The meaning and making of union delegate networks

Abstract

Networks are a subject of growing research interest. Yet union networks, particularly networks of delegates, and ways to build them, are still poorly understood. This is a study of the meaning that workplace union delegates assign to networks of support. It explores the characteristics of effective delegate and union networks and influences upon them. Effective networks are a combination of strong and weak ties, such that delegates sometimes do not recognise they are part of a network. Our three-stage research methodology involved delegate focus groups, a paper-based self-completion questionnaire of recently trained delegates (N=473) and a follow-up telephone survey (N=145). It found that organisers were key to creation of internal workplace networks (although they did not necessarily establish them) and in providing a bridge for delegates with external networks. They were the key support person for many delegates. Networks took a variety of forms. Only a minority were formalised. A majority were mainly internal to the workplace. Social media were rarely used, with little intention of using them more, and were, we suspect, underutilised.

Introduction

Network theory is well established, yet little attention has been given to building networks and social capital within union theory (Bailey and Brown, 2004; Jarley, 2005). Unions’ embeddedness in networks (Lévesque and Murray, 2010) and delegates’ involvement in networks (Peetz and Pocock, 2009) are accepted as being an important source of union power. Yet the development of delegates’ networks may be the role at which unions are weakest (Peetz and Pocock, 2009). This article seeks to address the gap in knowledge of union networks by drawing on Australian data to inquire into the nature of union delegates’ networks, what delegates understand by them, their characteristics and how they are built and maintained. We begin with a brief overview of literature on networks generally and the relevance to unions. We outline our three-phase empirical method and present findings from qualitative and quantitative research involving delegates in Australia.

Background and literature

Classical sociology (e.g. Simmel, 1955; Tönnies, 1955) has long focused on patterns of social interaction to understand power relations (Scott, 2012; Scott et al., 1985). Since the 1970s, attention has turned to interactions through social networks. Bourdieu (1977) spoke of four types of capital – economic, symbolic, cultural and social capital – the last one a function of social networks. R.D. Putnam (1995) emphasised the importance of social capital and of norms and trust in networks, while Portes Salas (1998) explained how the concept of social capital emphasises, among other things, the ways in which ‘non-monetary forms can be important sources of power and influence’ (p. 2).

At its broadest, a network is a set of interconnected relations among nodes (Newman, 2003), a term which has applications across almost any field (Lee et al., 2002; McCowan et al., 2008; Moretti, 2011), but which for us refers to human social interconnectivity. Network theory is premised on the idea that social life is relational and that relations form traceable patterns mapping the social world (Scott and Carrington, 2011). Social networked relations, between human individuals and groups, are typically ‘a densely knit clump of social structure’ or ‘cliques’ (Granovetter, 1983: 202; Tuti and Wieseb, 2015). Strong ties operate inside a clique (e.g. siblings or friends), while weak ties go between people in different social groupings that are linked by a ‘node’ (a person with multiple connections) or a ‘bridge’ (a tie between two people in different groupings). Weak ties are important in getting jobs (you gain access to information you did not have yourself – hence ‘the strength of weak ties’), and strong ties are more effective in situations needing emotional bonding (Granovetter, 1973; Tuti and Wieseb, 2015).

Giuffre (2013) looked at community networks and the ways that old and young individuals were educated, motivated and mobilised to commit to sustained activism against racism and racists, through use of a training process that built friendships and links that became the backbone of a group dynamic (Giuffre, 2013: chapter 6).

