, and the ‘offensiveness’ which is associated with the value systems of the audience. Image repair discourse, therefore, rejects or reduces responsibilities, thereby distorting beliefs about blame, and/or reduces the level of offensiveness, thus, altering value. Because an organization's image is directly linked to its reputation (Alvesson, 1998; Ashforth & Gibbs, 1990), reputation is seen as a resource (Barney, 2001) worthy of defending. What then are the image repair practices used to reject responsibilities or reduce offensiveness?

Dionisopolous and Vibbert (1988) argued that four image repair practices could be used, that is, denial; bolstering; differentiation; and transcendence. An organization could deny performing the wrongful or the occurrence of an incident (denial). Bolstering involves portraying a positive perception of the organization in light of a reputational damaging event such that the positive associations become the focal point. Differentiation uses a favourable explanation of the incident to offset negative feelings with positive characteristics. Transcendence provides a more abstract interpretation of the character attack of an organization by appealing to values and loyalties. Harlow, Brantley, and Harlow (2011) drew on the work of Benoit (1994) and assessed the image repair practices used by the BP following the Gulf of Mexico Deepwater Horizon spill. They listed 10 additional image repair practices to the aforementioned including: shifting blame (someone else did it); provocation (we did it, but were provoked); defeasibility (lack of information or control), accident (the incident was an accident); good intentions (the error was the result of good intentions); minimization (it's not that bad); attack accuser; compensation; corrective action; and mortification (Harlow, Brantley, & Harlow, 2011). These practices explain ‘what’ communications and actions can be taken by an organization in an attempt to correct negative public perception of themselves after a specific incident or series of incidents (Smudde & Courtright, 2008).

Image repair practices also mix ideas from rhetorical discourse with justificatory account giving. Benoit (1994: 5) defines image ‘as the perception of a person (or group, or organization) held by the audience, shaped by the words and actions of that person [or organization] ….’ This gives ‘image’ a sociocultural perspective. This perspective suggests that image repair practice is heavily reliant on communication and information dissemination such that audiences receive words from the organization and based on their judgement on such words. The communication could, therefore, go beyond the traditional communication to include subtle and often meaning-laden discourse which an organization can use to persuade its audience and shape their perceptions. To understand the dynamics of how persuasion could be achieved, we turn to IMT.

McCornack et al. (1992) asserted that in any given conversation there exists a set of basic, reasonable assumptions about how transmission of information occurs. This premise is based on the work of Paul Grice (Grice, 1989) on Cooperation Principle (CP) and its maxims which describes a variety of manipulated message forms. The theory views manipulation as emanating from covert violations of one or more of Grice's four maxims (quality, quantity, relevance and manner). According to McCornack et al., (1992), the violation of quantity maxim relates to altering the amount of information that should be provided in a message. Violation of quality maxim relates to prevarication of the veracity (truthfulness) of information that is presented whilst manipulation by evasion relates to the covert violation of the relevance maxim and manipulation by equivocation results from the covert violation of manner maxim (Anolli, Ciceri, Riva, Anolli, Balconi, & Ciceri, 2001; Yeung, Levine, & Nishiyama, 1999). As seen in literature IMT has been used to understanding the way information is manipulated in public consultation for project proposals (Hurlimann & Dolnicar, 2010); to assess whether managers manipulate the measurement information that is used to assess performance (Fisher & Downes, 2008); and in assessing the intentional use of ambiguity in strategic organizational communication (Dulek & Campbell, 2015). This opens an opportunity for using the theory in assessing the robustness of communication by an organization involved in safety-critical incidents in light of reputational tension, particularly, if such communications have safety and risk implications with respect to emergency response, regulatory sanctions and enacting social trust. Consequently, we present the paper's theoretical framing in Figure 1 below.

Figure 1. Twin theoretical conception of IRT and IMT

Through the twin theoretical conception of IRT and IMT, we were able to evaluate the meaning-laden discourse deployed by the organization in light of reputational concerns. This allowed us to assess, the image or reputational repair practices are deployed and how incident information is disseminated from the perspectives of veracity, manner, quality and quantity. In the next section, we discuss our theoretical conception of the reason an organization deploys rhetorical and metaphorical discourse in incident disclosure.

