Chapter 10: Discussion

10 Discussion: Overview

The first five sections of this discussion each relate to one of the earlier results chapters. Section 10.1 discusses the questionnaire data (Chapter 5) and also provides an overview of points raised during home visits which assist in placing the discussion in context. Section 10.2 reviews the PIPS Baseline assessments (Chapter 6), 10.3 Literacy (Chapter 7), 10.4 Numeracy (Chapter 8) and finally Section 10.5 considers findings from the Social and Psychological Skills assessments (Chapter 9).

10.1 Questionnaire Data:

10.1.1 How the data compared with previous research

This study found a considerable distinction between the North American 'homeschooler' (Gustavsen 1981; Gladin 1987) and the UK home-educator. UK home-educators were more likely to be town dwellers and were less formally ‘educated’ than their North American counterparts; families in the UK were less likely to have access to learning resources and although often religious, such families were, perhaps, less orthodox than in the USA. Of the US samples, Rakestraw’s (1988) 'homeschooler' (see Section 2.1.1) most resembled the UK home-educators in that most of Rakestraw's families followed some kind of plan, made use of various resources, used mixed learning and teaching techniques and assessed their children continually through informal observation. A notable difference between Rakestraw's homeschooling family and the UK home-educator was in the more traditional style of the US families whereby fathers were the main breadwinners (in the UK sample this often started as the model but then over time wage roles split as parents moved into part-time work), mothers undertook their children's education, homeschoolers were generally affluent, religious, made far more use of formal assessment than their UK counterparts and harboured expectations that their children would attend university. The three main reasons for homeschooling in the US also differed from the UK: in the US reasons given were; because parents considered it their responsibility to educate the children; to avoid negative peer influences (parents and children) and to control the instructional materials used. In the UK, the main motivations were freedom and flexibility so that children could learn in their own style and the family could maintain a close relationship with time together.

The current questionnaire did not include a question relating to income, nevertheless, based upon this researcher's familiarity with home-educators in the US and in the UK where the majority of UK home-educating families were in socio-economic Class II or below (Rose and O'Reilly 1998), it was possible to surmise that UK home-educators were generally less affluent than their North American counterparts. Ethnic minority home-educators in the present study appeared, prima-facie, to be in the minority (as they are in the US), although a glance through surnames and a review of the field-notes, indicated that whilst ethnic families were in the minority, mixed ethnicity families were relatively common. However, the absence of fully Asian families in the sample may well have concealed substantial number of such families who chose to keep their children at home (Rothermel 2000).

Home Education Advisory Service (1999, p.3) described home-educating families who home-educate from birth to be, 'very much in the minority'. However, although the research at hand found that only 27.57% of the school-aged children in the study had been home-educated from birth, a further 20.56% of the children in the survey (n=1,099) were pre-school age and 84% of their parents said they were planning to home-educate. In comparison, 25.53% of children had, at some point, been withdrawn from school. A common pattern appeared to be that most families withdrew a child from school but then home-educated subsequent children from birth. Having gained initial support from an organisation, such families were less likely to rely thereafter on organisations; thus, to say that ‘from birth’ home-educators are in a small minority is, according to the findings of this research, a distortion of the facts. This study found that the popularity of home-education was based equally upon poor experiences with schools (30.77% - perhaps including most of those 25.53% who had withdrawn their children) and family lifestyle (29.17%); however, bearing in mind the percentage of families that had earlier sent their children to school, Mayberry et al.’s (1995) observation that the home-education movement’s expansion was a response to growing dissatisfaction with school education was, perhaps, either directly or indirectly, the case in the UK also, where such concern was partnered with a increasing sense, for many, that they had a choice: this was evident from the number of families who were home-educating siblings after withdrawing one child.

In terms of motivation (Section 5.4) to home-educate, this study concurred with the 1983 findings of Grant (Webb 1990) and, to an extent, with the Scottish studies by Paterson (1995) and Brunton (1996). Brunton's (1996) Internet related findings mirrored the finding here, that computer technology, whilst useful to home educators, was not yet a leading motive behind the growing interest in home-education, but rather a useful post-decision support tool.

The motives for home-education given in this study, reflected some of the Gustavsen's (1981) findings, despite his conclusion that ‘moral issues’ was the most popular motivation to home educate, which here, was listed by just 13.14%. Mayberry (1988) had found the larger proportion of home-educators to have been motivated by religion, whereas here, ‘religion’ featured exclusively only amongst 4.17% of responses. So far as religion was concerned it appeared that religious families did not home-educate as a consequence of their religious convictions, but rather because the type of person who was religious was also the type of person who was likely to home-educate; this conclusion was supported by the study's interview findings besides reflecting Mayberry’s (1988) conclusion that ‘new-agers’ and religious families home-educated for similar reasons, i.e. wanting to take control over their lives.

Participants in this study stating that their children had special needs numbered 22.54%, contrasting with 19.87% of parents who cited 'special needs' as a reason to home-educate. Questionnaire commentaries and interviews with families of such children indicated that, often, they home-educated because of a family philosophy that would have led to home-education regardless, particularly, it seemed, where a child experienced mild difficulties such as dyslexia. Significantly, asked what home-education meant to them, not one family mentioned 'special needs', indicative, perhaps of the way that initial motivations were sidelined as the families found other benefits of home-education (Appendix 5.3).

Almost half the families in the sample said that home-education was not as they had expected (see Section 5.6), adding, in most cases, that it was either far more rewarding, far more demanding, or both. Although only a very small percentage of respondents said that their children had not turned out as they had anticipated, casual conversations with home-educating parents suggested this was a fairly common view. One home-educator with grown children told how, despite having 'given' to her children in every way she thought best, she could see that in retrospect, the children were destined to be who they were, regardless of her efforts to affect their development. Meighan's (1995) idea that home-education worked well was vindicated by the absence of negative questionnaire comments from parents in this study, some of whom, as the field-notes showed, were experiencing difficulties.

There was evidence that flexibility was much valued by parents and this came across, not just in the question relating to ‘meaning’ but in parents' descriptions of their routines, formal or informal. It appeared that in even the strictest families, learning followed the child's and parents' interests and although families did not always move towards the unstructured routine inferred by Webb (1990), Meighan (1995) and Thomas (1997), they did, as these authors also suggested, adapt to their children's changing needs. It was hypothesised that the more formal parents were those following the National Curriculum (Appendix 4.2 Question 11), whereas in fact, the converse appeared to be the case; it was the more confident parents (either formally or less formally educating) who shunned the national curriculum, while those parents wary of being under the watchful eye of family, neighbours, LEA inspectors etc. tended to follow the National Curriculum, or at least keep it in mind. Such parents might, perhaps, be classified as Blacker’s (1981) ‘compensators’ (see Section 2.2 - 'compensators' were parents who were trying to compensate for school). The findings contrasted with those of Gladin (1987); for example, here, the 'academic' day was generally no more than 2-3 hours, as opposed to the US 6.6 hours, and although religious families often tended towards routines, they were often very child- and family-centred, consistently conscious, it seemed, of the importance of 'family' and of fitting their religious agenda around the children's own interests, perhaps thus ensuring that the children's adherence to more formal aspects of their education would be maintained. Petrie (1992) had cited one LEA as believing that the religious families tended to pursue structured learning, however, although this may be so, there was evidence in the questionnaires of both these and other families of, 'letting the inspector see what he wants' but then returning to their own idiosyncrasies once the door was shut[1]. Science and Maths, contrary to what this author had anticipated (Appendix 4.2 Question 12), did not present major problems for the families. Parents in a considerable number of families were very confident of their own expertise in these areas and as Meighan (1995) reported, home-educating families adapted well to these areas of learning. Upon attaining levels of learning where specialist resources were needed, children tended to attend local colleges, having also reached an age where attendance was possible[2]. What was surprising was the minimal use of the computer for Maths and Science at a time when computer technology is thought to be leading the way to home-based learning for all (Hargreaves, 1997). Likewise, it was surprising to find at least 14.23% of families did not own a television.

Reviewing home learning in the families, it was clear that the space to develop non-academic 'intelligences' as described by Gardner (1993) (Appendix 4.2 Question 10) was one advantage of home-education. There was far more room for family activities, discussion, spontaneity. Respondents firmly believed that listening to their children, sharing experiences and involving the children in everyday responsibilities contributed to the children's education and growth in a way that school could never compete with.

