Sociological Research Online

Markets in Education and the UK Experiment

Stephen Gorard - Extract From Phd Thesis


This chapter examines the trend towards increased 'consumer' choice in school education, considering arguments both for and against this as a policy, in the context of recent changes in this respect in the UK. Its objective is to establish the need for further work in the area of school choice, and to explain the difficulty of predicting the likely outcome of marketisation for the UK state-funded school provision. The chapter is in three main parts, starting with an outline of some of the varied arguments that have been used to promote choice in education. There have been at least three different strands in the push for greater 'consumer' choice. One of these stems from the view that market forces will have a beneficial effect on the provision of what is seen as public service industry (James and Phillips 1995). Another is that allowing greater freedom of choice will end the reality of selection by mortgage, and give even the most disadvantaged sections of society a fair chance of using the best schools. The third strand is, perhaps, the simplest of all. It does not matter what the intended effects of a choice programme are, choice is a good thing by definition, according to views such as the libertarian perspective of Erickson (1989). Since the state only intervenes in the home life of a child - its clothing, shelter, and food - in case of neglect or abuse, it should behave in the same way with education. This is then followed by discussion of some of the criticisms of these arguments, and the chapter concludes with a summary of economic rational choice theories, and their relevance to the marketisation debate.

Choice in Education

There is some evidence that recent increases in population movement, have led to more changing between schools in many children's lives (Van Zenten 1995), so there are more choices of school to be made than in the past, and choice of schools looks set to become more and more important in the future. There are good theoretical arguments both for and against increased control of school choice for the individual, and it is unlikely that the debate on this issue will be decided by theory alone. Some education writers quite reasonably take a political stance on this issue but many of the reasons given, even, or perhaps especially, by politicians for the recent legislation in the UK, are educational. In this light, the recent reforms in the UK, as elsewhere in the world, can be seen as a massive educational policy experiment, although the UK is perhaps in a minority in using an entire generation of the nation as its 'guinea pigs'. It is too early to tell what the full effects will be, but it is beginning to look as though the reforms are likely to be both less effective than originally suggested by some, and less damaging than feared by others.

Although this chapter is chiefly concerned with parental choice of state-funded schools in the UK, several other liberal democratic states have introduced legislation relevant to increased school choice in the last 15 years (OECD 1994). These schemes are all different in important respects, and highlight the need to distinguish between market theory as a general concept, and each version of its implementation in education. In the USA, choice of school can mean selection from several mini-schools on one site, or between educational programmes affecting entire states (Levin 1992). Many of the supposed advantages and disadvantages of market forces that have appeared in the UK can be seen as arguments for or against the specific set of legislation enacted here, rather than dealing with choice per se.

All school choice schemes, however, have given the de jure role in choice to the parents alone, without much obvious consideration of the alternatives. Although parents may be beneficiaries of the custodial role and income-generating potential of their child's education, and the majority will genuinely have the best of altruistic intentions, it does not necessarily follow a priori that they have the necessary competencies to decide on an appropriate school. Should parents really have the right to select a school that protects their minority sub-culture, and so deprive the child of a more normative education? Should they be able to select a school based around a narrow and fundamentalist religion, and risk the charge of indoctrination? And what happens if parents do not value any kind of schooling? Children from rural families used to be regularly kept away from school, during harvest and other crucial agricultural periods, by their parents, as evident from this complaint by an earlier Amish farmer, 'once he goes to school and gets a little education, he's no good on the farm any more' (Coleman 1990 p. xii).

Closer to home, pressures to keep children out of school, or to leave early, are not merely due to poverty, and the need for the child to help in the family business. In Wales, one researcher claimed that ' the main reason was the lack of appreciation by parents of [its] importance for the future welfare of their children' (Evans 1971 p.262). Another pressure comes from not wishing to over-educate children, or prepare them for a type of job that would require leaving a rural area. Truancy, still worse in Wales than England (Reynolds 1990), is perhaps condoned, and maybe even encouraged by some parents. Where do they fit into an economic model of rational choice theory, and can a free market be fair to their children?

'Life contingency' is a concept originally used in occupational choice, to explain why people may make a choice for reasons other than its manifestly desirable qualities (Ginzberg et al. 1951), but it can be seen to apply to the much newer field of school choice. Thus debate about the value of school choice is really about whose values should apply, and can only be answered by deciding on an answer to the preliminary question - what is the problem that choice is a solution to? A large number of reasons have been used to argue for, or justify, the increase of parental choice of schools. A few are strictly educational, some are economic, but many are based upon notions of justice, equality, or equality of opportunity.

Justice for All

Perhaps the first, and most obvious, argument is that school choice has always been available for parents in the UK, to some extent, as it has in the USA (Witte 1990a). In a system of allocating places to schools based upon a catchment area, parents had the right to move house, and so come within another school's catchment area. Some cases were even reported of families renting pseudo-accommodation in another area to their home for this purpose. Parents could also opt out of the state-funded school and send their child to a fee-paying school. All of these choice strategies required finance, and so the catchment system provided choice only to those who could afford it. The Education Reform Act 1988 (ERA), and associated legislation such as the Assisted Places Scheme, has however given every child a de jure right to attend any school in the country including a number of fee-paying ones, and so can be seen as extending the luxury of choice, and diminishing the privilege of the elite, or at the least, sharing those privileges more evenly.

Central, or even regional control of local schools, imposes the policies and values of the state. Even where that control is established democratically, the prevailing majority view represents a dominant interest group which is potentially damaging to minority groups, whether based on religions, languages or cultures (Ball 1993). Parental choice of schools, on the other hand, can be seen as a local form of referendum, with prospective parents as the electorate, but with the constituencies so small that local voices can be truly effective, and each school can, in theory, cater for a different set of 'clients'. In this way, the 'dictatorship' of uniform public monopoly schools can be broken. Evidence that dissent to majority norms was being suppressed in the public school system was seen in the USA in debates surrounding censorship of books which tore some schools and communities apart (Arons 1982). Some parents wanted their children exposed to, and others wanted their children protected from, certain books, facts, and courses. This is reflected in Wales today in stories of parents choosing a fee-paying Catholic school, so that their child will not be taught about contraception and of schools cutting material about such matters from their textbooks (personal observation).

