In recent years there has been much interest in the impacts on children, both positive and negative, of different patterns of parenting after separation, especially where the care of children is shared equally or substantially between both parents. This article summarises key findings from two recent Australian studies of outcomes for two potential risk groups: school-aged children living in separations characterised by high inter-parental conflict (Study 1), and infants and preschoolers in the general population of separated families (Study 2). Both studies were commissioned by the Australian Government Attorney-General's Department.

In Australia, the Family Law Amendment (Shared Parental Responsibility) Act 2006, together with other aspects of family law reform - most notably recent child support reform,¹- have ushered in an era wherein a child's experience of care by his/her parents post-separation has become sharply defined by the amount of overnight time spent with each parent (McIntosh & Chisholm, 2008; Smyth, 2009). The Act now stipulates that in courts with family law jurisdiction in Australia, in dealing with cases where the presumption of equal shared parental responsibility is not rebutted, judicial officers "must consider" the merits of making orders that the child spend "equal time" - or, if not equal, then "substantial and significant time" - with each parent. In addition, all "advisors" in the family law system (dispute resolution and legal practitioners, and family consultants) also have an obligation to inform parents that, in developing a parenting plan, they could consider that the child spend equal or substantial and significant time with each parent if reasonably practicable and in the best interests of the child.

One impetus for the current studies arose from concerns about the rapid progression of family law reforms supporting more widespread shared parenting arrangements post-separation, ahead of evidence about potential risks for specific groups within the family law population. While the question of how shared overnight care supports, disrupts or otherwise influences the development of very young children would seem central for policy-makers, practitioners and parents alike, enquiry into the efficacy of shared parenting to date has not had a strong developmental focus. Both studies reported in this paper sought to generate data that might assist parents and those from whom they seek assistance (mediators, lawyers and judicial officers) to reflect, from a more informed perspective, on what kinds of living arrangements may or may not support the developmental needs of the children concerned, and what factors could usefully guide the decision-making process about those arrangements. The infant study brings a fine developmental lens to the practical questions being asked about infants in shared overnight care. For school-aged children involved in high-conflict divorce, our longitudinal study traces the place of children's living arrangements in their development over time.

This paper distils the original published synopsis of these two studies. That synopsis and the reports of the full study are available online from the Australian Government Attorney General's Department.² In the remainder of this article, we briefly summarise the studies, key findings and limitations, and touch on some possible implications.

Study 1: School-aged children in high conflict separation

About Study 1

Study 1 drew on data originally collected in an intervention study that compared outcomes for families who participated in (a) child-focused mediation, and (b) child-inclusive mediation. Data were collected from respondents at four points in time across a four-year period: (a) at intake into divorce mediation, (b) three months post-mediation, (c) one year post mediation, and (d) four years post-mediation.

Children, mothers and fathers from 169 families were involved in face-to-face interviews at as many of these time points as possible. For the purposes of the current analyses, the two intervention group samples were combined into a single high-conflict sample, yielding complete parenting pattern data over a four-year period for 133 families (including 260 children). Complete repeated measures data were available at all four points in time for 106 mothers, 93 fathers and 144 children.

These data were used to explore the following questions:

• What was the demography of various parenting patterns over time in this high-conflict sample?
• How satisfied over time were parents and children with their respective care patterns?
• In what ways did care patterns account over time for children's closeness to their parents, perception of and reaction to parental conflict, and their psycho-emotional wellbeing?
• How did the flexibility or rigidity of arrangements influence the above outcomes?

Cases were grouped in three ways. The first group was determined by the pattern of post-separation care over four years, yielding four categories:

• continuous primary care (children always spent between one overnight per month and 35% of overnights with each parent);
• continuous shared care (children always spent at least 35% of overnights with each parent);
• changed arrangements (one or more substantial changes to the care schedule since its inception); and
• no or rare overnight contact with one parent by the fourth year of parental separation.

The second grouping described the way in which the most recent care arrangement evolved. This also resulted in one of four patterns:

• a continuous, unchanging schedule;
• a change from shared to primary care;
• a change from primary to shared care; or
• a loss of regular contact.

Finally, the parenting arrangement was classified according to the flexibility of the arrangement in response to the changing needs of family members (as defined by parents), resulting in one of two forms: sometimes/usually flexible arrangements, or rarely/never flexible arrangements (the latter described as rigid arrangements).

