Sampling by Ethnic Surnames: The Case of American Jews

Transcription

1 Sampling by Ethnic Surnames: The Case of American Jews HAROLD S. HIMMELFARB, R. MICHAEL LOAR, AND SUSAN H. MOTT BECAUSE members of minority groups constituting a very small population of a national population are difficult and costly to locate using standard probability sampling, social scientists interested in studying small minority groups have frequently had to rely on nonprobability sampling methods. Some of the most common ways of obtaining rare population subjects have been by screening households in known minority communities, recruitment from lists of persons known to belong to the group, such as minority group organization lists, and selecting subjects on the basis of definitive demographic characteristics, such as physical appearance, language, and names. Yet the use of such techniques makes the representativeness of the selected sample and the research results questionable. If research projects on small population groups must rely on nonprobability methods of sampling, therefore, an attempt should be made to determine which of those methods yields the most unbiased Abstract It is often difficult and costly to locate members of numerically small minorities using standard probability sampling. Consequently, nonprobability sampling techniques of various sorts are commonly used. This paper analyzes the differences between samples chosen by two such techniques the use of ethnic surnames and the use of organization lists and compares them with probability samples. Using data from the National Jewish Population Study, we find that Jews with one of 35 so-called distinctive Jewish names do not differ substantially from the general population of Jews in demographic characteristics or indicators of Jewish identification, and that this technique produces a sample which is more similar to the general population of Jews than does the organization list sample technique. Harold S. Himmelfarb is an Associate Professor in the Department of Sociology, The Ohio State University. R. Michael Loar is a Systems Analyst for the State of Ohio Department of Highway Safety. Susan H. Mott is a Senior Research Associate at the Center for Human Research at The Ohio State University. The authors wish to thank the Council of Jewish Federations for lending the data to do this study. Public Opinion Quarterly Vol. 47: by the Trustees of Columbia University Published by Elsevier Science Publishing Co.. Inc. 0O33-362X/83/0O47-247/S2.5O

2 248 HIMMELFARB, LOAR, AND MOTT estimates of actual population characteristics. This paper examines the differences between a probability sample of a small population group (American Jews) and two nonprobability methods of sampling them the use of distinctively Jewish surnames and the more commonly used method of organization membership lists. Samples of American Jews American Jews are a good example of a small minority group that has received considerable research interest and yet study after study is plagued with the problems of unrepresentative samples. Jews constitute a small proportion of the American population and in recent decades that proportion, now estimated to be between 2.6 percent and 2.8 percent nationally (Massarik, 1974), has been declining. The U.S. Census does not ask a question on religion, and since surveys by national polling agencies yield very few Jews, some of the major sources of population data used by social scientists to study Americans have not been adquate for the study of American Jews. Consequently, most studies in this area have used nonprobability samples of an accidental or purposive nature. The most common sample is usually taken from some organization list or combination of lists which contain Jewish names and addresses. In small Jewish communities where affiliation rates for Jews are as high as 90 percent, the use of organization lists does not present big problems. However, in large Jewish communities, where most American Jews live, these lists necessarily bias the sample toward affiliated Jews, which we now know are only about half of all adult Jews in the United States (Bock, 1976; Lazerwitz, 1979). Thus, there is serious doubt about the generalizability of many of the findings. In the last decade and a half, there have been numerous Jewish community censuses which use a combination of list and area housing samples. 1 There was also the National Jewish Population Study (NJPS), which is based on a complex multistage, national probability sample (Lazerwitz, 1973). These large-scale surveys have been expensive and typically beyond the financial means of individual researchers working in the field. Thus, there is still need for some relatively inexpensive, efficient method of screening the population in order to obtain a representative sample of Jews for various types of research concerns. One method which might be adequate for this 1 See Appendix in Elazar (1976) for a listing of community studies before that date. Contact the Council of Jewish Federations. 575 Lexington Ave., New York, NY 10022, for a list of more recent studies.

