Further comments on the use of DJN

Abstract

The use of Distinctive Jewish Names (DJN) in Jewish population surveys is becoming more popular. It is therefore necessary to inquire to what extent the method biases results. Our work on DJN has concentrated on the usefulness of a DJN sample versus an organization-list sample (Himmelfarb, Loar, and Mott 1983). Based on the assumption that area housing or phone exchange samples (e.g. random digit dialing) is beyond the means of most social scientific studies on American Jews, we asked whether the use of DJN would not provide a more representative low-cost sample of Jews than the use of Jewish organization lists (i.e. federation lists). The answer, in short, was yes. Using the National Jewish Population Study (NJPS) data, we found that persons with thirty-five distinctive Jewish names were not very different from the other Jews in the sample with regard to demographic characteristics and Jewish identification. The differences that existed were not consistently in one direction, e.g. greater identification for the DJN. Moreover, regardless of direction, the magnitude of differences was low. On the other hand, a comparison of the NJPS respondents chosen from Jewish federation lists and the NJPS respondents chosen via area housing probability methods showed more substantial and more consistent differences. The list sample was more Jewishly identified and had many of the demographic characteristics typically associated with greater Jewish identification. Thus, with certain caveats regarding the use of DJN to tap intermarried Jewish women and Russian and Israeli immigrants, we conclude that if researchers must rely on nonprobability sampling methods, the use of DJN should produce a reasonably representative sample with considerably less bias than the use of Jewish organization lists. Bernard Lazerwitz's article in this volume complements our previous analysis and findings. He also uses the NJPS data, but we worked with 99