Key findings of Study 1

Patterns of parenting over four years post-mediation

While it was not unusual for shared care to be the agreed outcome between parents following mediation of their parenting dispute, over time, shared care arrangements tended to revert to those in place prior to mediation (which were typically primary residence with mother) (see Table 1). Not surprisingly, given the many logistical and relational challenges, shared care was a less stable pattern than primary residence. Four years after mediation, 41% had maintained primary residence, 27% had maintained a shared care arrangement (at least 35% of nights with each parent), and 32% of families had changed their care pattern.

Parents who participated in child-inclusive mediation (where school-aged children's needs and views were assessed separately, and incorporated into the mediation) were more likely to maintain the same arrangement over time than parents who received child-focused mediation (where children's needs and views were not assessed), and were more likely to have remained in a primary care arrangement. In contrast, families who exercised shared care prior to mediation were more than twice as likely to maintain this pattern as those families who moved to shared care after mediation.

The demography of shared care in a high-conflict mediation sample

Analysis of the characteristics of families with different parenting arrangements revealed some consistent patterns. Families who sustained shared parenting over three to four years were more likely than other care groups to have the following profile: male children, younger children at separation, smaller sibling groups, fathers with tertiary education, mothers with high incomes and tertiary education, co-located households with close proximity between parents, fathers who had been active carers during their children's infancy, and mothers who had re-partnered. At intake, these families also reported lower levels of parental conflict and acrimony, higher levels of parental alliance (i.e., worked together and respected each other as parents), warmer father-child relationships, and higher levels of paternal parenting competence and paternal availability than other groups. In short, a cooperative parental relationship was found to be one of the key ingredients for sustaining shared care over time. This finding is in accord with many other studies (for overviews, see McIntosh & Chisholm, 2008; Smyth, 2009).

Families who sustained shared care over the four-year study period also differed from the other care groups in the following ways: fathers continued to report more positive regard for the mother, while mothers' acrimony toward the father remained stable (i.e., it declined over time in all other groups); and fathers in sustained shared care were more confident in their parenting to begin with, and remained more confident in their own parenting over time. By contrast, families who moved from shared care to primary care tended to be characterised by the following: mothers reported high acrimony towards the father at intake; children reported poor emotional availability of the father at intake; fathers had low formal education; and children were aged 11 or over at intake (i.e., were approaching their teens).

Some families sustained a rigidly fixed shared care arrangement; that is, the living schedule was "never or rarely flexible/accommodating to changing family needs". Relative to the other care groups, the rigid shared care group was more litigious (i.e., operating from a court or consent order), and was characterised by higher marital and post-separation levels of conflict and acrimony and lower levels of cooperation. Mothers in this group reported feeling more threatened by their former partners than mothers in other groups, while fathers tended to have low regard for mothers' parenting skills.

Almost all of the 18 families in which fathers lost contact with children were characterised by high, sustained levels of marital and post-separation conflict at all points in the study. This finding is consistent with prior work in which conflict and acrimony have been found to be important precursors to "father absence".

Satisfaction with parenting arrangements over time

Satisfaction with care arrangements was associated at a number of levels with type of care arrangement. Fathers with shared care arrangements were the most satisfied of all groups with their living arrangements, despite reporting higher levels of conflict about parenting and poorer dispute management. In contrast, four years after parents mediated their parenting dispute, children in shared care (whether rigid or flexible arrangements) were the least satisfied with the parenting arrangements of all care groups; they were also the most likely to report wanting a change in their arrangement. Where shared arrangements were rigidly enforced, children became significantly more dissatisfied with the arrangement over time than did the flexible shared care group; these children were the least satisfied of all the groups with their living arrangements. Mothers and fathers were equally content when primary care and shared arrangements were reported to be flexible. Rigidity in shared care arrangements significantly affected mothers' but not fathers' report of contentment with the parenting arrangements.