One example of the perils of a public monopoly dictatorship is provided by the debate over the merits of coeducation. In the UK most state-funded schools are coeducational, and in the USA they are so by law (Coleman 1990), but, according to some, there is mounting evidence that many girls perform better academically in single-sex schools. The public monopoly may therefore be seen, in this case, as favouring boys, and so denying equal opportunities to girls. Another well known example, educational 'reproduction', is referred to by Cookson when he states that 'the life arithmetic of the overwhelming majority of children is that their social class destination will be identical, or very similar, to their social class origin. This is a reality of American education' (Cookson 1994 p.90). This conclusion is backed up by the Institute of Economic Affairs (Maynard 1975) and Spring (1982) who both found no empirical evidence that compulsory state education, in either the UK or the USA, had been instrumental in decreasing poverty or increasing equality since its inception. One possible reason for this is the social stratification that routinely takes place in schools using a catchment system, and thus it can be argued that one part of the solution is to provide greater equity by a programme of increased choice and open enrolment.

Choice programmes in the USA have therefore been particularly popular, and seen as particularly relevant to, the poorer sections of society, such as immigrant, minority, and one-parent families, who have been deserting the large inner-city schools (Levin 1992). Ironically similar motivations may lie behind some families' use of fee-paying education, in opting out of the majority norms, or rejecting the inferior urban education available. A scheme to provide compensatory education for low-income families, particularly from minority backgrounds was rejected as 'elitist' in the USA in 1984, as it involved the use of private schools (Peterson 1990). There is however, evidence that low-income families benefit from this approach, since 'within the inner city there are private schools, including religious schools, that deserve public support because they are educating disadvantaged students at a time when public-sector schools are in disarray' (Cookson 1994 p.128). The private Catholic schools in the USA take a very high proportion of children from apparently disadvantaged backgrounds (Gaffney 1981, Coleman and Hoffer 1987), and like the grass-roots schools described in this thesis, the pay-what-you-can Steiner schools, and home-schooling materials in the UK, they are within the budget of most families who are prepared to make the commitment. In Milwaukee, the strongest pressure for an increase in parental choice is coming from the ethnic minorities and the poor (Cookson 1994), and in Scotland, a much higher proportion of one-parent families, who are frequently economically disadvantaged as a group, exercise their choice by making placement requests (Willms and Echols 1992).

Some of the original resistance to one proposed choice programme in the USA, a compensatory voucher scheme, arose from the fear that it might be used in part to counter the effects of the end of racially segregated schooling in 1954. However such an argument is not valid as such in the UK, which has never had compulsory racial segregation, or integration, in schools. In any case, a liberal programme might have to accept the possibility of 'racist' choice as an integral part of any free-market. To argue that parents of minority religions and languages, which often correlate very strongly with racial differences, should be allowed to dissent from majority schools, is to permit voluntary racial segregation in schools. Separate Muslim schools have been seen as 'sexist' in nature, as well as encouraging even further racial segregation in schools (Dooley 1991). The right to choose over-rides the Race Relations Act in the UK according to Dooley (1991), as evidenced by the 'white flight' from Dewsbury middle school in the 1980s. This freedom to choose on possible ethnic grounds is one trade-off that policy makers thus have to consider when creating neo-liberal choice programmes. It cannot really be countered by racial classification and admission quotas, as attempted in Singapore, since these are messy, relying on the unrealistic concept of 'pure' race and are, anyway, contrary to the spirit of the right to choose.

Perhaps the most common reasons used to advocate both choice and diversity have been religious/philosophical, but even these can be seen as a reaction to denial of equality by the majority on a class basis, since 'religious involvement is and has been stronger on average for those who are less advantaged' (Coleman 1990 p. xx). Some dissenters have a religion other than the majority one sponsored by the state in the UK, and enshrined in the National Curriculum, while others have objected to theistic practices altogether (Bentley-Ball 1982). In the USA, where the constitution prevents religious worship in schools, and thus enforces secular education, some dissenters have objected to 'liberal progressivism' on moral grounds (Arons 1982). In the past, liberals have objected to 'attending a daily act of worship', 'taking the pledge' or 'saluting the flag', which still takes place in many countries, such as Singapore, where attendance at a state-funded school entails a daily vow of loyalty.

An Engine for Diversity and Improvement

According to Keith Joseph in 1976, leaving things to market forces is always more effective than state intervention, provided of course that the free interplay of supply and demand is not impeded (Edwards et al. 1992), and to some extent this claim is now being tested in education. The claim that parental choice of schools will lead eventually to an overall improvement in standards is seductively simple. It can be summarised as follows. Public monopoly schools, mostly large comprehensives in the UK, have been too similar to allow a real choice, and they have been alleged to differ chiefly in their standards of attainment. Greater diversity is needed. According to Coleman (1990) the demand for diversity can be seen in the range of summer camps available in the USA, and the lack of coherence in the private sector which has military, Quaker, and progressive schools as well as the traditional boarding schools. However, this demand has clearly not been reflected in state controlled schools. In the UK, the terms of the Education Reform Act 1988, so the argument runs, allow greater diversity in provision, and have the capacity to create a situation where schools differ in their organisation, their curriculum, their mission, and their specialisms, as well as in their standards. 'Good' schools should be popular, and since funding is basically linked to pupil numbers, a form of natural selection will cause popular schools to thrive, and poor schools to close. With the continuance of each school linked directly to the satisfaction of clients, it should be more responsive to their wishes, which will eventually lead to an overall improvement in both provision of choice, and educational standards, since 'schools that did not provide students with skills would not last long' (Cookson 1994 p.125). Choice may also allow better matching of students to types of schools.