Incident Disclosure, Rhetoric and Metaphor

Prior research has already illustrated that self-disclosure of incidents can be beneficial to organizations (Arpan & Roskos-Ewoldsen, 2005). However, incident disclosure whether self-disclosed or disclosed by a third party will undoubtably raise questions about the safety performance of the affected organization. This is especially true for safety-critical operations as incidents emanating from such operations can be of devastating consequences to safety.

When an incident happens and the consequence of such an incident affects human and environmental safety, the condition of safety is said to be violated. Hence both individual and societal risk perceptions are shaped by the knowledge of such events. Moreover, because maximum level of safety corresponds with zero risk and a low safety level guaranteed corresponds with a risk of almost 100% (Aven, 2009; Suddle, 2009), the psychological elements of safety are shaped by incident knowledge obtained from incident disclosures. Indeed, an organization's audience will judge its safety performance based on such knowledge. Similarly, knowledge of incidents caused by deficiencies in the organization's sociotechnical functionality may result in the imposition of regulatory sanctions. This suggests the rationale behind the use of rhetorical and metaphoric discourse to achieve image repair during incident disclosures by such an organization. We define incident disclosure as the dissemination of incident-related information by an organization to its audience for the purpose of safety management and risk governance including regulatory compliance, emergency communication and response, and liability management, and for ensuring social trust.

There is a recognized link between the conscious uses of certain metaphors with the conveyance of biases (Tinker, 1986) in order to build ideological attachments (N. Fairclough, 2014). Amernic & Craig (2017) discussed the use of metaphors in the camouflaging of ideology and its nuanced power during communication. Of similar application is the ‘art of persuasion' called rhetoric (Larson, 2012). Rhetoric has been used as a means of cajoling audience to have similar viewpoints with a communicator, covertly. Whether the communicator's viewpoints are true and valid or not, the use of rhetoric creates an informed appetite for good (Booth, 2004). Hence, in order to understand the application of metaphor and rhetoric during incident disclosure, we adopt a similar approach used by Amernic & Craig (2017) for analysing discourse by focusing on individual instances of metaphor and rhetoric in incident disclosure text (i.e., press release and incident reports) and also focusing on an underlying rhetoric and metaphoric structure of the text. To achieve this, we turn to discourse analysis as an analytical tool to aid in the analysis of collected archival incident reports and press releases following incidents involving NNPC's operations. We describe this in our method in the succeeding section.

Method: A Case Study Approach

We adopt Yin's (Yin, 2009, 2012) explanatory case study methodology as it is particularly suitable for answering ‘how’ and ‘why’ questions. For this research, we set out to develop an understanding of how NNPC discloses and communicate safety-critical incidents in the face of reputational tension. We purposively focus on the single case of NNPC because of the records of safety-critical incidents involving the organization (Ambituuni et al., 2015, Ambituuni, Ochieng, & Amezaga, 2018; Anifowose et al., 2012).

As a state-owned National petroleum company, the NNPC owns and operates safety critical upstream petroleum assets, via its six joint ventures with Shell, Chevron, Mobil, Agip, Elf and Texaco (NNPC, 2017). NNPC is also a major downstream player through ownership of all four (4) existing refineries in Nigeria, distribution pipelines, depots, oil import jetties, and retail stations. It also controls over 23% of the retails market subsector in Nigeria (Ambituuni, Amezaga, & Emeseh, 2014). Over the years, the organization has been implicated in many safety-critical incidents that has had a devastating impact on human and environmental safety For instance, Ambituuni et al. (2015) calculated the failure frequency of the entire 5001 KM pipeline network using 13 years' incident records (from 2000 to 2012) and found that in 2011 alone, for example, the NNPC reported a total of 2,787 pipeline failure. Such failures were attributed to operational deficiencies including poor practices and bureaucratic bottlenecks, and interdiction or sabotage of safety-critical assets (see e.g.,Ambituuni et al., 2015; Anifowose et al., 2012).