Children's adjustment and wellbeing

Over four years, the type of care arrangement had different impacts on the wellbeing of the children involved and on their experience of their parents. After adjusting for initial levels of conflict, children in the shared care groups reported higher levels of inter-parental conflict four years after mediation than children in the primary residence or changing care groups. Reports of inter-parental conflict over time were similar to those from children in the "no or rare contact" group. Children in the sustained shared care group were more likely than those in other care arrangements to report ongoing feelings of being caught in the middle of their parents' conflict. Over the four-year study, the greatest decrease in children's scores for feeling "caught in the middle" was for children in the primary parenting group. By comparison, children's reports of distress about their parents' conflict did not vary according to the overnight care pattern. Similarly, neither the nature of a child's living arrangement at any single point in time, nor the pattern of care across time, independently predicted the child's total mental health scores after four years (as measured by the Strengths and Difficulties Questionnaire, Goodman, 1997).

After four years, stable living arrangements and greater amounts of overnight time were independently associated with the child's report of greater emotional availability by his/her mother, but not by his/her father. Fathers' availability was predicted by a history of warm and responsive parenting pre-separation and not by the amount of time he currently cared for the children overnight. We note that in this sample mothers were the primary caregiver in the children's early years, and this may well have a bearing on these observed outcomes.

Children's experience of living in shared care over four years was associated with greater difficulties with attention, concentration and task completion by the fourth year of this study. Boys in rigidly sustained shared care were the most likely to have hyperactivity/inattention scores in the clinical/borderline range; however, children who were already vulnerable to hyperactivity/inattention tended to remain that way over time, regardless of the overnight care arrangement.

Strengths and limitations of Study 1

The strengths of Study 1 lie in its prospective, repeated measures, multiple perspectives design, enabling us to tap into family life at different points in the separation process, and to look across time at the developmental trajectories of the children concerned. Large, all-inclusive studies are typically broad and shallow, and are not well placed to obtain detailed information on family dynamics and child outcomes. Moreover, cross-sectional or retrospective data alone cannot provide the same long-range view or analytic power. Uniquely, Study 1 collected extensive data over time from children and their parents, affording the opportunity to explore the study questions from the vantage point of all family members. That said, the data are from a small non-random select group of cases - high-conflict families seeking help from community mediation services.

Study 2: Overnight care patterns and the psycho-emotional development of infants and preschoolers

About Study 2

Study 2 draws on data collected as part of the Longitudinal Study of Australian Children (LSAC).³ LSAC follows the development of 10,000 children and families from around Australia, exploring the interaction of children's social, economic and cultural environments with their ongoing adjustment and wellbeing. The study began in 2004 with two cohorts - families with 4-5 year old children (K cohort) and families with infants up to one year old (B cohort). The samples are followed up every two years. Three age groups were examined in Study 2: infants under two years (B1 cohort), older infants 2-3 years old (B2 cohort), and 4-5 year olds (B3 and K1 cohorts combined).

Four research questions guided the study. Relative to rare overnight and primary care overnight patterns, and controlling for related contextual variables (including low socio-economic status [SES], single-parent status, social support, parenting qualities, and co-parenting conflict and cooperation):

1. Does higher frequency shared overnight care parenting differentially affect the infant's/child's growing ability to self-regulate his/her emotions and behaviours, and to focus and attend?

2. Does shared parenting differentially affect the infant's/child's physical and psychosocial health status?

3. What is the demographic profile of families who largely share the care of their very young children?

4. What parenting qualities, co-parent relationship characteristics and socio-demographic characteristics moderate or mediate relationships between parenting arrangements and the above outcomes?

Three patterns of overnight care were examined. We distinguished higher frequency overnight stays from lower frequency overnight care, and included a third group of children who had some daytime contact, but rarely if ever had overnight care. Consistent with current Australian policy, we adopted the terms "shared care" to reflect the highest frequency of overnight-stay groups, and "primary care" to reflect situations in which the young child lived primarily with one parent while having steady but lower frequency overnight contact with the non-resident parent. Tables 2 and 3 show the sample sizes for the groups of interest.

Key findings of Study 2

Infants less than two years old

For this infant group (see Table 2), overnight time with the parent living elsewhere (PLE) was defined as:

• rare (if any) overnights - overnight stays occurred less than once per year but with some daytime contact;
• primary care - an overnight stay occurred at least once a month but less than once a week; and
• one or more nights a week - with the non-residential parent. This category was used as the reference category in the statistical modelling.