3 SAMPLING BY ETHNIC SURNAMES 249 purpose is a sample of individuals with distinctive Jewish names (DJNs). The Development of DJN Research The DJN method has been around for quite some time and has been used by demographers to identify geographic clusters of Jewish populations. While there have been several working papers and unpublished reports on the method, to our knowledge there has been only one published paper on the subject (Massarik, 1966) and it does not fully explain the development of the DJN list. Moreover, there is only one copy available of the original report (Kohs and Blumenthal, 1942) on the development of the method. Therefore, a brief history is in order. In 1942, Samuel C. Kohs compiled a list of the most common names appearing in the Los Angeles Jewish Federation files. He found that 106 surnames accounted for about percent to percent of the names on various Jewish Federation lists. Overall, his list accounted for percent of the total names on all lists. Upon further analysis, Kohs found that there were 35 names 2 which appeared 10 or more times on each of the lists. These 35 names accounted for percent of the total names on all of the lists. Generally, about 90 to 92 percent of persons with these 35 names are Jewish, although this might drop as low as 70 percent in some locales. Thus, by using only 35 names, the percentage of all Jews included is not much less than by using all 106 names, but the probability of actually sampling a Jewish person is very high. In a sense, then, these particular names are not only "distinctive" they are also "common." There are names not included in the list which might yield even higher proportions of Jews but persons with those names are fewer in number. Massarik (1966) reports that with some local variation the 35 DJNs constitute a relatively constant proportion of total Jewish persons, which he sets at 12 percent. Therefore, Massarik has argued, the DJN can be used to get a fairly accurate estimate of the total size of the Jewish population, achieving greater accuracy in areas with dense Jewish populations than in low-density areas. Lazerwitz (n.d.) reports that in the 1971 National Jewish Population Study (NJPS), 15.5 per- 2 Berman, Bernstein, Caplan, Cohen, Cohn, Epstein, Feldman, Friedman, Ginsberg, Gold, Goldberg, Goldman, Goldstein, Greenberg, Grossman, Horowitz, Kahn, Kaplan, Katz, Levin, Levine, Levinson, Rosen, Rosenbaum, Rosenbloom, Rosenthal, Rothman, Rubin, Samuels, Shapiro, Siegel, Silverman, Weinberg, Weiner, Weinstein.

4 250 H1MMELFARB, LOAR, AND MOTT cent of the Jewish respondents had the 106 DJNs. In fact, the Jewish identification sample of the NJPS to be used in this paper had 14.5 percent of the sample with 106 DJNs and 11.4 percent with 35 DJNs. Thus, the proportion of Jews with DJNs has indeed remained relatively constant overtime and nationwide. Of course, as everyone who has used the DJN method has noted, estimates of Jewish population size require some knowledge of the proportion of the DJN in the general population and also the proportion of DJN and general population who might not be included in whatever name list is being used (e.g., phone directory). Over the years a number of studies of American Jews have chosen their samples by using a list of names considered by the researchers to be distinctively Jewish but they have not always stuck to the original DJN list exclusively (Rosenthal, 1948; Himmelfarb, 1974; Mayer, 1979; and Jaret, 1978). Most recently, DJN was used as a sampling strategy for the New York Area Jewish population. Ritterband and Cohen (1978) suggest that by identifying phone exchanges where Jewish households are clustered (with clusters identified by Jewish names), the number of exchanges needed for random-digitdialing could be cut by three-quarters. This achieves great economies while excluding only about 10 percent of those Jews whose phone exchanges are not in the cluster of Jewish households. Thus, sampling by names has been fairly common and continues to be advocated as a cost-saving method. It has even been used for demographic studies of Jews in other countries (DellaPergola, 1977). A More Direct Use of DJN Samples Past research with DJN has concentrated on using the method to identify where Jews live and estimating Jewish population size. Rather than using DJNs to locate other Jews to be sampled, for some purposes it would be more economical, and perhaps as efficient, to actually sample only persons with DJNs. Such a sample might be less biased than the commonly used samples of Jewish organization lists. To date, there has been no research on the extent to which Jews with DJNs are representative of Jews generally. Two limitations seem obvious. First, it is quite clear that Jewish women who are married to non-jewish men are likely to be unrepresented in a DJN sample. With rising intermarriage, this might be of some concern. However, this problem is minimimzed somewhat by the fact that Jewish females are only one-third as likely as Jewish males to marry someone of a different faith (Lazerwitz, 1981). Second, we might assume that persons with DJNs are more likely to be foreign-born, older, and have

