A new study questions the concept of the ''good divorce'', a term popularised in the 1990s to describe a break-up in which the parents are co-operative, the children remain close to both and emerge, apparently, unscathed.

''Creating a positive post-divorce family environment, although worthwhile, is no guarantee that children will be unharmed by marital dissolution,'' said Paul Amato, the lead author and professor of sociology at Pennsylvania State University.

The study found the offspring of ''good'' divorces were no better off on a range of wellbeing measures than those of divorced parents who did not get on. In measures of self-esteem, satisfaction with life and school, experimentation with drugs, alcohol and cigarettes and school marks, the children of ''good'' divorces scored no better.

As young adults they were as likely as those with less co-operative divorced parents to have under-age sex and substance abuse problems, and did not have a better relationship with their mothers.

But the ''good divorce'' group had better relationships with their fathers and, as children, fewer behavioural problems.

Professor Amato, a highly regarded researcher with a background at the Australian Institute of Family Studies, said in the absence of domestic violence it made sense for couples not yet fully committed to separation to try to rebuild their relationship.

For parents who did separate, counsellors should help them reduce stress factors for children, such as a decline in living standards, new home, new school, new partners for parents and step- and half-siblings.

''Helping parents to have good divorces may be insufficient,'' he writes in Reconsidering the 'Good Divorce', published in the journal Family Relations.

The study has provoked spirited responses, including from the author of the classic 1994 book The Good Divorce, Constance Ahrons, an emeritus professor of sociology at the University of Southern California.

Professor Ahrons cited as important the findings that the children of co-operative parents had the fewest behaviour problems and the closest relationships with the non-residential father. The findings ''add further support to the idea that a good divorce acts as a significant buffer for some divorce-related stress children experience,'' she wrote in a commentary in Family Relations.

She also argued children's problems often preceded the

break-up. Professor Amato has previously shown that children's wellbeing improves when divorce removes them from high-conflict marriages but nosedives if the marriage was not overtly marked by conflict.

In his latest study, based on 944 parent-child pairs, families are grouped into three clusters: co-operative parents who get on well and share the parenting; parallel parents who don't co-operate, have some conflict but the children have ties to both; and ''single'' parents who do the parenting alone with the other parent having little contact or influence.

Compared with a control group of children with married parents, children in all three clusters had lower levels of wellbeing on the 12 outcomes tested.

But, while the children of co-operative parents were not as advantaged as expected, children of parallel parenting seemed worse off, said Jan Pryor, of the Victoria University of Wellington, co-author of Children in Changing Families: Life after Parental Separation.

In her commentary, Dr Pryor said children with unco-operative parents - and intermittent contact with the non-resident parent - did worse on a range of measures than those who rarely saw the other parent. ''There's evidence [that] contact every other weekend is the least beneficial,'' she said.

''Kids see the non-resident parent enough to want a relationship but not enough to really build a relationship.''