Spending one or more nights each week with the non-resident parent had an independent effect on infant irritability. Examples of irritability include the infant being fretful on waking up and/or going to sleep, difficulty amusing self for a length of time, continuing to cry in spite of several minutes of soothing, and crying when left to play alone. Infants in the reference category had higher levels of irritability than those in primary residence arrangements. Infants primarily in the care of one parent had the lowest irritability scores of the three overnight care groups, according to resident parent reports.

In addition, infants in one or more nights a week care arrangements demonstrated more vigilant visual monitoring and maintenance of proximity with the primary parent than was the case by infants with rare (if any) overnight care. This effect held when parenting and SES were taken into account. They also demonstrated a trend towards higher rates of wheezing than infants in primary care (p = .08). Frequency of overnight care was unrelated to differences observed in global health, global developmental concerns, or degree of negative response to the LSAC interviewer.

Young children aged 2-3 years

For children aged 2-3 years (see Table 3), overnight time with the PLE was defined in the following ways:

• rare (if any) overnights - (as for children aged under two);
• primary care - an overnight stay occurred at least once a month but less than five nights a fortnight; and
• shared care - based on the policy definition, in which care occurred five or more nights a fortnight - that is, 35% or more overnights a year. This category was used as the reference category in the statistical modelling.

In the 2-3 year old sample, after parenting, parent relationship and SES controls were included in the statistical model. Broadly, frequency of overnight care was unrelated to differences observed in conflict with day carers or degree of negative response shown to the LSAC interviewer. However, with respect to relating to the primary carer, two independent effects of shared care arrangements were identified.

First, children in shared care engaged in lower levels of persistence (i.e., the ability to play continuously, stay with routine tasks, examine objects thoroughly, practice new skills, and return to an activity after a brief interruption) compared with children in the other two care arrangements. Second, there was evidence of more problematic behaviours on the Brief Infant-Toddler Social Emotional Assessment (BITSEA) Problems Scale than the primary care group (and a trend with respect to the rare (if any) overnights group; p = .08). Specifically, the shared care group relative to the primary care group showed more distressed behaviours in the context of parent-child interaction and caregiving. Examples of distressed behaviours include crying or hanging on to the parent when he/she tries to leave; worrying a lot, or seeming very serious; not reacting when hurt; often becoming very upset; gagging or choking on food; refusing to eat; hitting, biting or kicking the parent.

Children aged 4-5 years

For children aged 4-5 years, overnight time with the PLE was defined in the same way as for children aged 2-3 years (outlined previously): rare (if any) overnights, primary care and shared care. The final category was used as the reference category in the statistical modelling. In the 4-5 year old sample, after parenting, parent relationship and SES controls were included in the statistical model, results indicated that independent effects of care arrangements on emotional and behavioural regulation outcomes for children were not evident. The vast majority of variation between overnight care groups in the 4-5 year old group was accounted for by factors other than overnight care patterns, with particular emphasis on the impact of inter-parental conflict and lack of warmth in parenting on children's self-regulatory capacities (e.g., ability for a child to calm him or herself) at this stage.

Strengths and limitations of Study 2

The LSAC is currently the most comprehensive data set of child outcomes in Australia, and was designed to allow the exploration of pressing policy questions about children growing up in Australia. It is important to remember that shared care in Australia remains a minority pattern of post-separation parenting. As a consequence, obtaining detailed information from a large, representative sample of separated parents sharing the care of infants and very young children is a formidable challenge, even for surveys of substantial scientific rigour such as this. In the context of a general population sample, the numbers of infants and young children in shared overnight arrangements in our analyses were inevitably small, particularly at the policy definition of 35% of nights per year. As a consequence, some findings are treated speculatively. Moreover, LSAC data from non-resident parents were scarce and therefore excluded, while longitudinal tracking of infants' care arrangements over time was also not possible because there were insufficient numbers for a meaningful analysis.

Integrated findings of the two studies

These two studies targeted different age groups and different populations, and the specificity of each set of findings is important to retain. These results are largely consistent with the backdrop of literature detailed in the full report of each study; nonetheless, the findings have important points of correspondence in what they say about the "equipment" involved in translating a shared time arrangement post-separation into a durable and developmentally supportive experience for the children concerned. This "equipment" comes in several potentially mutually reinforcing forms.

