1. Introduction
The Sampoong Department Store (Seoul, Republic of Korea) collapsed on 29 June 1995, killing 502 and injuring 937 people. Lee Joon, the store owner, modified the building during its construction by adding a fifth floor when it had only been designed to support four. The store collapsed due to a structural failure—more than 1500 people were in the luxury department store when it collapsed. The collapse was primarily blamed on shoddy (poor-quality) construction and corruption. Lee Joon was charged and found guilty of criminal negligence and sentenced to ten and a half years in prison. His son was arrested, found guilty, and received a seven-year prison sentence. The defendants’ relatives wept softly as the verdicts were read. Throughout the trial, prosecutors painted a chilling picture of a store owner more concerned with maximizing profits than customer safety and of city officials willing to take bribes in exchange for allowing illegal design and construction. "Therefore, they are responsible for the collapse" said the head of a three-judge panel before he read the sentences [1].
The Sampoong Department Store disaster should never have happened, but ill-considered decisions, turning a blind eye to poor quality and corruption provided the conditions for its collapse. Despite the sheer number of buildings, bridges, and dams, and the like collapsing since the Fidenae amphitheatre in 27AD (Fidenae, Italy), engineering failures are an ever-present reality causing significant economic and societal harm.
Examples of well-known calamities that similarly should never have happened include the Westgate Bridge in 1970 (Melbourne, Australia), the Rana Plaza in 2013 (Dhaka, Bangladesh), and the Siji Kaiyuan Hotel in 2021 (Suzhou, China). Harsh lessons can be learned from such events, as engineers can uncover, document, and change their designs to enable improvements and technological innovations to emerge and be adopted [2], [3]. Reinforcing this point, Petroski [3] cogently explains that failures “always teach us more than success about the design of things. And thus, the failures often lead to redesigns—to new, improved things” (p. 63). Repeatedly, engineering errors, poor-quality construction, miscreant behavior, and violations of standards and regulationsare typical contributors to disasters [4]. Regardless of the countless number of studies addressing engineering failures in construction, we still struggle to mitigate them, as we tend to overlook the conditions (i.e., so-called ‘pathogenic’ influences) that result in their occurrence [4], [5], [6].
Setting aside corruption† [7] (including bribery, extortion, fraud, and cartels), as it is secretive and difficult to detect [8], [9], we have been unable to make headway toward questioning the engineering design decisions and mitigating people’s (in)actions during construction that contribute to the occurrence of engineering failures. When errors and violations (also referred to as active failures) are identified, rework‡ [10], [11], [12], [13] may be required, negatively impacting an organization’s profitability and reputation, as well as a project’s productivity, safety, and environmental performance [10]. Thus, if we can mitigate errors and violations during the construction of an infrastructure asset, rework can be reduced and inroads can be made to prevent engineering failures [6].
Calls for greater investment in infrastructure, for example, have been made in the United States as a consequence of the Pittsburgh Bridge collapse in late January 2022, which is perceived to have occurred due to deferred maintenance [14]. In response, clarion calls have been made to throw money at the problem, although this is not the immediate solution. Indeed, there is a need for funding to upgrade and maintain bridges in the United States, but such assets need to be designed and constructed with ‘error resilience’ in mind. However, this has not been the case for infrastructure assets worldwide, which we bring to the fore in this paper.
The need to perform rework is a pervasive problem in construction [6]. Accordingly, an extensive body of work has been undertaken to determine the costs and causes of rework and to propose strategies to prevent its occurrence [11], [12], [13], [15], [16], [17], [18], [19], [20], [21], [22], [23], [24]. Such studies have focused on determining proximal (i.e., singular) rework causes, using prefixes such as ‘poor,’ ‘lack of,’ ‘inappropriate,’ and ‘inadequate,’ and thus have neglected the interdependency of events leading to its occurrence. Moreover, there has been a tendency to view errors (e.g., lack of skills or knowledge) as causes rather than consequences of systemic factors [5].
The perpetual reporting of singular variables (e.g., poor communication, lack of coordination, and improper material handling) that have an absence of a context has led to the creation of artificial narratives of rework causation that undermine the complexity of the problem, which has been described as being ‘wicked’ [25]. In sum, rework studies that have focused on identifying proximal causes provide an over-simplification of causality.
