1 Introduction

Expert groups occupy an important role in advising policy makers, producing guidance of all kinds, and writing standards—all things that impact modern life. The context to this thirst for knowledge is a desire on the part of decision makers, and a social expectation, to base decisions on the most robust information in the face of uncertainty (Kropp and Wagner 2010). The mechanisms for seeking expert advice vary by domain. Within European Union systems of governance, expert committees are ubiquitous (Gornitzka and Sverdrup 2008; Kohler-Koch and Rittberger 2006). Committees also direct the conversation within bodies such as the United Nations and World Health Organization.

Standards committees are another case in point, where groups of technical specialists collaborate to develop advice for those professionals involved in the design, construction, and operation of artifacts big and small. In the case of engineering in the energy sector, the importance of the role of these expert groups was highlighted by the Deepwater Horizon blowout in the Gulf of Mexico in 2010, where the formal inquiry into the disaster found that commonly used industry-based engineering standards “failed to reflect ‘best industry practices’ and have instead expressed the ‘lowest common denominator’—in other words, a standard that almost all operators could readily achieve.” This shortfall “undermined the entire federal regulatory system” (National Commission on the BP Deepwater Horizon Oil Spill and Offshore Drilling 2011, p. 225).

The use of expert groups raises questions about their constitution and operation, and how this impacts the advice such groups generate. To date, focus has been concentrated on the effect of expert groups on democracy (Joerges and Neyer 1997; Rhinard 2002), the challenges in producing “useable knowledge” in policy contexts (Kropp and Wagner 2010, p. 831), and techniques for elicitation of expert views around a set question (Linstone and Turoff 1975). Taking the case of the UN Food Systems Summit, critics raised concerns about the composition of the summit’s “top table,” noting that the group included experts in some aspects of nutrition and not others, serving to “skew” the advice that the group could produce (Nisbett et al. 2021). For knowledge workers such as practicing engineers, whose expertise forms the standards that they are working from is of great concern (Sandholtz 2012, p. 657).

The constitution, and function, including the leadership, of groups involved in the development of “how to” practice guides like standards has been subject to less inquiry. Outstanding questions include: Who needs to be on standards committees and subcommittees to get the best outcomes? How are decisions taken by such groups of experts? How is group leadership managed on standard committees? What are the challenges in getting a group of specialists with different varieties of expertise to communicate as part of the standard development process?

Situated in the literatures on expertise, elicitation of expert judgments in small groups, and the involvement of technical specialists in standard formation, this paper examines the tactics and negotiations involved in writing expertise into an engineering standard. In light of calls for understanding of standards through the concrete settings in which they are developed and applied (Timmermans and Epstein 2010), we draw on interview and observational data collected during a major update to one part of Australian Standard 2885 Pipelines—Gas and liquid petroleum. We address the constitution of the subcommittee, the qualities and role of a subcommittee chair, and the nature of the negotiations including the public consultation. We argue that the successful operation of the group relies on a Chair with three types of expertise—contributory, interactional, and decision-making—which together are critical to resolving debates and, in effect, agreeing on the collective wisdom of the group.

2 Eliciting expertise and the standards process

Much of what we know about expert-led decision-making comes from the decision sciences, which has focused on the case of eliciting expert judgements to guide policy makers, as opposed to standard formation. One of the major themes in this research is the value of a diversity of views. Groups that include people from diverse backgrounds have greater experience to draw on which has been shown to be beneficial in achieving the best results. The maximum gain per person added to the group occurs at around five people, although the effect continues as group size increases (Kosinski et al. 2012). With this in mind, science researchers have taken to publishing complex and intractable problems online and inviting contributions to development of a solution (Cranshaw and Kittur 2011; Khatib et al. 2011).

In bringing such a diversity of people into a group, decision scientists maintain that there also need to be “structures” in place to guide the discussion. A key focus is thus elicitation techniques, i.e., how to get the best judgement from a group of experts. The best-known structured technique for eliciting experts’ views and developing a consensus is the Delphi method, developed by the RAND corporation in the early 1950s (Linstone and Turoff 1975). Over more than half a century, Delphi techniques have been used widely for developing factors in decision support systems (e.g., Hassan et al. 2015), in developing input data for computer modeling purposes when no real data are available (Cadham et al. 2022; Westerhof et al. 2022) and in scenario development for both policy purposes (e.g., Nowack et al. 2011) and corporate strategic planning (e.g., Förster and von der Gracht 2014).

The heart of the Delphi technique involves putting a specific question (or set of questions) to a carefully constituted group of experts in several rounds. There is no expert leader of the group, only a facilitator of the process appointed by those seeking a solution (Linstone and Turoff 2011). Experts make an initial estimate individually and are then given feedback in the form of a summary of the group responses. Individuals are then given an opportunity to revise their initial estimate. Additional rounds are sometimes used, although Linstone and Turoff (2011) emphasize that their original intention was not to produce consensus, but rather stability in the range of responses obtained. Some forms of Delphi put significant weight on not linking interim judgements to specific experts when they are shared. In such cases, the expert group never meets face to face. Other forms of the Delphi method involve open sharing of judgements and group discussion, once initial estimates are developed individually. Some methods include specific training for experts on the (statistical) methods used to express the answer to the subject question and associated uncertainty (Meyer and Booker 2001).

More recently, multicriteria decision aid (MCDA) techniques have been developed for complex decisions where multiple defined options are identified and the role of decision-making is to choose between them (Delesposte et al. 2021; Merad and Trump 2020). While this method of arriving at an optimal choice has been applied in many domains, the technique only applies to problems of this nature, requiring experts to define and weight criteria and then rank alternative solutions based on these criteria (Mirzaee et al. 2019). Another recently developed method involves defining one or more “synthetic” models of the phenomenon being studied and seeking expert views on their viability as with McCarty et al’s (2021) study of low-energy retrofitting of domestic housing.