At the core of unions is the scope for collective action – without it, political activity is hollow. Networks are the circulatory system that pumps blood between the different elements of collective action. This is not to privilege networks over other elements of collective action, such as shared values, a sense of efficacy and identification of collective needs (Peetz, 2010). Equally as important as connections between people are mobilisers who activate collective behaviour within a network (Peetz, 2006) and thereby create coordinating capacity. We thus seek to adopt a wider context than Jarley’s (2005) conception of a social capital model as a basis for union renewal. He argued unions should organise around people, not issues, and build dense social networks among members by promoting activities that reinforce generalised reciprocity norms. Johnson and Jarley (2005) argued that building these dense social networks is difficult but likely to trigger union participation and engagement, with special appeal to workers who lack social capital and financial resources to purchase effective substitutes in the marketplace. As Bailey and Brown (2004) pointed out, however, this focus downplayed the multiple roles and activities of unions, and it is better to see networks as one important aspect of collectivism by workers and unions, not an alternative. Saundry et al. (2012) take a different tack, describing Jarley’s account as ‘optimistic’ (p. 282) and arguing that social ties are formed within very constrained social, economic and political contexts. Unions, they argue, can link into networks, rather than offer workers a network that they can ‘borrow’. Thus, union involvement could be seen as a threat to the collegiality of ‘purely’ worker networks. Fiorito (2001) suggests that one of the reasons for union decline is human resource management practices that facilitate, inter alia, social and other network characteristics that enact a ‘substitution’ effect for unions (see also Waddington, 2014). Alternatively, with networks’ ability to reduce employees’ intention to leave (Friedman and Holtom, 2002) and promote informal information sharing (Scully, 2009) and social comparison (Shah, 1998), there may be a synergy for the individual worker between these non-union social networks and a union network.

Other research relevant to union networks has focused on links outside the workplace. These may be links with other unions (Ellem, 2003), with community groups (Erickson et al., 2002; Tattersall, 2010) or with the state (Howard, 1977; Korpi and Shalev, 1979; Pizzorno, 1968). Or the focus may be on international union networks (Wills, 1998), transnational grassroots coalitions (Hossein-Zadeh, 1997; Tilly and Klausen, 1997) and even unions’ own use of external social network platforms, such as the Dutch union federation FNV Bondgenoten’s social network site for young workers (Kloosterboer, 2008). Theories of workplace power emphasise the importance of unions having external networks (Lévesque and Murray, 2010). Our focus is narrower but adds to this literature by looking more closely and empirically at networks directly involving delegates, some of which are external to the workplace, but many of which (as we shall see) are principally within the workplace.

Further theoretical works have practical implications for the role and operation of union networks. For example, Putnam (2000) distinguished ‘bridging’ social capital, which fostered broad and inclusive associations, from ‘bonding’ social capital, which risked insularity and fragmentation rather than unity. But for unions, ‘bonding’ social capital may be important if it helps mobilise against an employer. Our interest is less in whether unions are (or are not) seen by workers as providing ‘bridging’ capital or being the best ‘brokers’ (Burt, 2002), than in how well union networks function within their own logic. A related issue in the literature is whether diversity in groups leads to lower levels of cohesiveness (Thatcher and Patel, 2011), consistent with ‘balance theory’ (Situngkir and Khanafiah, 2004) which suggests that networks will more readily form within homogenous groups (Young and Wilkinson, 2004). For example, some (but not all) migrant workers may have weaker social ties within the worker context (Wu and Zhang, 2007). So the increasing diversity of workforces has generated what some see as a ‘crisis of interest aggregation’ (Müller-Jentsch, 1988: 177–178), not improved by the (white and male) characteristics of delegates and other union officials (e.g. Healy et al., 2004). On the other hand, Garcia (2002) argues that formal or informal identity caucuses within unions (e.g. of women within unions) are not a source of additional division in the labour movement, but act as a supplement, as these sub-groups may preserve a strong faith in the union.

Another focus of practical interest is methods of communication within networks. Unions, Lucio and Weston (1995) argue, have their origins in informal relationships and networking structures. These networks increasingly occur through virtual technologies (Manago et al., 2012). Some authors describe the Internet as the new ‘workers’ hall’ (Aalto-Matturi, 2005; Diamond and Freeman, 2002), and some union officials advocate much greater use of social media (White, 2010). Potential members can be identified through workplace ‘mapping’ using such technologies (Manago et al., 2012; Traud et al., 2011). On the other hand, the quality of information passed through networks may be poor (Marshall and Goodman, 2013), although this may be the case for other forms of transmission as well. Privacy conflict with social media is another issue with which unions need to deal.

Overall, the research points to the importance of networks for unions but a lack of data on what, for unions, networks are, especially at the workplace, how they are formed, what they consist of and what promotes or hampers their existence. This article seeks to redress those gaps. Our key question is, What are the characteristics of union delegates’ networks? Within that, we seek to investigate a series of issues including the following:

• • What are networks of support for union delegates? What do delegates understand by them?

• • How important are formal versus informal networks?

• • What methods (such as social media) do delegates use and seek to use more?