Since the link between the budget and satisfaction of individual consumers is more clearly broken in non-choice public schools, the organisation often appears to aim for 'profit in kind' such as prestige and avoidance of conflict (Boyd et al. 1994), all of which can detract from effectiveness. The efficiency of allowing 'consumer' dissatisfaction with such schools to be expressed by what Hirschman (1970) defined as 'exit', is described by Witte (1990b) - 'one does not have to say exactly what is wrong or what should be done to correct it; one does not have to argue, persuade, or cajole teachers or administrators; one only needs to find a suitable alternative'. In the USA the rapid decline of some urban schools rejected by the middle-classes shows the power of choice in those areas in which it has been implemented. Chubb and Moe (1990) argued that state democratic control of schools leads to an extra tier of bureaucracy in education, which can be unwieldy and impersonal, making schools unresponsive to their users. A market involving parental choice, and local control of schools by parents is one answer to such a problem. Such a scheme allows parents to express their dissatisfaction by 'deselecting' a school, but also to use their 'voice' on the governing body (Hirschman 1970). This is necessary because a market in schools is considered different to many other commodity markets, since conventional wisdom has it that it is not good for a child to move school frequently. This common convention creates the equivalent of 'loyalty' to a school for many parents, which is said to keep 'exit' at bay and so activate 'voice'. Parents are more likely to remain loyal, it can be argued, if their chances of changing the current situation at the school are favourable, compared to the risks of exit.

Choice advocates also argue that choice, by itself, can create a dynamic for the culture of the school as a society (Cookson 1994). Teachers, parents, and pupils in such schools feel that things are better than in non-choice schools. They may adopt some of the cultural characteristics of private institutions, such as a sense of mission for the staff, a sense of ownership for parents, and greater community participation. However, if most schools are to improve there must be a link between choice and school responsiveness. Schools must be prepared to make changes as a response to the newer demands of the market, and some early evidence suggests that they are. Woods (1992) studied three schools, and found that a number of changes were made as a direct result of perceived pressure from parents, including policies on school examinations, homework, safety of possessions, breadth of curriculum, banding, more exclusions, and attention to uniform. Although having implications for marketing, these changes are substantive, not simply cosmetic. Whether these changes constitute improvement is, as yet, unclear, but they do show that some schools are prepared to respond positively to market forces.

Perhaps the final argument for the implementation of choice in education is that such a policy is generally popular. Gallup found a clear 2 to 1 majority of parents were in favour of school choice in a 1993 poll in the USA (Riley 1994), and the early impact of the 1981 choice legislation in Scotland was maintained, with the number of placing requests doubling from 1982 to 1985 (Adler et al. 1989).

Problems with Marketisation

Opposition to the notion of using market forces to drive improvements in education is as varied as the approaches of advocates. Whatever the long-term consequences, for good or ill, of the recent changes in the UK, some children will find themselves stranded in failing schools. Although such schools may eventually close, any improvement in standards through market-driven evolution will require 'casualties'. It is therefore possible to argue that the children in these schools are the forgotten victims of the choice process, although it is also possible to claim that the schools they attend were failing anyway, as evidenced by their decline in roll since the advent of choice, and so, although there has been no impact for the rump, choice has actually reduced the number of casualties. Some opposition is based upon predictions of likely consequences other than those intended by proponents, such as an increase in social stratification, but much of it meets the economic model head on, and argues that choice simply does not work. In the Netherlands, a country with a very high proportion of private schools, no standard numbers, and considerable freedom of choice, unpopular schools do not disappear. They simply become very small. In fact, most private schools are very small (Cookson 1994). In addition, some evidence of the benefits of markets in education, such as that provided by Chubb and Moe (1988), has been questioned (Bryk and Lee 1992).

When Adam Smith developed the proposition that an economy of selfish agents making choices for their own benefits can be organised and made to work for the common good, he did so merely as a statement of logical possibility (Hahn 1988). This does not mean that it must be so. Many pre-conditions would need to be met. Assumptions include that the agents must be economically rational, knowing what they want, and how to get it. Above all the market must be genuinely free. These and other assumptions are usually not met in practice in any real market; witness the abolition of slavery, or the concept of industrial patents. The possibility of rational choice is examined in the third part of this chapter, after a discussion of some of the other limitations to a free market in education.

Education is not Marketable

One problem for the model of consumer choice in education is that it operates maximally in urban or suburban environments. Low population density and travel restrictions mean that many families in rural areas effectively have no choice at all. Early reports suggest that active choice, or deselection of the local school, is consequently much higher in urban areas than rural (Adler et al. 1989), and in areas where there are more schools (Echols et al. 1990). Some rural areas are in fact still running a pure catchment area system, allocating places by proximity where a school is over-subscribed (Hammond and Dennison 1995). Menter et al. (1995) found little evidence that patterns of enrolment in primary schools have been affected by the market since 1988. This is further complicated by administrative rules relating to school transport. Although there is free transport provided by the LEA for any pupils living more than three miles from school, this only applies to transport to the closest school, which makes a mockery of the concept of free choice for many families. Some parents are of course still prepared to pay for transport, or to drive themselves, but such a strategy is likely to have social class implications. It is clear that for choice to work, travel costs must be taken into consideration in any plans to retain equity of provision.