There are those, anecdotally, who describe home-educators as abandoning external learning assistance and shunning the 'outside'; however, this study found that over 50% of families used learning support (tutors, dancing groups, sports etc.) and this, combined with the lack of community support felt by many home-educators, demonstrated a need for home-educators to be involved and accepted by their communities: the majority of families did not choose or relish their 'isolation'. Few families could be described as insular, despite over 25% of the sample saying that they did not mix with other home-educators: many reported that they mixed with other kinds of people, perhaps a reflection of this researcher's growing sense that home-educators did not always get along with each other, often mixing out of duty rather than preference. Isolated or not, families did not generally believe that this was reason enough to [re-]adopt the school ethos they had abandoned, although as the research found (through the interviews), a number of families did succumb to this pressure.

Interestingly, 47.84% of respondents said they did not assess their children, with a further 28.64% reporting using very informal assessment such as discussion and observation. This non 'formal assessment' approach was supported by the findings from the assessment programme used in this research, where one of the overarching conclusions was that formal school style assessment was inappropriate for home-educated children whose learning took a very different pathway to that of schoolchildren.

10.1.2 Field-notes

As a result of the questionnaire survey, and as a precursor to a follow-up study, some of the families were visited, interviewed and field-notes taken. The following notes summarise the main points raised and provide an informative backdrop to the research. Where there is specific mention of some families the text attempts to be self explanatory in its reference.

Categorisation of Home-educators

Insofar as it was possible to classify home-educators this research suggested that they could, perhaps, be classified on four levels; first, as a homogenous group; second as a collection of diverse groups with home-education in common; third, as families; and fourth, as individuals. The difference between this taxonomy and those previously noted is that this classification is by stratum rather than by type.

Level One - A Homogenous Group

There was evidence that the categories of home-educator discussed by Blacker (1981) and Van Galen (1991) (see Section 2.2) existed only in the broadest of terms, the reality being far more complex than that implied by these authors, with families moving between classifications to such an extent that attempts to classify families might be unhelpful. For example, whilst some religious families home-educated, they did not necessarily do so because of their religious beliefs any more than the New Age families could be described as home-educating because of their anarchistic or spiritual leanings. As Mayberry (1991) had inferred, these two groups had in common their desire to have control over their children's education. Families from these groups, and others too, were characterised by a commonality that had often led them into home-education rather than their denominational beliefs. It might be said that what these people had in common was a concern for the environment, the community and the belief of absolute responsibility in respect of their children. These families might perhaps have differed from those who turned to home-education after trying school, although families who withdrew their children soon after starting school tended to have considered home-education at the pre-school stage. Many families withdrawing older children soon 'succumbed' to the influence of other home-educators, gradually homogenising with them in respect of their new 'social conscience' that extended beyond their parenting role to concern about society and the planet generally. It appeared that home-educators were people who took parental responsibility seriously; this was as applicable to the lady running from social services as for the family who used 'loving discipline' (corporal punishment).

Level Two - Group Differences

Whether various factions would view themselves as having anything in common was a different matter, since a further characteristic of home-educating families was the antipathy they often felt towards one another. At the level of family orientation, differences such as religion or secularism, formal or informal education, Jehovah's Witnesses or Catholics, mattered very much indeed. It was at this point that groups splintered, since membership of these categories was not directly connected to the causes of home-educating but rather came about as a result of families' need for a sense of 'community'; hence religious families who home-educated might have found fellowship at 'open house' evenings and families from more secular backgrounds often amalgamated at 'EO' gatherings even though the relationships therein might well have been frail.

Not everyone felt the need to belong. The family of a girl (described in the field-notes) who had been in care were not members of any organisation or church: it might be said of them that their strength was in their fight against the way that, as they viewed it, those in authority had victimised them. Some families, as the interviews revealed, may not have been members of a home-education organisation but they did, nevertheless, have other affiliations.

Level three - Inter-family Differences

Within groups there also appeared to be considerable differences. The Thistle family were very different Jehovah's Witnesses from many of their 'brothers and sisters': most Witness families were child orientated whilst the Thistles were more doctrine driven. The Thistles were one of the few families who might be said to have home-educated because of their religion, despite their children having spent considerable time in school before being withdrawn. The secular families likewise, often had little in common with each other: formally educating families conceived their educational responsibility very differently from those who opted for the autonomous approach and whilst both sets may have sympathised with those secular parents who experienced psychological and social problems, they probably shared little mutual ground. Within groups too, there were those who home-educated from birth and those who withdrew children from school, those who were in conventional families and those who were not, those with older children, those with younger children, those with one or more children in school and those with none. There was some suggestion that traditional parents preferred their children to mix with those from similar family compositions; a secular family with married mother and father might prefer their children to have friends from other such families, religious or secular, at school even, rather than with children from a non-traditional family structure even though they were fellow home-educators. Furthermore, there was a kind of snobbery about who were the 'real' home-educators: hence comments such as:

'They aren’t really home-educating. They use tutors'

or

'Well we hardly have anything in common with travellers! They don't home-educate!'

Level four - Intra-family Differences

Even from the few families visited there was evidence of friction within families, although this was most often between immediate and extended family. Only one family, the Thistles, included children who specifically had not wanted to be home-educated. Many families spoke of periods where one or other family member had preferred the school option. Often it appeared that one or other of the children wanted, temporarily, to try school, but that this often passed. Conversations during home visits, with post-16 year old previously home-educated children of the families involved in this research, suggested that, younger, they had sometimes felt the need to be with other children quite desperately, but overall were pleased that they had been home-educated and glad that their parents had pursued home-education with them. Webb (1999) and Knowles (1995) also reported similar satisfaction (see Section 2.2).

Nevertheless, there was the father who went along with his wife's decision whilst he would perhaps, have preferred his children to have been at school: the conflict was repressed but nevertheless present. Differences between spouses were common although most often the sceptical parent was persuaded in some manner or other. Further, there was plenty of evidence of problems with grandparents and parents' siblings, who disapproved of the family's decision to home-educate. It can be assumed however, that a number of families currently with children in school also have a parent who might have preferred home education but had been convinced that school was the better option.

Most of the children who were interviewed spoke of their pleasure at not having to be in school. Sibling conflict existed but this was, perhaps, less evident amongst the home-educated children who had been together from birth than within those whose separation from the family had brought them different experiences.

Intuitively, it appeared that those children who had access to large spaces should have been the most content; however, it was surprising to find that even children living in the most cramped conditions were satisfied, and that space available bore no relationship with the level of harmony within the family. Similarly counter-intuitively, home-educating family structure and lifestyle seemed unconnected with the children's apparent well-being. Being well down the socio-economic ladder did not appear to unduly affect families' contentment with home-education and as the assessment results showed, neither did it affect negatively their academic performance (see Section 6.4.2).

Summary of points noted in the field-notes

Besides the types and levels of home-educators the other main issues raised concerned alienation from the wider community, a sense of 'searching for something', the way in which families changed over time, 'working' to the school year, and finally, the sheer joy of so many families at their decision to home-educate.

Alienation from the Wider Community

This was a topic that touched most of the families. Non home-educators often looked upon the families as 'strange' and unconventional, often feeling as if their decision to send their own children to school was inherently being questioned. Generally, families complained about the way in which children's out of school activities were geared towards school style learning and the assumption that the children were at school. Families often remarked that they and their children were questioned in the street during school hours and their children were 'tested' by other adults with questions relating to maths or spelling. Indeed, 'not accepted in the community/isolating' was cited as the main disadvantage to home-education (Chart 5.7). Several parents commented upon the way that group activities for children often began with the leader asking children about their day or week at school, thus alienating the home-educated children who would be singled out by such questions. For traditional families who home-educated the situation was often uncomfortable; however, those families who were also travellers or religious (e.g. where women wore covered heads) or whose parents were lesbian, often suffered criticism from the assumption that they home-educated because of their different life-styles. Despite the many comments concerning alienation however, there were families who believed that they were accepted by their communities and whose children were happily involved with local activities and friendships; indeed, increasingly this seemed to be the norm.

Searching for Something

Families often appeared to be searching for something indefinable. This may have accounted for much of the change seen in some families over time. Most probably this was connected with the families' sense of identity, itself connected with the families' isolation from the community, imagined or otherwise. School life is fairly consuming for both parents and children; there are social events, meetings, school runs, after school events and the PTA; however, without these 'pegs' to arranged family life, some families were faced with a need to fill in the spaces. For some this was attained by joining a religion whilst for others moving to a community resolved the issues. Searching for child-friends was a fairly consuming occupation as parents often worried about their children's social life - almost 19% of parents believed that their children might suffer socially without parental effort (Section 5.11). Where children were withdrawn from school this was particularly profound as parents worried about their children being 'left out'. Discussions with the home-educated children suggested that this was more a problem for the parents than the children: if the children were lonely they soon adapted and generally appeared to prefer their own company and that of a few friends, to being in school.

