Specific resource factors or assets may have the potential to buffer or ameliorate the detrimental effects of adversity and lead to resilient outcomes. A considerable body of research has identified a core set of resources that are associated with resilience across various adversities and developmental outcomes. These are grouped in three domains: (a) children’s internal characteristics and strengths, e.g., self-esteem, self-efficacy, self-control; (b) family characteristics and relationships, e.g., child–parent closeness, parenting styles; and (c) characteristics of children’s social (particularly school) environment, e.g., student-teacher relationships, school-quality[17, 22–24].
Internal child characteristics such as self-concept, including self-esteem and self-efficacy, have mostly been associated with resilience in older children and adolescents. In longitudinal studies, Werner and Smith[25, 26], Masten and colleagues, and Elder and Conger found that positive self-worth (self-esteem and self-efficacy) was longitudinally predictive of psychosocially resilient outcomes in adolescents within the context of family adversity and stress. In the Rochester child resilience project[28, 29], self-esteem and perceived self-competence were associated with resilient adjustment for school-aged children experiencing stressful life events, but this was not the case in a similar study conducted in Australia. However support for a relationship between positive child self-concept and psychosocial resilience is also provided by other studies involving at-risk children and adolescents exposed to specific adversities such as socio-economic disadvantage[31–33], family disintegration[31, 34, 35], and maternal depression.
Children’s self-control or emotional regulation may also buffer adversity and promote adaptive outcomes by enabling children to respond positively to stressful circumstances[37, 38]. In two cross-sectional studies of socio-economically deprived preschool children attending Head Start, greater emotional regulation was associated with fewer internalising problems, fewer conduct problems and more pro-social behaviour. Longitudinal studies of at-risk young children growing up in poverty have found that toddler emotional/behavioural regulation and attentiveness/persistence on tasks is predictive of fewer behavioural problems 3 to 4 years later[41, 42]. Emotional regulation (including lower negative emotionality and greater inhibitory control) has demonstrated both concurrent and longitudinal associations with adaptive mental health outcomes in other high-adversity samples, including children exposed to domestic violence, maternal depression, impoverished minority youth, children experiencing maltreatment and cumulative family adversities, and homeless children[37, 43–51]. Further studies with school-aged children found that self-regulation moderated the association between socio-contextual family adversities and mental health outcomes, signifying the potential role of self-regulation as a protective factor in the context of family adversity[43, 52].
Supportive child–parent relationships characterised by warmth and closeness have been found to consistently predict mental health resilience in children. For example, in children from the Kauai longitudinal study exposed to cumulative family adversities, the resilient youth had more supportive relationships and interactions with capable parents than non-resilient youth[25, 26]. The association between the quality of child–parent relationships and positive mental health child outcomes is demonstrated in several other longitudinal studies within different adversity contexts including socio-economic disadvantage[27, 33, 42, 53–59], parental death and divorce[35, 60, 61], stressful life events[12, 62], and child maltreatment[31, 33, 63] some of which focussed on early childhood outcomes[42, 53–55, 57–59, 63]. In longitudinal studies of young children examining family relationships as a moderator of the association between adversity exposure and child mental health symptoms, O’Grady and Metz found that greater family support provided by parents for children buffered the effect of stressful life event exposure on children’s emotional and behavioural problems, Malmberg and Flouri found that mother-child relationship quality buffered the effect of socio-economic disadvantage on children’s emotional symptoms, and Maughan and colleagues found that maternal negative parenting moderated the effect of maternal depression on young children’s perceptions of social acceptance. In direct contrast, Calkins and colleagues found that a more responsive relationship between parent and toddler was associated with more externalising and internalising behaviour in 5 year old children exposed to high family adversity.
There is also a small amount of evidence that close supportive relationships with teachers are associated with resilience in children faced with adversity. For example, the resilient adolescents from the Kauai longitudinal study frequently had a favourite teacher who became a role model for them[25, 26]. In a longitudinal study of children aged 4 to 8 years, Peisner-Feinberg and colleagues found teacher-child closeness was more strongly related to lower levels of behaviour problems among children identified at-risk due to low maternal education, compared with their low-risk peers. Similar findings have been obtained from cross-sectional studies of preschool children at-risk due to socio-economic deprivation[40, 68]. In Qualitative studies of Australian and South-African children experiencing adversity, those identified as ‘resilient’ by their teachers frequently made positive comments about special caring teachers who had a positive impact on their wellbeing[69, 70].
