Manajemen | Fakultas Ekonomi Universitas Maritim Raja Ali Haji 2003 1 (29)

PUTTING THE GROUP BACK INTO
UNIONS: A SOCIAL PSYCHOLOGICAL
CONTRIBUTION TO UNDERSTANDING
UNION SUPPORT
LEDA BLACKWOOD,* GEORGE LAFFERTY,** JULIE DUCK* AND DEBORAH TERRY*

I

ndustrial relations research that attempts to grapple with individuals’ union-related
sentiments and activities often draws on one of two traditions of psychological research—
the individual-level factors tradition (for example, personality and attitude-behaviour
relations) and the social context tradition (for example, frustration-aggression and
relative deprivation). This paper provides an overview of research conducted from
within these traditions to explain union-related phenomena and identifies some of
the limitations that arise as a consequence of a shared tendency to treat people in an
atomistic fashion. The paper argues for an understanding of the psychological processes
that underpin group-based action. To this end, it elaborates a theoretical framework based
on social identity theory and self-categorisation theory that would allow us to examine
the dynamic interplay between the individual, their cognitions and their environment.
The paper concludes with a brief discussion of a specific case of union mobilisation, to
indicate how this theoretical framework might aid empirical analysis.


INTRODUCTION
The decline in union density in Australia is, in the main, attributed to structural
change in the economy and associated shifts in employment into less unionised
industries, occupations, and types of employment status (Australian Bureau of
Statistics 2000; Department of Workplace Relations and Small Business 1997;
Griffin and Svensen 1996; Hartley 1995; Peetz 1998). A conservative Coalition
government has further eroded the position of unions through industrial relations
‘reforms’, including the prohibition of compulsory unionism and preference
clauses, the encouragement of individual contracts and enterprise bargaining, and
reduction in the powers of the Australian Industrial Relations Commission
(Davidson and Griffin 2000). In addition to contributing to real constraints on
union subscription and participation, these changes have been accompanied by
the rhetorical entrenchment of individualist values associated with free-market
liberalism (Griffin and Svensen 1996; Hartley 1995). This is reflected in what
some peak unions in Australia and overseas perceive to be a decrease in people’s
ideological motives for joining unions (Peetz 1998).

* School of Psychology, University of Queensland, St Lucia, Queensland 4072. Email:
leda@psy.uq.edu.au ** School of Management, Victoria University of Wellington, New Zealand.


THE JOURNAL OF INDUSTRIAL RELATIONS, VOL. 45, NO. 4, DECEMBER 2003, 485–504

486

THE JOURNAL

OF

I N D U S T R I A L R E L AT I O N S

December 2003

While acknowledging that macro-level issues have precipitated a decline in
union density, not all commentators agree with the sense of inevitability of this
decline that has been expressed in some quarters. In a comprehensive review of
the industrial relations landscape, (Peetz 1998) revealed that while many countries
other than Australia are experiencing a decline in union density, this experience
is not ubiquitous and is particularly severe in Australia. He argued that this points
not only to the depth and rapidity of structural change in this country, but also

to the struggle that unions have had in formulating a response to these changes.
Importantly, Peetz (Peetz 1998) emphasised that while macro-level factors
are the driving force behind changes in the environment within which unions
operate, unions are responsible for how they choose to respond to these changes
and can have an impact on the outcomes. From this perspective, an understanding
of the dynamic relationship between the macro-level factors governing the
industrial relations context and the micro-level factors influencing the psychological processes that drive employees’ responses to that context, should be
able to assist unions in thinking about and preparing their responses to the
challenges that lie ahead. With this in mind, the present paper provides an
overview and analysis of the main psychological contributions to understanding
this problem.
The psychological research that has looked at the reasons for people subscribing
to, participating in, and resigning from unions (Klandermans 1986c) has taken
two main forms: (a) research that looks for enduring, individual level variables
associated with collective action including support for unions (for example,
personality factors, attitudes/beliefs, and individual decision-making processes),
and (b) research that examines the social contexts of conflict between unions
and management (for example, frustration-aggression and relative deprivation
approaches). These traditional approaches are reviewed before turning to a
discussion of a more dynamic theoretical framework. In recent years, a number

of studies from an interactionist perspective have focused on the relationship
between the individual and society (for example, social identity theory and self
categorisation theory) in the area of collective action and in relation to unions
specifically. This latter area of research—perhaps unfamiliar to an industrial
relations audience—has the potential to sharpen our understanding of the
psychological processes that influence the dynamic interplay between the
industrial relations context and the actors involved.