The safety and risk implication of such failures has been reported in Emeseh (2006), where, for instance, a pipeline explosion at Jesse community resulted in large-scale environmental pollution and killed over 1,500 people including women and children. Omodanisi, Eludoyin, & Salami (2014) also reported 14 different incidents including the December 26, 2006 pipeline explosion in Ilado-Odo around Lagos Nigeria that killed over 250 people. Over the years, this trend of safety critical-incidents has consistently occurred involving NNPC's assets including the 2018 Aba pipeline incident that killed 30 people. NNPC is also known for its operational secrecy and practices that suggests the lack of transparency of information disclosed by the organization (Emeseh, 2006). The case of NNPC, therefore, offers a rich context to explore incident disclosure and communication versus reputation tension in light of this reported incident.

The use of case study approach provides a robust method particularly when a holistic, in-depth investigation is required (Simons, 1996). The choice of a case study allowed us to study ‘how’ NNPC discloses and communicate these incidents in the face of reputational tension, identify the image or reputational repair practices deployed and explain the dynamic of incident information disseminated from the perspectives of veracity, manner, quality and quantity. Furthermore, the method allowed us to draw our theoretical conception of IRT and IMT in order to provide new generalizable perspectives of accessing the incident information released by organizations, particularly those involved in safety-critical operations. To achieve a robust case-study methodology, we applied Yin's five components of an effective case study research design. First, we established our aforementioned research questions by drawing of relevant literature. Second is the propositions or purpose of the study which we have discussed in the introductory part of the paper. Based on this, our unit of analysis focuses on incident disclosures by the case of the NNPC. Fourth, the logic that links data to propositions was achieved following the data collection phase, as themes emerge. We analyzed the data whilst building explanations about incident disclosure practices that appear in the data and linking these practices to the theoretical conception of the paper, thereby, answering the research questions. Finally, the interpretation of findings was achieved through careful extraction of meanings and theoretical links from the analysis (Yin, 2012).

Framework for data collection and analysis

The unique data collection method in this study provides a more holistic context of incident disclosure and the image repair practices used by an organization as the study benefited from the communication of many incidents within a single case organization. Hence, although the study drew inspiration from the methods used in existing literature where the focus is on a single incident (e.g., Harlow, Brantley, & Harlow, 2011; Muralidharan, Dillistone, & Shin, 2011), our method was designed to overcome some of the observed limitations of focusing on a single incident. Consequently, we obtained historic incident disclosure and communication data from two key sources. First, 170 incident reports were collected from NNPC's Group Health, Safety and Environment (HSE) Department.

In collecting these reports, we first wrote and obtained approval from the Group's Chief Executive through the HSE Department. Notwithstanding, obtaining comprehensive data was especially challenging due to the secretive nature of the petroleum industry in Nigeria, and particularly the NNPC. For example, we experienced deliberate deletion of some details from the pipeline failure reports due to confidentiality claims. This, however, did not present analytical challenge in this study as the obtained reports contained the data we needed to address our research questions. Nevertheless, there is the need to presume that since the obtained reports were written by the HSE department of NNPC, such reports are vulnerable for company management preferences, corporate culture, and specific investigation methods and work instructions. The HSE Department presents what happened in the company way. This could mean that even larger manipulation exists within the obtained reports. Ideally, the ‘absolute truth’ established by a third party would be the best reference basis.

Next, we carried out a systematic internet search of the press releases and press statements by NNPC in light of the incidents reported. We reviewed 573 press releases/statements on the NNPC website and further downloaded and reviewed 1114 online news articles using Nexis to identify statements issued by the NNPC in response to incidents involving their assets and operations. Through this review, 30 relevant press releases were identified as shown in Table 1.