However, it has been observed in the USA that it is largely the inner-city schools that have been deserted by participants in choice programmes (Fowler-Finn 1994), suggesting that it is in fact these schools which are leading to the greatest dissatisfaction. Such a possibility is backed up by the observation of Echols et al. (1990) that use of private schools is proportionately four or five times more likely in cities than small, or single school, communities. Although there is a financial component here as well, relating to the payment of fees, there is again the suggestion that the chief dissatisfaction is with city schools in the UK. Perhaps parental choice is an urban solution to an urban problem? Families in some rural areas may be no better off in terms of choice, but they will also be no worse off. They will have the restricted choice of home schooling, which is a growing trend in some rural areas (Meighan 1992), and which could in extremis force a local school to close. It is not clear then, whether the opposition to choice based upon the uneven distribution of schools and population density assumes that school choice is a 'good thing' of which some are being deprived, or whether it is a 'bad thing' in which case those in rural areas, where it is impossible to implement, are better off, and can hardly be used as the basis for an objection.

Other suggestions are that a free market in schools is not possible because of the likelihood of a monopoly being established. Schools are not 'for profit' organisations, their product is unlike many others, and it is not the parents who are the main beneficiaries of schools anyway (Maynard 1975). A producers' monopoly of schools could be established in a free market, since the market itself is of fixed finite size, and unlike the markets for 'genuine' consumer goods, it cannot grow without a population increase. There are also economies of scale in the provision of education services, so that larger schools can provide cheaper services per capita, using the money saved to fund the improvements over their smaller rivals. Thus, the early successful schools will be able establish a 'natural' monopoly. Such an outcome is possible, although it is also possible that assumptions about market size and economies of scale do not apply here. The length of time spent in education is not necessarily a constant, and successful schools may persuade more pupils to stay on at 16+ for example. Economy of scale may only apply within the scale defined by the existing plant, and if a school expands beyond that to become split site, its character as well as its unit costs may change. Above all, there is no reason to assume that a state monopoly, as existed before 1988, is preferable in any way to one established in a market by popular schools with low unit costs.

A more serious objection to the marketisation of schools is that they do not provide typical consumer goods, since their quality is chiefly determined by the quality of their customers (National Commission on Education 1993). Education as a produced commodity is not homogeneous in the same way as many others, such as breakfast cereals (Garner and Hannaway 1982). There is imperfect competition between schools (e.g. because of travel constraints), and imperfect information on which 'consumers' may judge them. As described above, 'exit' is a mechanism of control in a market organisation, but this is not an option that consumers of education are likely to use very much. The disruption caused by moving school is too great for it to be very effective. Perhaps the biggest obstacle is that schools are not run for profit, and so if demand outstrips supply in any school, it will lead to problems. A school is not like a mainstream business, and cannot necessarily expand to meet demand, nor can it raise its prices since state education is a free service at the point of consumption. In the UK, 85% of a school budget comes from age-weighted pupil numbers. Rather than raising prices or expanding its customer base, an over-subscribed school may simply raise admission standards. Parents can choose a school, but do not have a right to use the school of their choice, unless it has spare places. In this situation, the school may become more selective in its pupils entry requirements. Such a policy would presumably allow the school to produce better public examination results, by increasing the ability of its intake, but without actually improving teaching standards. As the relative supply of pupils dwindles, power of selection moves to schools, who can drive up the entry requirements, and reduce the effort required to maintain superiority over other schools in terms of raw indicators. Thus a market can lead to complacency and demotivation for successful schools, emulation of their conservatism by those less successful, and so to dis-improvement overall. The theory that choice in an imperfect market leads to selection by schools, and so to social stratification, is an important one, discussed more fully in the next section.

Market theories of education see the consumer as the clear beneficiary, but may do so to the neglect of the external benefits. A whole society could benefit from the custodial or socialising roles of schools and this benefit could be greater than any private value. Market forces would thus be inappropriate since they would not preserve the interests of society as a whole. At a local level, parental governance of schools leads to problems, since the parents are assumed to be 'selfish agents' in many versions of choice theory. If they are only concerned for their own child, they may act in such a way that short-term benefits override the collective good of the organisation for which they have been elected (Witte 1990b). It is also administratively more convenient for the state to simply assign a compliant population to local schools, as exemplified by the complaint that some LEAs found planning impossible because choice in Scotland produced such volatility with unpredictable swings in enrolment (Willms and Echols 1992).

One of the major reasons for the imposition of universal state education in the USA may have been the need to foster a sense of unity in a new culture through a sense of shared experience (Coleman 1990). Society values cohesion, which is one reason why racial integration is practised, and why there is suspicion of 'bilingual education programs that appear intended to maintain a non-English linguistic subculture' (Coleman 1990 p. x). State education in the USA has also been used to break down the power of families, and to liberate children from narrow horizons, religious dogma, and the 'slavery' of poverty, and above all to mitigate the potential divisions in society arising from the diversity of immigrants, who once formed the majority of the population. The situation in the UK is somewhat different today, but there are parallels in the desire for devolution and 'own language' education in some regions, as well as the voluntary racial segregation discussed below. The creation of a market, by allowing diversity, may 'splinter' the schooling of this country, and lead to greater disunity. The balance between the 'benevolent dictatorship' of the state, and the promotion of individual liberties is, as ever, a delicate one.

One of the pre-requisites necessary for competition to lead to an improvement in educational standards is the correct mixture of 'alert' and 'inert' clients (Hirschman 1970). Those parents who are more alert to educational rights, problems, and opportunities provide the stimulus for change either by exercising their 'voice' to the governors of the school, or by signalling their dissatisfaction through exit. Unfortunately it is likely to be the parents most likely to bring about change who are also those most likely to leave (Willms and Echols 1992). It therefore falls to the more inert parents, those perhaps less attuned to the situation, to provide the basis of loyal clients, who give the school time, and a 'dollar cushion', for any improvements to take effect. Both are necessary. If most parents are 'inert', choice does not stimulate competition. If most are 'alert', it may cause problems for schools trying to expand, and so suffering in the short term, leading to another exodus, and not enough stability for any substantive measures to be seen to take effect. Improvements in schools can take years to become obvious in the form of quantifiable outcomes (McPherson and Willms 1987), which will be too long to wait for a selfish agent with only one shot at education.