The need to forecast something specific or choose the preferred option from a pre-defined set is somewhat different to the case of standard-writing (Rowe and Wright 2011). In this case, we have groups of experts, with a nominated expert leader, writing for other experts in the same field, for a purpose that is only broadly defined—a “how to” question such as how best to engineer a particular technology. Standards are increasingly being produced under the auspices of standard development organizations (SDOs) such as the International Organization for Standardization (ISO) (Higgins and Tamm Hallström, 2007; Jain 2012). Early standards committees run by professional societies were typically composed of engineers who “saw themselves precisely as the translators of scientific knowledge into industrial effectiveness” (Higgins and Tamm Hallström, 2007, pg 688). More recent studies show that committees can be prone to capture “with large firms volunteering resources that influence the outcome, insiders manipulating procedural rules and players lobbying for their economic interests” (Jain 2012, p. 1644).

Despite these potential problems, given the largely technical nature of standards, standard committees are still dominated by engineers who function as “rationalizing agents who lend legitimacy and authority to the content of a standard” (Brunnson et al. 2012, p. 619; see also Hayes et al. 2022). The committee meetings in which groups of engineers meet repeatedly to thrash out the content of a given standard become “the principal arena in which socio-technical orders are negotiated” (Dokko et al. 2012, p. 685).

This way of working has both strengths and weaknesses when it comes to producing a robust and legitimate outcome. Such frequent interactions within an expert group create trustful relationships and reinforce professional values (Jain 2012) but “while social capital enhances coordination within the process of standard-setting, it can also undermine the role of expert knowledge if standards are shaped on the basis of personal favours and friendships” (Brunnson et al. 2012, p. 619). Repeated meetings also act as “a venue for the involved parties to educate each other about emerging technologies as well as settle disputes that may arise” (Jain 2012, p. 1643). These processes are often protracted with negotiations taking place over years with progress being distinctly non-linear (Dokko et al. 2012; Jain 2012).

One rare exception to this style of standards writing is the recent work on an Italian standard on animal welfare for indoor reared beef cattle. The process involved expert elicitation with 14 veterinarians following a modified Delphi technique over three rounds to characterize hazards and promoters of animal welfare (Lorenzi et al. 2023). The expert participants were chosen via a public application process. The initial list of animal welfare issues put to the individual experts was established based on a literature review. It is reported that the output of the exercise has provided a “starting point for the creation of the first national public protocol for the on-farm welfare assessment of indoor reared beef cattle” (Lorenzi et al. 2023, pg 156) but this protocol does not appear to have been produced as yet. This exercise had no formal expert leader, only a group with expertise in expert elicitation to design, coordinate, and manage the overall process.

Given the role of a leader in typical standards committees, when compared to the setting in which the Delphi and MCDA methods are implemented, we also need to understand the traits that may be needed to facilitate expert peer groups. Expertise comes in different varieties. This is important in the context of understanding the constitution of expert groups because it speaks to the skills and knowledge that prospective group members may or may not possess. Collins and Evans’ schema of specialist expertise recognizes both “contributory” and “interactional” forms (Collins and Evans 2007).Footnote 1 Contributory expertise refers to the possession of embodied competence in a given activity. This is the form of expertise that we most readily recognize, and in our case refers to the technical skills involved in doing an aspect of engineering work. Interactional expertise refers to fluency in the language of a domain (or multiple domains) of specialist knowledge, without necessarily having the embodied competence. These two forms of expertise are not necessarily exclusive. It is possible for a person to possess both forms of knowledge. As we will see, this combination greatly aids in the leadership of small groups of experts.

The leadership literature argues that the skills of a good leader go beyond specialist expertise with one study arguing that leaders of small expert groups “should be experienced members, who have the skills to facilitate a discussion, can manage conflict effectively and can be impartial” (Walker and McLeer 2004, pg 282). While acknowledging that such decision-making expertise is critical, the leadership literature has little to contribute more broadly on the topic of specialist expertise and leadership in a small group setting. The leadership literature typically assumes that leadership qualities are inherent in individuals regardless of the specific circumstances with a large-scale review finding only three twentieth-century studies that explored “how the situation itself may influence who emerges as leader” noting that such research assumes that leaders emerge “in part because the capabilities or traits of the individual are well suited to the task at hand” (Badura et al. 2022, pg 2086).

Despite the research interest in expert groups and acknowledgement that processes of standardization are increasingly important, studies that combine these two domains to study what goes on within standards committees are rare. The leadership literature also has little to say in this context and so this study aims to contribute to that small but important area.

3 The Australian case study

The Australian high-pressure gas transmission pipeline sector in Australia has an excellent safety record. No member of the public has ever been killed or injured by a failure of a transmission pipeline and pipeline failure rates are around 20 percent of those seen in the US and Europe (Tuft and Bonar 2009; Tuft and Cunha 2013). Having said that, several significant failures have occurred in remote areas that have resulted in interruptions to supply. Forty thousand people in regional South Australia were without gas for several days following one pipeline failure (Fedorowytsch 2015). A major pipeline failure in remote Western Australia in 2008 resulted in losses to the economy of approximately $3 billion (Bills and Agostini 2009). This record still stands up well compared to international data which includes the San Bruno pipeline failure in California in 2010 which killed eight people and destroyed 38 homes (Hayes and Hopkins 2014) and the pipeline failure at Ghislenghien in Belgium in 2004 which killed 24 people and led to an additional 150 hospitalizations (Mahgerefteh and Atti 2006).