1. Socio-economic equipment

Both studies highlight the conditions and socio-economic factors that help to make shared care work. Consistent with prior research (e.g., Johnston, Kline, & Tschann, 1989; Smyth, 2004; Steinman, 1981), shared parenting appeared to confer benefits to children where it is supported by resources linked to education and employment, and a host of interconnected relationship factors. The data suggest that parents who had good shared care arrangements lived near each other, tried to respect the competence of the other parent, and were flexible and accommodating rather than rigid in their approach. The sum of these component parts is likely to create an important "domino effect" for children's contentment and wellbeing.

2. Relationship equipment

Children read their parents' emotions as they move between households, and experience each parent's emotional availability and capacities. The relationships within each household, and the space between, become the soil within which children develop post-separation, with outcomes significantly determined by the richness or toxicity of that soil. Consistent with two decades of international research from the high-conflict divorce arena, these two new Australian studies show that, for school-aged children, nurturing relationships with each parent and supportive relationships between parents had greater bearing on many outcomes than the actual parenting arrangement itself. While children in shared care arrangements reported more inter-parental conflict than children in other arrangements, and lower contentment with their arrangements, neither a child's living arrangement at any single point in time, nor their pattern of care across time, independently predicted total mental health scores after four years.

3. Maintenance equipment

The manner in which children's living arrangements were maintained, however, did have an impact on children's emotional wellbeing over time. Rigid arrangements, often fuelled by acrimony and poor cooperation and set out in court orders, were associated with higher depressive and anxiety symptoms in school-aged children, as reported by both parents, and this form of living became something children often sought to change. Many of the above themes are encapsulated in a conclusion reached by Ahrons (2004), built on interviews with children reflecting back on their parents' divorce:

Parents agonise, argue, negotiate and litigate over the minutia of how much time their children will spend with each of them ... But ... especially as they get older, children want flexibility in their living arrangements ... They want to have their needs considered more by their parents and be able to transition between households on their schedules, not their parents' ... [They were] far less concerned about the specific number of days per week or month they spent living with one parent or the other than ... about how their parents' relationship infused the emotional climate surrounding their transitions between parental households ... At whatever developmental stage, children want to know that their parents will care for and love them while they continue their daily lives with as few interruptions and stresses as possible (pp. 66-67).

4. Developmental equipment

As important as the above three factors appear to be for children's outcomes in shared care arrangements, a key contribution of the second study is in identifying the "developmental stage" as a factor that, in many respects, may trump these other influences during the preschool years. Thus, regardless of socio-economic background, parenting or inter-parental cooperation, shared overnight care of children less than four years of age had an independent and significantly deleterious impact on several emotional and behavioral regulation outcomes.

Central to the infant study were questions about the impact of parenting patterns on the degree of confidence a very young child develops about the care he/she receives, and the resulting extent to which the child settled into a self-regulating pathway; was able to physically thrive and engage in stage-appropriate relatedness and to regulate their emotions across a number of psychosocial domains. Consistent with the findings of Solomon and George (1999), young infants less than two years of age living with a non-resident parent for only one or more nights a week were more irritable, and were more watchful and wary of separation from their primary caregiver than young children primarily in the care of one parent. Children aged 2-3 years in shared care (at the policy definition of five nights or more per fortnight) showed significantly lower levels of persistence with routine tasks, learning and play, than children in the other two groups. Of concern, but as predicted by attachment theory, they also showed severely distressed behaviours in their relationship with the primary parent (often very upset; crying or hanging on to the parent; and hitting, biting or kicking), feeding-related problems (gagging on food or refusing to eat) and not reacting when hurt, which are consistent with high levels of attachment distress. The second report details this body of work as an important context for understanding the pathways of disruption indicated by these findings.

By kindergarten or school entry at around age 4-5 years, these effects were no longer evident. Thus, once children can self-soothe and organise their own behaviour, are capable of representational thought and anticipation, have adequate receptive language, anticipate and communicate about past and future events and emotional states - in other words, by the time the child truly "knows what tomorrow is" and can manage themselves within it - then they are better able to straddle households in a frequently shared overnight arrangement. Schore and McIntosh's (in press) perspective from the neurobiology of attachment further explains this finding:

Attachment in the first year of life, when the brain circuits for attachment are still setting up, is different from attachment in the third or fourth year of life, when the system is going, so to speak. That is, to stress a developmental system while it is organising in the first year will have a much more negative impact in response to the same stressor than if you did it when the child was four.