5 SAMPLING BY ETHNIC SURNAMES 251 higher levels of Jewish identification than Jews generally. If these or other differences exist between DJN Jews and Jews generally, it is important to know the magnitude of the differences, because the bias resulting might be less than is typical in other list samples of Jews. It is our intention in this study, therefore, to investigate just how representative DJN Jews are of the American Jewish population generally. SAMPLE Method The data used for this study were taken from the National Jewish Population Study collected between the spring of 1970 and December The full explanation of the sampling method has been described by Lazerwitz (1973 and 1978). Briefly, households were selected in a two-stage process. Organizational lists were obtained and sampled for each primary sampling unit. Then area probability samples were selected in each PSU and integrated with list samples so that there was no overlap between the list and area samples. The households were screened for Jewish occupants, and ultimately 7,179 Jewish households were sampled nationwide. Of these households 5,790 were successfully interviewed for a 79 percent response rate. Despite the length of time that has elapsed since the survey was taken, the NJPS is still the most representative national data available on American Jews. Of the 5,790 housing units, our concern lies with 3,997 respondents who answered a special section of the survey dealing with Jewish identification. One adult Jewish resident per household (per two households in New York City) was selected via the Kish technique to complete the identification section. The weighted N for this section of the survey is 8,544, which is the N used here except for variations on particular variables due to missing responses. MEASURES To investigate the differences between DJN persons and other Jews, we selected eight measures of background characteristics, nine single-variable measures of Jewish identification, and eight Jewish identification scales. The specific single-variable measures were chosen because they are commonly used indicators of Jewish identification and because they range across the diversity of modes of identification in the religious, communal, social, and cultural spheres (see Himmelfarb, 1975): rituais, synagogue membership and attendance, Israel, charity, friendships. The Jewish identification scales were also developed to encompass a diversity of types of identification and to give more reliable indica-

6 252 HMMELFARB, LOAR, AND MOTT tors of those types than can be obtained from single measures. The development of the scales was rather complicated. Briefly, 31 items measuring Jewish identification were factor-analyzed using a principal component solution. This yielded eight orthogonal factors. One of those factors measured contributions to Jewish charities, but had too many missing values to be used. Thus, we were left with seven factors from which scales were formed: Devotional ritual observance. Associationul Jewish organizational participation. Fraternal living near, and associating with, other Jews. Ideological attitudes toward Israel. Intellectual knowledge of Jewish culture, history, religious teachings, and ethics. Esthetic enjoyment of Jewish music and literature and possession of Jewish art objects. Affectional attitudes about the Jewish people and the religion. Finally, the seven scales were rotated and yielded only one factor. Therefore, an overall measure of Jewish Identification (JI) was formed by weighting the separate scales by their factor score coefficients and adding them together. (See the Appendix for a more complete explanation of the construction of these scales as well as the survey questions from which scales and the single-variable measures of Jewish identification were derived.) The background variables were as follows: Number of children in household number of persons in household under 21 years of age. Age from 19 on up (only adults were used in the Jewish Identification subsample). Sex (1) male; (2) female. Education^-in completed years. Family income in 17 categories from low to high with numbers between 0 and 50. Generation American (1) foreign-born; (2) respondent Americanborn but had at least one foreign-born parent; (3) respondent American-born and both parents American-born; (4) respondent and parents American-born and at least one grandparent American-born Jewish education total hours of Jewish schooling, estimated by multiplying the estimated hours spent on Jewish studies in each type of Jewish school by the number of years in that school type. (For the hour estimates of each school type see Bock, 1976.)

7 SAMPLING BY ETHNIC SURNAMES 253 Denomination (1) Orthodox and Traditional; (2) Conservative and Reconstructionist; (3) Reform; (4) other, agnostic, atheist and "Just Jewish." Findings Previous analysis (not shown here) found that the 35 DJN were more different from the rest of the Jewish population than the entire list of 106 DJN. However, in general the differences between each of the DJN lists and the rest of the Jewish population were quite small. Therefore, for purposes of this paper, we will present only the comparison for the DJN group with largest differences, i.e., the 35 DJN and all other Jews in the sample. Table 1 shows the mean differences in background characteristics between Jews with one of the 35 DJN and all other Jews in the sample and between the list sample and area sample Jews. It also shows the significance levels for: the /-tests of the differences between the means, and the Spearman rank-order correlation between background and name-type (DJN=1, others=0) and between background and sample-type (organization list=l, area sample=0). Looking at the background variables, we see that DJN Jews are: a little younger, more likely to be female, less secularly educated, more wealthy, and, surprisingly, more native American than are other Jews. They also have a little more Jewish education and are less liberal denominationally. However, a look at the significance levels shows that only the differences in age, generation, education, and Table 1. Mean Differences in Background Between Jews with DJNs and All Other Jews and Between List Sample and Area Sample Jews N children in household Age % female Secular educ. (yrs.) Income Generation Jewish educ. (hrs.) Denomination (Orthodox to other) 35 DJN , All Others* * * * 10, * Spearman r List Sample* , Area Sample".81* 46.6* 53* 13.7* 12.3* 2.4 8,401* 3.0* Spearman r * r-test for differences between means significant at least at.05 level (two-tailed). + p ss.001. a N ranged from 875 to 948 except for income where N dropped to 637. b TV ranged from 7,373 to 7,596 except for income where N dropped to 5,665. c N ranged from 2,896 to 2,987 except for income where N dropped to 2,191. d N ranged from 5,395 to 5,549 except for income where N dropped to 4,106.