Pipeline safety and reliability is regulated in Australia at a state level. Each state has a unique set of regulations for this purpose all of which are goal-based in structure (Engen and Lindoe 2019). When it comes to pipeline integrity and public safety, all states require pipeline operating companies to take a risk-based approach to management of public safety and security of gas supply. A key risk control in Australian hydrocarbon pipeline engineering is the AS 2885 series of standards which sets a mandatory norm in that all state regulations require compliance.

The Standard sets specific requirements for engineering new pipelines based on the identification of possible threats to pipeline integrity. The overall requirement is to reduce risk from the pipeline to a level that is as low as reasonably practicable (ALARP). In seven parts, AS 2885 sets out a series of measures and methods to ensure the safety of the public and workers, the protection of the environment, and security of supply including:

  • Detailing how some specific technical engineering tasks should be done;

  • setting key engineering parameters such as design factors;

  • describing how critical engineering work should be organized;

  • detailing requirements for how risk is to be assessed and reduced to ALARP;

  • detailing who is responsible for key documents and processes; and

  • describing how key facts and decisions are to be recorded.

The Standard integrates both top down and bottom up approaches to risk management (Linkov et al. 2014). The first two dot points in particular cover approximately a dozen areas of pipeline engineering including corrosion, materials selection, metallurgy, stress analysis, and welding. Standard arrangements and methods are set out in each area with criteria provided to help those without specialist skills understand when a specific expert should be consulted. The role of those on the standards committee is to specify those standard methods and criteria.

The standards are developed and issued under the auspices of Standards Australia (SA), Australia’s primary standards development body. SA is a private, not for profit organization, a member of ISO, and currently maintains a suite of approximately 7000 standards. The work of standards development is largely undertaken via specialist volunteer committees which operate under SA rules and structures. In the case of AS 2885, there are two relevant committees. The standing ME-038 committee is made up of representatives of nominating organizations, i.e., those who have a stake in the content of the standard as chosen by SA.

Prior to the commencement of the major update of Part 1 of the Standard in 2013, the ME-038 committee formally approached a senior member of the pipeline engineering profession and invited him to take on the volunteer role as Chair of the ME-038-01 subcommittee. The remainder of the ME-038-01 subcommittee was then appointed by the Chair as described in the Findings section below.

The research examines the work of that committee undertaking a major revision to AS 2885 Part 1 between 2013 and 2018. The main changes to Part 1 in this update were driven by industry research outputs in two areas—pipeline fracture control and risk-based decision-making criteria. An invitation was also made through the industry association for comments and suggestions to be submitted regarding changes to Part 1 of the Standard for consideration by the committee. One hundred and forty comments were received from the wider pipeline engineering profession. Developing the revised Standard based on these 140 comments, research outputs, and the views of subcommittee members themselves was a complex process. Initially, the Chair grouped the factors to address in the update into a list of over 60 issues. Subcommittee members were asked to self-nominate into small groups (with a nominated lead) to consider all the factors around a given issue and prepare an issue paper regarding if/how the Standard should be changed. These papers were then considered in a series of 16 face-to-face multi-day workshops of the subcommittee. Much drafting and redrafting of potential changes and multiple interactions between the committees led to the production of a new draft Standard. Following a formal public comment period, the proposed new revision of the Standard was ultimately approved by the ME-038 committee and published by Standards Australia in 2018.

4 Methods

In order to investigate how the content of the Standard is determined, the experiences of members of the ME-038 and ME-038-01 committees were investigated immediately following the publishing of the new revision of the Standard. This research included a survey to elicit the views of members of the ME-038 main committee and ME-038-01 subcommittee in relation to the AS2885 series of standards, follow-up semi-structured interviews with a smaller group of subcommittee members, and review of other artifacts linked to the production of the revision to AS 2885 Part 1. In this paper, we focus primarily on the qualitative interview data which most clearly demonstrated the processes of group formation, the dynamics of the decision-making processes, and the role of the chair. The topics covered in the interviews are summarized in Table 1. Human research ethics approval was obtained for this research from the relevant University ethics committees.

Table 1 Areas addressed in data collection
Full size table

A total of fourteen subcommittee members were interviewed. Interviewees were, on average, just over 50 years of age, with approximately 18 years of experience in the pipeline sector. All interviewees identified their jobs as mainly technical or a mixture of technical and managerial duties. Eleven interviews were conducted face to face and typically at the interviewee’s place of employment, with the remaining three interviews conducted via telephone. Interviews were audio-recorded and transcribed by a professional transcription company for analysis with the consent of participants. Male pronouns (he, his) are used in all discussion about interviewees’ comments to avoid potentially linking specific comments to the small number of women involved in the committees.

A key methodological consideration in design of the study is the participatory action nature of the research, in which the first author was also an ME-038-01 subcommittee member. Trust in a fellow committee member is possibly part of the reason for the willingness of committee members to complete the survey and participate in interviews but these dual roles also presented a possible point of confusion for interviewees and bias in the data collected. As a result, the study design was for interviews to be conducted by this researcher and one other. This occurred in all but three cases meaning that this situation was effectively managed.

We also considered these dual roles in the data analysis process. The second author led the data analysis process for this paper, with both authors then collaboratively developing the key themes and findings. In this way, the study design maximizes the benefits on one author’s insider status while effectively mitigating the possible down sides (Berger 2015). In conducting the analysis, we sought to examine key issues linked to the functioning of the expert group/s, specifically the process of selecting committee members, the skills and role of the Chair, the character of committee meetings, and the public consultation process. In the coding process, we looked for evidence of the forms of expertise needed for committee membership, including that of the chair, accounts of the process of forming agreement including the number and skills of individuals involved in specific debates, points and resolution of conflict, and the treatment of contributions beyond the subcommittee.

