Later studies focused on specific populations of resilient children and adolescents (Garmezy 1974; Anthony 1987; Werner and Smith 1987; Garmezy and Rutter 1983). In these studies the subjects were classified as being at risk of psychiatric disorders, delinquency and other negative life outcomes because of a variety of individual, family and environmental factors. According to Thornberry et al. (1995: 230) these factors include low parental education, parental unemployment, family receipt of welfare, family transience and family members experiencing trouble with drugs and the law. Yet rather than focusing on those children and adolescents who were casualties of these negative factors, the studies focused instead on those who had not succumbed. The questions this work investigated were:

Instead of focusing on individual deficit, the new approach focused on individual and community strengths and thus, the concept of resilience emerged in the psychological literature. According to Masten, Best and Garmezy (1990) 'resilience' is 'the process of, capacity for, or outcome of successful adaptation despite challenging or threatening circumstances'.

In the literature, the identifying characteristics of resilient adolescents are such things as social competence, problem-solving skills, mastery, autonomy and a sense of purpose and future (see Waters and Sroufe 1983; Garmezy 1985; Rutter 1980, 1984, 1985; Werner and Smith 1987; Masten, Best and Garmezy 1990; Gore and Eckenrode 1994; Consortium on the School-Based Promotion of Social Competence 1994).

Protective factors have been described in relation to three primary systems in the adolescent's world - family, school and community. In relation to the family, many of the protective factors identified by research clearly relate to the consistency and quality of care and support the individual experiences during infancy, childhood and adolescence.

The work of Rutter, Maughan, Mortimore and Ouston (1979) in Britain shows that another source of protective factors can be the school. Young people in discordant and disadvantaged homes are more likely to demonstrate resilient characteristics if they attend schools that have good academic records and attentive, caring teachers. Studies conducted in the U.S. have also shown the important role that individual teachers can play in resilient adolescents' lives (Geary 1988; Werner and Smith 1987; Coburn and Nelson 1989).

In relation to the community, young people in disadvantaged areas are generally considered more 'at risk' than those in more affluent areas. However, certain community characteristics seem to operate as protective factors. The strength of social support networks provided by kin and social service agencies, for example, is one such factor (Pence 1988).

These clusters of protective factors were confirmed and extended by qualitative research into child and adolescent resilience in South Australia (see Howard and Johnson 1999, 2000a, 2000b; Dryden, Johnson and Howard 1998). Table I summarises the findings of this research. Briefly, it suggests that child and adolescent resilience is influenced by:

• Life Events

Rutter (1987, 1990) describes four types of protective factors or processes: those that reduce risk impact or reduce a person's exposure to risk; those that reduce negative chain-reactions that follow bad events or experiences; those that promote self esteem and self efficacy through achievements and, finally, positive relationships and new opportunities that provide needed resources or new directions in life. The first and last of these are captured within the life events category in Table I.

Repeatedly, in our study of the lives of children displaying resilient or non-resilient behaviours, critical life events served to either ameliorate or intensify individuals' exposure to risks to their well-being. Serious injury or the death of a close relative increased individuals' exposure to risk; the departure of a bullying de facto parent from the household proved to be protective.

Opportunities to join and belong to supportive groups contributed to individuals' resilience, while high mobility due to changes in parental employment contributed to individuals' disconnectedness and their vulnerability. These life events were largely beyond the capacity of individuals to control, and serve to reinforce Garmezy's caution about using the notion of resilience to blame individuals who, for a variety of complex reasons, do not achieve positive life outcomes (Garmezy 1994: 13). Acknowledging the influence of these sometimes indiscriminate and haphazard life events, challenges socially naive and simplistic explanations of success and failure based on the liberal-humanist view which individualises social issues and leads away from broader social and cultural considerations (Cormack 1998).

• Personal Factors

While serendipitous life events can either protect or threaten adolescents' well-being, a significant cluster of protective factors focuses on the personal attributes and skills of individuals. The combination of positive dispositional characteristics, personal coping strategies and beliefs about personal efficacy and agency contribute to individual resilience.

• Family Factors

Consistent parenting practices that promote attachment and emotional bonding promote resilience. So too, do siblings and members of the extended family who provide emotional and material support and who model social problem-solving. These relational requisites can be provided in many types of families, not just in conventional nuclear families (husband, wife and children).