Implications for policy and practice

Legislative reform is often a blunt instrument for shaping human behaviour - although the radiating message(s) transmitted by such reform should not be underestimated in the context of parenting disputes over children (Smyth, 2009). Anecdotally, there is little doubt that a number of separating parents in Australia (particularly non-resident fathers) have interpreted the 2006 family law reforms to mean that 50-50 care is the new default (e.g., see Kaspiew et al., 2009).

While the "best interests of the child" continues to be the paramount stated consideration for judicial decision-makers, children's needs at different developmental stages appear to remain relatively under-acknowledged by policy and legislation. Education and information have important roles to play in bringing developmental issues to the fore in the crafting of child-responsive arrangements - with or without the help of professionals. The findings set out above point to some key learnings:

• As with all relationships, parent-child contact after separation takes work (Trinder, Beek, & Connelly 2002). Shared care, as one of many possibilities, involves many logistical and relationship challenges; it is a skilful undertaking.
• Shared care is likely to be especially developmentally challenging for infants and preschool children. While a cooperative parenting relationship can make many things possible, the developmental needs of the young child and the additional demands involved in meeting those needs means that the challenges are even greater.
• Neither the existing literature nor our recent findings support using shared care (at least 35% shared overnights) as the starting point for discussions about parenting arrangements for infants and young children under four years.
• For older children - where parents can work together, are attuned to the child and can respond to their needs - the benefits of a shared overnight arrangement can be more evenly weighed.
• All possibilities in relation to developing child-responsive arrangements should be re-evaluated at regular intervals in the context of each child's developmental progress and emotional needs.
• Flexibility and responsiveness, and the corresponding capacities they entail within parents, are key to children doing well. These qualities have benchmark relevance for deciding post-separation living arrangements.

Where some families are ready for shared care, others may need time and support to evolve towards this care arrangement. For a smaller, but nonetheless significant group, shared parenting will never be appropriate. It follows that the legal and social science professions need to operate from a clear map of the known challenges involved in shared parenting, identifying the extent to which shared parenting arrangements provide the child concerned with a developmentally sound and viable vehicle for specific phases of their journey. Signposts are needed to assist professionals and parents to identify pragmatic, developmental and relationship equipment needed for the skilled task of shared care. Flowing on from this is the need to develop sound interventions that assist some families to mend faulty equipment, or to grow the necessary prerequisites, to "prepare to share". Equally important are support and educative resources for families who may have tried shared care but wish to move to another arrangement, or for whom shared care is not likely, short- or long-term, to be a viable option. In other words, pathways to, from and around shared care need to be carefully mapped, and supported by developmentally sensitive legislation.

The promotion of more positive relationships, and the creation of age-appropriate, child-responsive parenting arrangements through educational dispute resolution appears paramount, and we hope that existing services and programs can be further tailored to incorporate new learning about shared parenting identified through these two studies. Child-inclusive family dispute resolution (McIntosh, Long, & Wells, 2008) remains a promising tool across the family law arena for providing early screening of school-aged children's needs and views with respect to post-separation living arrangements.

Infants and very young children are among the least able in society to articulate their needs, desires or experiences of the world. In the study of their outcomes, standard ways of assessing their wants and wellbeing do not apply. The challenge for practice, research and policy is to be able to find ways of hearing the voices of very young children. There remains significant need for data sources that help to articulate the sum of the parts of early care-giving experiences that most affect the developmental security of very young children in separated families, and thus enable the infants' pre-verbal experiences to be better understood and acted upon within the family law arena. There is much still to be understood. Effective models of developmental consultation for infant and preschool family law matters are still needed.

Taken together, the results of these two new studies return the focus squarely to the importance of the questions we ask on behalf of children about post-separation living arrangements. The task continues to be to determine those arrangements and attitudes that will maximally support each child within his/her unique developmental context. While many questions remain to be solved, these studies have made a beginning with two vulnerable populations - very young children, and children in high-conflict divorce - in addressing the question of when a shared living arrangement becomes developmentally challenging rather than supportive.