Inheriting the City: The Children of Immigrants Come of Age

Methodological Appendix

Overall Approach A pilot study to test the feasibility of surveying the second generation in New York began in July 1996 with funding from the Russell Sage Foundation . Any such survey faces the immediate challenge of deciding whether to sample a cross section of the whole second generation, which would include members of many groups for which we could not economically gather a statistically reliable subsample, or to sample specific groups in order to explore how group membership might or might not influence individual differences, even though the result would not be a true cross section. Since new immigrants to the New York area come from world regions that differ strongly in terms of racial, ethnic, economic, and cultural traits, and since they enter a resident population that is also sharply differentiated on these dimensions, we opted to stratify our sample in terms of race and ethnicity as well as region of origin. In other words, we sought to sample the most important groups, not develop a random cross section. The pilot study sought to determine the feasibility of completing inperson interviews with members of important immigrant groups in each major racial-ethnic category: Chinese and Koreans among the Asians; Dominicans and Colombian, Ecuadoran, and Peruvian Hispanics; and West Indian and Haitian blacks; as well as the native control groups—whites, blacks, and Puerto Ricans. This experience taught us that the Korean and Haitian second generation populations were too small and too dispersed to be found in a costeffective manner. Moreover, it became clear that trying to reach people in person would yield far worse response rates than would be obtained through telephone interviews, in part because people were often reluctant to open their doors or were not at home. This convinced us to switch to Methodological Appendix 371 screening and interviewing respondents by telephone. As a result, our interview instrument would have to be much shorter than if we were attempting in-person interviews. Since we felt, however, that we could only capture many complexities of the respondents’ racial and ethnic identifications and personal experiences through in-depth interviews, we decided to ask a subsample of the telephone respondents to complete an in-person interview to probe these issues. To evaluate and refine our draft questionnaire, we commenced a pilot telephone survey of respondents in fall 1997. That experience and some preliminary in-depth interviews yielded a revised telephone questionnaire that we used with a supplemental sample gathered in distant suburbs beginning in fall 1998. For this supplemental sample, we called households with Chinese and Hispanic last names with listed telephone numbers and dialed randomly into known suburban concentrations of black residents. This pilot phase yielded 657 suburban respondents. Since these data were not gathered by the same method as the main sample, we do not analyze them in this study. The Telephone Survey With the validity of the questionnaire confirmed, we initiated the random digit dialing (RDD) of the main telephone survey in late 1998, completing a total of 3,415 additional interviews by early 2000. We used an innovative two-stage sampling methodology. A screening telephone call enumerated members of the household, identified its racial and ethnic background, and determined if an eligible 18 –32-year-old child of immigrants from one of our groups lived in the household. Because our interviewers were often speaking with immigrant parents and household members , we used multilingual interviewers who could conduct the screening interview in English, Spanish, Cantonese, Mandarin, and Russian. (Once the main interview began, we also offered the second generation respondents the opportunity to continue in another language, but only three chose to do the interview in Spanish.) Once the screening call identified an eligible person, the main interview began or was scheduled. The screening for the telephone survey covered the four boroughs of New York City (excluding Staten Island), two inner city counties in northeastern New Jersey (Hudson and Passaic), and four inner suburban New Jersey and New York counties (Essex, Hudson, Nassau, and Westchester ), where the 1990 Census Public Use Microdata Sample showed that 372 | Methodological Appendix our target respondents were present in at least 3 percent of the microdata area’s households. This sampling frame eliminated two-fifths of the total population of the New York metropolitan area but retained four-fifths of the target second generation groups and two-thirds of the blacks and Puerto Ricans. (The sampling area had a total...