• School Factors

Schools that are safe, positive and achievement-oriented help adolescents develop a sense of purpose and autonomy and promote connectedness. They can also teach valuable life skills such as social problem-solving as well as social competence. Perhaps most importantly, schools can ensure that every student develops the foundation academic competencies needed for further learning and the development of positive self esteem. In these ways schools can 'teach for resilience' by promoting academic competence and attending to the social and emotional needs of students.

• Community Factors

Individuals and groups within the community can provide opportunities for adolescent involvement and participation in social, sporting and cultural activities. These activities promote feelings of belonging and connectedness that are central to the development of resilience.

The aims of the study were threefold. First, we wanted to compare the life strategies typically employed by young people who had been identified as demonstrating either 'resilient' or 'non-resilient' behaviour' at the time of the study. In other words, we were interested in how young people (both 'resilient' and 'non-resilient') respond to and (crucially) deal with life events, crises and problems.

Secondly, using Bronfenbrenner's (1979) ecological systems theory as a framework, we wanted to get a clear picture of how each of his microsystem settings - home, peer group, school and community - afforded particular kinds of protective processes and factors. Obviously, different settings afford different sets of possibilities and resources.

Thirdly, we were keen to identify whether cohort variables such as age, sex, location (rural/urban) and ethnic/SES background had any influence on 'resilient'/'non-resilient' young people's life strategies.

With such aims, we were clearly looking at a qualitative research design. As Leininger (1985: 5) says:

Using a modified screening device that we had developed from our previous work with primary school children, we asked teachers in 5 state high schools (2 in rural areas and 3 in disadvantaged metropolitan areas) and 1 metropolitan private Catholic college to identify students 'at risk' aged 13 - 16 years who were displaying 'resilient' and 'non-resilient' behaviour at the time of the study. This procedure produced 35 girls and 36 boys, 38 of whom were judged to be displaying 'resilient' behaviour (20 girls, 18 boys) and 33 displaying 'non-resilient' behaviour (15 girls and 18 boys).

These participants engaged in individual semi-structured interviews using the following Key Questions. These questions were designed to elicit information regarding the availability and/or use of protective factors and processes in the young people's lives:

The 71 interviews produced more than a thousand pages of transcript for analysis. Our method for managing this huge quantity of text data was to enter all 71 interviews into NUD•IST (QSR 1995), the software program for managing and supporting the analysis of qualitative data. This software tool enabled us to identify answers to the main interview questions and gather them together (i.e. code them). Major themes in these question categories could then be identified and further coded. Key ideas and words signalling the location of significant concepts occurring elsewhere in the transcripts were then tracked down through exhaustive searches. Finally, the coded data was interrogated using demographic information gathered about the participants. In this way it was possible to identify patterns and trends in the different categories of participants.

Protective Factors and the School

The complete analysis of this study is available elsewhere (Howard and Johnson 2000b). Young people spend a good deal of their time in school and many protective factors and processes can be embedded within schools' routine practices and programs, so we will focus here on some of the study's key findings and link them to the school setting. In this way we will indicate how schools may help develop or improve the protective factors in young people's lives.

In this study, 'resilient' and 'non-resilient' students talked in markedly different ways about their lives. 'Resilient' students talked about accomplishments and personal achievements, skills and competencies of which they were proud; 'non-resilient' students rarely did. 'Resilient' students expressed a sense of belonging and connectedness to individuals, groups and institutions whereas 'non-resilient' students talked less and/or less confidently about these things. 'Resilient' students demonstrated a sense of autonomy and personal agency when they talked about their lives whereas 'non-resilient' students were more inclined to indicate a sense of powerlessness and fatalism. 'Resilient' students had definite plans and positive views about their futures. 'Non-resilient' students, on the other hand, had limited and less enthusiastic visions of the future and what it held in store for them.

• Personal achievements and accomplishments.

The first major difference between 'resilient' and 'non-resilient' participants in this study lay in the way those who were classified as 'resilient' talked, often with pride, about personal achievements and accomplishments.

Self concept theory emphasises how important a sense of self worth is in the development of functional individuals and how this self esteem can only be built on real achievements. Seligman (1995), in particular, warns us that 'feeling good' about ourselves can only be predicated on 'doing well'; thus, if we give students the opportunity to achieve, develop competencies and master skills in a wide range of areas (i.e. 'do well'), then 'feeling good' (i.e. self esteem) will automatically follow.