Evidence on the existence and proportion of these two types of clients in society is conflicting. On the one hand, the Carnegie Foundation report claims that 'in states where school choice has been adopted, less than 2% of parents participate in the program; and parents who transfer their children to another school do so mostly for non-academic reasons' (Cookson 1994 p.71). There is, however, an indication that the number of parents prepared to become involved in choice is growing in both the USA (Cookson 1994) and Scotland (Echols et al. 1990), and that the poorer families are becoming just as involved as any other section of society. It may be that some families are natural consumers, used to making choices, and these are the minority, or perhaps the elite, reported in early results, but that others are not necessarily 'inert'. Some families may simply be slower than others in becoming aware of their rights to choose and appeal, in which case the proportion of 'alert' clients would be expected to grow continuously, and, as was shown above, there is already evidence that this is happening in the UK.

'Choosers tended to select schools with higher mean socio-economic status and higher mean levels of attainment' (Cookson 1994 p.92), which is a rational attempt to boost a child's attainment (Echols et al. 1990). Unfortunately, in terms of pupil ability, a 'favourable school context' measured by mean pupil SES is a 'zero-sum resource'. The policy of parental choice may therefore benefit choosers in relation to everyone else without necessarily improving standards overall, and as the number of alert families increases, the proportion gaining any benefit may drop. Since much of the information available on schools necessarily relates to past, not present, performance, the feedback to schools based on migration could be wrong, and might encourage the wrong changes. Effective schools in disadvantaged areas may be tempted to change to prevent losing pupils, while less effective schools in better areas might be made complacent by the attractiveness of their high mean SES (Willms and Echols 1992).

A major defect in the UK school market envisaged by the ERA 1988 is contained in that same legislation. Although parents may now express a preference for a school, and even help to run it, they are compelled to use a curriculum, testing policy, and funding strategy set by the government (Hartley 1994). Although schools can deploy their own budget, its level is determined by the number of pupils it attracts. This condition is set by the government, and implemented using national benchmarks. Teachers may retain their own pedagogy, but only within the constraints set by national curriculum, testing and assessment. The introduction of choice reforms at the same time as others giving greater central control over education, has inhibited the very diversity felt necessary by some for an evolutionary market strategy to work.

Witte (1990b) does not agree that publicly funded 'state monopoly' schools have necessarily been uniform, and unresponsive, which is one of the charges made by choice advocates. Schools in the USA vary by: size; years of intake; location; racial composition; income; plus testing and curricula, and are complemented by the fee-paying sector. On the other hand, choice schools are now engaged in rivalry for expansion and survival, which may be leading to a 'dull uniformity' of provision in the UK (Tomlinson 1994). All schools are aiming to follow majority trends, and none is responding to the diversity among parents by providing a distinctive kind of school, and then targeting their potential consumers (Woods 1992). Parental involvement in schools is more to do with the policy and vigour of the school than school choice (Riley 1994). In fact, parents who have made a choice are less likely to be involved in school, perhaps because voice is more often used when exit is impossible.

Perhaps the reason that schools are being so unadventurous and unresponsive is that they do not have good processes for learning or responding to changes from outside (Levin and Riffel 1995). This is not due to ill-will or incompetence but 'long-ingrained patterns of thought and behaviour' (Levin and Riffel 1995 p.1), which is why it might be that the advent of choice will be both less beneficial than advocates suggest, and less harmful than critics fear. Thus it can be argued, 'any evidence that this... strategy [of competition through choice] leads to the expected improvement is hard to find' (Brown 1994 p.55). School effectiveness research suggests that the true effect of a school can only be judged after its intake, and the socio-economic character of its catchment area is removed from or controlled for in later results (Brown 1994). Intake achievement remains the single most important factor affecting subsequent achievement. Any slight improvements found in some studies could, anyway, be temporary and in the nature of a 'Hawthorne' effect.

Choice Leads to Stratification

One of the other findings from early choice research that does not seem open to dispute is that there is a danger of even greater social stratification. Choice may have led to a larger elite, but they are still enjoying a better education at the expense of others (Tomlinson 1994, National Commission on Education 1993). In both Israel and the USA, students of 'higher' social class are more likely to be enrolled in the schools of their choice (Goldring 1995), and this result is true across all ethnic groups. In Scotland, choice is more common among better educated and higher social class families (Echols et al. 1990), being linked to home-ownership and socio-economic status (Adler et al. 1989), and more prestigious occupations (Willms and Echols 1992). This trend is likely to be increasing the segregation between schools in terms of mean socio-economic status, and all of the pupil characteristics that have been found to be associated with it.

In addition to the problem of only some parents being alert to choice, there is the additional one of schools being perhaps too conscious of it. The introduction of a market could change the values prevalent in education (Ball 1994), so that schools may no longer be run primarily in the interests of their users (if they were before). A policy of 'improvement' through selection of intake may make sense for a school in the market, but it is only cosmetic, not making any school more effective. It must also be reflected in a decline in results elsewhere. In reality the schools would be making the choice, and not the parents. A policy of selection by schools can therefore also lead to segregation by first language, gender or social class, as schools use profiles of an 'average' successful candidate in order to predict future success. Children from homes whose first language is English, or whose parents are university educated, for example, might be seen as statistically more likely to be of value to the school, because of an established link between these characteristics and examination performance. The recent rise in exclusions, and associated appeals, suggests that schools are trying to screen out what they see as problem pupils (Ball 1994).

