Children who obtain alcohol from people other than their parents are three times more likely to binge drink by 15 or 16, compared to those given a drink by their parents, a major new study has found.
But parents have been urged not to give their children any alcohol before they turn 18, as the research also found that practice doubled the risk of a child drinking alcohol at the same age, compared those given no alcohol.
Those children who got alcohol from sources other than their parents were three times more likely to binge drink.
As well as being less likely to binge, the adolescents given alcohol by their parents also typically drank less on any drinking occasion than those supplied by their peers or others.
The UNSW National Drug and Alcohol Research Centre study, published today in the British journal Psychological Medicine, followed almost 2000 children and their parents over four years in Sydney, Perth and Tasmania.
It provides a deeper understanding of what is often called the 'European' method of introducing a child to alcohol, providing regular small sips of alcohol over a long time, in what some believe may help "protect" them from harm later in life.
But study author Professor Richard Mattick said the most important finding from the study was that providing alcohol to any child under-18 increased the likelihood of drinking later.
"Providing any alcohol actually makes them more likely to drink than not," he said.
"The second key finding is that if kids get alcohol from other sources, not parents, they are more likely to drink alcohol and they are more likely to binge on alcohol."
A "binge" is defined as consuming four or more standard units of alcohol in a day, while national guidelines recommend adult women and men each have no more than one and two drinks a day at the most.
As well as being less likely to binge, the teenagers given alcohol by their parents also typically drank less on any drinking occasion than those supplied by their peers or others.
Prof Mattick said the study also shed more light on the difference between young male and female teenagers' drinking.
"Young women are more likely to drink alcohol at the tender ages than young men, it's quite surprising really," he said.
"If you look at the four years of data, and we've now got six years, there's a very strong trend about underage girls drinking more alcohol than boys.
"We're not really sure why that is yet, and I don't think it's helpful to speculate until we've got more years of data and analysed it."
The study "controlled" for a range of variables, including parenting techniques, whether the child was aggressive or "acting out", peers' perspectives on alcohol and religiosity.
But it did identify links between some of those factors and higher or lower odds of drinking among Australian children.
The factors linked to, but not necessarily causing, higher odds of drinking included: peer substance use and child externalising (acting out), the latter of which was also linked to higher odds of binge drinking.
However, the factors linked, but not necessarily causing, lower odds of drinking included: parental monitoring, responsive and consistent parenting, religiosity, child social problems (such as being shy), and peer disapproval of substance use.
Prof Mattick said the shyness was probably linked a lower risk of drinking as a shy children might be less likely to be out with friends or others who were drinking.
"Some of those factors are really pointing to kids who might be aggressive or likely to be troubled or troublesome, which is really the kinds of relationships we've been observing in relation to alcohol for a long time," he said.
"I think the (religiosity factor) is because they are more likely to set rules and monitor the child's activities and be, I'll say, more conservative.
"But that could also say something more widely about the nature of religious communities, as they are often out doing community activities together, going to church and the like."
Despite it being legal for parents to give their underage children alcohol, but not others, Prof Mattick advised against it.
But he could not yet say if there was a preferred age, if at all, for children to be introduced to alcohol.
"I should say we are following these kids right into adulthood and we can't really answer this question until the kids are in their 20s at least," he said.
"But I can say that there are still a range of acute harms that underage drinking can cause, from unwanted sexual activity in young women, getting involved in other problematic situations like fights and trauma, down to the more serious events that preoccupy the media like one-punch attacks.
"We don't yet know whether, in the long term, kids given alcohol in their teens are any different to those not given alcohol."
He said the study, which started when the children were about aged 12, would continue to follow them into their 20s, possibly longer if it continues to attract funding.
The study was funded by a range of grants including funding from the Australian Rotary health, National Health and Medical Research Council and an Australian Research Council Discovery grant.