Table 1. Data collected over the period of 2007–2018

As we have theoretically established earlier, image repair is achieved through discourse (Daly & Wiemann, 1994; Muralidharan, Dillistone, & Shin, 2011). Hence, the paper adopted discourse analysis for analysing the obtained data. Discourse analysis sets out to answer questions about language, about writers and speakers, and about sociocultural processes that surrounds and gives rise to discourse (Gee, 2010). It involves paying close and systematic attention specifically to texts and their contexts. Indeed, we observe the different variation to discourse analysis, (N. L. Fairclough, 1985; N. Fairclough, 2014; van Dijk, 1996) with different analytical objectives. Hence, for this research attention was given to the use of discourse analysis in alignment to the research questions. Consequently, we adopted Gee's definition of discourse that is, ‘way of combining and integrating language [including rhetoric, metaphors], action, interventions, ways of thinking, believing, valuing, and using various symbols, tools, and objects to enact a particular sort of socially recognised identity’ (Gee, 2010: 21). This allowed us to consider three key elements of discourse during our analysis by paying attention to how NNPC is disclosing and communicating incident information, ‘who’ NNPC is or trying to be and ‘what’ NNPC is doing or trying to do in incident reports and press releases. We provide the list of press releases cited in our analysis See Table A1 in the appendix.

Discourse analysis involves ‘a constant movement back and forth between theory and empirical data’ (Wodak, 2004: 200). A similar approach was adopted. We analyzed our research data and developed empirical findings whilst constantly refining our theoretical conception. This was achieved through the development of content categories (Robson, 2015). The content categories were then used to analyze NNPC's incident disclosure model in relation to the tension between reputational concerns and risk communication objectives. The first step we took in our discourse analysis was to outline our categories and themes based on our research questions and the theoretical underpinning that emanate from our review of the literature. Using NVivo, we then went through the obtained reports and press releases to see if it contains any of these themes and categories. At this point, necessary adjustment was made to include new themes and irrelevant themes were excluded. Furthermore, broader themes were broken down into sub-themes. At the end of this process, a list of coded discourse strands themes was obtained and reviewed. In the second step, we began looking at the structural features of the texts. We identified the sections within the text that dealt with particular discourse strands and the different discourse strands that overlapped. We further identify how the NNPC structure their incident disclosures in light of our theoretical conception of the research question. This led us to obtain a good idea of the macro-features of the text in our data.

Finally, we zoomed in on the individual statements and discourse fragments by collecting all statements with a specific theme, and examining what the statements have to say on the respective discourse strand, vis-à-vis the context that informs the argument. This allowed us to analyze the function that intertextuality serves in light of the overall argument. Hence, we were able to identify and analyze the practices used by NNPC to achieve image repair vis-à-vis any violation of conversational maxims observed within the themes generated. Also, to achieve validity and reliability, Yin's tactics (construct validity, internal and external validity and reliability) were carefully integrated. Construct validity was achieved by the use of multiple sources of evidence and the establishment of a chain of evidence. Internal and external validity were addressed mainly through explanation building and replication logic, respectively. Reliability was achieved using case study protocols and database such that data collection procedure can be repeated (Yin, 2009).

Result

We set out to understand how NNPC discloses and communicate safety-critical incidents in the face of reputational tension, identify the image or reputational repair practices and explain the dynamic of incident information disseminated from the perspectives of veracity, manner, quality and quantity. We found that in the process of disclosing incidents to its audience, NNPC pays great attention to its reputation, hence using image repair as a means of reputational defence whilst ignoring the principles of effective incident and risk communication even when human and environmental safety are at risk.

Intertextual analysis and the development of organizational discourse

The intertextual totality of the analyzed incident reports and press release provides an insight into the approach used by NNPC to violate conversational maxims whilst disclosing incidents in the face of reputational tension. It further provides comprehension of NNPC's inclination towards addressing reputational concerns against safety in its incident reports and press release. For instance, NNPC claims that ‘the safety of our people and the protection of our environment and assets are the main focus’. Paradoxically, we found the word ‘safety’ is used only 24 times in the entire 15,272 words from the press releases and incident reports analyzed. The text is dominated by statements concerning operational performance, reputational discourse, and organizational efficiency. Indeed, the focus on performance, efficiency, and reputation and the lack of use of the keyword ‘safety’ discursively indicates that safety is not a desired organizational state. Notably, also, the context in which the word ‘safety’ is used, and the missing emphasis on the risks associated with the incidents, as well as the implications for objects at risk. Instead, the emphasis and the context in which the word safety was used centred on portraying that the organization as capable, effective, efficient, and therefore, focused on the image of the organization.

