Your design will depend on the specific objectives of your proposed study. Looking at your general topic of study, I would suggest mixed methods approach where you use both qualitative and quantatative data collection tools such as in-depth interviews and administering questionnaire. That will give you a broader picture on the impact of advertisements on children.

I agree with Gabriel that a mixed method study may work well with your objective because and you first need to identify the types of variables you will be analyzing. You have not identified any behavioural variables correlated to TV advertisements. You likewise have not yet determined the parameters around the intervals of TV advertisements or the frequency of which may have any sort of correlation to children's behavior.

By using the mixed method approach with a combination of questionaires and factor analysis, you may find some variables which are correlated from which you can perform subsequent studies to conclude on stronger predictive models.

Thanks everyone for sharing your views. Well, I am a bit unclear about the data collection part, please, suggest.

Thanks again.