5 Findings

5.1 The subcommittee Chair

When a subcommittee is constituted by the main committee to work on an update to a standard, typically, the previous subcommittee chairperson has some influence over the selection of the new Chair. In interview, participants reflected that there tended to be an “obvious choice” (Interviewee 11). The Chair needs to be a technical professional with considerable expertise in the required area/s as recognized by their peer group. The candidate/s are then discussed within the main committee who approve and then approach the person in question.

Subcommittee members’ views of the Chair appointed on this occasion were very positive. Survey comments about the Chair reported that he “did a fantastic job” as a result of his “great experience and passion for the standard.” Further he took a “sensible approach” and “did a fantastic job for setting it up, creating the framework, refreshing the [sub]committee.”

Subcommittee members spoke directly about what is required for the Chair to be respected in this way, converging on a senior person with the right mix of technical and non-technical skills. Participant 7 was of the view that such respect comes from experience with both the standards process and the industry in general: “I would have thought there’s an element of you’ve been through the process at least once, you’ve got some runs on the board and you’re recognized in the industry, you’ve got some sort of track record, and people are going to respect you.” Another participant was of the view that ideally the Chair should have a “certain level of seniority, certain level of understanding of industry issues – broader picture type thing” (Interviewee 1). As another interviewee explained, “you’ve got to have someone who’s kind of an elder statesman or they just wouldn’t be accepted” (Interviewee 8). One survey respondent recognized that perceived expertise is as important as actual expertise, noting that this has pros and cons for decision-making overall:

The Chair was very influential due to his expertise and other people’s perception of his expertise in areas that he is not an expert on. The Chair should be that influential as that is a role that requires power to control the room, but the negative is that others may not challenge enough due to his level of influence.

In addition to seniority, it was generally felt that having a broad technical appreciation, as well as “fluency” in the language of the Standard as a whole was critical, rather than holding a narrow technical expertise. Reflecting on his own practice, the Chair commented: “I don’t understand every bit of the Standard, but I can talk the language of every bit. I don’t necessarily have full technical expertise, but I can talk the language of every part of it”. In thinking of potential future leaders, the Chair was considering those subcommittee members with a similar generalist–specialist skill set. He commented on one possible future leader: “he’s probably got both the greatest breadth and depth of technical expertise that I’ve seen around the industry. … Every time you ask him a technical question, he has a very solid answer. … So, yeah, I thought he was to me, a fairly clear choice”.

5.2 Selection of the group

Who is on the subcommittee matters? Standards are the holder of collective industry experience as interpreted by those who literally sit together and write the words. As one interviewee described:

A lot of the changes we make [to the standard] are because of something that didn’t quite go right. But, it’s being on the committee that – you really understand why those words were put there. And, that’s something we have to be careful about, as an industry. Who’s on the standards committee? Because, if you take all the people away who understood why the words were put there, in the first place, it’s easy to have those words disappear, in the swipe of a mouse, these days. They have a really important meaning. (Interviewee 4)

As mentioned above, the members of the subcommittee were chosen at the beginning of the update process by the Chair of the subcommittee. An invitation inviting expressions of interest for committee membership was circulated through the industry association and sent directly to state technical regulatory agencies. One hundred and thirty engineers expressed interest—many more than could be accommodated by the process. The Chair had various considerations in deciding who to include. The first consideration is getting the requisite technical experience, as opposed to employees or representatives of any organization. In a Noah’s Ark like scenario, the Chair was reported to have the view “that you need two of every expert” (Interviewee 6). The reason for seeking multiple specialists in each area was to avoid the Standard being “flavored by one person” (Interviewee 6). The Chair explained: “I needed experts, you know, people who had genuine in-depth expertise in each of the very wide range of areas that are covered by the Standard. And I wanted to have at least, two if not three experts in each of those areas.” He further described his reasoning:

Because one of the, it might have been either the 1987 or the 1997 edition of the Standard, there was only one corrosion expert on the committee. And he had some views that … mainstream people thought were very idiosyncratic. And all this got written into the Standard (maybe the public comment process wasn’t as good then as it is now). … and the rest of the industry said, “What, this is bizarre.” … So, this is kind of a sort of thirdhand story, but it alerted me to an issue that you needed some diversity among the experts on a topic.

The aim was consensus, or “counterpoint,” not a single vision. Subcommittee members were generally in agreement with this approach, noting that “the better you’ve got your spread … the better the outcome will be” (Interviewee 7).

Ultimately, the Chair chose a group of 27 people with technical expertise in areas such as materials selection, facilities design, instrumentation and control, corrosion, and pressure testing. The broad makeup is similar to the main committee, but with even greater representation for expert consultants. These people are all local (Australian) experts. Some work for large engineering firms, whereas others are self-employed experts with a long track record and strong reputation in the pipeline industry for particular technical expertise.

As described earlier, the Standard has a regulatory role in addition to those holding specialist expertise. The Chair chose to invite regulatory representatives to be involved and every state with gas infrastructure responded by nominating someone for inclusion on the subcommittee. The Chair explained his decision as follows:

The best engineers within the industry are not working for the regulators. So, I think the regulators benefit enormously from participating in the [sub]committee work and getting, picking up the knowledge that they do from issue papers and discussions and all that sort of stuff. And I think it’s wonderful that we have this real positive relationship with all the regulators.

Another interviewee described the regulators’ contribution as follows:

We’ve had people involved in standards who are regulators who made a substantial contribution and have improved the way that we manage pipelines. There have been other regulators who probably didn’t have the technical competence to be on a Standards Committee. … Some have sat and learned and really absorbed and developed a lot, through that process, which is great. (Interviewee 4)

From the regulators’ perspective, their interest is in “trying to keep the industry at best practice” (Interviewee 3) and part of that involves being present during discussions so that they can see how decisions have been reached.

