Foundations of character: methodological aspects of a study of character development in three- to six-year-old children with a focus on sharing behaviours

Abstract

This article focuses on methodological issues arising in a study of character development, using illustrations of ‘sharing behaviours.’ Based primarily in six early years settings in southeast England the research records naturalistic observations of peer interactions for 55 children aged three to six years. Applying grounded theory to the processes of observing, analysing and interpreting evidence required a cautious and collectively reflective approach. The methodology sought to moderate the influence of the researchers' prior knowledge of ‘grand theories’ of moral development and assumptions about relevance to the observation records. The study's originality lay in the exploration of moral development without reference to any particular grand theory as an explanatory framework; and in the reluctance to be drawn to potentially simplistic rationalisations of the children's intentions on the basis of their observed behaviours. Exploring young children's subjective experiences, this research provides insights into the intricacy of this process, steering away from ‘neat’ findings and attempting to reflect the sophistication of the children's skilful and sometimes surprising negotiations of moral dilemmas. Implications for practice relate to the complexities involved in attempts to unravel the developing moral characters of young children and the practice through which this may be nurtured.

RÉSUMÉ

Cet article se centre sur des questions méthodologiques issues d'une étude du développement du caractère, illustrée par le “ partage de comportements “. L'étude s'appuie sur des observations naturelles d'interactions entre pairs (55 enfants âgés de 3 à 6 ans), principalement recueillies dans six services préscolaires du sud-est anglais. Appliquer la grounded theory dans le processus d'observation, analyser et interpréter des preuves a exigé une approche prudente et collectivement réfléchie. La méthodologie a cherché à modérer l'influence des connaissances antérieures, chez les chercheurs, des “grandes théories” du développement moral et de ce qu'ils pensent de leur pertinence sur le recueil d'observations. L'originalité de l'étude repose sur l'exploration du développement moral sans prendre pour référence une grande théorie comme cadre explicatif; elle repose aussi sur la réticence à la tentation d'une rationalisation simpliste potentielle des intentions des enfants, à partir de l'observation de leurs comportements. En explorant les expériences subjectives des jeunes enfants, cette recherche fournit un aperçu de la complexité de ce processus et tente, en évitant les résultats “propres”, de refléter le raffinement des talents des enfants et parfois leurs étonnantes négociations des dilemmes moraux. Les implications pour la pratique reposent sur les complexités impliquées dans les tentatives de dégager le développement du caractère moral des jeunes enfants et la pratique par laquelle il peut être nourri.

ZUSAMMENFASSUNG

Dieser Beitrag befasst sich mit methodologischen Fragen in einer Studie zur Persönlichkeitsentwicklung und illustriert diese am Beispiel des Teilens. Vornehmlich auf sechs Vorschulsettings in Südwestengland begründet, erfasst die Forschungsarbeit naturalistische Beobachtungen von Interaktionen unter gleichaltrigen von 55 Kindern zwischen drei und sechs Jahren. Die Anwendung des Ansatzes der gegenstandsverankerten Theoriebildung (Grounded) beim Beobachten, Analysieren und Interpretieren der Befunde erforderte eine vorsichtige und vertiefte Vorgehensweise. Durch die Methodik sollte sowohl der Einfluss der Vorkenntnis bestehender Großtheorien zur Moralentwicklung auf Seiten der Ausführenden minimiert werden, als auch jegliche vermutete Relevanz für die Beobachtungsergebnisse. Somit bestand die Originalität der Studie aus der Unabhängigkeit von bereits etablierten Erklärungsmodellen bei der Untersuchung der Moralentwicklung; sowie dem Widerstreben, sich zu möglicherweise vereinfachenden Rationalisierungen der kindlichen Absichten auf Basis von Verhaltensbeobachtung verleiten zu lassen. Durch die Erforschung der subjektiven Erfahrungen von Kindern, liefert dieser Beitrag Einblicke in die Komplexität dieser Prozesse. Entgegen bisheriger ,,makelloser” Befunde soll das kindliche Vermögen eines geschickten und mitunter überraschenden Umgangs mit moralischen Dilemmata reflektiert werden. Praktische Implikationen beziehen sich auf die Schwierigkeiten bei der Erforschung der kindlichen Moralentwicklung, sowie auf das Vorgehen, diese zu fördern.

RESUMEN

Este trabajo se centra en las cuestiones metodológicas surgidas en un estudio del desarrollo del carácter, utilizando ilustraciones sobre el tema “compartiendo comportamientos”. Está basado principalmente en la experiencia de seis centros de atención a la infancia en el sureste de Inglaterra. La investigación registró con observaciones naturalistas las interacciones entre compañeros para 55 niños de entre 3 y 6 años. La aplicación de la teoría fundamentada a los procesos de observación, análisis e interpretación de pruebas requiere un enfoque colectivo prudente y reflexivo. La metodología buscó moderar la influencia de los conocimientos previos de los investigadores de las “teorías de base” sobre el desarrollo moral y las asunciones acerca de la relevancia del registro de la observación. La originalidad del estudio reside en la exploración del desarrollo moral sin referencia a ninguna teoría de base particular  como marco explicativo, y en la reluctancia a ser atraído por potenciales racionalizaciones simplistas de las intenciones de los niños sobre la base de los comportamientos observados. Para explorar las experiencias subjetivas de los niños, esta investigación ayuda a comprender la complejidad de este proceso, dirigiéndose lejos de las resultados “netos” y tratando de responder a la complejidad de las negociaciones hábiles y a veces sorprendentes de los niños sobre dilemas morales. Las implicaciones para la práctica de este trabajo se refieren a las complejidades implicadas en los intentos de desentrañar el desarrollo del carácter moral de los niños pequeños y la práctica a través del cual este proceso puede ser alimentado.