While it might be thought that teaching for achievement and mastery was the core business of schools, some recent research (Howard and Johnson 2000a) suggests that for some (primary) schools, especially those in disadvantaged areas, these aims have slipped somewhat down the list of priorities. Too often, it seems, teachers are forced to abandon their central focus on teaching and learning because many troubled students present with social or behavioural problems that absorb available time and energy.

There are a number of ways in which schools can re-focus on the goal of helping students achieve mastery and competence:

• In this study, those students who claimed pride in personal achievements referred to sporting success and other skills largely developed outside the school. Not only should schools refocus on academic mastery and achievement for all students, but the range of ways in which students can demonstrate achievement also needs to be expanded. Achievements in art, drama, music, vocational education, work experience, leadership, social skills and so on, all need to be publicly recognised and valued.
• Although we gathered no hard evidence, it was clear from our use of printed materials during the interview process that many 'non-resilient' participants had literacy problems. Mastery of appropriate literacy skills is essential for most students' sense of self worth and for their ability to progress successfully through secondary school. Schools should not assume that students come to the high school with literacy skills at an appropriate level, nor should they assume that different subject teachers have the special skills necessary to teach remedial literacy. Special programs to facilitate achievement in these areas are required.
• What individuals judge as 'achievement' is relative. One young 'resilient' participant knew that he was less competent than his classmates in reading skills. However, his involvement in a structured program that was gradually but perceptibly developing these skills, was giving him a real sense of achievement and pride. Criterion-based assessment practices should be utilised more extensively in order to provide low-achievers with a genuine sense of progress and achievement.
• In this study, many students were not academically oriented but rather were firmly focused on the world of work beyond school. Vocational education and work experience programs that are rigorous and well-structured can provide students with a real sense of achievement and a future orientation.
• Many 'non-resilient' students claimed their teachers 'picked on them', ignored them or favoured the 'kids who were good at stuff'. Teachers need to be careful to hold high expectations for all students and not allow the presenting behaviours of some 'non-resilient' students (e.g. low self esteem, victim orientation) to turn into self-fulfilling prophecies.
• Many 'non-resilient' students in this study appeared to have poor study skills and few useful learning strategies. All students need to learn the relationship between achievement and good study strategies. They should be taught explicitly how to study and the specific values of goal-setting, effort and practice in relation to mastery need to be emphasised.
• More 'resilient' than 'non-resilient' students reported being involved in organized sporting and non-sporting clubs and associations but few of these groups were associated with the students' schools. Schools should support such clubs and associations on school premises and during school time. Many of these (e.g. army cadets, computer clubs) enable young people to develop skills and competencies that are beyond the range possible in most school subjects.

• A sense of belonging and connectedness

The second important difference between 'resilient' and 'non-resilient' students' talk concerned the extent to which they expressed a sense of belonging and connectedness to people, groups and/or organizations. Throughout this study, 'resilient' students consistently demonstrated a sense of belonging to family, school, peer group, sporting club and so on. 'Non-resilient' students on the other hand were less inclined to express a firm and confident sense of attachment to other people or institutions (often for what appeared to be very good reasons).

Schools can engage in specific practices that will help develop a sense of belonging and connectedness.

• Teachers are well placed to act as the significant, caring adult in a young person's life. In each school in this study, the same teachers' names cropped up time and time again as people the respondents valued and respected; people they felt took an interest in them; people with whom they felt comfortable talking about problems. Teachers need to be prepared to undertake this important role when necessary.
• Schools should have access to extensive resources to serve the personal, social and counselling needs of difficult or troubled students. Many 'non-resilient' young people in this study clearly had complex needs that were beyond the expertise of ordinary teachers. Without specialised help these students are unlikely to feel a sense of connectedness to the school - and indeed they didn't. In addition, their troublesome behaviour often made it difficult for other students to experience a sense of connectedness to school and teachers.
• It is important for schools to realise how important young people's social needs are especially in Years 8 and 9. In this study, learning how to make and keep friends - to become connected to peers - appeared to be a major preoccupation among the participants to the extent that many of them talked of school principally as a social centre rather than as a learning centre. The school needs to support young people in this regard. Again, school-based clubs and associations that support organized sporting and non-sporting activities are important here. Not only do they provide opportunities for connecting with competent caring adults, they also provide plenty of opportunities for learning and practising social skills. School-based clubs and associations that provide opportunity for leadership and service are good ways of encouraging connectedness to others.