Table 2 represents the categories developed from the initial content analysis of the incident reports and press releases as a measure of reputational inclination. Throughout, there is an apparent privileging of reputational concerns over the safety of objects at risk and no acknowledgement of the sociotechnical deficiencies within the organization or disclosure of the complexity of safety culture as an ongoing organizational challenge.

Table 2. The measure of reputational concerns in incident disclosures

These emergent categories were further analyzed into three organizational discourses: deflecting incident disclosure; the unavoidable incident syndrome and the unapologetic metaphor. These categories are discussed in the succeeding sections, focusing on how conversational maxims of quantity, quality, manner and relevance are violated to support the deployment of image repair strategies in alignment with the observed inclination toward reputational concerns.

Deflecting incident disclosure

Disclosure of accurate incident information illustrates a way of accepting safety deficiencies within sociotechnical systems of an organization. It suggests that such an organization is transparent to safety, accepts responsibilities for its deficiency and is ready to learn from the incident. On the contrary, we found NNPC deflecting attention whilst disclosing incidents to its audience. This is mostly done by violating the relevance maxims of the information provided during such disclosure. NNPC issue statements that lack focus on the deficiencies which led to the incident it is disclosing but also attempts to engage its audience by assuring of its aptitude to deliver what it deems a more relevant issue to them. An example is seen in NNPC's press release following the October, 2013 Warri refinery fire incident, cited below.

‘The Management of the Nigerian National Petroleum Corporation, NNPC on Tuesday confirmed that there was a minor fire incident in the Topping Unit of the Warri Refining and Petrochemical Company, WRPC… We wish to seize this opportunity to reassure members of the general public that the NNPC continues to hold sufficient stock of premium motor spirit [petrol] and other petroleum products and therefore there is no cause for panic buying….' (Press release document 2)

We noted two traits of maxims violation from NNPC's rhetoric above. First is an attempt to deflect attention from what may potentially be a dangerous occurrence to what NNPC termed a ‘minor fire incident’. We called this act of manipulation of information deflection by trivialization. In doing so, NNPC deflected its audience attention from the potential catastrophe that may have occurred. The word ‘minor’ is used in this context as a way of deflecting the seriousness of the incident. The focus on reputational concern appears to pressure the organization into reducing the veracity of the information communicated by trivializing the magnitude of the incident. Indeed, a fire incident in the Topping Unit of an oil refinery should be considered a very serious incident and a dangerous occurrence. We found similar use of terms like ‘minor challenge’ and ‘inconsequential incident’ used by NNPC to describe dangerous occurrences and serious incidents in 121 of the 170 incident reports. We also observed how NNPC trivialized incident communication in their press releases by excluding the causality information. For instance, even though four people died in September, 2017 NNPC Apapa jetty fire, the press release by NNPC excluded this detail.

Second is the attempt to shape perception by deflecting the attention of its audience from the entire event to what NNPC considers a more relevant issue. Because of its complete dominance of the refining, supply and distribution of petroleum product in Nigeria, NNPC knows that failure in its system will almost certainly result in a nationwide fuel scarcity, with the potential of crippling economic activities on a national scale. Hence, in this instance, it uses this knowledge to deflect attention from the ‘minor incident’ to its aptitude in averting fuel scarcity. By asserting this, NNPC showcases itsself as a solution to ‘panic buying’ despite the ‘minor fire incident’. We consider this an act of deflection by rationalization. The trait of rationalization in this act of information manipulation is evident in the way NNPC attempts to deal with potential stressors, away from safety concerns, that may have emanated from the incident through the elaboration of reassuring explanations. We cite another example below.

‘The Nigerian National Petroleum Corporation, NNPC, on Friday explained that routine gas supply for power generation is not affected by the fresh fire incident which engulfed a segment of the Escravos-Lagos Pipeline System’ (Press release 21).