The Chair was also of the view that it is important to have some younger people involved, largely as an exercise in generational change and succession management. ME-038-01 subcommittee members reported joining the ME-038-01 subcommittee at different points in time. While half of the participants had been on the subcommittee during previous revisions to the Standard in 2007 and 2012, many of whom were going through their third update, the other half of the subcommittee members were new to such processes and so typically more junior. The Chair explained his reasoning in selecting the group in this way:

As people like myself … and a whole bunch of others when you ride off into the sunset, there’s a whole lot of other people who’ve been through a distribution process, understand the background to it … I think [sub]committee culture is a really important thing. … They will provide that continuity as it goes on.

Another subcommittee member who joined in his thirties described his early experiences like this: “I remember going to that first meeting and then being terrified. I never said a word. All those legends!” Over more than a decade his attitude has changed and now “I find myself at these meetings talking a lot” (Interviewee 6). At least one of the younger subcommittee members could see that perspective too, but also said that the mix of experience levels is what makes it work. Junior people benefit from the learning opportunity, but experienced people guard against poorly thought through changes to the Standard:

You know the continuity is important and experience is important, so I thought it was a good balance and you kind of as a younger person then you’ve got that understanding of why the Standard is where it is today, so you know you can consider that before you make any other changes. (Interviewee 1)

Less experienced subcommittee members were initially labeled as “understudies” and yet that distinction between experienced and more junior was typically short lived. As the process of revising the Standard extended over five years, some of the less experienced members grew significantly in professional terms over the life of the subcommittee and were then in a position to provide input to the Standard based on their increasing technical experience.

Typically, once appointed, the subcommittee remains as constituted for the duration of the update. Making a good selection for the subcommittee is important because “it’s very hard to remove people” (Interviewee 3). Adding in, as needed, is possible, but the opposite was not true. The Chair explained that stability in subcommittee membership is important, and changes were only made where necessary:

I didn’t want to have a constant influx of people in and out of the [sub]committee. So, with rare exceptions there were very few changes to the constitution of the [sub]committee which lost one person who was simply not performing at all. And we gained another one or two along the way for various reasons. (Interviewee 9)

The cohesion of the group was fostered by the inclusion of solely professional engineers. Interview questions about including more diversity of expertise prompted a comment that perhaps there needed to be specialists on the engineering of liquid hydrocarbon pipelines included next time or more construction engineers as these areas were underrepresented. Expertise is so specialized that including another engineering specialty is considered an exercise in expanding the expertise of the group. Other disciplines (such as organizational safetyFootnote 2) and other stakeholders (such as land use planners or property developersFootnote 3) were not considered as appropriate to include. Subcommittee members broadly agreed with the Chair’s decision on these matters.

5.3 Dynamics of meetings

While some research and debate took place out-of-session, decisions over wording in the new Standard were typically taken in a series of workshops held quarterly over several years. In between meetings, subcommittee members often worked in small groups on sets of issues allocated to them by the Chair. Significant debate between specialists took place in some of these groups in an attempt to come to an agreed position that was then brought back to the whole subcommittee in the next workshop. This process of ever narrowing technical groups worked to resolve many technical questions, leaving debate in workshops of the full subcommittee to address things that were most contentious. As each quarterly workshop approached, there would be a flurry of activity within the small groups to move debate along and prepare materials for presentation on the day, ideally a recommendation for a wording change to the Standard or a finding that no change was needed. For issues where there were significant conflicting viewpoints within the small group, debates were brought to the full subcommittee. As one subcommittee member put it: “You can do a lot of work, but at some point, the rubber’s got to hit the road, and you need to kind of—meet up” (Interviewee 7). The Chair agreed: “When you’ve got a slightly contentious topic, there’s just no substitute for getting the key players in the room and thrashing it out on the spot”.

The language of “thrashing it out” suggests vibrant debate but sometimes only a small number of people were involved. In such cases, a sub-set of the subcommittee could be actively involved in a debate over a technical point while others watched on. Participant 1 reflected:

Very hard to have meaningful conversation and a lot of times some people just zoned out. So, there are people sitting in the room for no reason, and even that brings an element to the atmosphere of the room that we’re just – it’s like two or three people just talking about one point for 45 minutes. (Interviewee 1)

Decisions were generally made by consensus guided by the Chair, with the Chair sometimes breaking deadlocks as described above. Some participants suggested that, despite the technical character of the content, agreeing on the wording was a social negotiation in which interpersonal factors affected the outcome. Participant 7 said that, in addition to getting the right people in the room, “the personal relationships and all those sort of things, help as well” (Interviewee 7).

Whether people agree or continue to argue, or alternatively zone out is significantly an individual trait: “I think that’s also their personalities. Look, I mean, this is a social context, really” (Interviewee 1). Another interviewee was surprised that, as a group with a common profession, the way people operate in the group setting varies: “But the people are different as well, which I never realized how different engineers would be. I often thought we all worked the same way” (Interviewee 2).

