Several studies have assessed the AIDS-related knowledge, attitudes, and behavior of Dutch students (e.g., Vogels & van der Vliet, 1990; Brugman et al., 1995; Schaalma, 1995). Far fewer studies have been conducted outside the school setting, and have generally used different methodologies with regard to sampling, data collection, and analysis (e.g., Nabben et al., 1993). Comparing the findings for students and dropouts is therefore nearly impossible.
In the Netherlands, schools take an active role in AIDS and safe-sex education (Mellink, 1989; Mellink & Gijtenbeek, 1996). Thus dropouts, as compared with students, are expected to be less knowledgeable about AIDS and safe sex and to show less favorable attitudes toward condom use. In addition, because the majority of dropouts are employed, they are assumed to have more money, making bars and discos more accessible. Sexual activity has been shown to be highly related to visiting bars and discos and to drinking alcohol (Vogels & van der Vliet, 1990; Danz & Vogels, 1991; Brugman et al., 1995), leading to the assumption of greater sexual activity among Dutch dropouts. Such assumptions are supported by studies in other Western countries (e.g., King et al., 1988; Centers for Disease Control, 1994). If true, dropouts should be considered at elevated risk for HIV and other sexually transmitted diseases (STDs).
In 1995, a large survey was conducted on the sexual behaviors of Dutch adolescents. The samples consisted of 7,299 secondary school students aged 12 to 19 years and 571 dropouts aged 16 to 18 (dropouts younger than 16 are nearly nonexistent in the Netherlands). Using data from this survey, the above hypotheses were tested.
A stratified sample of 85 secondary schools was selected, based on type of education, location (urbanization), and type of administration. The latter sampling criterion was applied because a large number of Dutch schools are administered by nonstate, religiously associated school boards. In each school, data were collected in every grade, helping to achieve representativeness with regard to student age and gender. Schools that refused to participate were replaced by comparable schools in terms of the sampling criteria.
Two thirds of the schools in the initial sample agreed to participate. The reasons cited for refusing involved mostly organizational and time problems. However, the refusal rate was higher among Protestant schools than among Catholic or public (state) schools.
Parental consent was obtained prior to administration of the questionnaires. Students were also informed that participation was voluntary.
The response rate was exceptionally good. More than 91% of the students completed the questionnaire. Nonresponses were attributed mainly to absenteeism due to sickness or truancy (7%). Only 2% of the students or their parents refused to participate.
Ultimately, 7,288 questionnaires suitable for analysis were obtained. The sample was representative of the Dutch student population, with one exception: students from the small group of strict Protestant schools were probably underrepresented due to the larger refusal rate of those schools (compare Brugman et al., 1995). For the analyses in the present study, a subsample of older students (ages 16-18) was drawn to match the dropout sample. In addition, because adolescents from ethnic minorities were grossly underrepresented in the sample of dropouts, students were restricted to those born in the Netherlands whose mothers were also born there.
It is far more difficult to obtain a representative sample of dropouts. Some are employed, but contacting them via their employer was considered impractical and bound to lead to a large nonresponse rate. Some live with their parents, but others do not. Some are homeless and/or drug addicted. Considerable effort, therefore, had to be made to collect data among a varied sample of dropouts. …