REVIEW

OF TRADITIONAL APPROACHES

Personality factors
Personality focused research has attempted to show that people with an internal
locus of control (Rotter 1966), those with a sense that they can have an impact
on the political process (political efficacy: Campbell et al. 1954), and/or those
oriented towards collectivism (Smith and Bond 1993) are more likely to
become involved in collective action, including union-initiated action (Kelly
and Breinlinger 1996). Although we would expect that such individual-level
differences should have an effect on people’s readiness to act with others in the

pursuit of common goals, the evidence has tended to be inconclusive with such
factors proving of limited value in predicting who will engage in particular

PUTTING

THE

GROUP BACK

INTO

UNIONS

487

instances of collective action (Kelly and Breinlinger 1996).1 This is not surprising
when we consider that collective action, by definition, requires acting as a
member of a group, not as an individual, and thus what might be of more
relevance is a person’s perceptions of qualities of the group rather than of the
self. It is important that attributes of the individual that can contribute to agency,

continue to receive attention. However, these are attributes (e.g. perceived access
to relevant power, skills and resources) that are more specific to context and to
the action under consideration rather than enduring qualities of the person.
Attitudes and individual decision-making
Of greater interest than personality research, has been research that investigates
people’s attitudes towards unions and the antecedents and behavioural consequences of these attitudes (including subscribing, participating, and resigning).
Measures of attitude have largely been derived from a scale of union commitment
developed by Gordon et al. (1980), in which commitment is defined as (a) the
desire to maintain membership of an organisation; (b) willingness to make an
effort for the organisation; and (c) a belief in the values and goals of the organisation (Gordon et al. 1980, p. 480). Since union commitment has been conceptualised as an attitude (Sverke and Kuruvilla 1995), much of the work focusing
on the relationship between union commitment and union-related behaviours
has been informed, either explicitly or implicitly, by value-expectancy theories
of decision-making. The two most commonly used models are (a) one derived
from the value-expectancy theories of reasoned action (Ajzen and Fishbein 1980),
and its extension, the theory of planned behaviour (Ajzen and Fishbein 1980;
Ajzen and Madden 1986); and (b) Klandermans’ (1986c) value-expectancy
model. Klandermans’ (1986a; 1984a) model contains two elements: consensus
mobilisation, which is akin to consciousness-raising as the process by which union
members’ support for the union’s objectives is attained; followed by action
mobilisation, where members are persuaded that the benefits of participation will

outweigh the costs. It is the latter element that has received the most research
attention.
According to both the theory of reasoned action/planned behaviour and
Klandermans’ (1984b; 1986c; 1986b) model of action mobilisation, a person’s
intention to support his or her union will be influenced by his or her belief
that support will lead to a desired outcome, the expected reaction of significant
others, and the expected costs and benefits associated with the action. In their
influence on intentions, each of these beliefs and expectations is weighted by the
value the person places on them. In addition, the theory of planned behaviour
(Ajzen and Madden 1986) developed in response to the focus placed by the
theory of reasoned action (Ajzen and Fishbein 1980) on behaviours under
complete volitional control—proposes a role for individuals’ beliefs about their
personal ability to perform a specific behaviour in a particular context. It is likely
that perceived behavioural control (Ajzen and Fishbein 1980; Ajzen 1991) will
be of particular relevance when considering the constraints placed on unionrelated activity in some workplaces.
Value-expectancy models have had some success in predicting a range of
union-related behaviours including strike action (Brett and Goldberg 1979), union

488


THE JOURNAL

OF

I N D U S T R I A L R E L AT I O N S

December 2003

campaigns over work reorganisation, staffing agreements and the shorter
working week (Klandermans 1986c; Klandermans 1984b), and attendance at
union meetings (Flood 1993). Other research, however, has been more equivocal.
These models predict that if the costs of union activity outweigh the benefits—
for instance where the individual perceives their union to be ineffectual or
powerless, or where there are sanctions against being involved—then people
should withdraw their support for the union. This prediction, however, has not
been supported by empirical research in instances of, for instance, protracted
industrial strike action where employees appear to base their decisions on beliefs
about the justness of the cause (e.g. Fantasia 1988; Gouldner 1954; Turnbull et al.
1991).
There is now general support for there being two main routes to union

commitment and support—the instrumental route and the ideological route (e.g.
Freeman and Medoff 1984; Goldthorpe et al. 1968; Newton and Shore 1992;
Sverke and Sjoeberg 1995). The impact of the distinction between these two
routes to union support—and the perceived tension between the two—is evident
in divisions within the Australian trade union movement about strategic
direction. Specifically, these divisions are about the long-term sustainability
of a servicing model of unionism, where loyalty to the union depends on the
delivery of services (instrumental) versus an organising model of unionism, where
loyalty to the union depends on a belief in collective action. According to some
commentators (e.g. Hartley 1992), the evidence points to the two sets of beliefs
having quite distinct outcomes with instrumental or pragmatic beliefs being more
important in the area of union subscription, but ideology being more stable in
the long-term, and predictive of effortful forms of participation (Kelly and
Breinlinger 1996; Veenstra and Haslam 2000).
Value-expectancy models have an intrinsic, instrumental bias in the way
that they treat people’s beliefs. For this reason, some researchers have
claimed that the substantial body of research framed by these models does
not adequately capture the role of ideology. Specifically, there is a concern that
because most people prefer to present rational rather than emotional reasons
for their actions, what is actually being tapped is people’s rationalisations for