Keywords:

• Character
• behaviour
• naturalistic observation
• sharing

Introduction

In the period 2006–2010 we conducted a major research project involving five separate studies, mostly multi-method, looking at different aspects of character development in different age groups in different populations. The results were to provide not so much pieces of a jigsaw as a range of disparate clues to what happens in character development. Five research studies cannot provide a complete basis for an overall theory but were designed to provide a range of solid evidence with which any overall theory would have to be consistent. An overall account has been published on the project website for each study separately (Learning for Life). The study reported in this article was in many ways the most ambitious. Very young children were studied mostly in preschool settings either in relatively deprived areas of London or in relatively affluent areas of rural Kent (Arthur, Powell, and Lin 2010).

Overall, our research was based on a number of working assumptions.

• Firstly, there is such a thing as character, an interlocked set of personal values and virtues that normally guide conduct. Character is about who we are and who we become and includes, amongst other things, the virtues of responsibility, honesty, self-reliance, reliability, generosity, self-discipline, and a sense of identity and purpose.

• Secondly, there is no fixed set of values, easily measured or incapable of modification.

• Thirdly, choices about conduct are selections about ‘right’ or ‘wrong’ actions and thoughts.

• Fourthly, character does not develop within a vacuum; in order to develop as a person, an individual needs to grow up within a culture, and the richer the culture, the more mature a person he or she has a chance of becoming (Arthur 2003, 3).

• Fifthly, education is concerned with active character development, not simply the acquisition of academic and social skills.

• Lastly, at a conceptual level, it is important to distinguish between the qualities of character that define virtue from other qualities of the self and/or person which we are more inclined to associate with such notions as personality.

Clearly, when dealing with early childhood (three- to six-year-olds), some of the assumptions of the research had to be treated with caution:

• An open mind has to be kept on whether the term ‘character’ has any useful application to such young children. They do have personalities; but research has to ask whether this can sensibly be partitioned in the way that applies to adults.

• An open mind has to be kept on whether the notions of ‘right’ or ‘wrong’ actions aid or hamper an understanding of their behaviour.

• It must not be assumed that a ‘rich’ culture is necessary for developing a ‘mature’ character.

Whereas there is every possibility that the research might produce evidence to support the Learning for Life assumptions, in the early years context they could not properly be taken as starting points.

Research Ethics was a major focus in the design, avoiding coercion and being fair to and respectful of parents and teachers as well as children. Detailed information was provided for all who returned completed consent forms. The identities of all participants (practitioners, children and their parents) were protected through anonymised data storage avoiding the use of data in a manner that might inadvertently identify individuals. All participants retained the right to opt out of the project at any time. The field researcher sought the children's permission to observe or talk with them on each occasion. With the children, the researcher always tried to be alert to the range of cues and signals that the children demonstrated, either to give ‘permission’ to be observed or show displeasure or unwillingness to take part or continue. The field researcher was also sensitive and attentive to the ways of interacting with the young children and their peer group. The study was designed with a view to the Revised Ethical Guidelines of the British Educational Research Association (BERA 2004) and was approved by our University Research Ethics Committee.

Exploring young children's moral development

In this study of the foundations of moral character, the researchers were acutely aware of existing theories concerned with young children's moral development and reasoning. Any investigation of the developing moral character of very young children has to be seen as, to some extent, an exercise in relativism. The point was neither to see whether children exhibited evidence for any particular theory of development nor to look for the first signs of adult values and moral thinking. Rather, the aim was to look at children's behaviour and to try to elicit their thinking in contexts where an adult might consider moral considerations to be relevant. The research team leader had successful experience in designing and conducting studies which elicited the subjective views of children in their own terms (Hogan 2005). In this case, however, any success in avoiding the imposition of adult perspectives arose as much from critical discussion within the team as from the nature of the evidence collected. The study was designed as an exercise in grounded theory. On the large scale, every effort was made to avoid favouring any particular theory of moral development or character. On the small scale, every effort was made to ensure that conclusions reached were drawn from what was seen and heard at the study sites and not from the presuppositions or predilections of any individual researcher.

