ABSTRACT
While we have described many of the specific details of our survey procedures in appendix A, it is important to outline briefly several key aspects of our research strategy. First, because we wanted to be able to generalize from our example, we chose a stratified random sample based on various locations, different fields, and differing statuses in the United States (i.e., visiting scholars, current students, and people out of school and in the workforce). For each locality where we carried out interviews, we compiled a list of names of students in several schools and then chose randomly from each list. We did the same for those in the workforce, asking Chinese friends to give us lists of names of their friends. As we explain in appendix A, we were unable to do this for visiting scholars. The numbers were too small and changed yearly; instead, we simply ensured representation from both sexes and from different fields, age groups, and different lengths of time in the United States. Second, rather than using a mailed questionnaire, we employed face-to-face interviews based on a 105-question questionnaire, which we pretested in Toronto and then revised for use in the United States. While we hoped that using face-to-face interviews would increase our yield—a rejection rate of above 50 percent, which can happen with a mailed survey, would call the representativeness of our sample into question—we still experienced a refusal rate averaging 25-30 percent. Nevertheless, our findings, the frankness of the responses, and the variety of people we interviewed make us confident that we tapped views that represent those of a cross section of Chinese students, scholars, and former students in America.