their behaviour. In support of the view that research may not be capturing
the role of ideology, Peetz (1996) cited a 1990–91 survey in which the most
common reasons for joining a union given by 910 Sydney employees were
instrumental (57 per cent, compared with 5 per cent ideological). However, in a
subsequent logistic regression of the data, Peetz (1996) found that union
sympathy (ideology) was the variable most strongly associated with union
subscription. Furthermore, the sharp distinction that has been made between
the two sets of beliefs can be questioned, particularly in terms of the assumption
that people can be neatly divided according to which set of beliefs they support.
Based on her longitudinal study of union activity across five workplaces,
Fosh (1993) concluded that the conception of unionism held by most members
of a union will contain elements of both instrumental and ideological beliefs,
and that their relative importance will change depending on what is happening
in the workplace.

PUTTING

THE

GROUP BACK


INTO

UNIONS

489

Social influence
Contextually related changes in the valence of people’s beliefs is a reminder
that collective action involves the individual thinking and acting in concert
with others, and thus raises the need to consider how others influence our
thoughts and behaviours (Hartley 1992). Support for social influences on
union-related attitudes and behaviours—particularly joining behaviour—has
been found in relation to four predictor variables (a) family background;
(b) workplace social identity; (c) social networks in the workplace; and (d) extent
of social approval from managers and colleagues (e.g. Bain and Price 1983;
Fullagar et al. 1994; Fullagar et al. 1995; Kelloway and Newton 1996; Youngblood
et al. 1984).2
In the theories of reasoned action and planned behaviour (Ajzen and
Madden 1986; Ajzen 1991; Fishbein and Ajzen 1975), social influence is
operationalised as the subjective norm—the person’s assessment of the
likelihood that relevant others such as co-workers expect the person to support
the union. This is weighted by the person’s motivation to comply with these
expectations. The social motives component of Klanderman’s (1986c) action
mobilisation model maps directly on to the subjective norm. The proposed
role of the subjective norm/social motive has received little support in
research on union-related behaviours (e.g. Kelloway and Barling 1993;
Sverke and Sjoeberg 1995). Although this lack of strong support is consistent
with general findings from the theory of reasoned action and planned behaviour
research (see Ajzen’s 1991 meta-analysis), it is counter-intuitive and contrary
to the findings reported above from the wider social psychological research
into unions.
The lack of empirical support for the role of social influence can be attributed
to a number of limitations in the treatment of norms in the value-expectancy
models of attitude-behaviour relations (Terry and Hogg 1996; Trafimow and
Finlay 1996). Specifically, (a) the description of the ‘subjective norm’ in terms
of the extent to which people perceive that others want them to perform the
behaviour, rather than as the perceived attitudes and behaviours of others who
are contextually relevant (see Brown 1988); and (b) the assumption that the
normative and attitudinal components are based on different belief structures
and, therefore, are cognitively independent of each other (Liska 1984). Contrary
to this latter view, structural equation modelling has revealed ‘crossover’
relationships (Oliver and Bearden 1985) between attitudes and beliefs about
relevant others’ attitudes (normative beliefs) suggesting that norms may be
influential in the formation of attitudes (Terry and Hogg 1996; Vallerand
et al. 1992).
The social influence research leads us to the second psychological research
tradition that has contributed to the study of union-related phenomena: research
that looks at the social-contextual factors influencing collective action.
Frustration–aggression theory and work-related dissatisfaction
Research that has looked at the influence of the broader social context on
union support has tended to be focused on industrial action—mainly strike