Readers are invited to consider, however, the major academic propositions about young children's moral and character development as the theoretical frame within which the researchers sought to challenge their own and one another's beliefs and generate grounded theory from the project's data. Drawing attention to a cognitive perspective, Piaget (1947) emphasised the active nature of children's brain development, happening in an ordered sequence based on genetically determined stages. While the stage proceeds, children gradually learn to cognitively construct and structurally organize their thought and action. The successful cognitive development of one stage leads to the occurrence of the next. Piaget also stressed that the acquisitions of increasingly useful and more complex cognitive operations are achieved by children's interactions with their environment. He believed that social life among children is a necessary context for the development of intelligence, morality and personality (Piaget 1947; DeVries 1997).

Piaget (1932) described two types of morality: heteronomous morality and autonomous morality. Heteronomous morality is based on unilateral respect for and unquestioned obedience to authorities and the rules they prescribe. Autonomous morality is based on mutual respect, reciprocity and equality among peers. Two types of morality develop according to children's social context. A constrained environment characterised by conformity to the dictates of a hierarchical authority may lead to the formation of heteronomous morality. A cooperative environment characterised by mutuality and equality may result in the emergence of autonomous morality. It is crucial to note here that Piaget was cautious about calling the two types of morality ‘stages’ (Snarey and Samuelson 2008). There was no clear shift between the two types of morality in terms of the satisfaction of cognitive developmental criteria for a stage theory, which Kohlberg (1984) later on established in his work of moral development.

Kohlberg (1984) worked within a cognitive developmental paradigm and identified six moral reasoning stages. He postulated that moral reasoning proceeds through an invariant sequence of stages towards an increasingly adequate understanding of what is just or fair. In the early years of life, young children begin with a primitive morality guided by fear of punishment or desire for gain (pre-conventional stages). They initially focus on the self and do not realize that the interests of others may differ from their own. What is moral is understood by young children as respectful obedience to the authority. Moral behaviour or action is justified by avoidance of penalties and the superior power of authorities. When young children's moral reasoning moves to the next level, they start to recognise that other persons may have differing interests. What is moral depends on the person's immediate interests. The right action is that instrumentally satisfying the self's needs and occasionally others' by a process of exchange.

These brief examples of influential theorists illustrate how cognitive developmental theories, in particular the contributions by Piaget and Kohlberg, ascribe young children's moral development in ‘cognitive’ and ‘structural’ ways. Both have shaped the ways in which scholars, psychologists, educators and even parents conceptualize young children's growth in their moral character. However, both claims have not gone unchallenged. Among the critiques, the two most relevant to young children are discussed here.

The first is the idea of egocentrism. Piaget (1947) used it to describe young children's characteristic of preoperational thought. The term does not imply selfishness. It simply emphasises that young children have not yet grasped particular concepts including that there are other selves, with their own needs and feelings, their own beliefs, and their own perspectives. The concept of viewing young children as egocentric has also been involved in Kohlberg's early stage of moral development. In a social situation, egocentrism may lead a young child to project her/his own feelings and thoughts onto others. Sometimes, it can also form a unilateral view of rules and power relations. A young child can accept and obey the rules of others without any doubts. With an egocentric mind, young children may easily focus on the consequences of actions rather than the intentions behind them. For example, the amount of punishment for a wrongdoing, in a child's egocentric mind, should correspond to the amount of damage, regardless of any consideration of intention (Piaget 1932).

However, many investigators remain sceptical about young children's egocentrism (e.g. Harris and Kavanaugh 1993). In particular, some challenge the use of egocentrism in young children's understandings of the social world, which closely links with their development of moral character. Unlike the idea of egocentrism, the studies in young children's ‘theory of mind’ suggest that even three- and four-year-olds are able to understand why other people do the things and feel the things they feel. In general, the term ‘theory of mind’ refers to the set of interrelated concepts and beliefs that we employ whenever we try to make sense of our own behaviour or that of other individuals (Premack and Woodruff 1978; Leslie 1992; Fodor 1992). Although studies have also evidently found the limits of the young child's theory of mind (Wimmer and Perner 1983), the data showing that young children do have theories about the minds, beliefs, and desires of other people is still strong (Gopnik 2010).

The implication of young children having a theory of mind is widening the scope of studies in early years development of moral character. Young children are not hopelessly egocentric and merely the receivers of rules and regulations imposed by authorities. Researchers have begun to study, for example, young children's understanding of intentionality and moral judgement (Leslie, Knobe, and Cohen 2006), friendship and moral awareness (Peterson and Siegal 2002), and cognitive and social development (Rogoff 1990, 2003).

The second discussion is concerned with structural assumptions of the moral stages particularly set forth by Kohlberg (1984). Although some studies support the validity of these developmental stages (Colby and Kohlberg 1987; Dawson 2002), others have also demonstrated that older adolescents and young adults appeared to regress from stage four conventional moral reasoning to earlier pre-conventional stages (Holstein 1976). The problems lead researchers to look more deeply into the definition of morality, in particular the connections between reasoning about social conventions and moral concepts about fairness and human welfare (Turiel 1977). Thus, alternative conceptualizations of development in moral character were established, including Carole Gilligan's ‘Two Orientations’ – which suggested girls may reason differently from boys (1982) and Domain Theory developed by Turiel and his colleagues (Turiel 1983).

