Spanking children can make them temporarily more compliant but causes more problems than it cures by raising the risk that children will become aggressive, antisocial and chronically defiant, according to new research.

The discipline technique is also associated with delinquency, a failure to learn right from wrong, and an increased risk that the spanking might turn into child abuse, according to the author of one of the most comprehensive examinations of the subject.

"The bottom line is that corporal punishment is associated with numerous risks for children," said Elizabeth Gershoff, a researcher at Columbia University's National Center for Children in Poverty. "I would argue parents should to the best of their ability avoid using corporal punishment and instead use nonphysical and more positive types of discipline that we know are effective."

While many studies have tried to assess the effects of spanking, Gershoff's study, based on an analysis of 88 studies over 62 years, quantifies the effects of spanking on 11 child behaviors. Apart from immediate compliance, Gershoff found that spanking had negative effects on the other behaviors.

The results could generate controversy among parents and researchers -- who often disagree on every aspect of spanking, starting with its definitions, its consequences and whether anything should be done about it. Already, several researchers have contested Gershoff's conclusions, and some have argued that how parents deliver a spanking is more important than whether they do.

The controversy over spanking runs along a societal fault line where the rights of parents conflict with the rights of children. While society wants parents to have flexibility in raising children, no one wants discipline to turn into child abuse. The trouble is, few can agree on when exactly that happens. Researchers generally define spanking as two swats of the bottom with an open hand -- with the rule being that spanking does not cause lasting injury.

Spanking, a form of corporal punishment, is legal in the United States. Even as several western European countries have outlawed spanking, surveys suggest 94 percent of American parents spank their children by the time they are three or four years old, although that number hides great variability in the regularity and severity of the punishment, as well as the context in which the punishment is delivered.

Severity and context, said Robert Larzelere, an associate professor of psychology at the University of Nebraska Medical Center in Omaha, are much more important than whether parents spank their children. Parents, he said, should never resort to spanking as their initial technique in disciplining their children. But if reasoning does not work, and neither do nonphysical punishments -- like time-outs or taking away desserts -- a spanking could be beneficial, he said.

"Spanking is not only more effective when used that way, but the child is going to cooperate with the reasoning and time-out next time so parents can work themselves out of the need to use spanking," he said.

Gershoff's conclusion, Larzelere argued, ignored the fact that adverse outcomes like aggressiveness or delinquency might not have been caused by the spanking, but might be because the child had problems to start with. In the absence of spanking, in other words, the child might have turned out even worse.

On whether there were conditions when spanking could be done properly, Gershoff said, "I would say no. There is no situation I can think of where a child should be spanked . . . my counter to everyone who says, 'I was spanked and I turned out okay' is that children turn out okay in spite of spanking, not because of spanking."

Gershoff's research -- and Larzelere's response -- are being published this week in Psychological Bulletin, a publication of the American Psychological Association. The APA recommends against spanking in schools but is silent on whether parents should spank their children at home.

Gershoff said her research had not provided data that showed causative links between spanking and bad behaviors, but said everything she had studied indicated that the effects of spanking were mostly harmful.

Differences in how spanking is perceived might affect its consequences. Larzelere said that children in black families tended to see spanking as motivated by their parents' concern, while white children tended to see spanking as a sign of rejection by their parents.

Children under 7 tended to accept spanking as a part of parenting responsibility, whereas older children did not. Gershoff found that spanking was not associated with increased aggression among black children, but it was associated with that risk among white children. Similarly, said Larzelere, aggression in older children tends to rise with spanking.

Both researchers agreed that a spanking delivered in the heat of anger, as a visceral reaction by a parent, was less effective and much more likely to slip into abuse than a spanking that was carefully calibrated and used as a technique of last resort.