490

THE JOURNAL

OF

I N D U S T R I A L R E L AT I O N S

December 2003

activity. A widespread public belief is that frustration, dissatisfaction, and/or
alienation lead to strike activity. One psychological approach looks at internal
psychological processes and identifies frustration as the cause of participation in
collective action through trade unions. According to the frustration-aggression
hypothesis, people strive for a state of intra-psychic equilibrium and thus, in
work-related situations where they feel frustrated, dissatisfied or alienated, they
are driven to restore their sense of equilibrium by participating in trade union
activities. Although there is some support for a relationship between job
dissatisfaction and union-related behaviours, overall, the findings have been
inconclusive (DeCotiis and LeLouarn 1981; Farber and Saks 1990; Getman
et al. 1976; Guest and Dewe 1988; Kochan 1980; Schriesheim 1978; Snyder
et al. 1986; Youngblood et al. 1984). Participation in trade union activity is
not the only course of action that a person can take to combat feelings of
frustration or dissatisfaction (Bluen 1994). Thus, inconsistent findings may
be the result of some employees either suppressing their discontent or
expressing it in forms other than support for the union (Klandermans
1984b). Furthermore, strike decisions often arise in the context of tensions
between employees and management over employment conditions that may not
be related to employees’ general sense of (dis)satisfaction with their job or
their workplace.
Notwithstanding the above limitations, there are sufficient real-world
examples to suggest that dissatisfaction will, in some circumstances, underpin
collective protest such as union activity (see Premack and Hunter 1988). In order
to understand what these circumstances might be, we need to turn to theoretical models that move beyond intra-psychic processes to look more at the nature
of the social context. One of the most influential of these social context models
has been relative deprivation theory.
Relative-deprivation theory
The focus of relative deprivation theory is the individual’s perception of inequality
or deprivation in interpersonal and intergroup relations (Runciman 1966; Walker
and Pettigrew 1984). According to this model, when a person experiences
personal relative deprivation (relative to other individuals), this will lead to an
individual response and where a person feels collective deprivation (relative to
other groups), this will lead to a collective response. Thus, what underpins
collective action is the sense of inequality or deprivation experienced by
individuals as members of a group (see Walker and Pettigrew 1984). Research
reported within the prejudice literature (see Abeles 1976; Tripathi and Srivastava
1981; Vanneman and Pettigrew 1972), has confirmed that collective, rather than
personal, relative deprivation has the greater impact on participation in collective
action. This impact, however, is thought to be dependent on three critical
factors: (a) there is a categorical, intergroup focus; (b) the person identifies with
the group; and (c) the person attributes causation for the relative disadvantage
of his or her group to external factors. Similar factors have been identified in
other models, including Klandermans’ social movement participation model
(Klandermans and Oegema 1987; Klandermans 1997).

PUTTING

THE

GROUP BACK

INTO

UNIONS

491

Summary
Research into union-related sentiment and activity has been reasonably fruitful
and much is now known, or suspected, about the personal and environmental
factors that have an impact on these attitudes and behaviours. For example,
people’s instrumental and ideological attitudes, their sense of agency (both
personal and collective), their perceptions that there are grounds for action
and that others support such action must all be implicated. Additional factors
not elaborated here include aspects of social movement organisations such as
access to resources, existence of networks, relations of respect and trust between
members. However useful, running through all this research is a tendency to
treat people in an atomistic fashion. As a consequence, when taken alone,
they contribute little to understanding changes in union support across place
and time.

INTERACTIONIST

PERSPECTIVES

Social identity
Collective action is contingent on seeing the self as part of a group. A recurring
theme in previous research is the failure to account for the sense of group
solidarity that arises in particular contexts and how this experience of solidarity
implicates group-level processes of social influence. In accord with Klandermans’
(2002; 2000; 2002) more recent work, we believe that identity can disentangle the
knot of collective behaviour explanations. There is now considerable evidence
for the effect of group identity on union-related outcomes. Allen and Stephenson
(1983) found that increased identification with either union or management
led to a more stereotypic perception of the relevant outgroup. A number of
studies have found support for a relationship between union identification and
preparedness to participate in union activities (Frege 1996; Kelly 1993; Veenstra
and Haslam 2000). Furthermore, in comparison to other social psychological
variables, such as political efficacy, collectivist orientation, perceived intergroup
conflict, outgroup stereotyping, and relative deprivation, Kelly and Kelly (1994)
found that union identification was the best predictor of engagement in union
activity, and the only predictor of engagement in difficult forms of union
activity (e.g. speaking at a meeting). The above point to a direct relationship
between identity and collective behaviour. We propose that identity also has
an interactive effect with the variables that have thus far been discussed in
this paper.
Social identity and self-categorisation theories
The above findings are important in confirming what one would expect—the level
of attachment people have to their union influences their union-related attitudes
and behaviours. Thus, an integrative approach is required. Social identity and
self-categorisation theories are well positioned to provide this by allowing us to
consider the dynamic interplay between the person, their cognitions and their
environment. There are two basic premises of social identity theory. The first is
that an important part of our sense of who we are and of our self-worth derives
from the groups to which we belong. The second is that these groups exist in