Domain theory posits that social knowledge is constructed within three distinct domains: moral, social conventional, and personal (Turiel 1983; Nucci 2001). The moral domain includes issues of justice, human welfare, reciprocity, and individual rights. The social conventional domain involves rules and regulations serving functions of social systems and social organization. The personal domain is comprised of issues that are not matters of right or wrong bur rather, issues of individual discretion, preference, and personal choice. In contrast to Piaget's (1932) and Kohlberg's (1984) assumptions that young children only abide by social norms and authority and do not differentiate morality (fairness) from social convention until more advanced stages, domain theory suggests even very young children reason differently about moral actions that affect the welfare of others, and conventional matters imposed by social norms or the dictates of authority (Turiel 1983; Nucci 2008). These three distinct domains, therefore, provide research in young children's moral character development with an extended framework to work on.

Regarding moral issues, young children are concerned with the issues of harm to the self and others (Nucci 2008). The examples in early childhood education would be physical harm, e.g. pushing and hitting; psychological harm, e.g. teasing and name-calling; and justice or fairness, e.g. snatching and destroying one's toys (Hildebrandt and Zan 2008). With respect to social conventional rules, the examples would be issues like: table manners, e.g. waiting for others to finish their meals; classroom rules, e.g. being quiet in the book corner; and modes of dress, e.g. skirts are for girls. Regarding personal issues, young children might consider the matters for instance, choices of friends, recreational activities, and other activities designed as ‘free choice’ (Nucci 2001). It is worth noting here that concepts of personal issues are crucial in domain theory. Nucci (1996) suggested that the establishment of control over the personal domain can satisfy the need to establish boundaries between the self and others. It is then critical to establishment of personal autonomy, individual identity, and a sense of moral agency. A sense of personal identity and agency contributes to interpersonal relations in various social contexts, in which actions such as those of reciprocity and cooperation occur (Turiel 2001).

Proponents of domain theory suggest that other dimensions collectively contribute to the formation of a child's moral character; they argue that judgements in social convention and the valuing of personal issues have a place together with justice reasoning. From domain theory researchers' points of views, caregivers and teachers should consider how aspects of a caring environment and practices differentially map onto the forms of social experience associated with the developments of young children's moral and conventional understandings. Another strand of theory relating to moral development emphasises the importance of belonging and of reference groups (Noddings 2008) or the opportunity to engage in social interaction with people seen as of equal status (Horn, Daddis, and Killen 2008). It is quite possible that ostensibly moral behaviour in young children is founded less upon compliance with authority or with empathy than on the need to consolidate friendship groups.

Various influential theories point out different aspects in relation to young children's moral development and character formation. Cognitive developmental theories emphasise the degree of cognition maturation and the abilities of moral reasoning directing a child's moral character development. To provide opportunities for young children to practice and apply their moral reasoning skills is the key. The higher levels of cognitive development indicate higher possibilities of becoming a moral being. In contrast, Triune Ethics Theory (TET) (Gopnik et al. 1999) draws researchers' attention to the sphere of emotion in the formation of a child's moral character. The early emotional experiences influence brain structure building a foundation for further moral development. For young children's moral character development, the exploration of TET has two meanings. First, it seeks to explain individual differences in moral functioning. Young children's early emotional experiences can influence their character formation and behaviour in context. The emotional experiences interrelate with brain wiring and in turn affect information processing. The establishment of the emotional circuitry relates to the brain's architecture for morality as well as later ethical expression (Narvaez 2007). Second, TET suggests ‘the initial conditions for optimal human moral development’ (Narvaez 2008, 313).

Beyond the challenges concerning inherent features of the ‘grand’ cognitive developmental theories, there have also been debates about the epistemological positions and paradigms that shape research on children's development. Bruner, whose early work from the 1940s was influenced by Piaget's ideas about cognitive development, later went on to assert that Piaget's failure to consider intersubjectivity and the forms of culture on which its operation depends represented a surprising gap in his theories; but that Piaget's ‘early work on moral development reveals a sensitivity to the growing child's reinterpretation of extant cultural norms, even if he [Piaget] shows little interest in interpretation as a mental process in its own right’ (Bruner 1996, 8). As a founder of the social constructivist movement, Bruner highlighted how Piaget's work differed significantly in this respect from that of Lev Vygotsky (and indeed his own later writings). For the latter, he said, ‘the most central question…is how a culture's symbolic tools manage through social interaction to get from “outside” into our “inside” repertory of thought.’ (Bruner 1996, 9). As the twentieth century progressed, there came a shift towards more contextually situated or ecological (Bronfenbrenner 1979) explorations of children's development as well as attempts to understand the meanings they ascribed to their ‘real world’ experiences with increased attention to children's competence and agency. Haste (1998) has also ascribed these different stances to liberal rationalist versus communitarian ontological positions and suggests that the two viewpoints present very different implications and potential models for moral development and education.

