To explore the feasibility of a geographical analysis (mapping) of data from a national adolescent survey.
Teenagers 15 to 19 years in high school or apprenticeship (N = 5275 boys and 3993 girls) answered a self-administered anonymous questionnaire targeting health and lifestyles. Switzerland was divided in 14 subunits, each one including at least 400 subjects: large cantons were isolated by themselves whereas smaller ones were aggregated taking into account common cultural, geographical and health characteristics. To minimize the impact of sample differences across the cantons, subsamples were weighted according to age, sex, and professional background. For each variable, the discretization of the values in five classes was performed using the threshold approach which maximizes intergroup differences and minimizes the intragroup differences.
The analysis suggests different patterns of distribution depending on the type of variable studied: substance use differs mostly between urban and rural cantons, while mental health differs between French and German-speaking cantons.
Whereas most available atlases are derived from population-based data (i.e., mortality rates), this research demonstrates the feasibility of using self-reported data from school-based survey samples. The presentation of data on attractive maps provides a unique opportunity for generating hypotheses as to the causes of differences across cantons in a multi-ethnic, multicultural country.