In addition, some studies have focused on identifying single (or a few) root causes [10], [11], [12], [13], [14], [15], [16], [18], [19], [22], [23], [24], [26]. This approach promotes a reductionist view of causation, which we consider to be flawed, as multiple interacting contributions are often at play [25]. Consequently, this view has hindered scholars’ ability to understand the context and conditions that lead to the manifestation of rework and their means to reduce its incidence in construction [27], [28], [29], [30]. That is, the ‘reductionist’ perspective, which has traditionally been adopted to determine rework causation, relies on the use of ‘one-size-fits-all’ prevention strategies[15], [17], [20], [21], [23], [24]. However, various error types evoke different responses, suggesting that strategies to address rework need to be tailor-made to the context in which they occur [6], [7], [27], [28], [29].
While we know that rework in construction can adversely impact project performance, we still have limited knowledge about its initial conditions, its consequences, and how best to mitigate its occurrence, despite the considerable amount of research undertaken. Hence, the motivation of this paper is to shift the focus away from determining proximal and root causes, as seen in Asadi et al. [23], which reinforces a repeated discourse resembling ‘new wine in old wineskins.’ Thus, we require a new line of thinking in which the context and initial conditions matter [25], [26].
Without an understanding of the context, it is not possible to develop solutions to mitigate rework [30]. Thus, we draw upon our previous empirical research [27], [28], [29], [30] to provide a context to the rework problem and, in doing so, move the prevailing discourse forward from a position where people are viewed as being the cause (e.g., loss of situation awareness, procedural violation, and managerial deficiencies) to one where it is seen as a “symptom of trouble deeper inside the [organizational and project] system” [31, p. xii].
Our paper commences by introducing a nascent theoretical context to understand the etiology of rework that materializes during construction (Section 2). We then frame our paper around three fundamental questions to support the need for a new theoretical framing of rework. We first question why rework occurs, drawing on the error literature and our empirical research (Section 3). Then, we ask what the consequences of rework are (Section 4) and how it can be mitigated, drawing from best practices that we have observed in real-life projects (Section 5). Next, we identify the research (Section 6) and practical implications of our review (Section 7), before concluding the paper (Section 8). The contribution of our review to the contemporary error and rework literature is twofold: ①We present a theoretical context for rework causation and the role errors play in its manifestation; and ② build on the error-mastery culture theoretic proposed by Love and Matthews [29] by demonstrating how resilience (i.e., foresight, coping, and recovery) to errors can be incorporated into everyday practice in construction.
2. Theoretical context
The type of error culture within an organization and a project sets the tone for how people respond, share information, and deal with errors and their consequences [27], [28], [29]. An error-prevention culture dominates practice in construction. Table 1 presents the characteristics of such a culture [6], [25], [27], [32], [33], [34], [35].
Table 1. Processes and outcomes of error prevention (negative view of errors).
| Before an error | After an error | Interpersonal processes | Outcome |
|---|---|---|---|
|
|
|
|
Adapted from Ref. [32, p. 666].
Unwittingly, construction organizations have found such an error culture to be an Achilles’ heel: It has hampered their ability to learn and mitigate rework, as errors are viewed negatively and are often covered up [25], [27], [28], [29], [30], [36]. Having an error-prevention culture in place often results in rework becoming ‘uncomfortable knowledge’ (i.e., denied, dismissed, diverted, or displaced) or being explained away as a one-off event [34]. Nonetheless, errors enable organizations to learn and innovate, and thus should not be viewed in a negative light [2], [32], [37], [38]. Only a limited number of studies have examined how errors and violations result in the need for rework in the construction and engineering management literature [27], [28], [29], [30], [36], [39], [40]. Thus, we will briefly explain the nature of errors and violations, as they are a source of rework. We present a rework nomenclature in Fig. 1 [41]and identify real-life examples of errors observed in our studies that have resulted in the issue of non-conformances.