Perhaps in response to burgeoning social constructionist and postmodern perspectives in academia as well as challenges to and anomalies in the studies based on grand theories of moral development, Jensen (2008) proposes a cultural-developmental approach. This approach seeks to draw on and synthesise the earlier theories of Kohlberg, Piaget, Turiel and Gilligan mentioned above. Jensen emphasises the enormous variability in type and range of cultural influences to which different children (and indeed adults) may be exposed and with which they interact. The perspective has particular salience for this reflective account of how a group of researchers also sought to expose and examine their own socially and culturally constructed assumptions and theories about character development.

Research design

The early moral thinking of children can be manifested in what they say or what they do. In order to be more confident that the evidence gathered related to the way children actually saw things rather than to assume that the things they said necessarily indicated more than an ability to recycle words and phrases offered them by adults without anything resembling an adult understanding of the words, the study employed a grounded theory approach (Glaser and Strauss 1967; Strauss and Corbin 1990).

It is important to note that there have been criticisms of the grounded theory method (e.g. Bryman 1988; Bryman and Burgess 1994). Aull Davies (1999, 198) writes of a ‘naive assumption that data can initially be interrogated from a theoretically neutral position, as well as for not allowing sufficient development of more interpretative forms of analysis – keeping the emphasis on substantive as opposed to formal theory.’ These are very real problems; but rather than give in to them, this study attempted to minimise their effect, in part by being explicitly aware of theories that might steer the contextualisation of the research and, as far as possible to challenge such theories rather than privileging them. The development of the study's sample and methods aimed to encourage ongoing reflexivity among the research team and to ensure that theories were challenged. Davies' complaint about lack of formal theory is not seen as a weakness: to arrive at any such rigid conclusion, however tentatively, is not an appropriate conclusion to a single study. However, a range of speculative possibilities were raised. After an overview of the methods (more fully detailed in Arthur, Powell et al. 2009), the following sections deal with a detailed example of how this was attempted.

Six early childhood education settings were recruited. Each had been judged through inspection by Ofsted to offer ‘outstanding’ provision in relation to children's social and emotional development, which provided a theoretical degree of parity in terms of the settings' provision of early childhood education and care. A second criterion for selection was the classification as either urban or rural. This distinction was chosen to allow for subsequent analysis of data on the basis of differing demographic variables. Three urban settings were chosen from relatively deprived parts of London and three rural settings from relatively affluent parts of Kent (based on the English Indices of Multiple Deprivation [DCLG 2007]), with a view to possible analysis in terms of demographic data.

Children were sampled on the basis of their age, the initial requirement being simply that they were around 36 months at the start of an observation period. Up to two boys and two girls were then randomly chosen and consent sought from parents. In cases where parental consent was denied, the next child (boy or girl) on the list was selected. Three cohorts of three- to four-year-olds were selected in this way in May 2008 (seven children), September 2008 (24 children) and September 2009 (24 children). The first two cohorts (31 children) were tracked as they entered their (14) primary schools during the academic year in which they had their fifth birthday.

Within the six settings, 14 focus children's key workers and 24 (of the focus children's) parents were invited to take part in the study in stage two (cohort two). The head teachers or the managers of the settings also participated in interviews, informal discussions and the practitioners' survey (questionnaire). Other informants in school settings included 14 reception year teachers, three teaching assistants, and two head-teachers, with whom informal discussions took place.

Methods, which are described in full in Arthur, Powell et al. (2010), included:

1. Unstructured observations of children's everyday interactions and the ethos of the settings. There were 86 visits, lasting six hours each: 72 to preschool education and care settings and 14 to primary school reception classes.

2. Two discussion groups with the parents and one group with early childhood education and care (ECEC) professionals in the Kent settings.

3. Questionnaire to parents/carers (313) and ECEC professionals (50).

4. Interviews lasting 10 to 20 minutes with six parents in London settings, and 28 ECEC professionals, and reception class teachers in Kent and London settings (28).

5. Four discussions based on video-clips (of observations) with ECEC professionals and parents in Kent and London settings.

6. Discussions with children by means of stories, video clips, and children's own materials.

7. Semi-structured interviews with parents and discussions with the children (five) at home, once the children had progressed to their primary schools.

In this article, we present reflections on the design and methods using examples from the unstructured observations. The sampling strategy mentioned above was primarily designed to apply to the survey and interview methods. As such, the variables are mentioned in relation to this wider study (reported in Arthur et al. 2010) but were not applicable to the analysis of observation data and so are not elaborated in this article. The study was guided by an awareness of unnecessary and unhelpful distinctions between quantitative and qualitative research methods. Throughout, every effort was made to achieve the strengths of both approaches and to avoid the weakness of each.

Sharing behaviours observed and analysed

A major aim of this study was the identification of examples of morality and values exhibited through their play, narratives and interactions. Naturalistic observation was seen as useful in at least three ways (Dunn 2005). First, children grow up in social worlds, a complicated network which they need to adjust themselves to live with. It is within their social relationships that their characters, the understanding of a set of values, and the beliefs in certain moral frames are developed. Using naturalistic observations provides opportunities to gather data that may evidence how children use their power of understanding others and the value systems they embrace to interact with different people in a variety of social interactions

Naturalistic observations also occur in a real-life context. The data gathered record children's real-life experiences and their reactions to those experiences. Selecting a positive reaction in response to a cameo situation says very little about what a child would actually do in a real life encounter.

