# Sampling Design and Framework

We adopted a multi-stage stratified sampling method.

• 1. Area Sampling
• • Firstly, the whole country is divided into three regions, namely being the eastern, central and western areas, according to their economic statuses. And then, two provinces (or municipalities) were selected from each area as samples, including Beijing and Jiangsu province from the eastern area, Shanxi and Hebei provinces from the central area, and Yunnan and Sichuan provinces from western area.
• • Selecting cities from each area as samples (provincial capitals or prefecture- level city).
• • Selecting districts and county from sample cities.
• • Selecting schools from sample districts and county.
• 2. School Sampling
• • Determining the sample number of students based on the proportion of the students in reading in the junior schools from cities and counties.
• • Determining the sample number of schools based on the proportion of students studying in different kinds of schools (demonstration or general).
• 3. Student Sampling
• • The sample students are in grade 2 from junior school and high school.
• • Cluster random sampling method is used in the sample schools - that is choosing randomly a class as a sample.

# Distribution of the Number of the Samples in Different Areas

According to the Statistical Yearbook of the provinces in 2012 (the date for Jiangsu is 2011), the total number of middle school students from Jiangsu, Beijing, Hebei, Shanxi, Yunnan, and Sichuan provinces, includes 5,735,222 high school students and 11,088,310 junior school students. The proportion of high school students to junior school students is about 1:2. Then we distributed the numbers of the questionnaires for the students of high-schools and junior schools based on the percentage. The number of questionnaires delivered to students in cities is also based on the proportion of the population of the provinces (or municipalities). And for detailed datex, please see the table below.

The data of the questionnaires were analyzed by SPSS 16.0. The main statistical analysis tools are frequency distribution analysis, multidimensional frequency and distribution analysis.