The remainder of this section considers each causal theory in turn, with a narrative summarizing the evidence on the mechanisms operating within each, together with the contexts and outcomes. The narratives use consistent descriptors to indicate the relative strength of the evidence base: ‘strong’ is defined as from a range of studies, both qualitative and quantitative and spanning different service areas, countries or both; ‘adequate’ evidence is defined as a small number of robustly evidenced studies from a single country or a range of weaker evidence from across service domains or countries. Lastly, the descriptor ‘limited’ refers to one or two robust studies or to a larger number of weaker studies, but from a single context. The descriptors are intended to aid judgement as to the likelihood of an aspect of the theory being prevalent across policy domains or national contexts (Pawson et al. 2005). They should not be understood as indicating that a theory has predictive strength, nor as an assessment of the quality of the research.
The middle-classes and interest groups
The synthesis identified three kinds of mechanisms which appeared to underpin the overall theory that there was a relationship between middle-class collective activism and advantage in public service provision. The first related to the propensity of those with a higher socio-economic status to join groups and act collectively. The evidence for this mechanism was strong. The UK evidence demonstrates that middle-class people are more likely to be in groups and engage in formal voluntary activity (Egerton 2002; Li et al. 2003; Egerton and Mullan 2008). Evidence in relation to both the USA and England demonstrates how more geographically mobile middle-class mothers, in particular, rely on groups such as parents' groups to make friendship networks in new neighbourhoods (McGrath and Kuriloff 1999; Bagnall et al. 2003). There was also evidence that middle-class people are also more likely to join groups which are influential, particularly groups that have a direct say in policy, such as parent-teacher associations (PTAs) or parish councils. The evidence from England is adequate that rural parish councils – organizations often given official status as ‘consultees’– tend to be dominated by middle-class, educated men (Abram et al. 1996; Yarwood 2002; Sturzaker 2010). Thus, Abram et al. describe how, in a county with a large amount of middle-class rural in-migration, ‘[m]any of the village societies sought to use the planning system to shape the village in accordance with ideal images of the rural community’. Thus groups involved in these decisions became social networks representing ‘middle-professional, comfortable England’ (Abram et al. 1996: 361).
The second mechanism within the interest group theory relates to how middle-class groups behave. It suggests that the level of organizational sophistication in such groups, as well as the level of ‘noise’ they are able to produce in the policy process, explains their effectiveness. In land-use planning there is qualitative evidence that middle-class rural parish councils are vociferous, organized opponents of social or affordable housing within their villages and how they work to ensure it is retained for ‘local’ people (Abram et al. 1996; Yarwood 2002). Sturzaker quotes one parish councillor as saying:
In order to keep riff-raff from [nearby city] out of the community you need this s106 Agreement [planning policy] (Sturzaker 2010: 1014)
There is also qualitative evidence from England that middle-class communities are more likely to develop an ‘enraged response’ (Carroll and Walford 1996: 397) to school closures or local education policy and form groups to campaign against proposals (Bondi 1988; Carroll and Walford 1996). There is also evidence that they are effective – analysis of the process of drafting Regional Spatial Strategies in England shows those regions with a vocal, organized middle-class had a bigger reduction in housing targets as plans went through the formal consultation process (Sturzaker 2010).
These kinds of pressures are often dismissed as NIMBYism (not-in-my-back-yard). Although there is a broader literature on ‘bad neighbour developments’, such as polluting factories or waste facilities, particularly from the perspective of environmental justice in the USA (see, e.g. Bullard 2000; Faber and Krieg 2002), the focus has tended to be on whether and how these are concentrated in deprived neighbourhoods, not on whether middle-class communities actively resist them (Richardson et al. 2010; Faber and Krieg 2002). Like Yarwood (2002), Walker et al. (2010) provide qualitative evidence from renewable energy developers whose experience of interacting with middle-class NIMBYs leads them to ‘imagine’ an oppositional constituency through specific placatory strategies offering benefits to communities or neighbourhoods. In the case of rural housing, Yarwood (2002) reports one housing officer as saying that, although policies to keep affordable housing for ‘locals’ was not considered ‘good practice’, parish councils were not always told this as:
‘it was important to make parish councils think they have a say in housing allocation’ even though this was not the case in practice. (Yarwood 2002: 287, original emphasis)
For onshore windfarms in the UK, the provision of a community trust fund has become normalized as a way to offset local complaints (Walker et al. 2010).