Finally, naturalistic observations allow researchers to study the situations/events that are emotionally meaningful to children. The naturalistic approach enables researchers to record such information on emotional reaction shown in the real-life context by children's talk, behaviours, facial expressions and physical posture. The emotional dimension can enrich the interpretation.

Limits of this approach also need to be made explicit. The lack of control in naturalistic observations, compared with experimental approaches, prevents standardisation and limits the scope for general conclusions. However, purely statistical generalisation carries weight only as far as the population from which randomisation took place and any general conclusion relating to the world at large, especially in the absence of randomisation, must depend on similar considerations to those involved in drawing conclusions from naturalistic observation. Fawcett (2009, 56) has cautioned that naturalistic observation presents the researcher with potentially vast quantities of unstructured data, which may make comparisons difficult. By contrast, Rolfe (2001, 231) has suggested that observer subjectivity determines behaviours that are noticed and recorded; effectively, the researcher's unconscious bias may ‘structure’ the data. She recommends techniques to promote researcher ‘reliability’ (Rolfe 2001, 233), although phenomenologists and ethnographers have challenged such positions, suggesting layered interpretations of ‘real world’ experiences and embracing this ‘double hermeneutic’ (Smith and Osborn 2008, 53).

One topic that led to very complex discussion among the research team was sharing, an idea that many teachers were seen to emphasise and encourage. In one setting the children followed and talked about fairly complex rules on taking turns. Initial observations also showed that many incidents that might be seen as involving moral thinking did involve sharing or failing to share. It would be tempting to jump to the conclusion that children adopted the idea of sharing, either as a form of compliance with the teacher's authority or as a consequence of empathy, and that this notion would go on to blossom gradually into a notion of distributive justice at a later stage. Alternative hypotheses were considered. Did children actually mean by ‘sharing’ something other than what the teachers had in mind? Were they consciously sharing or were they going about their lives and incidentally doing things happened to be sharing? Were they simply doing what was necessary to play with their friends? Were they perhaps using sharing as a weapon, sharing with an ‘in group’ as a way of excluding others? After much discussion it was agreed that it would be very difficult to find evidence that counted either way in answering any of these questions.

It was essential to pursue a measure of reliability in the observations upon which discussions were to be based. It is all too easy for naturalistic observation to record atypical actions which are in fact misleading as a guide to the way children in general behave or think. The variability in children's behaviour and the enormous level of randomness inevitably involved in observing only some behaviour of some children in some sessions, demanded attempts to minimise any unnecessary sources of variability.

After initial observation of the children by the main field researcher and preliminary discussions, the field notes were coded by the field researcher and independently by another member of the research team. In addition to this, a senior researcher experienced (and thus perhaps prejudiced) in the field of early childhood was recruited from outside the team and unaware of the content of discussion thus far. When a level of agreement was reached, the second team member acted as a second observer in some sessions, not shadowing the main field researcher but exploring whether the two sets of notes on the two sets of interactions witnessed by the two independent observers bore enough similarity to suggest that what was being recorded was something of general relevance to the setting and thus capable of at least suggesting general interpretation.

Possibly because sharing behaviours are easier to notice than some other forms of ostensibly moral behaviour, sharing emerged from this process as a major focus of interest. Young children's sharing behaviours documented in some reports (Rheingold, Hay, and West 1976; Hay et al. 1991) suggest that children's early sharing may occur shortly before their first birthday. Rheingold et al. (1976) claimed that infants begin the process of sharing by offering food and other objects to their companions, e.g. mothers, fathers, siblings, peers and other adults. The sharing behaviours of these infants, which the authors called ‘partner play,’ also involved showing objects to other persons at a distance, and engaging in coordinated use of them.

Some studies have focused on the frequency of sharing behaviours (Rheingold et al. 1976; Hay et al. 1991; Rao and Stewart 1999), and the people with whom young children share, e.g. friends or acquaintances (Birch and Billman 1986). Others have explored children's understanding of sharing, the relationships between sharing and variables, e.g. young children's capacity in moral reasoning, and empathy, the reasons that young children share, e.g. need for approval or reciprocity (Staub and Sherk 1970), and the resources for learning to share, e.g. parenting styles (Radke-Yarrow, Zahn-Waxler, and Chapman 1983) and preschool education (using incentive social reinforcement, Doland and Adelberg 1967). However, there is a paucity of studies exploring the content (how children actually do) of young children's sharing behaviours. The main reason for this can be the use of methodology in studies of sharing behaviour. The majority of studies were experimental designs. Even though the data were collected in the school or home environment, a hypothetical scenario or a preset testing design was in place. In this study, employing naturalistic observations in the early childhood education settings provided a possibility not only to explore the possible contribution of environment to children's behaviours (in familiar contexts), but also while they were engaged in both adult-led and child-led activities.

A number of sharing behaviours occurred in subtle, non-verbal ways; in terms of language it was noticeable that in potential sharing situations the opening move was often that made by the child who stood to gain from the sharing. ‘Can I have it?’ was a common opener. Prompted by this and other observations, a sceptical line of thought was pursued. Sharing behaviours and potential sharing situations were looked at in terms of the following questions.