Sometimes, people express a view that others consider to be unfounded, specifically, “people who were dipping their oar in and being very strident about what should happen. I’m thinking – mate, you don’t know what you’re talking about, you know? But that’s what you’ve got to work through in a committee” (Interviewee 7). When it comes to making changes to the content of the Standard, there needs to be a balance between welcoming new views and making ill-informed or ill-considered changes as this interviewee explains,

people come in and say let’s change this, let’s change that but that might be a perfectly valid point and it’s good also to have a fresh pair of eyes on things to shake things up, to challenge things, all of that is good but I think if you know the history and know why things are there, then you can kind of consider all of that before you just throw it out the window. (Interviewee 13)

Coming to a resolution was also the function of decisions about when to dig in, and when to compromise. Most subcommittee members said that they pick those issues on which they have strong opinions and those on which they can compromise on their preferred way forward. As one participant described, all issues “have been debated and decided on – I won’t say agreed on” (Interviewee 5). This subcommittee member continued:

I think you put forward views on different aspects, and some of them are “die in the ditch” type stuff, that there will be a problem if these are not considered. Others are more minor, preference for terminology, that kind of thing, which – okay, it’s annoying that I didn’t get my way, but I’ve got to get over it. … Were all my views considered and agreed to? No. Were the important ones? Yes. (Interviewee 5)

While this subcommittee member was not alone in their stance of being heard and knowing when it was acceptable to back down, there were many instances in which participants held strong views which they advocated for repeatedly. One participant commented that some subcommittee members “hold out with religious fervor on certain topics” (Interviewee 11). Another, referring to his own behavior, said that he had continued to argue his position on a point over no less than 12 subcommittee workshops (of the 16 that were held) spanning years:

And I basically said that 12 times by going to 12 workshops here and trying to keep that in the forefront over and over. Because I had the impression we’re having that conversation every time and I was worried that if one time I didn’t go, that the next time I would turn up and they would have changed things, taken words out, reworked stuff. So it was just me – like the water on the stone. I’m going to every meeting and I’m going to say this again and again and again. (Interviewee 8)

The dynamic with the regulators is also important. They are less experienced in industry but they have to approve of what is written, as the Standard is called up in legislation in each state. When it comes to considering changes to the Standard, while the process is consensus-based, the regulators have special standing. One regulator was strongly against a change being proposed by technical experts and so his agency made it known to the Chair that if the disputed change was made, regulations would not be changed to reference the new version of the Standard but would rather continue to require compliance with the old version. Other subcommittee members were aware of this with one non-regulatory interviewee describing the dynamics of engagement with the regulators as follows:

If there’s some technical discussion, and the regulator says, “I’m not having it,” it’s the end, isn’t it? Because ultimately we’re writing a set of rules that they’re adopting, but if they choose – if they really want to say, ‘I don’t like that so much I’m not going to adopt it’, then I guess they end up holding the whip hand. For better or worse … they do wield a bit of power, as well they should. (Interviewee 6)

5.4 The role of the leader in standard negotiations

At one level, the Chair is simply a manager of a project made up of many small tasks to be completed by many people. This aspect of the role is about pushing people to deliver according to deadlines agreed and holding people accountable. As one subcommittee member noted,

And I helped [Chair] at one point to try and apply a bit of project management to the thing, when it was seen to be that we were never going to finish in time, I put together a spreadsheet with deadlines, and tried to make people accountable. Because we’re all volunteers, we’ve all got day jobs. … You can’t push people too hard, but at some point, you’ve still got to get the job done. (Interviewee 6)

The Chair has a structural role in setting up the processes whereby changes are considered and a social role in resolving contentious issues where consensus cannot be obtained. Subcommittee meetings are “an argument waiting to happen” simply because of the number of experts involved and their differing views (Interviewee 3). At its worst, the subcommittee could:

End up in this hyper-democracy where everyone has to have a say leading to nothing but a dictatorship of poor Mr [Chair] saying, “I’m making a call and everyone shut your mouths,” basically, because we’ll never move on, and therefore not achieving anything that we wanted to achieve in the first place. (Interviewee 1)

Successfully driving the group to consensus is a key role for the Chair in managing the conversation. “[The Chair must] decide the direction of the meeting, really. You’ve got to make sure that people stay on track. While there are other influential voices around, just by their nature. But it’s got to be driven.” (Interviewee 6).

The Chair structured the decision-making process to encourage committee members to come to agreement but this was not always achieved. There were some issues on which the subcommittee was effectively deadlocked. Formal voting was never used; subcommittee members were encouraged to come to a consensus. Where this was not achieved, the final decision ultimately rests with the Chair. While technical expertise is key to gaining the respect of the rest of the subcommittee, the extent to which the Chair should use their leadership role to impose their technical views is debatable. Subcommittee members commented on the different approaches they had observed with chairs, ranging from the “‘shut up, we’re doing this!’” approach, to “unfocused” conversations that suffered from being “too collaborative” (Participant 7). What the subcommittee members generally looked for in a leader was someone who could walk a middle line between directive and consultative. Participant 5 was articulating the view of many subcommittee members when he said:

I think what you need is a benevolent dictator. We’re not always going to agree with what he or she says or what the decision is, but the person needs to be in a position that you respect, a position of authority, that you accept – okay, it hasn’t gone quite the way I wanted it to, but that’s the decision, so we need to move on with it. (Interviewee 5)

In addition to the technical expertise, a core part of the role of the Chair is group management: knowing what the conversation needs to be, when voices need to be heard, or debates need to be had, and also when conversations can move on or need some assistance to reach resolution. Along these lines, one participant made the comment: “I think it’s good when the chairman can shut down that kind of noise when it’s not adding value anymore, and just move on” (Interviewee 5).

To those who are not involved, “some of [the discussions] just seemed interminable. You sit there and they’ve been discussing some sentence on some technical thing I don’t understand. And they’ve been discussing it for the past hour and you kind of think, oh [Chair], can’t you just make a decision?” (Interviewee 8). Sometimes in such cases the Chair has deliberately decided to allow debate to continue. The current Chair explained how he would decide whether to make a decision on a technical matter of disagreement. He would only make the final decision under two circumstances,

a) I thought I had the technical expertise to do that, confidently, and b) that I thought I’d be able to sort of carry the day. And that if I made a decision, people were not going to meet me and refuse to accept it. Because I wasn’t going to get in-between [three experts]. I don’t have the technical expertise and those guys with the technical understanding, just needed to sort it out between them. And I can’t contribute other than trying to keep them focused and keep a deadline and keep them working to a deadline. There are other things, where yeah sure, I’d make an executive decision.