Fig 1. Rework nomenclature. (a) Rework: errors and violations; (b) examples of non-conformances requiring rework. (a) Adapted from Ref. [41, p. 207].To reiterate, our definition of rework does not consider change orders, as these form part of a construction organization’s planned work when issued by a client [10]. More specifically, rework is an unplanned activity and is seldom identified as a risk; rather, it is often viewed as a zemblanity (i.e., an unpleasant yet unsurprising discovery) [25].
2.1. Action errors
We adopt Frese and Keith’s [32] notion of goal-directed actions to frame our definition of an error. Thus, action errors are defined as “unintended deviations from plans, goals, or adequate feedback processing, as well as an incorrect action that results from a lack of knowledge” [37, p. 1229]. When examining action errors and their consequences, the context within which they occur matters, as the environment within which people work influences their occurrence [38].
Making errors per se is not a problem. More often than not, in construction, errors are minor, as they are an unintended consequence of work activity [2, p. 256; 33]. Moreover, people “as part of their daily work activity commit errors routinely” [2, p. 256]. But, in some instances, errors can have serious consequences and thus need to be quickly identified before they contribute to a disaster. Most of the time, practitioners (e.g., design and project engineers, site supervisors, and subcontractors) discover their errors (and others) due to an array of procedures and systems (e.g., design audits and checks, Inspection Test Plans, and Last Planner®) that are put in place when designing and constructing an asset. However, there are occasions when errors remain unidentified due to constraints (e.g., production pressure) and the dynamic environment within which people work, which can lead to grave consequences [34].
Errors can occur due to impaired human cognition (e.g., slips and lapses of attention) and mistakes (i.e., rule- or knowledge-based) [41]. In the case of rule-based mistakes, a practitioner may misapply a rule that worked in a previous situation (e.g., using a different design) due to a changed condition. Relatedly, an imperfect rule may have remained uncorrected and formed part of a practitioner’s problem-solving toolbox [41]. Similarly, knowledge-based mistakes emerge when practitioners encounter a novel situation outside the range of their learned problem-solving routines [41].
At an individual level, several issues influence our ability to make errors, including fatigue (e.g., workload, time of day, and sleep deprivation), stress (e.g., workload and time constraints), boredom (e.g., repetitive tasks), and inadequate training and/or limited experience. Team and organizational errors are also common contributors to rework. Accordingly, team errors can “occur as a result of the joint effect of antecedents across individual and team levels” [42, p. 1322]. Several scenarios can result in team errors occurring in projects; these include cases in which [29], [43]:
-
•
The entire project team does not detect an error and work continues;
-
•
An individual commits an error that goes undetected, the team jointly decides on a course of action, unaware of the error;
-
•
An individual error is detected, but the team decides not to correct it and continues the work.
Organizational errors are defined as the “actions of multiple organizational participants that deviate from organizationally specified rules and procedures that can potentially result in adverse organizational outcomes” (e.g., accidents, litigation, and reputational loss), especially in high-stakes settings such as construction [44, p. 154]. Hence, a rudimentary “feature of an organizational error is that multiple individuals deviate from the expected organizational practice” [44, p. 154]. In our previous studies, a typical example of an organizational error was the non-reporting of non-conformances during construction, as there is a perception that senior management views these as indicators of a poorly managed project [29], [30], [33], [34]. As a matter of fact, non-conformances provide learning and improvement opportunities for construction organizations. But, as we will discuss below, such opportunities are often forgone due to an organization’s incumbent culture [28], [29], [36].
2.2. Violations: Rule breaking
In contrast to action errors, violations are the intentional breaking of rules and procedures that have been established to restrict self-interested behavior and protect “organizational members from the predations of others” [45, p. 36]. Moreover, violations may arise when there is a non-conformance to a standard, such as substituting a specified product/material for another and installing it without approval [34]. In this instance, we have observed that the motivation for such an action is to maximize profit (i.e., the substituted product/material is cheaper) or to adhere to a project’s program so as not to cause a delay [34].
The breaking of formal rules is typically associated with deviant behavior, especially within the context of safety in construction. However, “there is a longstanding antithesis to this view that rules are in some simple sense order-producing and violation order-destroying. Part of this critique is that rules harm individual well being” [45, p.37]. In doing so, they can adversely impact job satisfaction, contributing to stress and absence, undermining organizational functioning, and impeding the facilitation of organizational change and learning [45].