492

THE JOURNAL

OF

I N D U S T R I A L R E L AT I O N S

December 2003

relations of status and power to one another, and that because we derive our sense
of certainty and self-worth from our group memberships, we are motivated to
defend and enhance group status (Hogg and Abrams 1995). In the remainder
of this paper we explain how an understanding of the processes of selfcategorisation that underpin social identity can improve our analysis of union
support. We present propositions that can act as a guide to empirical research
and conclude with an illustration of theoretical application by way of a current
case study.
Social influence revisited
At a fundamental level, a group processes approach puts social influence centre
stage. In an extension to social identity theory, self-categorisation theory, Turner
and his colleagues (Turner et al. 1987) elaborated a single process model of social
influence in groups, called referent informational influence (see Abrams and Hogg
1990; Turner 1991). According to this model, when a valued group membership
that is attitudinally-relevant (e.g. the union or a group of fellow employees)
becomes salient, the process of self-categorisation results in a person’s thoughts
and behaviours being more in line with the group rather than with unique
properties of the self. It is this process that is thought to underlie collective
action (Kelly and Breinlinger 1996).
The referent informational influence model represents an important
departure from the more traditional models of social influence. This model is
distinguished by its emphasis on a single process of social influence based on selfcategorisation, rather than on the specification of dual processes of normative
and informational influence that traditional models suggest result, respectively,
in public compliance and private acceptance (Deutsch and Gerard 1955). For
instance, van der Vall’s (1970) explanation for approximately one-third of the
workers in his study identifying the influence of others in the workplace as a
reason for joining the union, was that these workers wished to avoid being seen
as parasites—that is, that they were simply displaying public compliance. In
contrast, referent informational influence specifies that ingroup normative
information is a source of both normative and informational influence and, for
this reason, it should produce more than simply compliance—even in the
presence of coercion, there should be a change in the person’s attitudes. This
explanation resonates with union movement experiences of the impact that
workplace cultures appear to have on workers’ more enduring attitudes towards
unions (Peetz 1998; Hartley 1995).
There are several factors that influence the referent informational influence
process. Most importantly, social psychological research suggests that social
influence will be greatest when (a) the individual identifies and therefore expects
to be in agreement with a group; (b) the individual experiences subjective uncertainty about their own perceptions or attitudes; and (c) the norms of a group are
unambiguous (Hogg and Abrams 1995). In relation to the second of these
factors, although people who have had a long association with a workplace (and
possibly some experience with their union) are likely to be reasonably confident
in their union-related attitudes, those with more limited experience are likely to

PUTTING

THE

GROUP BACK

INTO

UNIONS

493

be less certain about their attitudes and hence particularly susceptible to the
influence of their work colleagues. In relation to the third factor, research into
public opinion suggests that around highly contested values such as individual
versus collective action, there can be considerable variation in individual group
members’ or subgroups’ subjective representations of what is the normative position of the group (e.g. Giles et al. 1977; O’Gorman 1986). Thus, the employee
group’s normative position may be difficult to discern—particularly in a broader
political and economic environment that arguably has a very strong norm in
favour of individualism, and in a context of structural change that has produced
increasing fragmentation of the workplace.
Proposition one: When a relevant and valued group membership (e.g. union
or employee membership) is contextually salient, people who strongly identify
with the group will expect to be in agreement. Where they are not, they will be
motivated to bring their attitudes and behaviours in line with the group norm
(social influence will occur)—particularly where they are uncertain about their
own beliefs and the norm is clear. Where people do not identify with the group
or are strongly opposed to the group norm, they may choose to distance
themselves from the group and behave according to individual considerations
(including cost-benefit analysis and response to coercion).
Conditions for the activation of group processes
The above discussion implicates both individual-level variables and social context
in the activation of psychological processes of social influence. For a particular
social identity such as union member or employee to become activated a
number of conditions must first be met (Turner et al. 1987). To begin with, a social
category must be accessible to the individual—for example, if being an employee
or, indeed, a unionist is an important part of a person’s sense of self, he or she
will be likely to think and act in terms of that particular social category.
Contributing to this accessibility might be individual-level factors such as
previous experience with unions and pre-existing attitudes and beliefs.
Furthermore, there must be a match between a social category and the perceived
situation (Oakes 1987; Oakes et al. 1994). In a specific industrial relations context, this might mean that similarities within the employee group and differences
between employees and management must be consistent with a person’s
normative beliefs about the social meaning of the employee and management
categories and their relationship (Oakes et al. 1991).
In relation to the above conditions, there are three important criticisms of much
of the work that has examined the relationship between group-level identity
and union support (see Turner 1999). The first criticism relates to researchers’
tendency to operationalise identity as a chronic individual difference variable,
rather than as the degree to which a person’s identity is activated (e.g. Brown
et al. 1986; Kelly and Breinlinger 1996; Veenstra and Haslam 2000). While it is
important to know whether being an employee or a unionist is an important part
of a person’s sense of self—because this will help predict their preparedness to
self-categorise as such—it is the actual process of self-categorisation that leads
to social influence and results in people acting as group members. The second