• Was each child sharing only or largely to comply with the wishes of a teacher or a stronger child?

• Was each child concerned with what (s)he could gain from the transaction or with what others could gain?

• Was each child using sharing as a way to strengthen (power) relations with other children?

• Was each child using sharing to emphasise the exclusion of children not allowed to share?

A provisional list of types of sharing behaviour was also compiled. For children who stood to gain:

• Distribution;

• Taking turns;

• Asking;

• Demanding;

• Demanding with force.

For children giving up something:

• Distribution (with justice);

• Enjoying together;

• Taking turns;

• Giving away.

Examples from the observations as they were recorded by the researchers are provided below. These serve to illustrate some of the behaviours that were coded broadly as sharing. After these examples, there follows an explanation of how the team approached the analysis of this complex data.

Straightforward compliance with adults' wishes was sometimes easy to spot. In one observation it was first.

A teacher was supervising children playing with paper balls together in the playground. Children all wanted to try to throw the ball. A child came and was desperate to have a go. Teacher: ‘After A, B, and C you can have a go.’ The children took turns.

The absence of adult intervention might lead to a breakdown in sharing.

Four girls were cooking. They kneaded the dough and then started using a cup to make scone shapes.

D said to E: ‘I give it to her and then she gives it to me.’

At the same time, E was looking towards the teacher while she was talking.

The teacher did not notice/hear her, or just did not respond.

F gave the cup to G: ‘After you, I have a go.’

G was using it. From her posture F seemed to want it back for making the shape.

F: ‘Please.’

G: ‘No.’

F was waiting.

The teacher took it to the other child.

F was still waiting.

Z was using the model.

F was still waiting.

F got the teacher.

Teacher: ‘G, can you let F use the model?’

G gave it to F.

E to F: ‘Can I have that?’

F passed it over.

In one setting children were found to insist on the compliance of others when spontaneous sharing did not meet their needs. One boy who clearly did not like the idea of letting others share was keen to use the authority of nursery routine (of using an electronic egg timer to help the children take turns) when sharing was in his interest.

The same thing appeared to repeat every day.

H got the bike he wanted. He rode for a while. Then he asked his friend I to look after it.

Another child J came over and took the bike.

H was not happy and cried when he came back and found out his bike was taken. He complained to the teacher.

Teacher: ‘H, you had a turn already and now let J have a go.’

H was not happy. Immediately he went to pick up the timer and showed it to J.

On occasion the idea of compliance seemed to be absent. Sharing appeared to come from spontaneous good will.

K was preparing the snack. He was going to cut the apple. ‘Everyone wants to have snack. They can all have snack’ he said while he was doing it.

The apple was quite hard and he could not manage it.

He asked for some help from the adult.

Then he finished cutting it. He put the apple on the plate.

Two girls came over and sat down to have snack.

After them K had his snack.

Sharing was sometimes a means to consolidate friendship:

L carried on her water song and washing the paint area.

M tidied up her work and watched L.

M stood there for a while and looked at L.

L: ‘You want to try it?’

M: ‘Yes.’

L: ‘You can help me if you like.’

M: ‘I want to do it.’

M looked rather impatient and slightly annoyed.

L: ‘What's your dad's name? Does he have a name?’ … ‘Have you got brother and sister?’

M: ‘NO.’

L: ‘Wait for me finish?’ L kissed M.

M did not seem to feel interested in the conversation.

L still took time to tidy up everything.

L finished and M started to do her painting.

The sharing sometimes broke down:

N (male) and O (female) were very busy collecting various things: toys, road sign, books, chairs, etc. They were building a house.

N turned to Researcher: ‘We are going to play the game in a minute. Oh no, we are playing the game.’

N got more board from other places.

After a while, some other children came over and rushed into the house.

P (female): ‘I am home.’

O: ‘This is ours.’

P: ‘I can get through this.’

N: ‘No, that's ours.’

O: ‘That's our house. You can't have it. It's our house.’

P: ‘It's everybody's house.’

O announced: ‘We got to change.’

N ‘We are going to move house because the children are upset. They are stealing.’

N and O moved some stuff to a ‘caravan’ and both played inside.

Q (female) came and wanted to play in the tree house the boys were playing.

R (male) slightly pulled her shoes and wanted her to get down: ‘That's our ship.’

Adult: ‘Can you let Q.’

R: ‘That's our ship.’

Adult: ‘Why not?’

R: ‘That's our ship.’

Later on R threw the train out.

S (female) ‘R, I think T's got hurt.’

R: ‘That's our ship. Get off.’

T: ‘It's our ship. I'm coming to get you. Get off.’

Those asking to share were sometimes in control. In the following incident the nursery turn-taking routine acted as a restraint on the boy demanding rather than the boy in possession of resources.

Children were riding bikes in the garden.

U, demanding the bike, argued with V who was riding it.

V began to cry and toppled over.

U left.

V went to fetch the timer and approached W (male) who was riding another bike, demanding again.