• • What support do delegates demand and what barriers do delegates see?

• • Are union networks created independently or within existing workplace networks? Who mobilises delegate networks and how?

Methods

Our investigations were in three phases. The first comprised four focus groups – encompassing 60 delegates – from four unions in late 2011. These were coordinated by the Organising Centre of the Australian Council of Trade Unions (ACTU) and held in Sydney and Melbourne, Australia. Their purpose was to elicit delegates’ views on the meaning and purpose of networks and obtain qualitative data that could generate ideas to be tested quantitatively. The protocol for the focus groups was designed by the researchers, in consultation with the Organising Centre.

The second phase comprised a paper-based self-completion questionnaire, distributed in 2012–2013 to union delegates at several delegate training courses run by the Organising Centre or by a number of other trade unions affiliated to the ACTU. The questionnaire was designed by the researchers, in consultation with the Organising Centre, in light of the above literature and the preceding focus groups. It was distributed at the end of training sessions – so all respondents had been through some form of union training (which, overall, only around two-thirds or less of delegates in Australia have experienced; Peetz and Pocock, 2009). While some questions referred to ‘networks’, others referred to ‘links’ or ‘connections’ because these are what participants directly see. We could not assume that they would use the term ‘network’ in a standardised manner. In total, 473 delegates from 21 unions completed the survey. As it was distributed by multiple training officers in multiple locations, it was not possible to obtain an accurate estimate of the number initially distributed, but we expect it would be something over double that number.

Respondents came from all six Australian states plus the Northern Territory; 53% lived in metropolitan regions, 14% in provincial cities and 30% in rural or remote areas. One-third of respondents came from teaching unions (mostly in the public sector), while three-tenths came from other services (including health, public services and utilities). In addition, 16% were from the retail industry, 8% from unions representing low paid workers and 12% from unions representing other blue collar workers. The sample appeared to thus over-represent white collar delegates (although white collar unions account for two-thirds of union members; Australian Bureau of Statistics (ABS), 2013). Some 64% were females, whereas among unionists as a whole males slightly outnumber women. The median age of our respondents was 43years, slightly older than the average age of unionists, although delegates are typically older than unionists on average anyway; 21% of our sample were aged up to 30years and 17% aged 55years or above.

The third phase was a follow-up telephone survey, undertaken by the ACTU Member Connect call centre via computer-assisted telephone interviews (CATI) in 2013–2014 of those from the self-completion survey who agreed to be followed up (slightly under half did so). A total of 145 completed responses were obtained from these phone calls, two-thirds of those we followed up. (Of the remaining one-third, half refused and half were uncontactable.) Of those 145, only 118 (81%) were still delegates. The phase 3 survey repeated some questions asked in the second phase, and some new information was collected.

Among those 473 delegates participating in our phase 2 study, 82% had delegate rights, such as to union training or time off, through their enterprise agreement. While 39% had some coordinating or leadership role in relation to other delegates at the workplace, for 25% someone else had that role (suggesting senior delegates were more likely to receive training), for 13% no-one had that role and 24% were the only delegate in their workplace. (A handful self-identified as ‘new’ or said they were still training to become a delegate.) There was some diversity in how many members they had personal responsibility for, with median coverage of 25 members but half within the range of 12–60 members. Estimated median union density of the workplaces of delegates was 75%, with a quarter below half and another quarter at or above 90%.

There were a small number of differences in characteristics between our phase 2 and phase 3 respondents, the main one resulting from higher attrition (drop-outs between the first and second surveys) from females than males. So, females were only 51% of wave 3 respondents (close to their share of all unionists). There were also some differences in attrition between unions, with education unions tending to higher attrition than others. However, when defined by many key aspects of networks, most group differences in attrition were non-significant.

Analysis of quantitative data here was done principally through univariate or bivariate techniques, with significance tested through chi-squared or t-tests where appropriate. Respondents who appeared in both phases 2 and 3 we refer to as in our ‘panel’.

Findings

Understandings of networks

Delegates in focus groups mostly described networking as being about talking to co-workers, other delegates within their own union and the organiser (a paid official of the union). Some described networking as information gathering, being ‘in the know’ or keeping up-to-date with union activities. Many, at least in three unions, were unaware of networking as a separate concept; the fourth union had been heavily involved in a joint campaign with outside groups. That said, delegates identified networks as crucial to their ability to resolve issues.