In summary, considerable evidence exists for the role of each of the groups of child, family, and social/school factors in the development of mental health resilience. However, some limitations deserve mention. First, the majority of this research has focussed on resilience in middle childhood and adolescence. In comparison, few studies have investigated resource factors during preschool, or resilient outcomes in young children across the preschool to school transition, which is considered a critical period of rapid developmental change. Thus, it is unclear if promoting these factors during preschool will improve mental health outcomes in young children exposed to family adversity. Second, it is notable that the school environment, and particularly the potential role of teachers, has received far less empirical attention than other resource domains. As a result, few studies have examined resource factors from all three child, family, and social/school domains in the same study (notable exceptions include[25, 26] and). Without knowing what their unique contributions are, it is unclear whether one resource may be more important than another. This is an important omission, given that evidence already exists of the influence of resources from all three domains. Third, a considerable proportion of studies examine single adversity factors in isolation (e.g., maltreatment, poverty). Comparatively fewer studies[15, 25–27, 66, 71] have examined cumulative family adversities including combinations of socio-economic factors, stressful life events, parental mental health, and parental separation. This is considered problematic because “focusing on a single risk factor does not address the reality of most children’s lives” (p.367).
Finally, the vast majority of research on child resilience has been conducted in the US and UK. Conducting research in other countries such as Australia is important because resource factors relevant to resilience may be context and culture specific[72–74]. It is not known whether Australian children may demonstrate unique developmental patterns and responses to adversity. While these countries are all English-speaking multicultural western societies, the different distributions of socio-economic disadvantage, greater income mobility, less spatial concentration of public housing, and the nationwide universal provision of free preschool for all 4–5 year old children in Australia make it difficult to know how directly applicable findings from the US and UK would be to Australian children[72, 75]. Only a handful of studies have investigated mental health resilience in Australian children (e.g.,[30, 36, 51, 69, 74–79]), with the evidence for young Australian children limited to studies finding support for positive child–parent relationships and home environments as correlates of mental health in the context of family disadvantage and stress[51, 76, 78, 79]. There is much more knowledge to be gained in this context.
Multiple methodologies for measuring resilience
Resilience is a concept that is inferred on the basis of associations between the levels of (a) exposure to adversity and (b) positive adaptation or positive adjustment outcomes, and therefore it cannot by directly measured[18, 24, 80]. There is no ‘gold standard’ for operationalising the concept of resilience, and several different approaches are currently used to combine adversity and adjustment levels to measure resilient outcomes. When this occurs it can be difficult to compare results from different studies of resilience as it is possible they may not actually be measuring the same concept or phenomenon[24, 80, 81].
Broadly, methods of measuring resilience can be classed as variable-centred or person-centred approaches. Variable-centred approaches examine statistical associations between measures of adversity, hypothesised resource factors, and developmentally-relevant functioning, using regression-based analyses. If a factor modifies (i.e., reduces) the negative effects of adversity on functioning, then it is labelled ‘protective’, and it is implicated in resilience among the children for whom the risk and protective factors co-occur[82, 83]. Researchers typically test such modifying effects using a statistical interaction term between the adversity and hypothesised protective variables. The ‘statistical interaction’ approach draws on the statistical power of the whole sample. However, the children who meet the criteria for resilience are never explicitly identified, and thus which children are deemed resilient remains unknown. Additionally, statistical interaction terms within regression can lack adequate statistical power to fully and reliably detect real interactions, leading some researchers to caution against relying on statistical interaction terms[16, 82, 84].
Two other variable-centred approaches, used in combination, can address these two main limitations. First, the ‘residuals’ approach can identify resilient children who, in a statistical sense, are ‘doing better than expected’, while also keeping all data as continuous. With this approach, when regressing adjustment on adversity, the difference between a child’s actual adjustment score and his/her adjustment score predicted by adversity (i.e., the standardised residual scores) can be utilised as a continuous vulnerability-to-resilience score. Children with positive residual scores (i.e., falling above the regression line fitted) show ‘better than expected’ adaptation than predicted by their exposure to adversity, and are considered resilient (with the size of the residual indicating their level of resilience). This residuals methodology is a relatively innovative approach and variants of it have been used in several resilience studies[27, 85–88]. Second, the ‘residuals’ approach can be used with a ‘multiple-groups’ approach, where main-effects regression analyses predicting resilience residual scores are run separately for low- and high-adversity groups[89–92]. Subsequent effect sizes for each group can then be compared to examine the specificity of processes (i.e., whether a resource is a general ‘promotive factor’ associated with good outcomes in both low- and high-adversity children, or a specific ‘protective factor’ with unique benefits only for high-adversity children) while avoiding the statistical problems related to statistical interaction terms.