Another trait of deflection observed is in the way in which NNPC communicates its HSE strategies as an integral part of incident disclosures. As cited in the example below NNPC recognizes the importance of a good safety culture, and positions itself as an organization with a good safety policy whilst citing external examples of organizations involved in disastrous incidents to portray the implication of bad safety culture. In doing so, NNPC appears to distance itself from being involved in incidents thereby providing a false sense of good safety performance, even though the disseminated message is linked to an incident involving the organization.

‘The deployment of a solid Health, Safety & Environment (HSE) policy in NNPC's business and operations has been described as ‘critical' to the on-going transformation programme in the Corporation… According to the GMD (Group Managing Director), it is pertinent for Chief Executives in the oil and gas industry to be HSE-conscious so as to prevent incidents and minimize environmental impacts. He noted that recent happenings in the global oil and gas industry like the Gulf of Mexico BP Oil Spill and the Chevron's Brazil Offshore incident further necessitated the desire to inculcate HSE among all staff of the Corporation’ (Press release document 13).

Here, NNPC describes its HSE policy as ‘solid’ and ‘critical’ to its ongoing transformation. The ‘solid’ and ‘critical’ metaphor evokes a sense of effective HSE policy that is, indeed, very important. However, it is the context in which this press release was deployed that captured our attention. We noted the way in which NNPC swiftly deflected attention from the incident being disclosed by citing incidents from other oil and gas operators at a global scene. We consider this an act of deflection by distantiation. Distantiation was achieved through the characterization of other operators as actors with bad safety performance. In doing so, the organization created characters that audiences can link bad safety performance to, thereby, deflecting such characteristics from itself and distancing itself from an object of scrutiny. Indeed, NNPC has been involved in some of the worse cases of incidents and disasters the global oil and gas industry has seen including the earlier stated pipeline explosion at Jesse community on October 15, 1998 that killed over 1,500 people including women and children and the December 26, 2006 pipeline explosion in Ilado-Odo which killed more than 250 people.

Distantiation was also observed in incident disclosure practises that focused on demonstrating capabilities of the organization as a means of obscuring the seriousness of the incident. This practise is manifested in NNPC's strong defence of its safety systems and procedures despite failures that resulted in disastrous incidents. Typical examples are the use of phrases like: ‘the gallantry of a team of engineers and technicians’, ‘the effectiveness of the firefighting equipment’, ‘the efficiency of our fire fighters’ and ‘the gallantry of NNPC personnel’ whilst reporting on incidents that had occurred with serious levels of consequences. The rhetoric from these phrases attempts to deflect attention to the robustness of the safety systems and procedures within NNPC whilst ignoring the severity of the incident on the public.

The ‘unavoidable incident’ syndrome

The ‘unavoidable incident’ syndrome associates incident occurrence to factors that on the surface, seem like they are unavoidable and are merely part of the day to day safety critical operations. The aim of such communication maxims violation is to normalize incident occurrence and provide a metaphor that suggests that nothing can be done to avoid such incidents and that such incidents are associated with ‘normal mechanical or systems’ fault. We illustrate how NNPC uses this information manipulation strategy from the citation below.

‘We are committed to doing all within our powers… in order to eliminate avoidable incidents' (Press release document 14).

From the above citation, NNPC attempts to demonstrate to its audience that there is a limit to what the organization can do to ensure that safety is optimized from its operations and to mitigate any associated risks. The word ‘powers’ is used in this context as a means of demonstrating absolute commitment which, however, can be constraint by forces beyond that power. This means that some incidents are simply ‘unavoidable’. The term ‘avoidable incidents’ is further used to reinforce the suggestion that some incidents are, indeed, mere accidents. The overall strategy of communication and incident disclosure here adopts a positive self-presentation using a metaphor that constructs an ideology of absolute commitment and hinged on accidents which ‘simply just’ occur.