In the phase 2 survey, we asked trained delegates to choose from four statements. Among those who answered, 24% indicated no significant support, answering ‘I do my job as a delegate pretty much by myself’; 31% cited ‘a limited amount of involvement from time to time with other people who support me’; 41% were ‘part of a network of people that help me do my job as delegate’; and only 4% indicated multiple network embeddedness, being ‘involved in several networks that help me do my job as delegate’.

When asked about their most important network, half said it consisted ‘exclusively [of] people in my workplace’; another 30% indicated ‘mainly people in my workplace’; 19% responded ‘mainly people outside the workplace, but in the same organisation’; and only 1% said ‘mainly people from outside this organisation’.

In the phase 3 telephone survey, with a smaller N, we asked separately about internal and external links (again providing four options in each case) and obtained a different impression. When asked about their situation as a delegate in the workplace, 11% indicated multiple internal network embeddedness by saying ‘I am involved in several networks that help me do my job as a delegate’. When asked separately about external links, 9% said, ‘I am involved in several networks outside the workplace that help me do my job as a delegate’. These higher rates of multiple network embeddedness, compared to phase 2, are unlikely to reflect wording effects and are unaffected by attrition rates. Within the panel, we found a number of people who reported being in multiple networks (either in the workplace, or externally, or both) in phase 3 but not in phase 2. Far fewer showed lower rates of multiple networking in phase 3 than phase 2. It is plausible that higher networking was a result of training effects over time (respondents who, in phase 2, reported being trained in how to develop networks of support reported more network involvement) or even networking opportunities provided by the training events.

Another noteworthy difference is that although phase 3 confirmed the greater incidence of internal than external networks, the difference was considerably smaller than implied by phase 2 data. In phase 3, 49% explicitly identified internal networks to which they belonged, while 32% identified external networks to which they belonged. Further investigation showed that for 34% of respondents to the phase 3 survey, their internal networks seemed stronger than their external networks; however, for 19%, their external networks seemed stronger than their internal networks; and for almost half (47%) of respondents, networks inside and outside the workplace were of roughly equivalent strength.

The origins and personnel of networks

Focus groups suggested networks, in most cases, were self-initiated by the delegate. Few had been initiated by union staff. They saw themselves as being largely left to their own devices to form networks and to work out how to deal with the situation in their own workplaces. That said, discussions indicated that training courses attended by the delegates had given them new ideas and skills to create networks.

In the phase 2 survey, we asked whether the respondents’ most important network existed ‘before the union got involved, or was it created for a union purpose?’ While 55% said it was created for a union purpose, 21% said it already existed and 24% did not know. (In the phase 3 survey, 77% reported a union purpose. Union-created networks appeared to become more important over time than pre-existing networks.) We also asked, ‘Who initiated that network?’ In the phase 2 survey, 33% of respondents referred to the organiser and 53% referred to people in the workplace (including 31% in which the respondent had a role).

Separately, we asked, ‘Which of the following people are helpful to you in your role as a union delegate?’ and then who was ‘the most use to you in terms of helping you do your job as a delegate well?’ We also asked, ‘Are there any with whom you would, realistically, like to have more contact?’ Results are shown in Table 1. The most helpful person was clearly their union organiser, with by far the highest frequency of ‘very helpful’ responses to the first question and ‘most helpful individual’ to the second. Other delegates, most commonly senior delegates but sometimes at or below respondents’ levels, were also helpful. All up, paid union staff (mostly but not exclusively the organiser) were rated most important by 38% of respondents, delegates or members in various locations were rated most important by 23% and 3% gave non-union references (such as supervisors), while 37% nominated no-one, although they did not necessarily identify as being outside of networks.

Table 1. People helpful in respondents’ role as a union delegate.

Source: Phase 2 self-completion survey.

All numbers are in percentages.

Columns (1)–(5) are answers to the question ‘Which of the following people are helpful to you in your role as a union delegate?’, with the options on ‘how helpful’ as shown. Hence, each row across columns (1)–(4) sums to 100%.

Column (6) is answers to the question ‘who is the most use to you in terms of helping you do your job as a delegate well?’ Only one response is possible, so the column sums to 100%; 37% gave no answer.

