Livingstone and Bovill (1999, p. 55) point out: ‘not only are children and young people in many ways distinct from adults, they are also a diverse population – varying according to gender, age, social class, lifestyle, etc. and so are not readily reduced to simple categories’. Livingstone's and Bovill's comment reflects the paradox that the social subgroup ‘children’ is a homogeneous one with fewness of years of life as their common distinguishing factor; but they are also a microcosm of the whole population in their demographic diversity. This makes confident generalisations about them difficult, and without very carefully selected and scrutinised data, unwise. Our sample was not a scientifically representative one, but with 1,332 children in seventeen different schools in England and Wales, with nearly equal numbers of boys (631) and girls (645), and a reasonable spread of social advantage and disadvantage (as measured by the schools' statistics), it was comparable to Livingstone and Bovill's and was certainly much more representative than much research on children and media. More details of the sample are given in the Introduction and in Appendix 1.

Social science research generally takes the view that demographic factors function as independent variables, that is they make a difference to what people do. This is a view also held by scholars from other traditions. Whether one is a cognitive scientist or a feminist literary critic, one accepts that being male or female is likely to make a difference to how people behave and think; that socio-economic class makes a difference; and that ethnicity makes a difference.