The following sections introduce our sampling design and sampling results (final samples).

1. Sampling Frame

Due to the relatively high acceptance rate of Taiwan’s mandatory middle education, TYP includes the entire junior high 1st and 3rd year students (excludes night schools, tutor schools, and continuation high schools) as the sampling population.

2. Sampling Size

For the purpose of having a sampling size large enough for statistical power, TYP collects 1000 students in the 1st and 3rd year of junior high from both Taipei City and County, and 800 students in the 1st and 3rd year of junior high from Yi-Lan County. The total sampling size is 5600 students.

3. Sampling Method

TYP employs a multi-stage stratified cluster sampling method, with the division of county and city as the first stratum and division of town, county, city, and district respectively as the second stratum. After that, cluster sampling is applied to divide the second stratum into two sub-strata for random sampling. “School” is further used as the sampling unit at the first sub-stratum while “class” is used as the final unit at the second sub-stratum. The entire sampling process is demonstrated below:

The first stage is based on proportional stratified sampling. Since TYP is interested in examining the contextual role of community on youth's physical and mental health, we first divide the area of community into strata according to its degree of urbanization based on RCHSS's urbanized degree scale for town, county, and city as the standard. At this stage, TYP stratify our targeted city/county into 2-3 areas according to their urbanization degree, that is, this procedure reduces heterogeneity between sampled units across different strata while allowing sample-specific characteristics to be homogeneous in the same stratum. Furthermore, the number of samples for each stratum to be taken is determined by the ratio of the student number in each stratum to the total student number. Two classes are sampled in each school, and the number of schools for sampling is determined thereafter.

The second stage is based on cluster sampling. The sampling unit at this stage, “class”, enables all classes in each stratum to have equal chances of being sampled. Proper numbers of schools and classes are then selected according to the number of students needed for each stratum. To enable sampled students representative of the students in a particular school’s particular grade, two classes are selected from each school’s 1st and 3rd grade, respectively. The sampling procedure demonstrated above is applied to Taipei City/ County and Yi-Lan County, and the sampling in each area is conducted individually.

In order to include greater diversity into the samples, TYP takes the physical education class of the selected public junior high “B” in Taipei City into our sampled classes (one class from the 1st and the 3rd year each) in addition to the selected private junior high “A” in Taipei County (two classes from the 1st and the 3rd year each).

4. Sampling Procedures

The first stratum divides Taipei County, Taipei City, and Yi-Lan County into three sectors. According to the degree of urbanization within its administrative area, the three sectors are further divided into two to three sub-sectors. Then, the sampling is carried out in the following steps:

a. Calculate the ratio of each sector’s student number to the total student number in that county/city and use this ratio to determine the number of students needed to be sampled in each sector.

b. Calculate each sector’s mean number of students in each class. Then, calculate the number of classes needed to be sampled and use it as the number for the next stage’s cluster sampling.

c. After obtaining the number of sampled classes needed, the number of schools needed to be sampled in each sector is determined by the principle that in each school two classes would be picked.

d. Randomly select schools for sampling in each sector.

e. Pick two classes from the 1st and the 3rd year of the randomly selected schools. By the cluster sampling, all the students in those selected classes are then taken as our study subjects.

According to the sampling methods and steps mentioned above, TYP has randomly selected 40 junior high schools (16 from Taipei City, 15 from Taipei County, and 9 from Yi-Lan County). In each of the schools, two classes were randomly chosen and all the students and their parents in the 81 classes were surveyed (the number of 7th graders is 2696, 9th graders is 2890, and the total student samples is 5586).