Column (7) is answers to the question ‘Are there any with whom you would, realistically, like to have more contact?’ The scores shown here are, for each response, those who wanted ‘more’ contact (either ‘much more’ or ‘some more’) minus those who wanted ‘less contact’.

N = 473. Respondents answering specific questions range from 278 to 452.

^a Less than .5%.

The final column of Table 1 shows net proportions wanting more contact with each group. Despite high support from organisers, a slight majority of respondents still said they wanted more contact from them. Indeed, for most types of possible contacts, respondents were roughly evenly divided between those who wanted more contact and those who wanted the same level of contact. For most groups, very few respondents wanted less contact – although views were more divided on contact with other unions – and they seemed to want less contact with customers or clients of the employer.

Formal and informal networks

In the phase 1 focus groups, delegates indicated networks were largely informal and often somewhat accidental. Some expressed a desire to have more formalised networking opportunities, to share information with other delegates, particularly within their industry.

In the phase 2 survey, informal networks outnumbered formal ones by 61%–39%. A slight majority of informal networks met as a whole sometimes, the rest did not. A larger majority of formal networks (formal structures with a name) met regularly, the rest met occasionally.

The wording of the phase 3 survey invited more responses about informal networks (from people who would not have answered the question in phase 2 if they did not think of their connections as a ‘network’); 74% described their network as informal and 26% as formal. This reinforced the impression that informal networks were more common than formal ones, but that the definition of an informal network was rather fluid for delegates.

Examples of the formal networks given by survey respondents included ‘workplace organising committee’, ‘joint delegate committee’, ‘union council’, ‘Metropolitan South Forum’, the ‘[name of town] Association’ and even the union itself. Examples of informal networks included ‘talking with delegates on other shifts’, ‘other union delegates and the organiser’, ‘just colleagues and other union people from one of the larger towns’, ‘contact up the road’ and ‘I guess where we see a problem, we call a meeting after hours and work out the best solution’.

Network tools

In the phase 2 survey, we asked trained delegates to select, from a list of tools, which they ‘use for union matters?’ and which they would ‘realistically, like to see used more often, or less often, or not used at all?’ Table 2 shows the results. Delegates relied heavily on (informal) face-to-face meetings with individuals, email and noticeboards, and, to a lesser extent, telephone contact and formal meetings of their own members. Few made extensive use of social events outside the workplace and use of social media was very rare.

Table 2. Tools used for union matters.

Source: Phase 2 self-completion survey.

All numbers are in percentages.

Columns (1)–(4) show answers to the question ‘Which of the following tools do you use for union matters?’ Columns (1)–(3) sum to 100%.

Column (5) shows answers to the question ‘Which would you, realistically, like to see used more often, or less often, or not used at all?’ and depicts the number who said ‘Should use it more’ minus the numbers who said ‘Should use it less’ or ‘Should not use at all’.

N = 473. Respondents answering specific questions range from 425 to 445.

Many delegates wished to make greater use of traditional networking tools, with majorities supporting more face-to-face meetings and formal members’ meetings (see final column of Table 3). While there was considerable support for more social events, there was strong opposition to greater use of social media. This is despite several key union leaders advocating greater use of social media as a union organising tool. Yet in the phase 3 survey, self-reported changes in use of social media correlated positively and significantly with self-reported changes in how closely respondents worked with other delegates (r=.25*) and with how closely they worked with other delegates (r=.21*), although with small N and low variance it is necessary to be cautious.

Table 3. Barriers and facilitators to being part of an effective union network.

Source: Phase 2 self-completion survey.

This table shows responses to the question ‘How much are each of the following an impediment or a help to your having or being part of an effective union network?’

Columns (1)–(7) are gross or net percentages. Column (8) is a score.

Columns (1)–(5) show the options as presented to respondents and the frequency of responses. Columns (2) and (4) were unlabelled intermediate points. Responses sum to 100%.

Column (7) shows a measure of net impediment, being columns (1) and (2) minus columns (4) and (5). Some apparent discrepancies are due to rounding.

Column (8) is another summary measure, being an average ‘impediment score in which the responses of columns (1)–(5) are given values of −100, −50, 0, +50 and +100, respectively.

N = 473. Respondents answering specific questions range from 434 to 448.

