As French as Everyone Else?: A Survey of French Citizens of Maghrebin, African, and Turkish Origin

Appendix: Methodology

APPENDIX Methodology T HE DATA USED HERE are from the CEVIPOF survey Rapport au politique des Français issus de l’immigration (“Attitudes toward politics of the French stemming from immigration,” or RAPFI). The TNS-SOFRES Institute was chosen as a partner for conducting the unedited survey with a representative sample of 1,003 French people 18 years old and older with a personal or familial history of African or Turkish immigration (for whom at least one of the parents or grandparents has or had nationality in one of the relevant geographic areas cited above). We drew attention to the criteria of nationality, which was a first among the sampling surveys that focus on this population. Using birthplace would have included Algerian repatriates, who do not belong in the groups that principally interest us. Some surveys conducted before ours also made the choice to use religion to define the New French, but we realized that to define our population by Islam would give a distorted impression of reality. The RAPFI survey proved novel on other counts as well. First, in defining our population by the presence of at least one immigrant in two generations, we included mixed families. Second, the method we used to contact interviewees solved the biases of prior survey research, biases notably revealed by JeanneH élène Kaltenback and Michèle Tribalat (2002) and Claude Dargent (2003). Faced with a population that is numerically small, for which demographic knowledge remains fragmented, survey institutes often used methods that seemed unsatisfactory to us. For example, the use 118 ■ Appendix: Methodology of patronymic selection, while more economical, introduces a bias in the sample because it excludes all mixed families for whom the father is not from the target population. Also, face-to-face surveys pose numerous problems: for reasons of expense, they concentrate on geographic areas with large foreign populations, generally urban, and therefore tend to exclude the population of individuals living in rural areas or areas with a small percentage of foreigners. This was probably significant for several of our indicators, notably the social networks of people interviewed and the socioeconomic aspects of the population. That is why we made the decision to privilege the telephone and its regular method of constructing a survey population. Third, asking a question early in the interview about the nationality of parents and grandparents might have proved detrimental to the acceptance rate of the survey; several socio-demographic questions (sex, age, profession, nationality of the individual) were therefore asked first, before the filter questions. Thus, 28,000 people started the questionnaire before 1,003 among them were selected.1 TNS-SOFRES was also in charge of the mirror survey, which was based on a representative sample—the “control group”—of 1,006 French people 18 years old and older. The method was exactly the same for these two surveys: telephone interviews lasting approximately 35 minutes. In both cases, a collection of surveys representative of the two household categories was built according to the quota sampling method (the History of Families survey [EHF]2 and the Employment survey of the Institut National de la Statistique et des Études Économiques [INSEE]3). The survey of the New French sample took place between April 8 and May 7, 2005. The control group was surveyed between April 13 and 21, 2005, from a questionnaire as comparable as possible to that of the RAPFI survey, including the content and order of the questions. The two questionnaires are available from the authors on request. 1 This research phase was important for the duration of the survey fieldwork. We thank here all those who administered the survey and Éric Hély, their supervisor, for their excellent work. 2 For additional information on this survey, see Tribalat 2004a. 3 The Employment survey is available at http://www.insee.fr/fr/themes/detail.asp?ref_id =fd-eec09&page=fichiers_detail/eec09/telechargement.htm. Appendix: Methodology ■ 119 THE QUESTION of the quality of the data and of the way in which they “represent” the targeted population is clearly a central concern of the research project. In this section, we briefly present the socio-professional characteristics of the New French sample, compare them with our control group, and estimate the goodness of fit of these characteristics to those coming out of the EHF survey. First, the distribution by origin in the sample is as follows (Table 36): 40 percent of the people interviewed count among two generations of their family at least one immigrant from Algeria, 25...