There is already some evidence that school responses to market demands are discriminating against black families and groups (Blair 1994). The content of the National Curriculum encourages a narrow definition of national identity, in history, geography, literature and especially in the orders for Religious Education which emphasise Christianity. 'Britishness' underpins many of the changes, and schools may unwittingly pander to the racist sentiments of their ideal consumer. The free market denies that there are unequal starting points for families, and so does nothing to increase equity in education, appealing as it does to individual effort and 'bootstraps'. Muslim parents can query a definition of diversity that allowed Judaic or Catholic state schools, but denied them aid for state voluntary schools. Schools opted out of LEA control are allowed to draw up their own criteria for pupil selection, especially if they are performing well and become over-subscribed. In 1992 the Commission for Racial Equality found that Watford Boys and Watford Girls Grammar schools had indirectly discriminated against Asian parents in its entry procedures (Blair 1994). Parents were asked to write their reasons for choosing the school, but only those submitting six or more reasons were admitted. Some Asians may not have had the literacy, cultural capital, or simply the confidence to produce so many reasons. In Dewsbury, white parents removed their children from a primarily Asian school, and justified their 'exit' on racial grounds. Ball (1993) suggested that ethnic minority students will become more liable to exclusion from school, leading to a loss of real choice for them. Since the Reform Act, LEAs no longer have same authority over appointments in schools, and thus cannot impose a policy of equal opportunity. This is a particular problem, since most governing bodies are still predominantly composed of white middle-class males. Discrimination may also apply to black teachers not deemed acceptable to white parents. In the Dewsbury case, a court ruled that the right to choose overrode the Race Relations Act, and this may have set a dangerous precedent for the selection and promotion of teachers as well (Blair 1994).

The paradox of choice programmes is that, although they are more popular in the USA with poorer urban families, recent immigrants, non-whites, and less prestigious socio-economic classes, these may be the very types of families who are least likely to use their choice, with the smaller better educated families in fact being more active (Lee, Croninger and Smith 1994). Because the poorer families are the ones that generally have access to the urban schools with the worst reputation, choice is often seen as negatively rooted, as being choice away from something (Goldring 1995). This is more likely in the USA than the UK, where the disparity in educational expenditure between US districts, financed by a local property tax, can vary by as much as 10 to 1, with the wealthier residential areas contributing solely to the upkeep of their own schools (Lee, Croninger and Smith 1994). Perhaps one of the reasons that researchers in the UK are less enthusiastic about choice is that, before 1988, expenditure was already more evenly spread between schools than in the USA. Another may be that, whereas US visions of progress may be based upon individuals, their freedoms, and right to advancement, in the UK visions are focused more readily on whole segments of society.

Although Ball (1993) conceded that not all of the effects of the ERA were necessarily intended by policy makers, he also stated that 'the implementation of market reforms in education is essentially a class strategy which has as one of its major effects the reproduction of relative social class (and ethnic) advantages and disadvantages' (Ball 1993 p.4). Raw scores in league tables of results reinforce selection of pupils by ability in over-subscribed schools. As the market eliminates under subscribed schools, surplus places may be eliminated, reducing the pressure on schools to compete for students. Schools will seek to maximise their performance measures and minimise their costs per head, and thus will recruit more able students and turn away those with difficulties, and concentrate their resources on the most able. Therefore the indicators may have a political purpose other than to raise standards, perhaps to further a policy of 'cultural restorationism' (Ball 1994). Selection of pupils by schools becomes possible, indeed rational, and might be ''a selling point'; what you get is who you go with', so that unlike in most markets, who the customer is makes a difference. As the value-in-exchange of education depends on how much one holds relative to others, it leads to a system of winners and losers, so that the market is not neutral. It requires access to possibilities (time, transport, literacy, numeracy, child care facilities) which are unevenly distributed in society. Those with little education, or on low incomes, may only access the lower level of information concerning schools, and so the system could lead to social reproduction (Bowe et al 1994b). This reproduction could then be justified by claiming that non-choosers are bad parents, and reinforced by sending money to popular schools.

Several points can be made against this thesis, in addition to the evidence cited above that the proportion of alert clients is growing, especially among the poorer and minority families. Since income and wealth are unequally distributed in society this will inevitably lead to unequal access to education in a free market. Poorer families will have fewer funds for travel, and additional contributions to schools, and less opportunity to move. However this has always been true, even under a catchment system, and can be seen as an argument for greater equality in society, and not one against parental choice per se. Since inequality is undesirable in its own right, its existence should not be accepted as a 'given', and then used to argue against choice. Education, by itself, cannot be expected to solve major societal injustices.

The government could intervene in the financing of education to encourage greater equity by guaranteeing loans, or introducing negative income tax, but this would not necessitate intervention in the provision of education itself. It could still allow a market to operate, but one in which the 'players' are more equal, and it could enforce a policy of allocating places at over-subscribed schools by random draws. Alert 'players' may then apply to slightly less popular schools, to increase their chance of first-choice allocation in a satisficing school, which could increase the proportion of first-choice allocations altogether, as well as reducing social stratification. Allocation of contested places could even use a stratified random selection (i.e. with target proportions set by whatever strata seem appropriate, including ethnic background, gender, ability and socio-economic group). Anyway, although research on parental choice has not yet reached an agreement on how schools are chosen by parents, it is clear that public examination results are not the only, nor even the major reason for school choice, so a popular school would be misguided to select pupils on ability alone. There is of course a danger that parents themselves will select a school on the basis of its current social class, gender, or racial breakdown, but such an outcome must not be confused with selection by schools, which it has been by some previous writers.

A policy of selective intake by schools only makes sense in a market in which their places are in high demand. In terms of organisational survival, any pupils are better than none. The evidence that schools seen as successful will be overloaded with applications, suggests that there are many parents not happy with their current neighbourhood schools. This is, in itself, evidence of the need for improvement in schools, and of the enthusiasm of parents for their increased choice. It is not at all clear that a policy of not allowing the parents in the less desirable schools to express their dissatisfaction is preferable in any way. It is also not clear that demand must necessarily outstrip supply. Some surplus places must remain in the system anyway, to provide the incentive for schools to compete in the market. Many parents appreciate small schools, so the paradoxical situation might arise that although small schools are popular, popular schools are large. The market might then see a natural evolution of schools to an optimum size, above which they cease to be popular.

