A major aim of this thesis was to contrast, within a laboratory setting, the family systems of two potentially different groups of people; those families with an aggressive 8-12 year old child in their midst about whom there was some parental concern, and those families with a highly-socialized 8-12 year old in their midst. The aim was to study the relationships within those families with the emphasis broader than simply focussing on the relationship between target child and significant others. Equal status was given to all within-family relationships. An attempt was made to both gather information on as many within-family relationships as possible and to overcome some of the pitfalls which cast doubt on the ecological validity of some earlier studies.

The author has taken the perspective that if the current status of a family is to be understood then we must have some understanding of the perceptions family members hold of each other as well as the behaviours they choose to practice in each others' company. It is also held that the current status of a family is influenced by their relationship with the supports embedded in the community around them: work, leisure, extended family as examples.

To begin with, an assessment of parents' perceptions of family relationships was carried out on both family groups, recruited from schools and helping organizations around Christchurch. A modification of Kelly's (1955) Repertory Grid called the Dyad Grid (Ryle, 1970) was utilised without adopting Kelly's Construct Theory. Findings indicated that parents of aggressive children found more difficulty with conflict resolution and reduced their emotional support for problem offspring. Mothers of those children also reported most feelings of hurt and signs of stress. All adult groups viewed fathers as functioning peripherally in family relationships.

The relationships between the parents is a variable frequently mentioned in the literature on distressed and non-distressed families. An attempt was made to relate family perceptions to relationship satisfaction by grouping constructs from the grids into coercively and co-operatively associated constructs and by using the Locke-Wallace Marital Adjustment Test to measure relationship satisfaction. All parent groups reported satisfactory marital adjustment. However, parents of aggressive children reported a lower level of marital adjustment, lower inter-spousal and inter-familial co-operation (especially from husbands to wives) and higher use of coercion.

The perceptions of the children were measured by using the Dyad Grid (Ryle, 1970) with the same INGRID and DELTA analyses (Slater, 1972) and same grouping of constructs as used with their parents. Compared to their parents' perceptions, aggressive children reported more parent-child co-operation, and their nearest-age siblings reported more parent-child coercion. These findings suggested that aggressive children may not be as much at risk from emotional isolation from their parents who may, in turn, perceive other family relationships more positively, but inaccurately. All children were aware of co-operative and coercive processes associated with relationships involving their fathers. Surprisingly, highly-socialized children described much awareness of intra-familial coercion reminding us that even the best of families can still be intimate battlegrounds.

One of the potentially most interesting sections is that dealing with the supports available to each family group. Topics like stress and depression are currently being actively investigated in the context of their relation to family circumstances and perceptions (Wahler and Dumas, 1981; Belsky, 1984; Middlebrook and Forehand, 1985) but little definitive evidence is available to relate various kinds of support to different family types. So this section was largely exploratory. The area of supports and family types is very likely to be one where much exciting investigation occurs over the next few years. Findings suggested that parents of aggressive children, particularly mothers, tended to prefer separate leisure activities and to be less satisfied with their supports. These findings gave some support to those authors indicating links between parenting styles and isolation from supports.

The final empirical section of this study looked at the behaviours demonstrated by family members to each other and attempted to relate those findings to those from the studies of perceptions, in an attempt to find the best predictors of intra-family behaviour. Families with aggressive children showed consistent trends to use and reciprocate more coercion over time. Their parents demonstrated greater involvement within their families' problem-solving interactions, but those involvements appeared to be neither contingent upon nor more effective in managing their childrens' behaviours. Consistently, behaviours were the best predictors of coercion in aggressive children and their parents, whereas behaviours plus perceptions were the best predictors of coercion from highly-socialized children and their mothers. It is suggested that perceptions may be a crucial mediating variable which reduces the likelihood of families being involved in recurring coercive cycles.