From the researcher's home visits there emerged a sense that families were sometimes pursuing an ideal that was neither school nor home-education. Part-time school seemed not to be an appropriate substitute. As the law stands, there is the choice to either go to school or to stay at home. Many families would have preferred a mid-way point perhaps akin to the learning centre model discussed by Rothermel (2000).

Working to the School Year

One of the anomalies found from this research was the extent to which home-educators actually fitted in around the school year. Whilst some families did feel acutely the absence of the school related 'pegs' there were many who could not escape them even if they had they wanted to. Many of the home-educated children had friends in school and so were restricted to playing with them only after 3.15pm. There were families with children both at home and at school and home-education therefore, was necessarily organised around the schoolchild sibling's day. Home-educated children involved with after school activities - music, swimming, language and dance classes, Brownies, Rainbows, Woodcraft Folk etc. needed to be home from their home-education groups or to have completed their day's activities so that they could attend. It was clear that involvement with so called 'after school' activities was integral to the social life of many home-educated children and itself ensured that they were not isolated. It was the link between home and school-educated children and was often crucial for those home-educated children, in giving them a sense of involvement. The extent to which home-educators were influenced by the school time-table was perhaps hinted at by the 13% of home-educators who described their 'home-education day' as finishing by 3pm (Table 5.7).

This arrangement around the school timetable also applied to amenities that were better used whilst other children were in school, thus avoiding the crowds (ie. swimming, library, museums). Holidays too, were often taken, either during term-time to avoid high season costs or during holiday time to allow the children to meet and enjoy time with other children. After school activities involved half-term and long holiday breaks and this, inevitably, affected the home-educated children's own timetables.

Change Over Time

As families home-educated they often underwent a metamorphosis. Parents who had left school early themselves often found that learning in tandem with their children motivated them to take further or higher education courses themselves. Others found it appropriate to re-arrange their working lives to enable them to share responsibility for the children. Fathers had more involvement with their children's day to day education and family life than might normally be expected had the children been in school. Flexibility and fluidity characterised many home-educating families as they searched for, tried out and found, a way that suited them. There was also change from home to school: sometimes the children simply wanted to try school and the parents acceded, whilst in other families circumstances altered so that home-education was no longer feasible. Occasionally parents decided that they did not want to continue the commitment.

For some the rate of change was faster than others. Quite conventional families, after deciding to home-educate, appeared to change, sometimes, quite radically, as they became more attuned to their new circumstances. One family when first visited, lived fairly conventionally, in a brick town house: the following year they were living in a community on a huge tract of rural land. The Johnson's became Jehovah's Witnesses, whilst the divorced parents of the girl taken into care, reconciled as a result of their problems related to home-education.

These 'knock on' effects often involved attitudes to childcare and medicine, tending to bring about the questioning of many 'taken for granted' aspects of conventional society. Noticeable, and an example of actual practice, was the way in which home-educators tended towards long-term breastfeeding of their children beyond infancy. This tendency became increasingly apparent during the home-visits. At time when the WHO and UNICEF recommendations for breastfeeding are that it extends to a period of 2 years of age 'or beyond' (WHO 2000) and the natural age of weaning in humans, is considered to be from 2.5 years, with a maximum of 7.0 years (Dettwyler 1995), but when in the UK 44% of mothers do not breastfeed at all (Office for National Statistics 1995), home-educators were, perhaps, leading the way in child health and in protecting their children against infections through natural means. This finding was not about 'eccentric' home-educators choosing to nurse their children long term, but illustrated the way in which, because the mothers were at home, day-to-day practices altered and the parameters of what was possible broadened. Where parents were not restricted by work and school they could spend more time devoted to the care of their children. It was logical that if a mother was not preparing a child for school or getting ready to 'return' to work, there was no need to wean. Similarly, natural remedies that tend to involve frequent and varying doses throughout the life-cycle of the illness were far easier to administer to a child readily at hand throughout the day. In this respect, some home-educators reported previous problems with their children's schools because of the refusal by schools to administer non-prescription, frequently dispensed, remedies.

On the educational front, there were the changes made by families as they adjusted to the form of learning that best suited them. Some parents expected to become less formal with time, others more formal (see Section 5.6 for supporting evidence): quite often the reality was the converse of whatever they had expected, as children adapted in their own way to their situation. As Thomas (1998) suggested, the children often discreetly manipulated their daily routine to one that best suited themselves. It was, thus, apparent that within families the children usually benefited from a routine formulated for each individual.

The joy of home-education

This research focused sharply on questions about the details of home-education. This meant that more tangible aspects such as academic levels, socialisation, problems, statistical data etc. were stressed during the research at the expense of the many benefits that families emphasised. Nevertheless, the enthusiasm expressed during the home-visits had been hinted at through the questionnaire with almost 35% saying that they had not realised before commencing home-education that it would be so fulfilling and so much fun (Section 5.6). There was a sense that families on the whole, 'lived' home-education and were extremely content with it, seeing themselves not just as home-educators but as families living a lifestyle of their choosing. This aspect of home-education is important because it highlighted the extent to which home-education was a 'spin off' from lifestyle choices, as opposed to a 'statement' being made by some anti-school lobby. The questionnaire data showed families speaking of learning together about things that matter, the freedom and flexibility that home-education brought them to live according to their own agenda, the sense of doing 'what we want when we want', with children being permitted the space to 'develop naturally' (Section 5.4). Many families valued their closeness and the time to talk and be together. These were aspects of family life that, it seems, often fall casualty to the demands of everyday life where work, school and material standards overshadow the intimacies of a close and loving family life. The researcher observed children running free, enjoying their life and feeling fulfilled: and not just the children, but the parents also.

10.2 Baseline Assessment

The baseline assessment focused on children aged 4 to 5 years in what would have been their Reception Year. This discussion looks at baseline assessment in the context of home-education; the initial 'compulsory education' year in which, according to Tymms et al (1997) children make a 'leap' forward; and the predictive value of such assessments whereby there appeared to be a correlation between attainment and parental attention. The section also discusses the possible reasons for the home-educated children's poor value-added scores, the reasons why children from lower socio-economic backgrounds outperformed their more affluent peers, and finally, the section explores the issues surrounding school starting age and parental involvement.

10.2.1 Baseline Assessment in the Home-Education Context

Baseline testing is used to assist in the planning of children's learning and the measurement of their future progress. In the context of home-education, however, there are two ways to consider this. First, at home children's learning was generally not 'planned' in the way it might be in school, particularly at this early age and parents appeared not to think in terms of 'future progress' but rather of allowing the children to learn at whatever pace suited them. Second, with regard to assessing children's prior learning, the results suggested that the scope of the assessments was too narrow to be able to gauge, in any useful way, the extent of the children's knowledge and experiences. This narrowness might well apply to some extent with both home and school educated children. However, it was clear from the tests that there was an assumption about what children already knew, or should know, and it is likely that this was based upon what the test designers knew about most children's pre-school environments, that is, playgroup, playschool, pre-school. The home-educated children were different because so many had not attended formal pre-school settings and therefore, their experience prior to the baseline testing was, for many, very different from that of the schoolchildren. It was clear therefore, that in view of the individual and idiosyncratic learning experiences of the home-educated children, the assessment alone would not have been particularly useful to gauge the home-educated children's 'pre-school' knowledge.

The children's excellent scores on the PIPS 'Reception' assessment masked their diverse skills. In this respect the message conveyed through the assessment procedure was very much in agreement with the conclusions expressed by Tizard and Hughes (1984) who found that the work of early years teachers was limited by lack of knowledge about the children at home; the conversational input of the children, for example, was often meaningless to the teachers in the absence of knowledge about the children's home life. The situation they describe reflected that found from this assessment, namely that very little was revealed about what the children did and did not, know: the test focus was so narrow that the results data, whilst of interest to this study, was of little significance in respect of individual children. There was the skilled fiddle player with all the cognitive development that such a skill entails, whose vocabulary score was limited by not having encountered the word 'violin' and whose reading scores were poor, despite the competence of the child in reading music: there was also the child who could type, but who lost marks because his fluency was with uppercase letters. There was a poor rhyming score for the child who would not rhyme to order but whose expertise at rhyme bordered on the expert: and the children who failed tasks simply because of not understanding the language of the assessment or the inferences being made by the procedures. All manner of skills went unrecognised by the test, the designers of which, like Tizard and Hughes' teachers, had no prior knowledge of the children being assessed. The 'End of Reception' assessment was based upon the prior knowledge that schoolchildren ought to have amassed by that point in their educational career; however, for the home-educated cohort, none of whom had followed the path of the school style curriculum, the latter assessment was substantially more different from their learning than the initial assessment had been. The difference between the mathematical components of the 'Start' and 'End of Reception' assessments was, for the home-educated children, startling: they were unable to see continuity between the two assessments, one involving covering up one bike etc. (Image 10.1) and the other more formally laid out (Image 10.2). Schoolchildren would have been submitted to a smooth transition between styles over the course of the year.

