Sampling Method

This study was conducted in Beijing in the summer of 2011. A stratified survey design was chosen to randomly select neighborhoods and households from a sample frame, according to random number table or using draw lots [26]. Interviewers obtained information about households from neighborhood committees before a random sampling was conducted; thus, no repeat interviews were conducted in this survey.

Once a sample household was identified, a face-to-face interview was conducted at home with randomly selected household members who were 18 years of age or older and had lived in Beijing for at least one year [14]. Before each interview, the respondents completed a questionnaire without consulting other family members besides the respondents and researchers. Respondents were given a gift as an acknowledgement of their participation after the 30 to 40 minute interview.

Members of our research group conducted the survey. In addition, we recruited and trained investigators with a bachelor’s degree or higher and at least one year of work experience in a professional market research company, (preferably in a conservation-related area) [15]. To expedite this search, we also employed a polling firm to help us train investigators involved in the study.

The interviews were randomly checked by supervisors on the spot to assure sampling normalization. Interviews containing any omissions or errors were asked for a re-run. Interviews were not deemed suitable for a re-run if it was determined that knowledge of the survey content which respondents gain from the last round of survey would bias the results, for instance, respondents may have gotten the right answers to the questions from the last round of the survey. Supervisors randomly selected 30% of the questionnaires to validate by telephone return visit, to ensure that no fraud had taken place and eliminate unqualified questionnaires.