On tricky technical matters outside of his own specialization, he makes the call when he feels that the technical understanding is basically resolved, and the conversation is no longer productive. He does not seek to make a captain’s call on a technical matter when debate is active. Subcommittee members acknowledged the importance of this ability in a subcommittee Chair, noting those participants who needed wrangling:

[a recognized industry expert] really knows his stuff, but every time we’d go in there, we’d seem to be listening to him rehashing arguments that – I thought we’d all sorted this out. (Interviewee 7)

I mean this is a bit of a thing that some of the older people in the industry have especially. They sit at their desk and write these long, ranty emails to people. … I think in the end, for most of them … people agree to disagree and then at the next workshop [the Chair] will just say, “So there’s this case and this case and I’m deciding we’re going this way.” (Interviewee 8)

You know some people were vocal but I’m not sure how influential they were, they would just kind of take up a lot of time. I think if you did a graph of words said per value added, it wouldn’t look that great. (Interviewee 13)

The question arises as to whether this is simply a matter of efficiency of reaching the outcome or whether extended and cyclic conversations can impact what the outcome is. Participant 7 was of the view that the impact could be material, and so the Chair’s ability to move a conversation on was vital.

When you’ve got people who are dominant personalities – [one such person would] hold court for a long time … and that does suppress other people … there’s a lot of people who have sat there pretty silently. … whether that really affected the outcomes, I don’t know.’ (Interviewee 7)

In addition to those negotiations in the room, the standard revision process is opened to public comment. Public comments were managed via normal SA processes meaning that any member of the public is free to respond. In total, 894 comments were received, all from pipeline engineers working within the industry. Dealing with the comments fell mostly to the Chair who had to decide which to adopt, which to reject outright, and which to discuss with the rest of the subcommittee. He was of the view that despite the logistical difficulty, it was an important exercise:

It’s valuable input. It’s really important and the [sub]committee can become a bit of an echo chamber. There’s only 30 of us and not all of those are very active. So, it’s a pretty small cross section of the industry. So, getting opinions from people outside the [sub]committee who will be using the standard is just really important.

This is true of technical changes, but also where comments indicate that the Standard as written is being misunderstood and so needs to be redrafted for clarity. In his view, “there’s no doubt it improves the standard. It’s just a painful process for the Chair in particular and maybe for the subcommittee as a whole”.

In practice, the level of impact of those comments on the final Standard was limited and the final version of the revision was approved for publication in 2018.

6 Discussion

6.1 Building an expert group

In seeking to identify multiple experts in the sub-domains of pipeline engineering to serve on the subcommittee, the Chair made tacit use of some of the same principles as Delphi when it comes to diversity (Burgman 2016). The chair recruited two or three people in each highly specialized area, e.g., corrosion, fatigue management, welding. The Chair would maintain that the group as constituted has sufficient diversity to be able to challenge the status quo but not go so far as to provoke an institutional forgetting of accumulated experience that comprises the previous revision of the Standard. The level of epistemological debate required to reach agreement supports this view when it comes to the specific technical requirements.

In addition to different specializations, the Chair targeted a mix of skill levels. The purpose of the diversity in terms of experience is not in the service of having specializations represented but to build expertise in future leaders. Such diversity in membership also leads to a large group which research has shown leads to a better overall outcome (Kosinski et al. 2012). It is not the users of the Standard in the first instance that are feeling the effects of the new norms, but the subcommittee members themselves. The processes of cyclic debate and discussion are a form of expert education enrolling more junior professionals into a collective way of thinking.

This expert education and relationship building also extends to regulators. The Chair chose to involve regulators to normalize their perceptions of the Standard to those of the wider group. It might be seen that this arrangement is a form of “regulatory capture” (Ayres and Braithwaite 1992) but this reinforces the importance of the professional, rather than industry, alignment of the most respected members of the group. If regulators are captured by the norms of the group, it is to be captured by professional engineering values, not the values of the industry itself which might be assumed to be more commercially and business focused.

The Chair is exhibiting expertise in group decision-making, albeit tacitly, in his choices about how to constitute the group. Further, in building the expert group in this way the Chair calls upon significant interactional expertise in the sense of being literate in multiple disciplines and functions as they relate to the problem, a skill perhaps not readily understood by those who select such leaders and distinctly different from the contributory expertise they hold in their specific domain. This has specific implications for selection of standards committee chairs. A most effective chair will consider the skills required to complete the task at hand (producing a new/updated standard) and the longer-term goal of building a strong epistemological foundation for the technical discipline as a whole by enriching the relevant technical (and regulatory) professional group.

6.2 Generating the outcome

The decision-making literature focuses on the role of the leader in designing and facilitating expert interactions (Linstone and Turoff 2011), not their subject matter expertise. In contrast, our interviewees described how the Chair of their group did an excellent job due to his knowledge of the content of the Standard. Such technical knowledge was seen as critical for the Chair’s emergent leadership of the group whereby the Chair becomes influential over subcommittee members due to their leadership qualities, rather than the formal role held (Badura et al. 2022). Subcommittee members were also clear that other qualities were required. Key skills included managing the work of the group and bringing the group to agreement. Managing this process to reach a common corpus requires both interactional and contributory expertise for the Chair (Collins and Evans 2007). They need to know when to call time on the debate, which is not so much temporal (we’ve run out of time) as epistemic (we have reached a point where we have reached an agreed “knowledge”). The cyclic nature of the conversations requires from the Chair not so much a single “ruling” as a process of continual arbitration each time an issue re-emerges.