In contrast to variable-centred approaches, person-centred approaches involve identifying a group of resilient children (who experience high adversity but exhibit adequate adjustment), and comparing their characteristics with other groups of children showing different patterns of adversity and adjustment, in order to identify resource factors associated with resilience (e.g.,[12, 25, 90]). Using Masten and colleagues taxonomy as an example, if four groups of children with divergent outcomes are identified - two high-adversity groups identified as either ‘resilient’ (good adjustment) or ‘maladaptive’ (poor adjustment), and two low-adversity groups classified as ‘competent’ (good adjustment) or ‘highly vulnerable’ (poor adjustment) - it is possible to determine if a resource is truly protective rather than generally promotive by examining if resource levels differ between ‘resilient’ and ‘maladaptive’ children, but not between ‘competent’ and ‘highly vulnerable’ children. A key advantage of the person-centred approach is that it better reflects resilience as it actually occurs naturally within the whole child, rather than through associations between variables. Due to this, manifestly resilient children can actually be identified[17, 18, 23, 36]. However, reducing the vast individual differences present in early childhood development into broad dichotomous categories may be problematic, as valuable detail becomes lost, particularly if the sample size is substantially reduced by selecting more extreme subgroups only[84, 93]. Furthermore, if cut-points are somewhat arbitrarily defined (particularly a median-split) without a solid reason to suspect different effects between the groups created, then effects that occur within rather than between groups may be obscured.
Despite the considerable methodological variation in resilience studies, the fact that a common set of child, family, and social resources have been consistently recognised in resilience suggests that these resources are all implicated in the same underlying phenomenon, and support the validity of resilience as a construct[18, 23, 80]. Given their seemingly universal importance, these particular resource factors could be quite useful for further systematic exploration of the resilience construct, and critical examination of its measurement. However, researchers have rarely addressed whether similar variables emerge as significant resources while employing multiple resilience methodologies within the same sample. Inferences have needed to be made across studies, when many other factors could not be accounted for, such as sample characteristics. Given the relative strengths and weaknesses of both variable- and person-focussed resilience methodologies, it seems sensible to use both types of methods in combination in the same study (e.g.,[12, 84]).
As different resilience measurement approaches are rarely used in a single study, little information exists regarding how different methodologies may affect results (whilst holding constant the sample and variable measures). Masten and colleagues conducted both variable-centred analyses (examining whether resource variables buffered the negative impact of adversity using regression interactions), and person-centred analyses (examining whether the same resource factors distinguished between ‘Resilient’, ‘Maladaptive’ and ‘Competent’ groups of children in MANOVAs). However, the fourth ‘Highly Vulnerable’ group (low adversity + poor adjustment) was omitted because it was an ‘empty cell’, so the possibility that associations between resources and positive adjustment differed between high-adversity and low-adversity children could not be examined. Thus, although complementary, their variable- and person-centred approaches were not directly comparable (see also[46, 49, 94–96]). To our knowledge, only one study has assessed interactive effects within both variable- and person-centred analyses. Lengua examined whether resource levels discriminated not only between two high-adversity groups (e.g., ‘resilient’ vs. ‘maladaptive’), but also between two low-adversity groups (e.g., ‘competent’ vs. ‘highly vulnerable’), using logistic regressions. Findings were then compared with those from linear regression interaction terms. However, these methodologies were not fully comparable because they used a different adjustment variable – the adjustment variables were examined separately within variable-centred analyses, but were combined into a composite adjustment variable for person-centred analyses.
The present study
The aim of the present study was to investigate child, family, and preschool resource factors associated with the development of resilient mental health outcomes during the early childhood years. We hypothesised that (a) children’s characteristics (higher self-esteem, self-efficacy, and self-control), (b) better quality child–parent relationships, and (c) better quality child-teacher relationships during preschool, would be associated with greater mental health resilience in children two years later once at school. To achieve this aim, we utilised the four different methodological approaches for operationally defining resilient outcomes (as described above). This strategy allowed the investigation of whether similar resource factors emerged as predictive of resilient outcomes in young children when different methodological techniques were used. To our knowledge, this is the first study to analyse results from directly comparable techniques for operationally defining resilient outcomes.
There are several unique aspects to this study. We add to the relatively small body of literature on resilient outcomes in young children, and to the limited information regarding the various potential resources in the child, family, and school domains that children experience during the preschool year. This may inform early intervention efforts designed to maximise positive development in young children and intervene before mental health difficulties become entrenched[4, 5]. The present study also builds upon previous research by longitudinally investigating resource factors associated with resilient outcomes in a contemporary cohort of young children. Finally, the present study represents one of the first investigations of mental health resilience in the context of cumulative family adversities in Australian children. These aspects are important given that resilience is considered a contextually and culturally embedded phenomenon, and a multiply-determined and mutable developmental process[20, 74].