Image 10.1: mathematics at the 'start of reception'

Image 10. 2: mathematics at the 'end of reception'

A further factor, one which in effect penalised the home-educated cohort, was that at the 'Start of Reception' they had performed especially well, with most children able to complete the set tasks. This meant that 10 months later the home educated sample had very little, if any, of the 'Start of Reception' assessment to 're-do', before being presented with the second, differently constructed, 'End of Reception' assessment[3]. Most schoolchildren, however, having performed less well, were 'treated' to a re-run of the first, colourful and attractive assessment, many not completing very much of the second assessment at all. There was considerable difference, for the home-educated children, between having to point to the one word from a choice of four, that matched an accompanying picture (Image 10.3) and having to read a nonsense sentence such as, 'Nightmares are something I don't dream about' under a picture of a cat (PIPS 1998e), as illustrated in Image 10.4, below. Thus, whilst at the 'Start of Reception' the assessment was pretty much equal for both home and schooled children, by the 'End of Reception' this had dissipated, leaving the home-educated cohort with a poorer value-added score than the schoolchildren.

Image 10. 3: example of a reading task from the 'start of reception' assessment

Image 10. 4: example of a reading task from the 'end of reception' assessment

Whilst many elements of the second assessment had initially appeared to the researcher to be child friendly, during face-to-face administration it became clear that the home-educated children were reacting negatively, when contrasted with their positive attitudes to the 'Start of Reception' assessment. Schoolchildren had, by the time they took the 'End of Reception' assessment been trained into the style used in the 'End of Reception' tasks, whereas many of the home-educated cohort with their less identifiable learning experiences, floundered at the formal test presentation. Thus, to evaluate the home-educated children's learning, the interview field-notes and 'passive observation' were crucial accessories to the Baseline assessment. It was also the case that test reliability[4], validity[5] and predictive validity at such an early age was debatable, equally for the home educated as for the schoolchildren.

10.2.2 First Compulsory Education Year[6] in the Home-Education Context

The home-educated children demonstrated a slightly greater increase in points over the reception 'year' than had the children from Tymms et al. (1997), that is, 45.32 as opposed to 42. This higher score was curious in the light of the home-educated children's relatively poor value-added scores (for reading, over 50% of the children were in the bottom 10% band). Value-added progress, however, related to the progress made between children's 'Start' and 'End of Reception' scores, whilst the 'points' difference (see Section 6.2, Table 6.2) referred to the contrast between the oldest (n=8) and youngest (n=4) children's scores at the 'Start of Reception', and their progress from 'Start' to 'End of Reception'.

Tymms et al. had attributed what they termed, the 'massive difference' in points, to the first year at school. However, the first measure was between the oldest and youngest of a cohort of new-schoolers whilst the second measure was the score difference between the 'Start' and 'End of Reception' - that is, new and not-so-new schoolers. Therefore, it should be born in mind whilst considering Tymms et al's claim of a 'leap', that the 'comparative' measures used were very different. For a more accurate comparison there needed to be a measure of differences between school minus one year and 'Start of Reception', and between 'Start' and 'End of Reception'. Tymms et al's school based comparative measure approach does not account for the home-educated cohort's point gain. After all, the home-educated children's progress could not be attributed to school. During the 'Reception' year, many of the home-educating parents, as the commentaries illustrate, would not have claimed to have 'taught' their children at all, allowing the children instead to learn at their own pace with parents simply responding to their children's lead (i.e. answering questions as they arose), rather than adopting any parent initiated learning sessions. Nevertheless, it was conceivable that parents, despite protestations to the contrary, were actually observing national, compulsory schooling norms. However, in view of the findings, not just from this element of the research but from the research overall, this explanation is highly unlikely; the interviews and observations generally showed children learning in relaxed atmospheres that bore little resemblance to school norms.

In Chapter 3, it was suggested that schoolchildren's 'Start of Reception' scores were, perhaps, artificially low, owing to the upset, reported by Riley (1996), that some children experience when starting school. Assuming that the point 'leap' between 'Start' and 'End of reception' made by the home-educated children (45.32 points as opposed to 42 points for the schoolchildren, see Table 6.2) was the result of their having scored lower at the 'Start of Reception' than they might potentially have managed (in spite of their excellent scores), it was perhaps possible, that the children at school nationally and in this home-educated sample were disturbed by the 'Start of Reception' assessment (as opposed to any significant change in their lives) and, thus, produced 'low' scores (lower than potentially they might have been). These 'low' scores would then contrast with the higher 'End of Reception' scores to produce an apparent 'leap'. Observation of the home-educated children suggests this was unlikely, however, since the children generally enjoyed the 'Start of Reception' assessment.

A more reasonable explanation might be that the 'Start of Reception' was biased against the schoolchildren who, owing to (if such a concept exists) the 'waiting to start school' philosophy, were less prepared than their home-educated counterparts for any assessment of their knowledge, whilst the 'End of Reception' instrument was biased against the home-educated children who did not have the 'benefit' of being trained in more formal test instruments. This would account for the home-educated children's good initial scores, poorer final scores and poor value-added scores and also for the schoolchildren's depressed 'Start' scores and good value-added. In terms of the 'leap', it could be that the measures used were, in so far as the schoolchildren were concerned, too different to be useful and in respect of the home-educated children were the result of exposure to the instrument itself regardless of age difference and progress; at the start all the children were unfamiliar with the test instrument and by the end they were all more familiar, irrespective of being home or school educated.

Another possible explanation attributes the difference within the home-educated cohort's scores, to the test instrument. The contrast between the 'Start' and 'End of Reception' assessment design, from the viewpoint of the home-educated children, was substantial (see the differences between Images 10.1 and 10.2, and 10.3 and 10.4); the reactions of participants at the times of testing made this apparent. The difference in children's scores at the 'Start of Reception' was effectively measured on one instrument but the latter difference was evaluated using data from two different measures, those of the 'Start' and 'End of Reception' and for many just one very different measure ('End'). Thus, the difference in instruments may have accounted both for the poor value-added performance and also for the apparent 'leap' over the 'year', of the children involved with this research. As explained earlier in the 'Results' section, the difference between assessment instruments may not have been so apparent to schoolchildren who would, during their initial year have been accustomed to the style used in the 'End of Reception' instrument, thereby making it possible that their improvement was, indeed, a response to school, either positively (Tymms et al. 1997) or otherwise (Riley 1996).

10.2.3 Predictive Value of Assessments with Home-Educated Children

From the data it appeared that some predictions could be made about the home-educated children's progress. Their 'Start Total' correlated well with their 'End Maths' and 'End Reading' scores[7]. As the children commonly spent their time with either one parent or the other, parental influence appears to have been the cause of the children's exceptionally high 'Start of Reception' scores. In the same way that Pederson et al. (1978) had found that the effect on children of a good teacher was enduring, so the home-educated children's continued high scores may have been the result of the strong bond with their parents: the academic results discussed in Sections 10.3 and 10.4 (Literacy and Maths) indicate that the correlation between attainment and parental attention may continue, at least until 11 years of age. Tymms et al. (1997) concluded that the extent of schoolchildren's progression related to their prior attainment and this certainly appeared to be the case with the sample at hand also. Just as Tizard et al. (1988) found that teachers responded better to children whose company they most enjoyed, it would seem that the parents were those who had the most invested in their children's development; this pattern created the cycle of positive attribution, described by Georgiou (1999). Tizard and Hughes (1984) found that parents had the advantage of understanding the context of their children's lives, a finding strongly supported by the home-education 'Reception Year' data. With regard to Rich Harris' (1995) belief that peer groups accounted for the greatest impact on development, this may have been born out by the present findings, whereby parents and other family members, in effect, tended to replace the 'peer' group influence.

10.2.4 Value-added in the Context of Home-Education

The home-educated children's value-added scores did not reflect their high 'Start' and 'End of Reception' scores and overall, their value-added scores were fairly poor. One possible reason for this may be that the home-educated four-year-olds may have been accumulating knowledge on a gentle incline since birth with no foreseen alteration to that pattern of learning.