This is not to say that there are not also skills required to manage the group in terms of decision-making processes. The process to generate the final version took five years. The process may at first appear to be unstructured—what Burgman might call “naïve group judgement” (Burgman 2016, p. 122). However, the process involved structured processes whereby the Chair identified the topics to be discussed, the required subgroups needed to attend to these issues, and forums in which to debate and agree upon the relevant wording. Where experts cannot agree, they must air their disagreements in front of the Chair and the full subcommittee and find a way to compromise.

The process of negotiation over the content is therefore not conflict-free nor always smooth. This is common in debates within groups of experts. One biologist noted regarding interpretation of complex field data “major technical disagreements stemming from philosophical differences that seem to run as deep as religious beliefs are commonplace in such technical teams” (Ruckelshaus et al. 2002, p. 691). In the case we studied, there were times in which participants felt the debate may have benefited from greater intervention on the part of the Chair, though in general it was felt that, however painful, it was a necessary part of the process. The multilevel processes, interactional nature of the discussions, and the physicality of face-to-face conversations all form part of the relationship development required to make epistemic progress. As Brunnson et al. (2012, pg 619) note, “frequent and lasting interactions among a group of experts tend to create trustful relationships and promote common values”.

Overall, the process of producing the Standard sees individuals’ knowledge and experience come together being redebated, rediscussed, and recomposed into the final framework that is the new version of the Standard.

6.3 A new standard

The draft Standard created by the expert group is subject to formal review via a process of public comment. Theoretically, the consultation mechanism transforms the process of standardization of group knowledge into knowledge discussed with and agreed by external empowered third parties. Thus, the group’s expertise becomes the community’s expertise. In this way, the draft Standard that results from internal group norms is exposed to wider scrutiny and must be renegotiated to reach a new externally acceptable version of itself.

However, our case saw no participation of other professionals whose work impacts pipelines (e.g., land use planners and land developers). It is reasonable that these other professionals are excluded from the five years’ worth of technical debates that culminate in the draft Standard. This is a technical artifact in engineering first and foremost. However, in failing to more specifically target these adjacent professions at the point of finalizing the Standard there is a missed opportunity to arrive at a renegotiated version of the Standard that appreciates pipelines operationalized in their community context.

The public consultation process also failed to engage the general public. The many “public” comments received came from those pipeline engineers who were not in the room and, more often, subcommittee members who were in the room, but wanted to continue the fight. In some fields, as in synthetic biology, there is more public involvement in debates over standardization as they relate to manipulation of DNA (Epstein 2009). This is not the case in pipeline engineering where decisions appear to be as buried as the infrastructure itself.

Overall, the Standard produced by the expert group has been well accepted by their professional colleagues across the pipeline industry. It has achieved high input legitimacy (Botzem and Dobusch 2012) i.e., “the confidence that users hold in the standard that it will be operable and produce the right outcomes” (Hayes et al. 2022). This arises in part as a result of the perceived expertise of the group. Competition between industry professionals to join the subcommittee was significant with membership bestowing status (Hayes et al. 2022). The members of the group are known to their colleagues thereby providing a strong “chain of accountability” (Mayntz 2010; pg 10) that links those who set the Standard with those who use it.

7 Conclusion

Production of a new standard (or a major revision to an existing standard) is typically a slow and somewhat opaque process. This paper goes some way towards opening the “black box” of standards committees, looking to the constitution of the expert group, the nature of the decision-making process, and the role and the expertise of the leader, as well as how the Standard becomes socialized into the public domain. We conclude with some implications for future committees.

First, our findings highlight the value in having a leader with relevant subject matter expertise in addition to a basic knowledge of group decision-making. The decision sciences literature captures many instances in which expert groups contribute to decision-making processes with no expert leader, or indeed without conversation between group members. In such expert elicitation processes, there is typically a facilitator who is appointed based on their skills in group decision-making, not the specific subject matter. This would not have worked here. In the case of development of a technical standard, both expert leadership and group interaction are vital to reaching an outcome. The Chair must have specific technical expertise to contribute to the outcome, thus gaining legitimacy in the eyes of the group. But the best Chair is not the most knowledgeable in a specific area of practice. A good Chair also has interactional expertise, speaking the language of other specialists to be across the issues. The Chair must understand the technical debate to the point where they can recognize and “call” agreement and have that decision respected.

Second, our findings speak to the nature of the group constitution. In the case of the committee we studied, they had a strategy of covering off different disciplines in pairs. This strategy to get breadth and depth of voices has been recognized as important in expert elicitation (Burgman 2016) although the case we studied points not only to the value of this to decision-making in the short term, but the potentially positive impact for collective knowledge over longer temporal scales. The group interactions foster expertise within the industry as a whole, including within regulatory agencies.

Finally, this case highlights the potential conflict between a desire to socialize the standard more broadly and yet maintain the best functioning of the expert group that produces it. In our case, subcommittee members maintained that a technical industry standard is served by the inclusion of voices only from relevant engineering disciplines. While such limits on the inclusion of expertise may be in the service of the professional socialization function of the standard as well as the goal of debating and reaching epistemic agreement, it has the side effect of silencing other professional and public perspectives as well as leaving these “others” out in the cold. Perhaps the interests of civil society would be best served with a new reference group that could be consulted at regular intervals or at least engaged during formal public comment periods, in order to take account of the knowledge and needs of experts from other professions with an interest in pipeline location (e.g., land use planners, developers) and/or members of the general public (who are exposed to the risk posed by this infrastructure).