With construction subjected to many rules, people need to consider their applicability to specific situations [46]. Rather than rule-breaking being viewed as a deviant behavior, it may be deemed to be “pro-social” and “a way of testing rules and the truces around them” [45, p. 36; 47]. Thus, in the case of a violation, the context and intentionality behind the person’s (in)actions must be considered. People may intend to do the right thing but find themselves breaking a rule and vice versa [48]. It is notable that rules and procedures are often “written for the ideal situation,” yet in construction, “work situations are rarely ideal” and are subject to constant change [48, p. 298].
If there was no intention to commit a violation, then the act can be categorized as an “unintended violation” [41, p. 195]. If there was a prior intention to cause damage to the system, then the violation is deemed to be “sabotage” [41, p. 35]. However, intentions are not always so black and white. Certain violations may “have some degree of intentionality, but do not involve the goal of system damage” [41, p. 195]. In such cases, violations can be categorized (Fig. 1) as either routine—that is, “habitual, forming a part of an individual’s behavioral repertoire” [41, p. 195]—or exceptional—that is, “singular violations occurring in a particular set of circumstances” [41, p. 196].
Production pressure, the unavailability of skilled labor, pandemics (e.g., coronavirus disease 2019 (COVID-19)), incomplete design, and the like mean that rules may become problematic or may render it impossible for people to perform their work. In situations when rules are inappropriate, “alternative courses of action tend to be used to achieve the same ends” [48, p. 298]. As a result, this provides people who break the rules with the opportunity to concoct reasons for their actions. In doing so, people are viewed not as rational but as reasoning agents who create and convey their reasons through dialogue to others [49]. Thus, when examining deviant behavior, we need to consider the context within which it has occurred. As Pablo Picasso insightfully remarked, “learn the rules like a pro so that you can break them like an artist.” Indeed, rules can be changed by challenging them, which can result in positive outcomes. Thus, we need to take heed of this point when we consider the issues associated with rework mitigation (Section 5).
3. Why does rework occur?
To recap, the construction and engineering management literature has ignored why errors and violations transpire. Therefore, understanding the nature of error-making and rule-breaking provides the impetus to address the rework problem, which equally applies to accidents [26], [50]. In actual fact, a symbiotic relationship exists between quality and safety [51]. Thus, akin to the accident causation literature [41], where resident ‘pathogens’ (i.e., latent conditions) within a system engender error-making and violations, they can result in the manifestation of rework [52], [53]. Such pathogens arise from the strategic decisions made by a construction organization’s senior management and project clients [53].
Thus, pathogens tend to lay dormant, often for a considerable period, with people being unaware of their existence “within a system until an error [violation] comes to light” [53, p. 425]. As pathogens enter an incubation period, they become an integral part of everyday work practices [52], [53]. When pathogens combine with active failures, then rework is often needed. Indeed, active failures are difficult—if not impossible—to foretell. In effect, errors and violations are seen in hindsight; until then, they are actions just like any other. However, pathogens can be identified and remedied before rework is needed [53].
In Table 2, Busby and Hughes [52] identify eight types of pathogens that emerged from a study of errors in large-scale engineering projects. Examples of pathogens identified by Busby and Hughes [52] and from the rework studies undertaken by Love et al. [35], [53] are also presented in Table 2. Pathogens do not exist in isolation and can interact with one another. Therefore, being mindful of their interdependency improves the ability of organizations to redress resident pathogens holistically [36].