10.2.5 Affluence

All the children scored good marks, whatever their background and family structure. However, contrary to the findings of Tymms et al. (1997) the home-educated children from the lower socio-economic groups scored significantly higher than those with professional parents on the Registrar General's classification (Rose and O'Reilly 1998). The most obvious reason for their doing well, and one that is supported by evidence from other sources, is that home-educated children are, at least amongst their own ranks, free from the stigma of being poor, simply because they are not learning in an environment where affluence and labelling are an issue. Goldthorpe (1996) provided a scenario of how social class in school affects children's chances and this was born out by inference in the work of Galloway (1985) and by the writings of Tizard and Hughes (1984, pp256-257). Tizard and Hughes (1984, pp252-253) described the 'working-class mothers' as 'just as concerned' for their children as 'middle-class mothers'; they further indicated that mothers from the lower socio-economic classes were more likely to adopt 'traditional approaches' to maths and reading. The difference in approaches to learning adopted by the lower socio-economic home-educating parents and those more affluent parents, appeared to mirror that described by Tizard and Hughes and this may well have accounted for the high performance from the children of non-professional families in these school style tests. Children from religious families also tended to come from non-professional families and although such families were not necessarily strict with their children, their beliefs about duty and self-discipline are likely to have had an impact on the children. The children from families at higher socio-economic levels may have been learning under more liberal values and this method, as the results illustrated, was nevertheless a recipe for success. The key to performance irrespective of background was, it seemed, the availability of parent[s] to spend time with their children, since at least one parent in each family was continually present throughout this period of their child's life.

10.2.6 School Starting Age and Parental Involvement, in the Context of the Home-Educated Children's Results

As the work of Aubrey, Tancig, Magajna and Kavkler (2000) and Aubrey and Godfrey (1999) has found, starting school at a later age may well have no detrimental effect upon learning skills. The results from the present study suggested that a delayed school start date may have very positive effects. This does not detract from the suggestion that school makes a difference (Riley 1996), but questions whether such difference is academically and socially beneficial in the shorter term, as evidenced by the home-education findings from the PIPS Baseline research, and in the longer term, as indicated from the findings described in the other chapters.

The deliberations expressed by the PIPS Project (1996) over the usefulness of parental involvement and the remark made by Coulson (Appleyard 1998) that baseline assessment will assist parents, both give the impression that parents are somehow incapable of supporting their children's learning without external assistance. This attitude was summed up in the final lines of Tizard and Hughes' (1984, p 267) work:

'Indeed, in our opinion, it is time to shift the emphasis away from what parents should learn from professionals, and towards what professionals can learn from studying parents and children at home.'

Tizard and Hughes (1984, p 267)

This statement by Tizard and Hughes echoed the comments from the home-educating families involved with the research. The findings lent support to the idea that parental input was not just a useful support tool, but that indeed, it could substitute for professional assistance.

10.3 Literacy

The home educated children’s NLP assessments for Years 1, 3 and 5 produced mean scores for each year group in excess of one standard deviation (SD) above the national mean[8]. The literacy mean for the PIPS Year 2 assessment cohort (n=17) was also more than one SD higher than the norm[9]. Thus, the majority of the home-educated children gained scores that were at least one SD over the norm.

The differences between mean scores across all the reading tests were interesting, with the four and six-year-olds performing at a higher level for their age than the five, seven, eight and ten-year-olds. The perception gained by this researcher, through the home visits, assessment administration and data analysis, was that the two former age groups enjoyed their assessments far more than those in the latter age groups. The five-year-olds were disappointed that the PIPS 'End of Reception' assessment did not resemble the 'Start of Reception' one in terms of structure. It also was apparent that the NLP Year 1 (six-year-olds) test differed in structure from those for Years 3 and 5 (eight and ten-year olds respectively). Moreover, the differences between the PIPS 'Start of Reception' and the NLP Year 1 scores were not statistically significant, contrasting with the significant pair-wise comparisons between the PIPS 'Start' and 'End' of reception, and between the 'PIPS End of Reception' and NLP Year 1 groups.

The gender findings added fuel to the debate over differences in gender performance, namely, within the NLP assessments the difference was in favour of the girls (Section 7.1.3). This advantage in the girls' favour supported findings from Sainsbury's (1998) 'National Literacy Project' research. The PIPS Year 2 test however, revealed that the higher means in each of the four domains observed, belonged to the males (Section 7.2.2). Thus, girls attained the highest scores on the NLP assessments and the boys performed better on the PIPS Year 2 assessment. A possible explanation for this came from one father who pointed out that the PIPS Year 2 test was biased in the favour of boys, with most of the names used in the assessment being masculine. This emphasis on masculinity was absent from the NLP tests where a gender balance was evident (Section 7.1.3).

Whether the gender bias in the PIPS Year 2 (Section 7.4.4) and or the existence of a maths component spurred the boys on to greater heights is unclear, but the PIPS Year 2 test did, perhaps surprisingly, produce a homogeneity between maths and reading score components for all participants. Unlike school, home-education does not make for a gender biased environment; there is no gender peer pressure for example and in many families both parents share roles and responsibilities. It seems that within the UK at least, both parents play a significant role in their children's learning leading perhaps to a better understanding by their children of gender balance.

The value-added scores produced by the PIPS Year 2 Literature data were, in the absence of prior assessment, 'concurrent', contrasting contextual information with academic performance. The home-educated participants showed a more central tendency than expected from a normally distributed sample but nevertheless, performed in value-added terms, 'as expected'. This conflicted somewhat with the PIPS 'Start' and 'End of Reception' findings, discussed in Chapter 6, namely that despite the high percentages of home-educated four-year-olds achieving above average scores at the 'Start of Reception', their performance in terms of valued-added progress during the year, was rather poor: for Maths, they had maintained the expected learning incline over the ten month 'school year', whilst in English, only 27% of the group managed average, or above average, progress. The impact of these low value-added scores was mitigated by the finding that at the 'End of Reception' 21% of home-educated children still scored two standard deviations above the norm, compared with only 2-3% of children nationally. It had appeared from the evidence that there might, however, come a time when the home-educated children, with their lower value-added scores and more gentle incline, might be overtaken by the more aggressive upward slant of the schoolchildren's value-added performance. However, Tymms (1998) has reported evidence from various sources, that schoolchildren's performance actually slows, or even regresses, during vacation periods. Coupled with the disclosure by Galton in Galton, Hargreaves, Comber, Pell and Wall (1999), that following transition to secondary school 40% of 11 year olds make little progress in their first year and 7% actually 'unlearn' what they previously knew, it is possible to speculate, in view of the results, that the home-educated children's 0-11 year old value-added performance might, in fact, outweigh that of their school peers.

It is also possible to hypothesise that information acquired by home-educated children was more readily retained than that acquired didactically by their school peers, having absorbed their knowledge gradually by virtue of informal repetition, and assimilation through everyday learning. Karmiloff-Smith (1994) described the way in which children take on board information from the external world, reorganising it internally and combining it through conflict and, or agreement with previously internalised knowledge, eventually attaining mastery over the situation. She described this process as 'representational redescription' (RR).

Application of Karmiloff-Smith's model of cognitive development came to light unexpectedly during the research of Pine, Messer and Godfrey (1999) into the effects of didactic and guided teaching of a balancing task using symmetrical and asymmetrical rods with a fulcrum. Whilst the taught groups received instruction, the control group had none, being allowed to practise alone only. Results showed that before intervention all the children scored poorly on the task. Following intervention, all three groups improved significantly, the two taught groups demonstrating substantial improvement over the control group, with a significant difference appearing between the didactic and control groups. A week later, without further intervention, the didactic group had regressed in their on-task ability and the guided participation group had remained at virtually the same level. The control group, however, had improved to a level exceeding the guided participation group, showing a rise of 1.31 points contrasted with .07 for the guided participation group and -0.43 for the didactic group. The children tested were all school attenders but the experiment, nevertheless, lends weight to Karmiloff-Smith's theory of an incubation period within her representational redescription model and thus serves to illustrate the ability of children to organise their own learning.

Galloway (1982), in a study of school absenteeism, found that the absentees in his sample outperformed their school attending counterparts in reading ability. Despite Galloway's rationalisation of this event[10] it is possible, remotely perhaps, that in the absence of tuition, the children had benefited from their informal learning environment. Prior to the present study, specific evidence-based research on the formal assessment of children's informal learning ability has not it appears, been undertaken, therefore, the evidence taken from the work of Pine, Messer and Godfrey (1999) and Galloway (1982) is important[11]. Similarly, there may be other studies that have incidentally reported on informal learning without actually debating the implications of such findings.