The basic problem at present lies in the ability of schools to turn pupils away once their standard number, a completely arbitrary figure, is reached. It is this which may lead to selection, and could be prevented by supplementary legislation, such as that in the Netherlands, or that proposed by a former Secretary of State for Wales, allowing successful schools to expand. There are still spare places in the system, in the sense that there is provision for more pupils than actually exist. There are therefore more than enough teachers, buildings, and other resources to provide an education for all pupils in the school of their choice (this is not to say that there are, in fact, enough resources for a good education). The main problem with resources in the current parental choice scheme is that they are often in the wrong place. However this is also not a criticism of market theory as such. In a free market, popular schools would be allowed to grow indefinitely, and unpopular ones would close. Some of the resources from under-subscribed and closing schools could be transferred to popular ones. Teachers contracts would need transfer clauses, and schools would need to offer incentives to attract more and better staff. Most schools use some 'temporary' accommodation, and in future such accommodation could be organised so that it can be switched easily between sites. Many schools operate on split sites already, and have devised organisational and timetable strategies to minimise disruption. Such strategies could be used by a popular school which has outgrown its plant and takes over a nearby less popular school. It may be that as schools grow and begin to operate on split sites, in 'portakabins' and employ teachers from 'failing' schools, that their attractiveness may decline, and an equilibrium be reached without a single pupil having to be turned away.

The Limitations Of Rational Decision-Making

At the core of a programme of school improvement through market competition lies the notion of parental choice. If parents make good choices, the model works well, at least in theory, and if they do not, the market fails to deliver. The market includes a belief that people are rational choosers, and will select schools that are academically superior, and as yet 'there is little evidence from the private sector that parents will choose schools that.... are not in the best interests of children' (Cookson 1994 p.125). Market research reveals the stated preferences of respondents, and assumes that their choices are a kind of revealed preference (Sen 1982). The discussion here will concentrate on how people make decisions in general, and whether they are likely to make 'good' decisions about schools. Psychological research on decision-making has emphasised the study of actual decisions being made, and because of its methodology has tended to highlight the factors other than logic that lead to them. At one extreme, people can be seen as rational, violating optimality by simple errors, and at another as primarily emotional driven by unconscious motivation. The attribution work of Schachter and Singer (1962), and others, show how susceptible people can be to unconscious, perhaps even subliminal, cues when making judgements of people and situations. Courts are well aware of the strength of primacy and recency effects in the presentation of material.

In fact, 'people are usually seen as trying vaguely to be rational but failing frequently to appreciate normatively appropriate strategies' (Eiser and van der Plight 1988 p.76). This usually means that a decision maker is rational up to a point of 'good enough', or that they are using bounded rationality, or 'satisficing', in coming to a conclusion, as opposed to a formal cost/benefit analysis. There are many reasons for this, such as lack of time, subjectivity, computational ability, distractions, existing values and commitments, 'groupthink', stress, and notably a lack of understanding of probability. Most decisions involve uncertainty about outcomes, and even premises, and so to some extent all real-life decisions can be seen as being based on probability. Classical probability theory, however, is based on long-run frequencies of many 'trials', but most decisions are one-off cases, and they correctly involve subjectivity, in the sense that the value of an outcome may differ between individuals and over time (Eiser and van der Plight 1988). For example, a gain of 5 has a real value dependent upon the present need of the individual concerned.

One model of such decision making is the Subjective Expected Utility theory (SEU), which suggests there are five stages in a decision:

  1. list all possible courses of action;
  2. list all possible consequences for each;
  3. assess the attractiveness of each outcome;
  4. assess the likelihood of each outcome;
  5. calculate the SEU of each course of action as the sum of [the probability times the utility of each consequence].

However even this simple model has little descriptive validity, since studies show that most people ignore SEU in making decisions, even after the method has been explained to them (Eiser and van der Plight 1988). It may be that such a calculation exceeds most peoples' calculational ability, especially as very few personal decisions, even concerning choice of school or a house purchase, are represented formally on paper, despite the improvement in the quality of decisions that can be achieved by the use of balance sheets, cost/benefit analysis, and iterative computer simulations. Using a heuristic or short-cut leads to bias, partly due to the psychological factors outlined above, but the biggest flaw may be that the probabilities and the value of each outcome are not seen as independent in the way that theory suggests.

A common flaw in heuristics is the misinterpretation of probabilities (Eiser and van der Plight 1988). Any reduction in the probability of an outcome from 100% certainty produces a greater loss of its perceived attractiveness than an equivalent drop in probability from an originally lower figure, so that perception of probability is not a straight line function. Paradoxically, low probabilities are often greatly over-weighted in decisions, such as insuring for fire damage, or entering a lottery, but they can also be neglected entirely. One reason for this may be that risks from easily pictured causes, such as plane crashes, are more likely to be in the media than more common risks, such as diabetes, and so are exaggerated in subjective estimates. Decisions also depend on the phrasing of the problem, and the frame of reference of the subject. For example, losses loom much larger than gains, so that most people prefer to make 100 than to have a 50% chance of gaining 200, but the same people would prefer a 50% chance of losing 200 to definitely losing 100.