V: ‘W, I want your bike!’

V showed W the timer and waved it up and down in his face.

Discussion

Sharing behaviour is problematic largely because of the difficulty of devising a notion of sharing that is sufficiently precise to satisfy adults and sufficiently vague to capture what is going on when children seem to share. There are two fundamentally different types of sharing context. In the first someone in authority (or otherwise in control) distributes resources to all those who wish to have them. In adult situations questions of distributive justice and fairness arise at this point. Children can also be very alert to distributive injustice to their own disadvantage.

The second type of sharing context is more relevant to the present study. In this one person has resources that (s)he wishes to use but allows others who also wish to use them to have some. It has to be possible for the first person to use the resources given to the second and thus to suffer some loss when they are handed over. It may be that the second person can claim a right to some of the resources, in which case there is a (tacit) appeal to some authority and the context is really of the first type. Where this is not the case generosity, empathy and compassion are of more relevance than distributive justice.

Much of the sharing behaviour observed in one preschool setting was either prompted by an adult or forced upon the sharer by the beneficiary using a timer that the nursery employed to regulate turn taking with popular equipment. In both these cases it is possible that the sharer was doing no more than complying with adult instruction of one kind or another. Also, much of the behaviour recorded seemed more like a power struggle for resources than benevolence.

We cannot expect children's thinking, talk and behaviour to fit neatly into a pattern required to be a precursor of adult virtues. Perhaps the most promising piece of behaviour, in terms of preparation for adulthood, is the example reported by Lin (2010, 56).

X (male) had been using scissors for five minutes.

Y (female) approached and wanted to use the scissors.

Y: ‘Can I have some scissors, please.’

X: ‘There are some inside.’

Y messes up the work table.

X: ‘Look what you have done.’

X carried on cutting his work.

Y: ‘Can you cut this for me?’

X: ‘There are some like this inside’

Y ignored what X said.

After a while X went in to the classroom and brought some scissors for Y.

Y used them for a few seconds and left.

The example is intended to illustrate difficulties that are always present even in the simplest piece of research, but which become salient when more complicated designs, particularly those drawing on both qualitative and quantitative perspectives, are employed. The naturalistic approach to observing the children and the iterative and collaborative analysis of the evidence had advantages and disadvantages as described earlier. Although the researchers were tentative in claims about the children's moral awareness and development, the approach enabled the creation of a theoretical model for sharing on which future research may be developed and for which various research designs are possible.

The findings from the naturalistic observations were considered in parallel to those from questionnaires and discussion groups or interviews with parents and early years practitioners. Referring back to comments made in the group or individual interviews, as reported in Arthur, Powell, et al. (2010), some parents construed specific behaviours in complex ways that would be difficult to reduce to a simple response on a Likert scale in a questionnaire.

‘The ‘naughty’ spells are usually to do with tiredness or…'‘Naughty or just learning boundaries?’

Consequently, the questions in the questionnaires and the analysis of responses to these were treated with caution and the use of principal component analysis. Without the observations of the children's complex sharing behaviours that so challenged the research team, it might have been difficult to resist drawing simplistic conclusions from the questionnaire data. Clearly that would have been wrong, though Seale (1999) suggests that this is the way in which qualitative analysis often works. A quantitative approach would have suggested that no such conclusion should even be hinted at without further corroboration. Seale (1999) suggests that the absence of such considerations in much qualitative research can also lead to poor quality work.

Conclusion

The use of naturalistic observation (repeated and longitudinal) can produce a wealth of rich data that is possible to analyse in a variety of ways, including the way this article approached the analysis. The method employed, after initial visits when everyone becomes familiar with the researcher, is largely non intrusive because the researcher effectively becomes part of the environment. The main danger, as already highlighted in the article, is of researcher bias during data collection. Do we only notice some things and not others because of our beliefs, values, and prior experiences? What do we miss? Use of video observation was used as a check and the research employed more than one researcher to go in and the team as a whole explored the observation records in order to critique/share the analysis. Using observations and discussion groups/interviews at the beginning of the research emphasised the need for caution in developing the questions and response categories for the questionnaire. This article focuses on the collection, analysis and interpretation of data gathered through naturalistic observations of sharing behaviours and confirms the need for a cautious approach to any explanations of motives underlying the children's behaviours or associations with their moral development.

The overall conclusions about sharing are that the children's complex and varying approaches to what we identified as sharing behaviours are framed by adult beliefs, values and assumptions in terms of our/practitioners' assumptions about what sharing is and should be. These assumptions are themselves culturally constructed on a macro scale, but also on a micro scale; the cultures of early years settings, e.g. where the timer dictated behaviours and this being accepted as ‘right.’ It could even be said that it had become ‘best practice’ in terms of the staff's assumptions about how the children should behave with one another when the resources were to be distributed. It might therefore be wise to think about moral development in relation to more than one of the ‘grand theories’ instead of one or other of them. This would connect with ideas that aspects of young children's development should not be seen in isolation, but that that these are interconnected.

Acknowledgements

The authors wish to acknowledge the contributions of members of the research team to discussions that led to this article.