Barriers and facilitators of networks

In our phase 1 focus groups, the fear of management reprisal (against members, not delegates) for those in insecure work was mentioned frequently as a barrier to networking. Some delegates mentioned language and cultural barriers as issues that they were dealing with, but this seemed to be a positive aspect, as many reported that the union organisation within the workplace was breaking down barriers between cultural groups. Gender issues were identified by some male delegates as a reason to have both male and female delegates working in a team. Confirming earlier quantitative research (Peetz and Pocock, 2009), little formal training in networking was reported. Delegates asked for more opportunities to meet and share experiences in networks organised by the union, including both formal and regular networking sessions.

In the surveys, we asked trained delegates ‘How much are each of the following an impediment or a help to your having or being part of an effective union network?’ followed by a list of 14 items. The phase 2 results are in Table 3. The major barriers perceived by delegates related to time: its availability and shift work or working hours. The other, less frequent, barriers cited were the attitudes of supervisors and management. (Bear in mind that delegates who faced major constraints in time or from management would have the greatest difficulty in attending training courses and hence being part of the survey.) Notably, group diversity did not feature much one way or the other, with most saying each of gender, age and ethnicity made no difference.

However, whether or not delegates perceived these things as barriers said little about how important a barrier each was. As shown in Table 4, few of these variables were useful in predicting measures of network ability reported by panel respondents in phase 3. Our finding, that time was the most commonly mentioned barrier, was consistent with earlier research showing delegates perceive time as the major barrier to greater involvement in the union (Peetz and Pocock, 2009). That research, though, also showed the ‘time’ barrier did not predict workplace outcomes well, and likewise time constraints in our delegates’ survey did not predict network outcomes. Shift work appeared to make it harder for delegates to act as a bridge between groups, but not much else. Similarly, the barriers created by confidence or opportunities had low predictive ability, but this missed the importance of levels of, or changes in, confidence and opportunities in explaining networking utility (and other workplace outcomes – see Peetz and Alexander, 2013; Peetz and Pocock, 2009).

Table 4. Correlations between selected phase 2 and phase 3 variables.

Source: Phase 2 self-completion survey and phase 3 telephone survey.

Organiser mentoring = my organiser has taught me many valuable things about being a delegate; Time = availability of time; Shift/work hours = shift work or hours of work; Supervisor = attitudes of my supervisor; Ethnic = language or ethnic composition of the workforce; Meet other delegates = how many opportunities I have to meet other delegates.

Access time off = how much access I have to time off and delegate rights through my enterprise agreement.

* Significant at the 5% level; **significant at the 1% level.

However, a noteworthy aspect of Table 4 is how respondents’ phase 2 perceptions of organiser mentoring influenced network utility outcomes, not only in phase 2 but also in phase 3 (where common method variance was much less likely to influence results).

Some organisers may have changed between phase 2 and phase 3. In phase 3, there were very strong relationships between, on one hand, whether delegates felt that they met or talked with their organiser more or less often than in phase 2 and, on the other hand, whether they felt a range of key variables had gone up or down since then, specifically: involvement in supportive networks (r=.47**), how closely they worked with other delegates (r=.42**), opportunities to meet delegates or activists from other workplaces (r=.40**) and opportunities to meet activists from other organisations (r=.39**). While common method variance might strengthen the appearance of these results, they are consistent with the findings from the previous paragraph.

Discussion

First, we are not seeking here to gain a representative count of delegate networks. Delegates who participated in these surveys had all been through training (and by phase 3 had time to implement the lessons of this training). As networking is linked to training (Giuffre, 2013), the overall incidence of networks is likely to be lower than in these data. Our main interest, though, is in seeing how delegate networks differ and are driven.

Internal networks were more common – by how much differed between surveys: the difference was not as great in phase 3 as the phase 2 wording implied. One possible explanation is that while external contacts were numerically fewer than internal contacts within networks, their significance was greater than their frequency indicated.

It appeared from the data on perceived barriers that the more successful delegates were, the more they raised their expectations and recognised barriers to their achievement. For example, time or (on most items) shift work did not predict the utility of network involvement. Yet, they were the most commonly cited barriers facing delegates developing networks. Presumably time and shift patterns were real barriers to network activity, but there is not much unions can do about it. Unions can influence training and delegate confidence, and they matter (Peetz and Alexander, 2013).