A third mechanism identified in the research synthesis suggested that the social processes within middle-class groups facilitated the development of useful knowledge for group participants. There is strong evidence from both education and land-use planning demonstrating that the networking opportunities enabled by middle-class collective engagement enables the exchange of ‘soft knowledge’. For example, involvement in a PTA allows middle-class parents to engage in conversation at the school gate, providing deeper knowledge about the school, as well as access to information which might facilitate school choice (Ball 1993; Ball et al. 1995; McGrath and Kuriloff 1999; Bagnall et al. 2003; Crozier et al. 2008; Vincent et al. 2008; Archer 2010). That middle-class Black and minority ethnic parents, in the UK and the USA, often feel excluded from these networks adds qualitative weight to the evidence of their importance to parents (McGrath and Kuriloff 1999; Archer 2010). In land-use planning there is evidence that middle-class parish councils use experience of fighting development proposals to continue to resist development (Abram et al. 1996).
Overall, the evidence on the set of mechanisms identified as part of this causal theory is adequate or even strong. There are, however, evidence gaps in relation to the contextual factors which give these mechanisms their salience and in terms of the links between the mechanisms and eventual policy outcomes. However, the review suggests that the evidence base is at least adequate, and in some aspects strong, in terms of linking collective activity to the accrual of advantage for middle-class service-users, for example through ensuring school places, maintaining the ‘character’ of their village or generating a pre-emptive response for a service-provider. Interestingly, we found no evidence that these processes drive up service standards for a broader cross-section of service-users.
The middle-classes as individual consumers/activists with public services
The review hypothesis focused on middle-class activism, which suggests collective action. The literature search also found evidence the middle-classes can impact on public services as activist individuals via two main mechanisms. First, via co-production – the vast majority of public services are co-produced in some way, for example a good diagnosis requires discussion with a doctor. Second, individuals can be more active, complaining about things when they go wrong and working individually to ensure provision is tailored to their needs. The evidence which we have categorized as strong or adequate comes predominantly from the schooling and healthcare fields. The evidence from education tends to be located in the broader literature concerned with the reproduction of social inequality through education systems (see, e.g. the review in Ball 1993). Evidence from health takes its cue from the inverse care law (Tudor Hart 1971) but focuses on the micro-social relationships between practitioners and patients. There is a broader literature focused on how predominantly middle-class bureaucracies fail to engage effectively with poorer or working class individuals (e.g. Lipsky 1980; Wright 2003; Dubois 2010).
The ways in which middle-class people gain advantage through co-producing services can be theorized as either greater knowledge of problems and service-providers or cultural capital – meaning norms, behaviour, ways of speaking and deportment. The evidence for the latter is more limited than for the former. The evidence from health is strong that middle-class service-users are more vocal. Indeed, the evidence is adequate that this can make a difference to treatment or other interactions with health services. Thus being affluent leads to a greater likelihood of being treated as urgent rather than routine in heart surgery (Pell et al. 2000); and relates to other admissions rates – although the evidence is mixed over whether it increases your likelihood of being referred to a specialist (O'Donnell 2000). It also impacts on the nature and quality of information provided by doctors and others (Reid et al. 1999).
That this is due to middle-class patients' knowledge and social skills is often left implicit in analysis. There are a small number of studies in which health professionals allude directly to the impact on their decisions and practices of – for example – the threat of litigation (Somerset et al. 1999). Studies of GPs suggested that ‘[t]he patient's social status and ability to articulate verbally were put forward as tacit influences which affect the likelihood of referral’ (Somerset et al. 1999: 218; Mercer and Watt 2007). Qualitative evidence from midwifery also suggests middle-class mothers are more knowledgeable and better prepared to interact:
They will have like an A4 page of questions waiting for the midwife. So the midwife can't just go in and say, ‘Ah, that rash is nothing’. She has to explain what it is called, how long it will last, what colour it will turn, what cream to rub on, what cream to rub off, blah, blah, blah. (Hart and Lockey 2002: 491)
That the evidence is adequate that the ability of middle-class individuals to vocalize their needs leads to them getting better healthcare is concerning. However, the evidence seems to be mixed (O'Donnell 2000) and sits within broader debates on the redistributive qualities of universal healthcare systems, particularly the NHS in England, which has produced mixed evidence and shown that some redistributive efforts have been effective (Bramley and Evans 2002; Boyne et al. 2001). Our review suggests this debate needs to be supplemented with evidence on how services are utilized by different people on a day-to-day basis and the impact of this.