• Autonomy or personal agency

Another key difference between 'resilient' and 'non-resilient' students' responses to questions about their lives was the extent to which they indicated a sense of autonomy or personal agency. Young people displaying 'resilient' behaviour were far more likely to talk in ways that indicated they believed they could control their lives and what happened to them. They did not see themselves as victims of fate or circumstance; they did not accept personal weaknesses and faults as unchangeable givens that would prevent them from being successful in the future. 'Non-resilient' students, on the other hand, talked frequently in ways that betrayed a profound lack of personal agency. Across a range of questions these participants consistently expressed a sense of powerlessness, a fatalistic attitude and a victim-orientation towards various life challenges.

Psychological theory variously explains this phenomenon in terms of attribution theory or locus of control theory. Learning from observing how others behave is also implicated here and is explained by social learning theory. In any event, these theories suggest that one's sense of personal agency and autonomy is learned, largely from what others say to you, from observations of how others respond to life's events and from conclusions you draw from your own attempts to act autonomously.

• Many 'non-resilient' students in this study were quick to claim 'they just couldn't do' particular school subjects; they were prone to blaming others for their lack of success; they dismissed rather than attempted to solve problems. Teachers have an important role to play in both challenging defeatist talk and learned helplessness in students and in modelling problem-solving behaviour. In their feedback to students, teachers must take care to attribute failure to such things as lack of effort or poor preparation (things within a student's control) rather than lack of ability (things beyond a student's control).
• The 'resilient' students in this study had powers of self-reflection, the ability to hypothesise and plan, juggling a range of possibilities. These are higher-order thinking skills that can be developed through any school subjects be they traditional academic ones or vocationally-oriented ones.
• 'Non-resilient' students were far more likely than 'resilient' ones to talk about being involved in violent interpersonal conflict, particularly at school. Young people need to be taught how to take control of conflict situations in socially acceptable ways. Behaviour management policies should reinforce the conflict resolution skills that should be explicitly taught.

• A positive future-orientation

Finally, 'resilient' and 'non-resilient' students differed in the extent to which they expressed a positive future orientation. In this study, the 'resilient' respondents generally had very positive views and definite plans about their futures. The 'non-resilient' participants, on the other hand, talked about the future with less eager anticipation, with more apprehension and they had fewer and humbler ambitions.

• All students had absorbed the rhetoric about the importance of work and 'getting a good job' but the differences in career aspirations between the 'resilient' and 'non-resilient' students were markedly different and not necessarily a reflection of their academic records. Vocational and career advice provided by schools needs to be comprehensive, up-to-date and it should provide a range of challenging (but achievable) options for all students. Careers advisers, like classroom teachers, need to ensure that the presenting behaviours of 'non-resilient' students do not influence their judgements about appropriate career choices (i.e. creating self-fulfilling prophecies).
• The narrow work-oriented view that many 'non-resilient' students in this study have towards their education leads them to reject as 'useless' or 'pointless' such subjects as Society and Environment, Drama and Science. Obviously these subjects have intrinsic value for all students but may need to be 'sold' in different kinds of ways if they are to engage all students effectively.

Conclusion

While changes at any level of Bronfenbrenner's model of nested systems can have an effect on the development of children and young people, changes at the level of the microsystem are most immediately able to be undertaken (Bronfenbrenner 1979). The School is, of course, an important microsystem setting and is one that affords many possibilities and resources for assisting young people to develop resilience-promoting skills and attributes.

All the young respondents in this study were judged by their teachers to be 'at risk' but were either displaying 'resilient' or 'non-resilient' behaviour at the time of the study. Many key differences in the ways these two cohorts of students talked about their lives were evident and clear pictures emerged about the ways in which protective factors and processes actually work in young people's lives. As we have shown, several of these protective factors and processes are ones that the School is ideally placed to consciously and deliberately promote and reinforce.

Through a careful reassessment of many of the routines and practices that make up the life and culture of a school, changes like those suggested above will assist many more 'at risk' young people to develop resilience-promoting habits, attributes, characteristics and beliefs. In this way, schools can indeed make a difference to young people's resilience.