The violation of conversational maxims by NNPC is mostly observed in the disclosure of incidents which NNPC claimed are associated with technical faults beyond their control. For instance, in the obtained incident reports, NNPC associated four incident causation factors to ‘pump overheating’, one incident causation factor to ‘failed mechanical seal’, three incident causation factors to ‘electrical fault’, and two incident causation factors to ‘auto ignition’. This narrow association of incidents to technical faults ignores the links between technical systems with human factors, organizational and regulatory deficiencies, procedure failures and dysfunctional work environment developed at various sociotechnical levels of the organization (Rasmussen, 1997). In doing so, NNPC covertly exonerates itself from such deficiencies as though the failed technical systems existed independent of its procedures, people, work environment and organization.

The ‘unapologetic’ metaphor

We have seen many safety critical operating companies apologizing following incidents and disasters involving their assets. A typical example is the unreserved apology by BP's CEO following the Gulf of Mexico disaster. An apology during incident disclosure accords an organization the opportunity to admit failures or errors within their sociotechnical systems and express remorse to its audience (Coombs et al., 2010). Sincere and unreserved apology makes use of terminologies that expresses regretful acknowledgement of one's failure, typically integrating the phrase ‘I'm sorry’. We found NNPC using metaphor that provides an appearance of sympathy whilst camouflaging an unapologetic stance in its incident disclosure strategy. For instance, in a statement issued by the NNPC's Group Managing Director (GMD) following the Okogbe community tanker fire that killed 95 persons, the GMD avoided issuing an apology but provided metaphoric condolence that appears regretful by stating that his:

‘…heart and that of the entire NNPC family is with the government and people of Rivers State, especially those who lost loved ones, at this time of great sorrow and grief.’ (Press release document 15)

The ‘heart’ metaphor evokes emotion and emphasizes the sympathy of the entire organization towards the victims and their loved ones. Furthermore, by issuing a personalized condolence, the GMD expresses his sympathy which only provides goodwill and neutral sentiments but failed to express remorse, guilt, and self-criticism, thereby, evading taking responsibility for the actions of NNPC. This may be an attempt to disenfranchise victims from taking legal actions as an apology is perceived as an admission of liability, or a conscious attempt to evade responsibility considering all NNPC trucks tankers are operated via franchise ownership. Notwithstanding, his lack of apology ignores the appropriate moral remedy offered by an act of apology even for blameless harms.

Implications: Theoretical and Managerial Perspectives

We found that the sociocultural perspective of the image and reputational concerns drives an organization to deploy discourse in such a manner that shapes public perception of its reputation when involved in incidents that violate public safety. In such instance, however, the priority should be on safety (Coombs et al., 2010; Sturges, 1994) with risk communication focusing on public safety. Image repair communication should come afterwards. On the contrary, we found NNPC paying greater attention to image repair and ignoring the need for effective risk communication during incident disclosure. Table 3 below shows the discursive incident disclosure information assessment framework developed from the case organization as a model for understanding the interface between the deployment of image repair practices in the face of reputational tension and the way conversational maxims are violated.

Table 3. Incident disclosure information assessment framework

The evaluation of incident disclosures through the twin lens of IRT and IMT provides a micro-level assessment of image repair discourse that looked beyond what the practices are (e.g., Benoit, 1994; Dionisopolous & Vibbert, 1988; Harlow, Brantley, & Harlow, 2011) to understand how conversational maxims are violated in support of image repair. Consequently, we observed the predisposition and link between the violations of certain conversational maxims to deploy image repair practices as an outcome. For instance, trivialization achieved by reducing the quality maxim of information provides rhetoric that the incident is not that bad, hence supports minimization image repair practice. Similarly, deflection by rationalization supports transcendence, whilst deflection by distantiation achieved through the violation of relevance and manner maxims supports bolstering. Notably, also, the violation of certain maxims and the permutations of such violation can support or allow the deployment of a single or multiple image repair discourse.