Table 2. Description of pathogens and examples of errors.
| Category | Pathogen resource [52, p. 429] | Busby and Hughes [52, p. 429] | Love et al. [53, p. 429] and Love et al. [35] |
|---|---|---|---|
| Practice | Peoples’ deliberate practices | It was the practice for designs to be checked only for internal consistency, not consistency with external constraints and requirements | Failure to undertake design reviews and the distribution of tentative design documents to contractors |
| Task | The nature of the task being performed | Trace quantities of a contaminant had disproportionate consequences in a particular process design task | Engineers failed to detect and correct omissions in design documentation; furthermore, schedule pressure resulted in disproportionate time being allocated to tasks |
| Circumstance | Situation or environment the project was operating in | The firm procured services in a market where there was inadequate information about the quality of products | Low design fees meant that tasks were deliberately left out; schedule pressure resulted in some tasks not being recalled at the appropriate time |
| Convention | Conventions, standards, routines, and codes of practice | A person adhered to a company standard that had previously always been superseded by ad hocagreements—which, as a result, had unknowingly become obsolescent | Re-using existing specifications and design solutions; also, failure to adhere to new company policies |
| Organization | Organizational structure or operation | The slow ramp-up of projects led to delay in early tasks on which many others were dependent for information and which therefore had to proceed on tentative assumptions | Blocking of communicative action due to an error-prevention culture and absence of psychological safety |
| System | An organizational system | Latency in a change control system meant that a significant amount of engineering | A trade-off between quality and safety; when a trade-off arises, having more of one element means less of the other; safety is given preference, as it is bound by legislation |
| Industry | Some structural properties of the industry | Public contracting regulations required that the firm consider vendors with whom the firm had no direct experience | Ever-present symbolic representation of a zero-vision and the notion that errors can be eliminated by striving for a zero-error culture |
| Tool | A characteristic of a technical tool | A design tool provided a layering facility that encouraged people to simplify their tasks but allowed them to forget possible inconsistencies with other parts of the design | Interoperability with computer-aided-design software applications (i.e., no checking for consistencies); simplification of tasks and neglect of other aspects of design |
The practice pathogen has been identified as the most popular in projects [52], [53]. For example, the practice of design re-use is often used to improve productivity and drive down costs in projects [52], [53]. But this practice is “inherently vulnerable to unidentified differences between the context in which a re-used design first originates and the context in which it is re-used” [52, p. 431]. We have also seen design engineers failing to undertake detailed design reviews due to production pressure or minimizing costs [53]. The pathogen of circumstance is similarly ubiquitous in projects [53]. In this instance, two issues come to the fore [53], [54], [55]:
(1) The use of fast-tracking (i.e., overlapping design and construction activities) often results in commencing construction based on a tentative design.
(2) Traditional contracting (i.e., design–bid–construct), in which information asymmetry, adverse selection, opportunistic behavior, and moral hazards materialize, exists during the procurement process. The problems with traditional contracting are exacerbated when competitive tendering is enacted and the lowest bid is selected.
Besides the empirical studies of Busby and Hughes [52] and Love et al. [53], research examining the nature of pathogens and their associated incubation periods has not been forthcoming. Thus, further exploration is needed to garner an improved understanding and awareness of the implications of pathogens for decision-making and practice.
3.1. Functional stupidity management
Adding to the mix of pathogens and active failures, we have seen ‘functional stupidity’ at play in construction organizations, indirectly contributing to people making errors and committing violations [35]. Alvesson and Spicer [56]frame functional stupidity as an organizational issue and describe it as the incapacity and/or disinclination on the part of organizational members to exercise critical reflection about what they are doing (reflexivity), understand why they are doing it (justification), and determine what the consequences of their activities are beyond the immediate task at hand (substantive reasoning).
Akin Alvesson and Spicer [56], Love et al. [35] have observed that functional stupidity is linked to power and politics in construction organizations. In particular, Love et al. [35] observed that managers, working within their error-prevention culture, tried to shape the cognitive capacities and mindsets of their employees using symbolic manipulation to create “conformity and to limit critical thinking” [56, p. 1204].
3.2. Competing demands
Even though quality and safety are interdependent, construction organizations view them to be competing demands and have been unable to accommodate them equally, resulting in trade-offs occurring [8], [27], [33], [35], [50], [51]. When a trade-off arises, there is a gradual exchange in which having more of one element means less of the other. In this instance, it has been observed that construction organizations typically provide more resourcing to safety, as it is bound by legislation, with the consequences of not adhering to regulations and the code of practice being potentially costly and threatening to their competitive advantage and reputation [35]. To this end, quality is treated as subordinate to safety [34], [50].