A further problem arises when probabilities have to be multiplied, as they would in practice in most real life decisions. Even professional mathematicians, who have correctly stated that successive tosses of a coin are independent, have been found to believe that a run of HHHHTTTT is less likely than HTTHTHTH (Eiser and van der Plight 1988). In addition, misunderstanding of the 'law of small numbers' suggests that the probability of a maternity ward having 60+% male babies is not seen to depend on the size of the ward, because small samples are seen as equally representative as large ones when the results are expressed as proportions or percentages. This is good news for the small private schools regularly topping league tables of examination results in Wales with a 100% pass rate for seven candidates, for example. In another experiment, respondents rated the probability of two independent events happening together as greater than one of them in isolation. For example, when told a story about a strong independent woman, she was rated by participants as being more likely to be a feminist cashier than a cashier (Eiser and van der Plight 1988), which is, of course, absurd. For these, and several other, reasons there must be considerable doubt about the ability of parents to make decisions which involve judgments of future uncertainty.

Assessing the value of an outcome can be as difficult as assessing its probability, and combining the two in the fifth step of SEU theory is very complex. Accuracy and reliability decline as the information load increases, and in general, people 'limit themselves to one salient dimension whilst screening out dimensions that suggest a different solution' (Eiser and van der Plight 1988 p.99), particularly where they have already made a public commitment to one solution, or feel compelled to agree with a consensus. Since school choice is so complex, such findings suggest that parents will judge schools on only one major criterion only. This has important implications for the results described in this study.

Economic models of choice, including forms of rational choice theory, assume that all actors have preferences, and make rational decisions to maximise their realisation at lowest cost (Boyd et al. 1994) - that they are 'self-interest maximizing individuals'. These theories can be compared to real behaviour to see how well they predict or explain it, and some remarkably accurate predictions have been made in other fields. They are only just beginning to be applied in school choice research, but some recent findings can be derived from collective choice theories. Since public education is a collective good, a rational self-interested player should be a 'free rider', taking the benefit but not participating in the school nor contributing to its provision. This suggests that parents who have chosen a school will be less likely to participate in its running. Research also shows that the same actors, with the same preferences, can produce entirely different results depending on the 'voting' rules (Boyd et al. 1994), and so the effects of any choice programme are likely to be very sensitive to the specific set of legislation enacted. Game theory, especially its analysis of the prisoner's dilemma, shows how sensitive the outcome of some choices can be to the choices of others, and such a game may be being played by parents wishing to send their child to a school that will also be chosen by parents of higher mean socio-economic status. It also shows that the consistent pursuit of self-interest, assumed in rational economic models, can produce inferior results for all (Sen 1982).

In practice, to make a choice based on the effectiveness of a school, one needs to be clear what the objective of education is, but this is something that even professional teachers and research academics cannot agree on. Parents cannot be expected to make such choices according to Thiessen (1982), who illustrated this point by saying that 'just because... many families might prefer programs emphasizing... the classics... or the sayings of Chairman Mao, but do not find them served in the local public schools, it does not follow that anything is amiss in the school and that somehow the child's right to an education... has been abrogated' (Thiessen 1982 p.79). Even if parents are capable of bounded rational decisions in theory, they may not have sufficient information to make them. For example, Raven (1989) pointed out the difficulties for parents in acting out the role of consumer of education. In order to be able to choose effectively, parents need access to the options available, convenient geographical access to more than one school, valid information about the alternatives, help to articulate their needs, familiarity with present educational programmes, the capacity to understand the information, and the time to consider and review it.

A MORI poll in 1993 found that whereas most parents were happy with their child's local school, most also felt that the education system as a whole was poor (Young 1994c). Such a result casts doubt on the rationality of parents, in general, as judges of education, since this overall picture cannot be true. One explanation is that the finding is a result of media criticism of schools and teachers, leading to a public opinion that something was wrong overall with education, although parents' knowledge of their local school suggested otherwise. HMI has been surprised to discover that parents with children attending schools identified as 'failing' and liable to direct take over by the Secretary of State, have been satisfied with the school. Such findings also cast doubt on the competence of parents to judge the qualities of a school. Hughes et al. (1994) recalled a comment made by the then Education Secretary on national radio. He stated that the educational views of the National Confederation of Parent-Teacher Associations (claiming to represent over 8 million parents) were 'Neanderthal'. Such a statement shows little respect for the views of parents, and is hardly consistent with a belief that parental choice will improve standards in education. This attitude can perhaps explain some of the ambivalence and contradictions in Conservative education policy for the 1990s.

Willms and Echols (1992) analysed the relationship between choice and school effectiveness in the area of performance in national examinations. They divided the effect of a school into Type A contextual success, sensitive to the characteristics of the intake, and Type B, a truer comparative effectiveness with equivalent pupils. Parents making placement requests (choices) were not sensitive to this distinction, and tended to choose schools with Type A benefits, particularly a high mean SES. Willms and Echols (1992) concluded that 'parents' choices are rational in the sense that they increase their children's likelihood of success' (Willms and Echols 1992 p.340) but that the effects are probably not as great as they appear on the surface, since parents are using the raw performance indicators, and so are choosing the past pupils of a school, and not the school itself. There may be an element of superstition involved in sending a child of whatever ability to a school in which other children of unknown ability have previously done well in examinations.


This chapter has examined some of the points made in the debate over the use of market forces in education, and shown how difficult it is to decide on the likely outcomes, especially as the different schemes proposed vary so much, and the UK legislation appears to contain contradictory elements. Since the results of policy changes can take a long time to become clear (McPherson and Willms 1987), this provides further justification for using the established market of fee-paying schools in order to try and predict what the effects may be in the state sector. Also, given that choice is seen by some as an antidote to stratification and social reproduction in education, and by others as a major cause of it, it will be interesting to see to what extent fee-paying schools are elitist, and to what extent they are compensatory. Finally, as it is not clear whether parents are capable of making rational choices about schools, even when they behave as 'self-interest maximising' individuals, doubt must be cast on the possibility of using increased choice as an engine for improving school standards. Again, the fee-paying sector can shed some light on this, the quality of the choices made by families, the extent to which the schools are driven by their 'consumers'' preferences, and ultimately the comparative quality of the schools.