Table 3 also illustrates the safety and risk management implication of such practices. For instance, deflection as it relates to organizational safety and risk governance can impact on the organization's ability to accept responsibility for its poor safety performance and, therefore, unable to see faults in its safety systems and procedures. This practice continuously downplays the severity of risk associated with incidents and obscures the safety implications which could, otherwise, be a pivot for scrutiny and enhancement of the sociotechnical operating systems. Hence, deflecting incident disclosure will avert criticism and scrutiny of the organization's audience. Furthermore, deflection may also lead to a false sense of safety and risk management capabilities. Similarly, the unavoidable incident syndrome limits the organization's ability to link failures of technical systems to failures in sociotechnical functions.

Indeed, prioritizing the image of the organization through deflection during incident disclosure falsely influences reception and acceptance of risk information by potential risk receptors and objects at risk. This may have further safety implication on emergency response especially if the seriousness and potential consequence of the incident is downplayed such that it further exposes people to greater danger, hence diminishing social trust. A typical example is the pipeline explosion at Jesse community earlier reported where the NNPC adopted an aggressive image repair practice that downplayed the severity of the secondary risk associated with the incident and further threatened victims with prosecution. Consequently, many injured victims died whilst evading hospitals amidst the fear of being arrested. This reportedly contributed to the demise of over 1,500 people (Emeseh, 2006). This practice can be seen to be at conflict with Kasperson (2014) call for enhancing social trust in safety and risk communication by organizations. Organizations disclosing safety critical-incidents need to be open to better deliberative fair information dissemination process in order to produce acceptable outcomes locally and internationally.

We did not find data pertaining to NNPC's communication to an international audience and offering possible explanations owing to the relative lack of international media coverage of safety incidents in Nigeria when compared to for example, the BP Deep Water Horizon incident. This could be due to the fact that NNPC is a NOC rather than an International Oil Company (IOC), hence, its operations are local and it is not really subject to much international scrutiny. Consequently, it focuses on its national stakeholders and local media. This, really, is partly a failing of the international oil industry and press/media as well as NNPC IOC partners who do not apply pressure on NNPC either because they are also complicit in such safety failings or do not want to ‘rock the boat'. Improving such communication will require a close attention to be paid to the values of the people (Wardman, 2008) achieved through adequate and reliable information dissemination and engagement process (Arvai & Rivers, 2014; Kasperson, 2014; Lidskog, 2008; McComas, 2014) locally and internationally. The social performance and trust of an organization have been conceptualized in terms of its demonstrative responsible behaviour toward community, environment, and employees (Ghosh, 2017). This should not be eroded by manipulation of incident information even in the face of reputational and regulatory tension.

Conclusion

We set out to develop an understanding of how the effective communication of safety-critical incidents takes place in the face of organizational reputational tension. We focused on the case of NNPC and found the use of image and reputational repair strategies whilst also violating conversational maxims during incident disclosures that threatens the reputation of the organization. We empirically identified four categories where the violation of conversational maxims was used to deflect attention, portray a sense of capability, evade blame and avoided issuing apologies. We also found incident disclosures by NNPC consistently focusing on reputational concerns against safety concerns.

We argue that the paper makes the following contributions. First, by identifying NNPC's inclination towards reputational concerns during incident disclosure, we provide an insight into the actual industry practices which shows that amidst the tension between reputational concerns versus effective risk communication of safety-critical incidents, organizations tend to focus on their reputation. However, this practice has wider safety and risk management implications which we discussed in our findings. Organizations can, therefore, draw on the wider implications identified as rationale for developing effective incident communication strategies that pay greater attention to the safety of risk receptors.

Second, our assessment of incident disclosures through the twin lens of IRT and IMT provides a micro-level analysis of image repair discourse that looked beyond what the practices to understanding how conversational maxims are violated to support the deployment of image repair practices. Consequently, we provided a theoretical point of integration between the two theories in our discursive incident disclosure information assessment framework. However, the limitation of the study should be observed. The framework presented, is context specific to the disclosure of incidents by the NNPC, which is, indeed, a very unique but important case. Notwithstanding, the framework provides a nuanced theoretical model which scholars can use to support the assessment of the veracity, manner, quality, and quantity of information released by an organization whilst disclosing incidents. It also provides a means of understanding the safety and risk management implication of such practices to the organization.