Similar evidence of active co-production has emerged from research on education. Evidence from Norway suggests parents with higher socio-economic status are more likely to attend parents' evenings than other parents (Baeck 2010). It is important to note this pattern has emerged even though parents' evenings are a relatively new phenomenon in Norway. Overall, it seems that while working-class parents may be willing to trust educationalists (Crozier 1997) the anxieties of the middle-classes create a much more hands-on, interventionist approach to education.
If there's something that's not good enough – and my friends' parents were the same – they'd all get together and complain (Archer 2010: 463)
This is part of a broader literature on the concern of middle-class parents to invest time and energy in their child's education. This may involve: accessing educational childcare and buying-in extra classes such as French lessons or baby gym (Vincent and Ball 2007); working to ensure that their child attends the school of their choice (Ball et al. 1995); ensuring their child is kept out of special educational needs classes and in classes for able or gifted and talented children (Crozier 1997; McGrath and Kuriloff 1999; Crozier et al. 2008); or their child gets specific publicly funded special educational needs provision (Archer 2010). This evidence blurs the line between active co-production of services and the active process of complaining to get a service tailored to suit needs.
A great deal of the evidence on complaining dates from studies in the USA in the 1970s and is therefore outwith the review, although some dates from the 1980s. In this literature two possible relationships between socio-economic status and complaining behaviour were theorized. First, a direct relationship – those with a higher socio-economic status had greater resources and so complained more. Second, a parabolic relationship – those who complained least would be those with low socio-economic status who did not have resources but had need, and those with very high socio-economic status who had resources but did not need to complain. Those who complained the most would be those in the middle, with the resources to do so and a need that made them reliant on public services. Research in Wichita, Kansas, supports the direct relationship, even though those with low socio-economic status have greater dissatisfaction with service provision. The evidence suggests this is because they feel less effective when they complain (Sharp 1982).
Evidence from US longitudinal panel data supports the importance of efficacy as a driver associated with socio-economic status, as children who are stable middle-class (from a middle-class family and have remained middle-class) felt more politically engaged and more politically effective (Walsh et al. 2004). More recent data from Norway adds greater context (Aars and Strømsnes 2007). This study shows that socio-economic variables had minimal impact on complaining behaviour and in multivariate analysis almost diminished entirely in their effect. The most important factor in explaining complaining behaviour was previous political engagement of some sort, and this was associated with increased feelings of efficacy and trust in institutions. The lack of strong links between socio-economic status and complaining was theorized as being due to continued class-partisan alignment in Norwegian politics which enabled issues to be voiced collectively. This suggests greater socio-economic inequality may mean middle-class people are more likely to complain, are more likely to get a positive response, feel effective and are more likely to complain in the future.
As with NIMBY pressures, this propensity to complain is recognized by service-providers in the education, environmental services and health spheres who pre-emptively respond to it. In education policy, qualitative evidence from the USA and the UK suggests head teachers and other education professionals welcome complaints from middle-class parents to keep them on their toes (Crozier 1997; McGrath and Kuriloff 1999). In a more negative way, in environmental services there is evidence (although it is as yet limited to a single study) that street sweepers changed their activities in response to middle-class complaints (providing a better, or personalized service) or changed their behaviour around affluent neighbourhoods to make cleaning more obvious (Hastings 2009b).
In summary, the evidence base for individual middle-class actors gaining additional benefits as a consequence of the way in which they interact with public services is generally adequate, and in some respects strong. Where it is more limited is on whether distinct cultural capital – how things are said as much as what is said – enables the middle-classes to gain advantages. There is more evidence of this theory in the next section.
The middle-classes as bureaucrats
Lipsky's (1980) analysis of street-level bureaucrats provides a theoretical basis for assuming frontline workers are likely to reflect the biases of wider society when making decisions and allocating resources. With the middle-classes there is the added dimension that many public service occupations are service, or middle, class (Butler 1995). Therefore, middle-class people might get preferential treatment as they are interacting with other middle-class bureaucrats who are likely to empathize with them (Gal 1998). This would appear to be a key mechanism within this third causal theory. As one interviewee in Crozier's (1997) research explained, in terms of their relationship with teachers:
It's not . . . a differential relationship because of our own background . . . professionals in our own right; so you know we do have a set of expectations about the standard of education that the school provides. (Crozier 1997: 194)
Because of this, there is (adequate) evidence that head teachers or principals will support middle-class parents within schools, particularly if they face challenges, as it is expected the parents will keep the school on its toes (McGrath and Kuriloff 1999; Crozier et al. 2008). Further context is provided by research into Black and minority ethnic middle-class parents (McGrath and Kuriloff 1999; Archer 2010). This shows that when faced with racism from service-providers, these parents will use their middle-class identity as they are aware that this will trump their race or ethnic identity to obtain the better service they expect.
