James White from Cardiff University's School of Medicine undertook a three-year-study, involving about 3,500 11-15 year-olds, as part of the British Youth Panel Survey -- a self report study of children in the British Household Panel survey.
The children were asked about the frequency of parental communication, arguments with family members and the frequency of family meals.
Recognised risk factors for smoking, such as age, participant sex, household income, parental monitoring and parental smoking, were all taken into account during analysis of the study's findings.
After three years, the responses of children who had remained non-smokers were compared to those who said they had experimented with smoking at some point.
Results indicated that one of the strongest protective factors for reducing the risk of experimenting with smoking in early adolescence was how often fathers talked with their children, both boys and girls, about 'things that mattered'.
The frequency of family arguments and family meals did not have a significant effect.
White presented his findings to the British Psychological Society's Annual Conference, 14-16 April 2010.