494

THE JOURNAL

OF

I N D U S T R I A L R E L AT I O N S

December 2003

criticism relates to the tendency to focus exclusively on ‘union’ identification,
which has obscured what is important in terms of people’s perception of the intergroup context of union activity. Although unions are certainly organisations with
which some people will strongly identify, they are at the same time a vehicle for
collective action by employees. Given the importance that work-based identity
has in our society, shifts in support for a union may be less a consequence of selfcategorisation as a union member and more a consequence of categorising
the self with other employees in an occupation or industry. Finally, the third
criticism relates also to the assumption of context and the tendency for
researchers to focus exclusively on the path from social ‘reality’ to social categorisation (Reicher and Hopkins 2001). This ignores the active and purposeful
framing of social ‘reality’ that is central to political projects. A person’s decision
to pursue collective action through their union is dependent on the perception
of an ‘us’ versus ‘them’ relationship.
The above discussion indicates the difficulties inherent in field research
and the need for researchers to be more attuned to individuals’ perceptions
of the intergroup context. For instance, when looking at the potential role of
employment-related identity a consideration is the gradual shift away from
occupationally homogeneous, integrated communities that traditionally have
been associated with greater union support (Brett 1980). The greater tenuousness
and transience of work with increasing casualisation and outsourcing across
industries, has arguably contributed to many employees feeling little sense of
identification with their workplace. Similarly, the creation of ‘superunions’ that
cross over occupations and/or industries has the potential to leave some groups
of employees feeling marginalised. The point here is that it is not simply
physical distance (i.e. impediments to contact between unions and potential
members) but also psychological distance (i.e. the lack of a sense of shared
identity) that can be the challenge for unionists. Thus, it is important to assess
whether people respond to the categories or groups presented to them, and
whether they see themselves as part of the dominant group or as peripheral
to this group (e.g. part of a subgroup defined by a particular occupation or
employment status).
Proposition two: Contextually relevant groups that can form the basis for
self-categorisation do not exist prior to or outside of context. Researchers (and
unionists) can sharpen their investigation of contextual factors by being
clear about how this context is both represented and experienced by the actors
involved. Specifically, dimensions that will add clarity are (a) what are the relevant
categories or groups as framed by the various agents (e.g. the union may
frame an industrial dispute as being between employees and management while
management may frame the same dispute as being between the industry and
government); (b) what is the prototype for the group (i.e. who are typical
members/what is the normative position); (c) what are the parameters of
the groups (i.e. who are included and who are excluded); and (d) how is the
individual placed within the groups (i.e. ingroup versus outgroup; core versus
peripheral member). The central point here is that how people read the context
and their placement within it has a powerful effect on whether they self-

PUTTING

THE

GROUP BACK

INTO

UNIONS

495

categorise and hence, whether their response will be directed by psychological
processes operating at the group or at the individual level.
Social context and the role of perceived threat
In the current economic and political environment, many Australian workers
feel personally threatened in terms of job security and general levels of change
(Griffin and Svensen 1996; Peetz 1998; Peetz 1996). At the same time, the
ability of workers to bargain collectively through their unions is being undermined (Davidson and Griffin 2000; Griffin and Svensen 1996). There is considerable evidence that points to an increase in union support around disputes
between employees and management over wages and conditions (Bluen 1994;
Turnbull et al. 1991). Intuitively, this makes sense—we would expect that a
‘rational’ person would support their union if it was defending their conditions.
It is less easy to account for instances where threat to employees’ rights to
organise has lead to increased support for the union. One explanation, based
on the ‘rational’ actor model, is that under such circumstances support for the
union becomes more important as a form of collective insurance (Peetz 1998).
Another explanation, which does not necessarily obviate the first, is that the
increased support for unions under such conditions is, for some, an expression
of solidarity in an intergroup context of threat.
Various studies have found support for the proposition that perception of
group threat leads to increased self-categorisation, including studies examining
threat to the identity of women (Dion 1975), Jewish males (Dion and Earn 1975),
and national identity (Branscombe and Wann 1994). The implication is that
perceptions of group threat should, in turn, lead individuals to exhibit greater
conformity to group norms. Considered in these terms, the relationship between
perceptions of threat to the union or to employment conditions and engagement
in union activity needs to be understood in terms of a process whereby people
self-categorise on the basis of a group membership that has a normative position
supporting the union. Rather than a straightforward relationship between threat
and support for the group, Spears, Doosje and Ellemers (Spears et al. 1997) proposed that threat interacts with levels of (enduring) identification. For instance,
a person who identifies highly with a group is likely to show increased solidarity
when the group is threatened, whereas a person for whom group membership is
less important may choose to distance him or herself from the group and act
more as an individual.
Although there is some support for the above arguments (e.g. Doosje et al. 1995;
Karasawa 1991; Kelly 1989; Spears et al. 1997), findings from union research
suggest that the picture may be more complex. For instance, in a recent study
Veenstra and Haslam (2000) found that low-union identifying union members
distanced themselves from their union in an experimentally manipulated conflict
context, but showed increased support for their union in an experimentally
manipulated threat context. These findings are consistent with the industrial
relations literature that suggests that, for the most part, union members prefer
their union to have good relations with management, but when wages and
conditions are on the line (i.e. in times of threat) they do not want their union