The literacy skills of the home-educated cohort could most adequately be described as non-conformist and evidence therefore for Kress' (1997) demand that definitions be reappraised and literacy viewed in terms of connection, continuity and coherence with other skills. This appears to support the connectionist model submitted by Karmiloff-Smith (1994). The finding that the home-educated children had attained higher levels in their literacy assessments than their school peer group apparently supported Goodman's (1972) argument that we do not fully understand the acquisition of literacy skills and that intervention, contrary to assisting children, may in fact not be as beneficial as has been assumed (Riley 1996). Like Kress, Goodman held the global, holistic view that Karmiloff-Smith's model follows and which, to an extent, trivialises both the DfEE's (1998c) policy laden explanation of literacy skills and Beard's (1999) exposition of the importance of school. Meighan (1995) described the unplanned route to language competence followed by many home-educators, and whilst Meighan's work was based on informal qualitative intuition, the academic results of the present research, coupled with the qualitative fieldnotes, supported his view, bearing in mind that there were exceptions to the rule and that it was impossible to form judgements concerning those home-educators who did not receive and return initial questionnaires, some of whom will undoubtedly have chosen to isolate themselves from the wider community. Thomas' (1998) allusion to the 'apprenticeship' model as an active learning medium within home-educated families, also found accord within the present research, although it was impossible to know with these children, as with any children, the extent to which each influence is important; a view posited by Moseley, Merrell and Tymms (1998), commenting in the context of literacy skills.

10.4 Numeracy

The home-educated children's maths scores between the three PIPS assessments averaged 64 points, and thus were 14 points and almost 1.5 standard deviations above the national norm. As a group, over half the children scored, on each assessment, in a band where only 16% of children nationally score. The stability across the three age groups, 4, 5 and 7 years, implied that the home-educated children were able to maintain their level at least throughout these four years. The 'Reception' children in the cohort had been tested twice, once at the 'Start' and again at the 'End' of the 'year', whereas the PIPS Year 2 group were entirely different children. This similarity and difference between the three assessment groups may serve to emphasise the point that the home-educated children were performing to a high standard generally. Had just one group been followed, it might have been possible to identify this cohort as 'exceptional'; however, where two different 'Maths' cohorts were assessed (one twice over a ten month period) and produced similar results, this strengthened the validity of the results. The adoption of three tests produced by the same team further indicated that there was some level of continuity of assessment type.

In general accordance with the data from North American studies (Rudner 1999, Ray 1998, 1994; and Richman, Hornberger and Ebeling 1992), 88.8% of the UK home-educated PIPS Year 2 group scored above the 75^th percentile, 66.6% attaining a level exceeding the 90^th percentile.

The value-added scores endorsed the home-educated cohort's level of attainment by showing the predicted percentage of the 'Reception' cohort to have progressed satisfactorily in Maths over the ten month period between the 'Start' and 'End of Reception', and almost three quarters of the PIPS Year 2 group to have attained this level in concurrent terms, as judged in light of their 'Context' scores.

Although the PIPS Year 2 'Context' assessment (Self-Esteem, Attitudes, Non-Verbal Ability, Picture Vocabulary and Cultural Capital - see Section 7.2 and Appendices 4.9 and 7.6 for further descriptions) had not, perhaps, been intended as an assessment in its own right, the questions relating to home background were particularly pertinent for the home-educated cohort. The finding that 'Context' scores contrasted well with both 'Reading' and 'Maths' grades (Scatterplots 7.1 and 8.1) implied that the children were, perhaps, experiencing a valid curriculum in terms of at least some of the ‘broad and balanced’ demands made by legislation (Educational Reform Act 1988).

There did not appear to be gender continuity amongst the Maths scores, with girls performing better at PIPS 'Start of Reception' and boys doing better by the PIPS 'End of Reception' and PIPS Year 2. Interestingly, Aubrey and Godfrey (1999), in a three cycle mathematics testing programme of 300, 5 and 6 year olds within England, found that at mid-Reception Year and mid-Year 1, boys had higher mean scores than the girls, whilst the boys' end of Reception Year grades were not so good as those of the girls. Without reference to Aubrey and Godfrey's (1999) standardised score data, however, it was not possible to assess whether there had been a similar gender pattern of performance between assessments as those illustrated by Graph 8.1 of the present study's three cycle 'Maths' testing program. The national data is no clearer: whilst in Maths for the years 1999-2001 at Key Stage 1, girls outperformed boys, at Key Stages 2 and 3 it was the boys who tended to perform better (DfES 2001e). Whether the boys' better performance, at least for PIPS Year 2, was related to the apparent 'maleness' of that assessment instrument was unclear.

In view of the high scores generally for each home-education 'Maths' appraisal, it was, perhaps, surprising that so many of the children should have regarded mathematics with indifference (55.5%) with just over one in ten expressing unhappiness about the subject. However, as this research has found, home-educators' interpretations of various concepts, appear to differ considerably both from fellow home-educators and most certainly from those of other people generally. Quite why this is so is unclear, unless related to the absence of mass peer interaction at both parent and child levels with whom to exchange common inference and understanding. It may well be that some home-educators harbour a purposeful quest for 'distinctness' from wider society and that this dimension is, in their eyes, desirable. Hence, when a home-educated child makes known his or her opinion of mathematics, this should not be interpreted with the same understanding that one might have if the child had been attending school. This difference in perception appeared to lie at the heart of many difficulties between the home-education community and the wider society.

The subject of assistance recurred throughout the post ‘Reception’ age assessment assessments and discussions (see Sections 7.4 and 8.2). The ‘Reception’ cohort was researcher examined, whilst the NLP and PIPS Year 2 assessments were parent-administered, with several tests re-administered by the researcher (see Appendices 7.1, 7.2, 7.3). Parents were asked to make notes relating to the assessment procedures they experienced and it was clear from these notes that some assistance was provided. These instances were few and assessment scores excluded any items where help had been given. This was a far preferable method to having requested parents to withdraw altogether. A certain amount of trust needed to be given because, as this research concludes, home-educating parents are those who best know their children. Had the children been in school there would have been no question over the reliability and honesty of the teacher administrators and here too, parents needed to be treated with the same respect as would have been afforded to the children's teachers at school. Parents were aware that this was a test and therefore any assistance needed to be viewed by the researcher as having been the result of a parental decision that this was necessary with regard to the child in question. This, indeed, has formed a vital part of the research. Through attention to the procedures followed by families, it was possible to assess the situation in a more holistic manner than would have been possible with purely quantitative data, provided under strict exam standards. Through the annotation of factors such as levels of assistance deemed necessary, it was possible to see ways in which families coped with the assessments. Generally, it appeared (and as the researcher observed) that tests were administered in a relaxed homely 'test' atmosphere where children were not heavily monitored but did nevertheless have their parents close by. It seemed that at times the parent, having viewed the assessment at an earlier date prior to administration, chose to run through some mathematical procedures that had not yet been the subject of specific learning. This level of assistance, occurring prior to administration, simply provided the children with the tools to tackle the items within the test. Where parents noted that they had assisted with such and such an item, they supported their decision by explaining that their child could not comprehend that their parent would not supply such assistance as required. This emphasised the nature of the parent-child relationship whereby the parent provides such information as is required, at the time required. The idea of waiting to have questions answered at a later date when the context was forgotten, was not, it appeared, an attractive one, as Mrs Easter's remark aptly summed up, 'What's the point of the exercise if you can't learn?'

Several parents pointed out the difference between their children meeting a problem on a page of a book and encountering it in a real life situation. It was easy to see how a child of seven years or under might identify with computations related to pocket money or to the sharing of items amongst friends, yet find it difficult to transfer these problems on the written page.

Kahney (1983 p.63-65) identified the difficulty involved with transferring information in one form to another unless provided with a hint, even where the pattern represented is identical. Whether such an inability to make connections between similar problems is a symptom of poor mathematical literacy is unclear: the evidence from Kahney (1994 pp. 91-95) suggests that this is characteristic of novices, inferring that it is an essential part of the process towards becoming an expert, whereby declarative knowledge transforms itself into procedural knowledge[12]. In dealing with everyday mathematical issues, such as supermarket shopping, it is possible that some of the home-educated children were able to detect analogies between questions set on the assessment and those they met with in daily life; therefore, this may have assisted some children in answering mathematical concepts which, according to their parents, they had not covered. In the absence of having followed school curricula, these home-educated children were, perhaps, more dependent upon their ability to make analogies than schoolchildren; this may itself have been a skill that this class of children excelled in, although in the absence of comparative assessment on this ability such a proposition remains speculative.