Neither survey data nor focus groups provided evidence in support of the idea (Thatcher and Patel, 2011) that diversity created inherent problems for cohesiveness and networking. If any existed, delegates appeared able to overcome them, perhaps by promoting unity where others used divisiveness, perhaps using ‘like with like’ strategies and challenging the White male stereotype (Healy et al., 2004). Our study did not focus on this issue, so we are cautious in interpretations, but our implications are closer to Garcia (2002) than to Thatcher and Patel (2011).

Nor did our study show that union delegate networks relied on linking into existing networks (Saundry et al., 2012). It seemed unions largely created networks themselves rather than linking into existing networks. For this, the organiser was critical.

The importance of organisers as the most important contact person for delegates reinforced findings along similar lines from an earlier, larger survey of trained and untrained delegates (Peetz and Pocock, 2009). Interestingly, delegates still wanted more contact from organisers. Organisers were key to the links that delegates made with other delegates – and indeed outside.

We need to think of union delegate networks as somewhat different from many networks canvassed in mainstream network theory. Union delegate networks are often a result of conscious network-building. Differences in organisers’ engagement with delegates make a large difference to networks. When delegates talked about the barriers they faced, they focused upon what most easily saw – constraints on their time and other organisational issues. Not so visible to the participants is the role organisers play in passing on (or not passing on) the skills and understanding needed to develop networks in the first place.

We can think of two distinct types of networks to which delegates potentially belong: internal and external. Organisers are not key nodes between workers in delegate’s internal workplace networks. But they are the key to ensuring that delegates have the capacity to develop workplace networks. The focus groups and surveys provided conflicting information on the roles of delegates and organisers in formation of networks, but either way organisers appeared to initiate them less often than delegates. These inherently are delegate-centred networks, as organisers cannot maintain those strong ties with so many people. Weak ties between organisers and members, mediated (‘bridged’) principally through delegates, combined with strong (‘bonding’) organiser–delegate ties through which delegates receive training, support, resources and knowledge, enable delegates to build up their own networks of support among the membership, increasing the likelihood members will do things ‘for themselves’ and develop a sense of collective efficacy. Whereas in Granovetter’s (1973) study of job hunting, strong ties only provided information that jobseekers already had, strong ties in the workplace enable delegates to mobilise within the workplace.

In delegates’ external networks, to the extent they exist, organisers are a bridging node. Organisers can (but do not necessarily) ensure delegates have the ability and opportunities to meet equivalents from other workplaces and activists from different organisations. The ‘weak ties’ that delegates have with those outside the workplace can be as useful as the stronger ties within. In terms of earlier theoretical exposition on collectivism (Peetz, 2006, 2010), organisers are key mobilisers of union delegate networks. Delegates in turn are key mobilisers of worker networks, but dependent in part on the skills and opportunities provided by organisers.

Finally, delegates seemed reluctant to make use of social media for workplace networks. Preliminary indications are that this reluctance is likely to hamper network-building and utility, and the evidence from elsewhere is that social media hold much potential for unions (Aalto-Matturi, 2005; Diamond and Freeman, 2002; White, 2010). While the ACTU increases its profile in social media campaigning (Skulley, 2015), there is little indication of it being used effectively and consistently for local campaigns and networking. Given strong resistance, it requires considerable further research.

Conclusion

Networks for union delegates took a variety of forms. Only a minority were formalised. A majority were mainly internal to workplaces. Delegates sought extended networks, and they wanted to do things that would increase contact with members. Time and shift work were often mentioned as a constraint on delegates doing more in networking, and while this was partly an expectations effect that had limited predictive ability, it probably reflected a real constraint facing many delegates. Training was a major facilitator of networking. Social media were little used, with little intention of greater use; we suspect they are underutilised in light of calls in the literature and among some union leaders for greater use.

Effective networks are a combination of strong and weak ties. Sometimes these ties are so weak that delegates do not recognise they are part of a network. Hence, a slight majority of (trained) delegates in our paper survey did not see themselves as part of a network, although many had some support. In substance, delegates could participate in two distinct types of networks: internal and/or external. Organisers were important in providing the skills for delegates to develop internal networks, but by definition were not part of them. Organisers were also critical nodes bridging delegates to external networks with which they had weak but important ties. Organisers were the key support person for many (but not all) delegates, but delegates sought more interaction and more support.