Research in the health field, focused on the one-to-one interactions with professionals, demonstrates additional mechanisms which may produce advantage for middle-class patients. Studies from the UK show how health professionals take into account the social identity of patients when deciding treatment, particularly identities they can empathize with. Hughes and Griffiths provide the example of the husband of a potential patient who is a professor of pharmacology ‘a big wheel in stroke therapy’ and the work done by the consultant to provide some treatment, even though the patient is ‘in a category that would normally receive a low priority for admission’ (Hughes and Griffiths, 1997: 596). A less explicit benefit is identified in Hart and Lockey's research where midwives worked to help ‘Mrs Average’ and managers were ‘advocating for resources for women who were remarkably similar to themselves’ (Hart and Lockey 2002: 487). Neumann et al. (2009) demonstrate how ‘clinical empathy’ between practitioner and patient is likely to produce longer consultations, with more knowledge being exchanged and the patient being more enabled by the interaction (Mercer and Watt 2007). Importantly, empathy is produced in conditions of similarity rather than difference. If this is present throughout a bureaucracy, as Lipsky's (1980) study of street level bureaucrats suggests it will be, then this kind of behaviour implicated in the production of our fourth causal theory – that middle-class needs are normalized within policy and organizational processes.
Policies and organizational processes and the middle-classes
The final causal theory for which there is evidence is that policies and processes ‘normalize’ the needs of the middle-class and tend to cater for their needs. This is an important contextual factor for explaining the impact of the previous three theories and a causal theory in its own right. A great deal of the evidence here is from studies of those experiencing multiple deprivations, or from a working-class background, experiencing barriers and difficulties accessing public services (Duffy 2000; Dixon-Woods et al. 2005). This is supported by (the limited) research on environmental services in the UK which points to the importance of wider, stigmatized views of deprived areas (in which environmental problems such as litter are seen as a product of a cultural/behavioural pathology) as a mechanism which leads to such areas getting inadequate environmental services provision. The research contrasts the view that litter in non-deprived neighbourhoods is seen as normal, and deserving of adequate levels of service provision (Hastings 2009a). How this normalization is reflected in policies and processes is discussed below. Whereas the bias and changed organizational behaviour discussed in the previous sections is strategic and intentional, the type of bias produced by clinical empathy or the normalization of middle-class problems is socially embedded. Theoretically, it would therefore be much more difficult to challenge through organizational policies, systems and processes (Lipsky 1980).
At a macro-level, a larger body of evidence points to biases within resource allocation systems as another kind of mechanism which leads to the accrual of middle-class advantage. Julian LeGrand's econometric analysis of expenditure under the Conservative Government in the UK (1979–83) showed that those services most used by the middle-classes were protected from the expenditure cuts implemented as part of its austerity programme (Le Grand 1982; reviewed for this synthesis: Le Grand and Winter 1986). There is similar, more limited, evidence of similar expenditure patterns from the USA. Two separate studies on transport infrastructure (Boschken 1998) and education (Colburn and Horowitz 2003) demonstrate, on a state level, that areas with a higher concentration of high socio-economic status individuals spend more on high-technology transport infrastructure (light rail as opposed to buses) and education. In the latter case the relationship is slightly parabolic, as those areas dominated by people on very high incomes (over $75,000 per annum) have a reduced expenditure as people exit to the private schooling system.
As well as general patterns of expenditure, suggestive of unintended bias, policy analysis suggests a further mechanism in which some policies are explicitly tailored to the middle-classes or, through their implementation, benefit the middle-classes over other groups. The most apparent case of this is the move towards ‘choice’ in service provision in health and education. The evidence on schooling suggests a different facet to this. The qualitative evidence from England is adequate that middle-class parents have the most knowledge about the education market to get their children into the best school (Ball 1993; Ball et al. 1995). Further, the case study of Sutton Coldfield in the West Midland presented by Carroll and Walford (1996) demonstrated how wealthier parents can campaign to get policy changed when they do not get the school choice they want. Within the educational sociology literature there is debate as to the impact of the choice agenda and whether it has exacerbated spatial inequalities (see Gorard 1999).