In a similar vein, there were those home-educated children whose parents felt it important to initiate them into the world as young adults (Das Gupta1994) rather than as 'children' in the sense of their cognitive abilities being tangibly different from adults. Such parents described schools as creating synthetic representations to prepare children for 'real life' rather than expose them to real life itself. This concept appeared to underlie the reason why Paula's mother had introduced her daughter to the mathematical symbol 'n' rather than the use of a star, to denote a missing number (Section 8.2). It can of course be argued that children should understand school-based conventions and yet, if a child is not to enter school, what need is there to become acquainted with such representations as are used there? The language of school and thus, school mathematics was, for some, irrelevant. The concept of sharing was also viewed by some parents as having its source in school activity and was, therefore, not necessarily such a familiar idea to the home-educated children as to those attending school. This was illustrated in one child's inability to detect an analogy between sibling sharing and the written concept. During interviews, a number of parents pointed out that adults are not generally made to share their possessions beyond the family and that it was unrealistic to expect children to do so.

Despite high scores, identification of the children's actual levels of education was masked by assessment content that could not possibly relate to the diverse areas of their different learning environments: this may well have been true of schoolchildren also. It is notable that plans to remove key exams from the school timetable are being proposed in Hong Kong (Forestier 1999): the Hong Kong Education Commission Chairman proposed:

‘What we are hoping is to reduce the emphasis on public exams which tend to measure only one aspect of students’ ability and distort the entire learning process.’

Leung (1999)

The potential unsuitability of early testing has also been referred to by Aubrey and Godfrey (1999) in their international comparison of mathematical standards amongst five and six year olds. Of particular note was that the English reception aged children’s Dutch counterparts were still, owing to a later school starting age, pre-schoolers and yet performed as well as their English peers who had, prior to the assessment, experienced a period in school. Aubrey (1997) highlighted the way in which, within English schools, children’s pre-school mathematical knowledge is undervalued, despite research by Tizard, Blatchford, Burke, Farquar and Plewis (1998) showing that four-year-old’s pre-school knowledge was the best predictor of ability in school at seven years of age. Interestingly too, Aubrey et al. (2000) found that despite their later school starting age, by eight years of age Slovenian children were mathematically at a level with the English cohorts who by then, had been in school for four years in comparison with the Slovenians' one year. That ideas about mathematical knowledge as the domain of formal tuition persist, was evident in the work of Thomas (1998) and may, earlier, have been informed by research such as Rutter et al. (1979), which, although relating specifically to secondary age mathematics learning, promoted the view that mathematical learning was less likely to occur outside school than other subjects. Assuming that children in school are learning mathematics deemed appropriate to their age, the substantially better scores attained by the home-educated group bide well with Karmiloff-Smith's (1994) theory of 'representational redescription', whereby informal learning is beneficial for cognitive development: the work of Aubrey and Godfrey (1999) also appears to support this opinion, as do the findings of Thomas (1998). Aubrey and Godfrey (1999) found that amongst the English children to whom they administered mathematics tests at mid-reception year, at the end of reception year and again during mid Year 1, there was a ‘large difference’ between the end of reception and mid Year 1 scores. This difference, although on an upward incline, reflects the pattern detected in the present research whereby significant differences were apparent, between the ‘End of Reception’ and ‘PIPS Year 2 scores. Aubrey and Godfrey (1999) attributed their sample differences to the ‘leaps and discontinuities’ of mathematical learning. They further referred to the existence of an incubation period whereby maturation occurs internally during what appears from the outside, to be a period of stagnation, until the matured knowledge is externalised. This concept expressed by Aubrey and Godfrey appears to match that of Karmiloff-Smith (1994), cited in Section 10.3, and thus does not appear to be the sole domain of mathematical thinking. The existence of incubation periods also explains, perhaps, the reason why the home-educated children appeared to be continually surprising their parents with what they knew.

Generally, the home-educated children's learning and attainment appeared indicative of mathematical knowledge in the form suggested by Brown (1999) when referring to the definition of numeracy as perceived by the National Numeracy Project (Section 3.3.1). There did not appear to be the anxiety amongst the home-educators over mathematical issues that been conveyed through the work of Thomas (1998); in fact, in the under 11 age group targeted by the present research, mathematical learning appeared to be low key; and an area of learning with no especial difficulty attached.

10.5 Social and Psychological Skills

10.5.1 Home-Educated Children's Social Skills

The CABS questionnaire explored children's social skills in the five domains: 'Positives', 'Negatives', 'Requests', 'Conversations' and 'Feelings' (See Appendix 4.16). The home-educated children appeared far more passive than aggressive in each of the domains measured except in the domain focusing on the ability to make and receive complaints ('Negatives') where there were some very passive and very aggressive responses, indicative of some difficulties in this respect. The children were most able in terms of giving and receiving compliments, considered by Michelson et al. (1983) to be linked to good self-esteem, positive socialisation and confidence. Overall, the levels of passivity came within the same score band as those of Shyers' (1992) North American home-educated and school educated samples, who had themselves scored within the same, mostly passive, band.

The home-educated children's assertiveness score also overlapped the score band described by Michelson et al. (1979) as 'normal', although the home-educated children did tend towards the higher, less assertive extreme within that band. By scoring within the 'normal' spectrum, the children had, however, achieved Michelson et al.'s (1983) social skills criteria. His view that 'social skills' were not connected with school style socialisation was supported by the results of the present home-educated sample. In sum, the CABS data showed that the home-educated children were not in any way unusual.

The differences between the RRS and the SDQ instruments in the 'Prosocial' domain were echoed both across the groups and within the cross-scale sample. The SDQ differed from the RRS and CABS measures in that the SDQ 'diagnosed' most of the home-educated children as experiencing problems on the prosocial scale, that is, items concerning sharing, consideration for others and helping out. Moreover, the SDQ indicated that over one fifth of the sample exhibited 'peer problems'. Peer problems were considered to exist where children liked to play alone, preferred adult company, were picked upon by other children, did not have at least one good friend and were not generally liked by other children. In contrast, the RRS 'Prosocial' results showed the children to be outstandingly prosocial, a finding that cross-referenced well with the CABS data where the children had displayed 'normal' social skills. Why there should be such a difference between how these two scales perceived the home-educated children is certainly a matter for further inquiry. It seemed that whilst the SDQ (5 prosocial items) and the RRS (10 prosocial items) appeared to share five items, the wording perhaps differed enough to make a difference to the parent raters. One example would be the SDQ item 'Shares readily with other children' in contrast to the RRS 'Shares out treats with friends'. The SDQ question concerns sharing per se and applies only to children whilst the RRS items speaks specifically of treats and uses the non-age specific term 'friends'. In home-education terms these subtle differences in wording, as has been noted elsewhere (eg. Section 4.5.2) could substantially affect how the items would be perceived.

The RRS and CABS 'social' results agreed well with the researcher observations of home-educated children. The SDQ result therefore, may have misrepresented the sample's social side and this could, perhaps, be attributed to the SDQ/RRS item ratio of 5:10 in the 'Prosocial' domain. It was possible that the SDQ did not provide for the home-educators, who defined very different patterns of behaviour as 'normal'. If, however, the SDQ scale was the most exacting measure in this domain, then the results certainly reinforced the viewpoints expressed earlier, in Section 3.4.2 of the stereotypical home-educated child as shy, passive, lethargic and isolated (Aiex 1994; Wragg 1997; Hastings 1998). What appeared to be the case, however, taking into consideration the CABS, RRS and observation data, was that the SDQ represented a narrower vision of what social attributes are considered to be and that such a perspective reflected the prosocially desirable aspects for schoolchildren who, after all, form the majority of children in the UK and therefore, make up the 'norm' for whom the SDQ scale was designed.

The SDQ 'Peer Problems Scale' domain, where the home-educated sample also did less well than expected in terms defined by Goodman (1997), further reflected common expectations that children should prefer the company of other children, be liked by other children, have at least one good friend and prefer to play in groups. A number of home-educators, both through their initial questionnaires and during interview had described this image of peer groups as precisely the reason why they would not want their child to associate in large school-style groups with inherent peer pressure. Michelson et al. (1983) certainly found school style socialisation to be the cause of 'maladaptive behaviour' that negatively affected both academic performance, and personal and social development.