496

THE JOURNAL

OF

I N D U S T R I A L R E L AT I O N S

December 2003

to be acquiescent (see Peetz 1996). The finding that low union-identifiers show
heightened support for the union in the threat condition is also consistent with
Fosh’s (1993) longitudinal research in which she found that in a workplace where
there was traditionally low union support (and by implication, weak, enduring
union identity), 400 to 500 out of a total 840 employees turned out for a special
branch meeting when there was a crisis in employee–management relations. The
question that both Fosh (1993) and Veenstra and Haslam (2000) posed is whether
such increases in support—coming from a low support base—are a consequence
of people pursuing individualistic strategies to protect personal interests (i.e. where
‘bailing out’ is not considered an option) or group strategies to protect collective
interests (i.e. where threat heightens the salience and self-relevance of a unionbased category and increased support results from social influence). What the
Fosh (1993) and the Veenstra and Haslam (2000) studies demonstrate is the
considerable fluidity in people’s union activity across contexts. However, it is not
simply the ‘objective’ presence of intergroup threat that is important, but how
people interpret that threat.
Social context and the role of social-structural beliefs
For industrial relations researchers the idea that socio-political context matters
for union support is a truism and the description and interpretation of structural
relations, when attempting to explain instances of collective action, represents
the more familiar level of analysis. From a social-psychological perspective,
structural factors likewise are acknowledged, but the focus of analysis moves
to the level of people’s subjective representations of a contextually specific
intergroup relationship. Social psychological research examining contextual
factors that promote collective action as a strategy has found that a crucial
factor is a shift in causal attributions for lack of success from internal attributions
(e.g. insufficient ability) to external attributions (e.g. economic factors or
illegitimate practices). Thus, what contributes to collective action is the belief
that one’s position is a function of one’s membership of a particular group and
not of something intrinsic to the self.
This shift from individual to group-based processes has found support—mainly
in laboratory studies—and is consistent with the explanations offered for the mixed
results from studies exploring the effects of locus of control and political efficacy,
and with research into collective relative deprivation. It is also reflected in
how the social movement participation model conceptualises the process by
which people become part of the mobilisation potential. According to this
model, people must first share a ‘collective action frame’ for the movement’s
cause (Klandermans and Oegema 1987; Klandermans 1997). The three components of the collective action frame are: (a) explanations in terms of injustice;
(b) identity (us versus them antagonism); and (c) agency (social change is
possible). The ‘collective self’ is seen as playing a facilitating role in all three
components (Smith et al. 1994).
Although sharing similar elements, social identity theory proposes a more
elaborated model. What is important to how a person will perceive an intergroup
context of threat, for instance, is their beliefs concerning the nature of society

PUTTING

THE

GROUP BACK

INTO

UNIONS

497

and relations between groups within it (i.e. their subjective belief structures). Two
broad types of subjective belief structure are social mobility and social change.
Social mobility reflects the dominant myths (e.g. belief in a ‘just world’) and leads
to individual action to improve one’s personal position. Social change ideology
is one that we would associate with group-based action, such as through a
union. In accord with the social movement participation model, crucial factors
thought to influence which strategy people adopt are: (a) their strength of
identification with a behaviourally-relevant group (identification), and (b) their
causal explanation for the position of employees (perceived legitimacy). Social
identity theory proposes additional factors which are: (c) whether people
perceive that the boundaries between employees and employers/management can
be transgressed (permeability, e.g. through exerting influence in the workplace),
and (d) whether they can conceive a cognitive alternative to social-structural
relations between employees and employers/management (stability [Hogg and
Abrams 1995; Turner 1999]).
It is assumed that people with an ideological commitment to their union
have a social belief structure that supports social change through collective
action—that is, they are likely to perceive that structural relations between the
employee group and management are illegitimate, the intergroup boundaries are
impermeable, and power differences are unstable. For this reason, these people
are primed to read a context involving any level of conflict or threat in such a
way as to activate their union identification. That is, the perception of conflict
or threat should lead almost automatically to alignment with the union. At the
other end of the spectrum, there are those who have a deeply held ideological
opposition to unions and who are, therefore, unlikely to support the union, regardless of the provocation to do so. For those employees (arguably the majority)
who do not hold strong ideological positions either way, their relationship with
the union might be conceived as more ambivalent. It would therefore be expected
that for this group, changes in the local industrial relations context—and how
this context is subjectively interpreted—might be particularly important in respect
of their preparedness to support the union.
Although social identity theory treats social change (collective) and social
mobility (individual) strategies as distinct responses, field studies of women
(Breinlinger and Kelly 1994) and gay men (Cox 1998) have indicated that it is
not possible to separate people according to these strategies. In reality,
people can hold beliefs that support both social mobility and social change
strategies—thus, the distinction needs to be made between micro- and
macro-social perceptions in a particular context, rather than between people.
For instance, it is conceivable that an individual can believe that he or she
personally has the ability to successfully pursue social mobility in the workplace,
while at the same time having an ideological commitment to social change
through collective action via the union. That is, people who believe that they
personally can strike a better deal for themselves (permeability) or that the
union is likely to be ineffectual in protecting employees’ interests (stability)
may nevertheless be moved to support the union on principle (legitimacy).
This points to the need to return to ideological factors—considered here as

