[Dr. MacMillan and colleagues respond:]
We are pleased that our article1 generated such interest among professional and public audiences, because investigation of the relationship between physical discipline and emotional health is an important but neglected area of research in Canada. As outlined in our article, we hope that a longitudinal, prospective Canadian study will be carried out in the future to add to the information obtained from the 5 recent prospective American studies Murray Straus referred to in his editorial.2 Despite its cross-sectional design, the Ontario Health Supplement has nevertheless led to some important results on which to build further Canadian studies. Although a number of rigorous US studies have been conducted in the area of physical discipline, differences between the 2 countries such as demographic and cultural factors preclude a reliance by Canadians solely on American data.
As most of the letter writers are no doubt aware, researchers often need to apply a step-by-step approach when examining a scientific question about causation. The information about the hazards of smoking, for example, resulted from an accumulation of data over many years.3 The initial hypothesis-generating studies were based on case series, with far fewer subjects than in our survey. We do not suggest that ours is by any means the definitive study with regard to the relationship between childhood spanking and mental health problems in adults. Rather, it provides important information about the epidemiology of physical discipline on the basis of the largest Canadian survey to examine this question to date.
As we emphasized in our paper, "limitations of design and measurement in this study require caution in interpreting the findings."1 We stated that "... the cross-sectional nature of the survey precludes comment on the causal role of slapping and spanking for psychiatric disorder."1 It is therefore surprising that Joel Paris states that "the authors interpret their results with insufficient caution, leading to seriously misleading conclusions." We were very clear that the association between a history of slapping and spanking in childhood and adult psychiatric disorder could have several explanations. (We would also like to reassure Len Prins that in any information provided to the media, this point was consistently high-lighted. Whether or not the media chose to relay this to the public was beyond our control.)
Joel Paris and Philip Whatley suggest that the "linear association" between a history of slapping and spanking in childhood and adult psychiatric disorder is invalid. The statistical approach used to test whether there was a linear association between spanking and slapping and psychiatric disorder is a standard well-established technique. We acknowledge in the paper that this association could be due to other factors. Lee Isaacs suggests that we "isolated a single marker" and did not control for several confounding variables. Indeed we did control for 2 of the variables that he listed, including current family income and parental education. Determining which factors to control for is based on an understanding of possible confounders (variables that make 2 other variables appear to be associated when they are not).4 By excluding people who reported a history of either child physical or sexual abuse, we controlled for 1 of the most important factors that is associated with a history of physical discipline: exposure to maltreatment.
Joel Paris, Raymond Jacques and Derryck Smith suggest interesting hypotheses about our findings regarding the relationship between a history of exposure to spanking and slapping and psychiatric disorder. We welcome any attempts to explore these hypotheses in a longitudinal prospective study.
The letters from Victoreia Foss, Kathy Lasell, Arnita Lee and Teresa Leeth provide opinions about spanking and slapping of children based on personal observations. A survey of almost 10 000 people has the advantage of moving beyond personal experience to investigating a question using a sample representative of the general population. Our data do not suggest that most people who were spanked or slapped in childhood will experience emotional problems, but rather that a history of being slapped or spanked "often" is associated with an increased risk of psychiatric disorder.
We agree with Gabriel Leung that the issue of whether there is a threshold of harm associated with spanking is a crucial question. In contrast to what he states, however, we did examine the association between reported slapping or spanking "rarely" and "never" and found that the lowest rates of psychiatric disorder were associated with a history of "never" being spanked or slapped (this reached statistical significance for the association between a history of being slapped or spanked rarely, and externalizing problems).
In summary, we stand behind the results outlined in our recent article on slapping and spanking but encourage others to take up the challenge to understand further the relationship between exposure to these experiences and psychiatric disorder. Surely this issue is of sufficient importance to the mental health of children and adults that it warrants investigation in a Canadian longitudinal survey.