Despite the home-educated sample differing from the standardisation sample within the 'Prosocial' and 'Peer Problems Scale' SDQ domains when analysed in accordance with Goodman's (1997) instructions (adding each participant's domain scores together), these differences were not evident when an alternative, albeit contentious, analysis was used, whereby, instead of calculating domain scores for the individual, individual item scores were summed across the sample, on each item, before being grouped by domain (see Section 9.3.3 for a detailed explanation) .

Thus, there was no consistent evidence, at least for the majority[13] of the home-educated children, that they were lonely, isolated, anxious, phobic, socially inept and academically lazy (Desforges 1999, Webb, 1999; Shearer 1999; Hastings 1998). The majority of the children were socially adept and prosocially healthy, hence, it is questionable whether the children actually had any need for socialisation beyond that with which they themselves were content. The proposition that children need socialisation on the scale offered by schooling, was not substantiated. Any expression of discontent with the level of social skills and prosocial behaviour acquired by the children through their home-based education would most probably have emerged in the course of the psychosocial investigation and if not here, then at other points in the study.

This research found that at the very extremes of the spectrum, the level of emotion expressed about socially related issues, both by critics of home-education and home-educators alike, was extreme, the one group apparently feeding the other in a continuous round of attack and defence. It was argued in Section 3.1.3 that the rush to immerse children earlier and earlier in 'education' may not relate to any developmental benefit. From the questionnaire data, the field-notes and from the psychosocial results, it may be that in terms of social confidence, there is little, if anything, to be gained by mass schooling at an early age. The results from the RRS in particular, indicate that children can actually benefit from living without mass peer socialisation such as that experienced in school.

10.5.2 Home-Educated Children's Behaviour

The results for both the RRS and the Goodman SDQ showed that the group were generally normal in terms of behavioural difficulties. The RRS highlighted behavioural difficulties with ten children (23.8%) and whilst this percentage was high, it was less than the 25% found by Rutter et al. (1975) and only marginally higher than the 20% national figure published by the Mental Health Foundation (Ellis 1998). Where difficulties did exist, the children were split fairly evenly between those with conduct and those with emotional problems. A comparison with the home-educated sample's data and that provided by Ekblad (1990) relating to previous studies, revealed that the home-educated children were more aggressive than the norm and that the girls' levels of anxiety was higher than those found in other studies. This RRS based indication that the sample were aggressive, did not conflict with the passive CABS interpretation of their behaviour: CABS dealt exclusively with the children's social skills, whereas the RRS required that 'difficulty' items and 'Prosocial' items be analysed independently. The fact that the RRS results had confirmed the CABS data in finding the children socially adept, perhaps, validated the feature of the RRS which pinpointed where problems lay. Encouragingly too, the participant commentaries indicated the integrity with which they had completed this assessment.

The SDQ results showed the sample to be generally normal (92.7%[14]) in terms of their 'Total Difficulties' scores. Just 2.4% of the group were identified as exhibiting behavioural problems: this percentage represented the one child with Asperger's Syndrome, who was not, however, the only SEN child in the sample. Whether the RRS was over-rigorous in its 'Total Difficulties' diagnosis, or whether the SDQ was more liberal was difficult to say, just as it was impossible to ascertain whether the differences in the samples had caused the variance in outcome. Participant differences remained a possibility bearing in mind the 100% interscale agreement for the 29%[15] of each sample who provided cross-scale data. An interesting question that, in retrospect, might have resolved this issue, would have been to ask both the RRS and SDQ parents whether they considered their children to have behavioural problems per se.

The high percentage of children identified by the RRS as displaying behavioural difficulties might have related to the large quantity of time that the parents spent with their children, exposed as they were to the fullest range of their behaviour, and thus more likely to see cause for criticism. Galloway (1982), finding that parents of school absentees reported more behavioural problems than parents of children in school, may, perhaps, also have met with this phenomenon.[16] Moreover, Ekblad (1990) found mothers (no data for fathers) more critical of their children than others and certainly where the norm was for 25-30% of doctor's appointments to be related to behavioural problems (Ellis 1998), it would seem that many parents are overt in expressing their concerns. The present SDQ mother and child related data also appeared to confirm this opinion, the mothers finding their children's behaviour more problematic than the children themselves did. Whether these children would have been rated as 'difficult' by a detached rater cannot be known.

Earlier research has indicated a connection between problematic behaviour and academic achievement (See Sections 3.4.2 and 3.4.3) and in the current research six of the ten children scoring above the RRS cut-off point had participated in academic testing. One, a girl with SENs, scored below average on the literacy test, another two scored above average on literacy and three were above average on both literacy and mathematics (two of those scoring above average were graded as exceptional). It was, therefore, in this research, difficult to relate behaviour ratings to academic performance. Of the two high performers, both girls, one followed a fairly academic routine and the other, an autonomous educational style (Meighan 1995; see Section 3.4.4 for a definition) that involved only a regular Kumon[17] routine for mathematics. This finding, therefore, conflicted with previous research in this area (eg. Ekblad 1990; McMichael 1979) that described a link between behavioural problems and academic performance.

An important point to consider when faced with the results from these social and psychological assessments is that classifications of 'normal' are derived from national norms, ie. they represent the average. Since home-educators, by definition, differ from the norm, it is hardly surprising that they were found, at least in some areas, to vary from that norm. The real consideration lies in how society receives such deviation.

10.6 Methodological issues

This thesis has depended upon data derived from questionnaires, field-notes, academic tests and psychosocial assessments. At the questionnaire planning stage access was a major issue and knowing what questions might reasonably be asked without upsetting potential respondents was impossible. With hindsight it is clear that the questionnaire could have involved more direct questions relating, for example, to religion, income etc. It would also have been possible to have included questions directed at the children, relating to their emotions and ideas about the education they were receiving. However, with the knowledge that more questions were possible, comes the understanding that time and money were limiting factors and the inclusion of some questions would have meant the exclusion of others. The questionnaire version used latterly was far more precise and it is hoped that future researchers might usefully refer to this writer's experiences.

The samples in each assessment are small (35, 33, [PIPS Reception] 17, 15, 17, [NLP] 43 [CABS] 51 [SDQ] and 42 [RRS]. It is therefore important that results are treated with some caution. Larger samples would have been preferable but bearing in mind the access problems involved, the number of participants was the largest possible at that time. Notably, the sample sizes are appropriate for the statistical tests used and the results from those tests can therefore be taken as reasonably reliable.

Discussion Endnotes

---

[1] Advice given too by Pat Farenga, President of Holt Associates, USA at the 1998 Growing Without Schooling Conference in Massachusetts, attended by this researcher.

[2] Although Further Education Colleges are 16+ increasing numbers accept younger children.

[3] For 'Start of Reception' children achieving < the cut-off score, the 'End of Reception' assessment consisted of retaking the 'Start of Reception' assessment before being presented with the actual 'End of Reception' booklet.

[4] The reliability of a test is its consistency

[5] Test measures what it is intended to measure

[6] For some of the participants it was their pre-compulsory education year.

[7] However, the children's 'Start Total' scores did not correlate well with their 'End Total' scores, in contrast with the .76 correlation reported by Tymms et al. (1997). This may have occurred as a result of the standardisation process for 'Total' scores.

[8] From the 49 NLP cohort children: 43% scored > 2 SD above the norm; 41 > one SD above the norm; 12% scored within a SD and just 4% of the children produced scores < the mean.

[9] PIPS Year 2 'Literacy': 6% of the children scored > 2 SD above the norm; 70% > one SD above the norm; 12% scored within one SD and 12% of the children produced scores < mean.

[10] He believed that the difference in scores may have been attributable to a 'suggestion' that the absentee children had been placed amongst, and thus compared with, children from remedial classes where there may have been less able children.

[11] Although evidence from Fogelman (1978) does suggest that persistent absence from school is associated with lower attainments and that subsequent regular attendance leads to recovery.

[12] Declarative knowledge: following statements at a verbal level, (from books, instructions, people): Procedural knowledge: skilled performance (car driving) (Kahney (1994p.92).

[13] Between 76% and 92.7%

[14] The contrast between 92.7% found to be overall 'normal' and the 61% found to be prosocially 'abnormal' is that the 'Total Difficulties' score excluded the 'Prosocial' items, which, as highlighted above, appeared to interpret 'normal' prosocial behaviour as that displayed in large groups rather than social interaction with others per se.

[15] 12 children = about 29% of 41 (SDQ) and 42 (RRS)

[16] Galloway found that 83.3% of the primary aged children were with their parents whilst absent from school without authorisation.

[17] Kumon maths involves 10 minutes/day. Worksheets are used and work marked by an instructor. Students visit a centre twice weekly (not, however, the child mentioned above).