498

THE JOURNAL

OF

I N D U S T R I A L R E L AT I O N S

December 2003

people’s beliefs about the social-structural relations between employees and
management—and how this impacts on self-categorisation in an industrial
relations context.
Proposition three: Support for one’s union is not the only strategy that a person
who perceives employment-related vulnerability or threat can pursue. For
instance, a person might pursue social mobility strategies, either moving up or
out of the organisation, or they might psychologically disengage from their
work-place and from their colleagues. How individuals respond will be affected
by whether their social-beliefs about the context implicate group level or
individual level processes. Thus, contextually specific perceptions about (a) the
justice of the intergroup relationship (legitimacy); (b) the ability to transgress
group boundaries (permeability); and (c) the ability to imagine an alternative
scenario, namely the intergroup relationship (stability) are important factors.
The degree to which a person identifies with an ingroup in this relationship is
thought to moderate the influence that these beliefs have on union support. The
complexity of the interrelationships between these factors and how they are
mediated in the real world still requires considerable investigation.
Coming full circle—or individual-level variables matter, but they are constructed in
dynamic social contexts
How people perceive their world is of course influenced by the experiences,
thoughts and feelings that they bring to bear. For instance, we would expect
that some people would be ideologically predisposed to read the employment
relationship as an intergroup context requiring collective solutions. People’s
pre-existing commitment to their union—comprising their previous behaviours,
union-related beliefs and attitudes and their enduring identification with their
union—influences their preparedness to support their union, both directly
and indirectly via their reading of the context. For those with high levels of
union commitment, increased support for the union should follow as a direct
consequence of being ‘primed’ to self-categorise with the union. Here we find
that the union category is potently accessible. The environment is experienced
as saturated with cues to union category membership and the normative position
of this category is unambiguously in favor of union support. Hence, such individuals are relatively easy to mobilise. For those who are more ambivalent about
the union, possible readings of the context are more varied and other factors
will come into play. Here, support may occur as a consequence of instrumental
calculations at the individual-level or of identification with another relevant
ingroup (e.g. co-workers) where the norms of that group indicate support for
the union. Those factors that prior research has found contribute to enduring
union commitment or support (i.e. previous behaviours and ideological and
instrumental beliefs) remain important to investigate. So too, factors bearing
on personal and collective agency and specific expectancies in the particular
context. How each of these factors influences union support in the particular
context, however, is a function of whether people’s collective self is engaged (i.e.
they are self-categorising at a group level) or whether they are disengaged and
operating at a more individual level.

PUTTING

THE

GROUP BACK

INTO

UNIONS

499

To illustrate the complex interrelationships between individual, group and social
level factors we will describe the unfolding of a recent dispute. It concerns the
termination of employment of the National Tertiary Education Industry Union
(NTEU) Branch President at an Australian university. There had in recent years
been several conflicts between the university management and the union. For
example, in the previous round of enterprise bargaining, management had
attempted to introduce a non-union agreement (which went to a ballot, but
was defeated). The Branch President had also been outspoken on several
controversial issues in previous months.
Yet immediately prior to the termination, performed by a middle manager,
relations between the union and senior university management had appeared
reasonably harmonious. Therefore, the salience of the employee and/or union
group memberships—and by implication, the salience of collective goals—had
been relatively low for most of the union’s constituency. As a consequence, selfcategorisation on the basis of either of these group memberships was unlikely,
and we would expect that individuals’ support for their union was driven by past
behaviours and by individual-level calculations that included both ideological and
instrumental attitudes. However, this situation was transformed quickly, once the
NTEU represented to its constituency its understanding of the circumstances
of the Branch President’s unexpected termination of employment. There
emerged a context in which union and employee group memberships became
potentially more salient. We will not discuss the merits of the respective
positions here, but will concentrate on the impact upon union mobilisation.
The situation developed into a clear-cut union versus management dispute, as
senior management supported a decision originally made by a middle manager.
The union developed a campaign oriented around core values, in particular
free speech and freedom of association, coupled with a rejection of unrestrained
managerial prerogative and the potential threat to the job security of all
employees. The appeal to core values proved particularly successful and became
the main theme of the campaign, which was centred on a series of three
meetings/rallies, each of whi