What are the key assumptions on which the approach adopted in this study relies?
The first of these terms refers to assumptions about the nature of the phenomena being investigated. For example, randomised controlled trials almost inevitably assume that children and young people can be treated as objects having a set of independent properties, each of which can vary in level, such as ‘level of physical activity’ and ‘body mass’ (calculated on the basis of weight and height). Furthermore, it is assumed that there is either a relatively standard relationship among variables or no relationship at all. In other words, the possibility that level of physical activity may significantly affect body mass in some children but not in others, or do this in some localities but not in others, is largely ruled out. These, then, are ontological assumptions often involved in randomised controlled trials.
The second term, epistemological assumptions, refers to assumptions about how knowledge can be obtained. The key epistemological assumption is that relevant variables are measurable, or at least that we can allocate objects, such as nurseries or individual children, to one or another category in reliable ways, e.g. children’s parents to one or another social class category. It is assumed that the two main variables can be treated as scales. It is therefore assumed that fat levels in the body can be measured, indirectly, by measuring weight and height. [The body mass index (BMI) is calculated by dividing weight in kilograms by the square of height, measured in metres.] There are some known anomalies deriving from this indirect measure: in particular, it does not distinguish between fat, muscle, and bone mass. So, for example, many adult athletes would be judged obese on the basis of their BMI scores. This is less likely to be a problem with 4-year-old children, but it is still important to remember that what is being employed here is not a direct measure of levels of body fat.
Randomised controlled trials also assume that we can understand the relationship between level of physical activity and level of body mass, and thereby obesity, through varying what is taken to be a causal variable (level of activity) and observing whether the outcome of interest (level of obesity) changes, in much the same way that researchers in the physical sciences carry out experiments in which they manipulate variables such as temperature or pressure. Furthermore, Reilly et al. hypothesised that the effects of the treatment would be detectable after a few months of increased physical activity. In fact, this turned out to be false, although they suggest that this may have been because the ‘dose’ of physical activity was too small to have the desired effect.