8 254 HDHMELFARB, LOAR, AND MOTT denomination are statistically significant. The Spearman correlation is significant only for age and generation. In any case, all of the correlations besides age are very low. It is no surprise to find that the DJN sample fares better in approximating the composition of a more general sample than does the sample drawn from organization lists. A look at the last three columns of Table 1 shows that the list sample differs dramatically from the area sample drawn from organization lists. A look at the last three columns of Jewish education, and especially income. The Spearman correlations are significant for every variable except sex, secular education, and generation. The differences between the list and area samples are almost always substantially greater than the differences between the DJN sample and all others. Table 2 provides similar comparisons between the Jews with DJNs and all others and between the list and area samples with regard to the various measures of Jewish identification. We expected to find that Jews with DJNs would be persons with stronger Jewish identification. However, our expectations were confirmed in only a few instances. The first three columns of Table 2 show that the DJN Jews are a little more ritually observant than the others, a slightly higher percentage gave to a Jewish charity, and more of them are members of a synagogue. On the identification scales, DJN Jews score higher than the others on the devotional and intellectual identification scales; however, they score lower than the other Jews on all of the other scales, including the overall measure of Jewish identification (JI). A few of the mean differences and Spearman correlations are statistically significant for the single-variable measures of Jewish identification, though in most cases the mean differences are quite small (and significance levels are easily achieved with such large samples). Though the differences are greater for the scale measures, the Spearman correlation is only of notable magnitude for the fraternal scale. Apparently, DJN Jews are less likely to have Jewish friends, neighbors, and business associates. A look at the last three columns of Table 2 shows that the organization list sample less adequately approximates a more general population of Jews than does the DJN sample. On most measures of Jewish identification, the list sample differs from the area sample to a greater extent than the DJN sample differs from the "all others" sample. (The only notable exception is the fraternal scale.) Affiliated Jews (the list sample) are considerably more observant and have a stronger Jewish identification than the others, particularly on affiliational measures, i.e., synagogue membership and activeness in Jewish organizations (associational scale).

9 SAMPLING BY ETHNIC SURNAMES 255 Table 2. Differences in Jewish Identification Between Jews with DJNs and all Other Jews and Between List Sample and Area Sample Jews Selected measures' 1 % for whom all or most friends are Jewish % who have been to Israel % who gave to a Jewish charity % who attended Seder last year % living in household that observes kosher % who said Kiddush last Sabbath % who observed Sabbath in some way % who attend services on all or most weekends % who belong to synagogue Jewish ID scales (means) Devotional Associational Fraternal Ideological Intellectual Affectional Esthetic Total JI 35 DJN* All Others* 68* 15 95* 82* * 6 47* -.02*.01*.03*.01* *.01*.02 s Spearman r List Sample* Area Sample* 64* 13* 93* 81* 27* 18* 36* 5* 39* -.03* -.14* -.02* -.10* -.02* -.08* -.05* -.09* Spearman r * /-test for differences between means significant at least at.05 level (two-tailed). * p = "Non selected measures ranged from 735 to 876 except for charity question, where N dropped to 502. N on scales ranged from 909 to 948. h Non selected measures ranged from 6,000 to 7,659 except for charity question, where N dropped to 5,212. N on scales ranged from 7,053 to 7,596. c N on selected measures ranged from 2,409 to 2,983 except for charity question, where N dropped to 2,294. N on scales ranged from 2,794 to 2,987. " N on selected measures ranged from 4,320 to 5,547 except for charity question, where N dropped to 3,414. N on scales ranged from 5,760 to 5,549. ' The percentages shown are actually the mean values of variables dichotomized into the classification of 0 to 1. Even these differences, however, are not as great as one might expect, and this might encourage some researchers to continue using

10 256 HIMMF.LFARB, LOAR, AND MOTT organization list samples. We would argue that the differences in Jewish identification coupled with the differences in demographic characteristics (particularly in income) are sufficient to caution researchers against the list method, especially in light of our findings that DJN Jews are a readily available sample and generally quite similar to representative samples of the American Jewish population. Conclusions Common sense would lead us to believe that Jews with distinctive Jewish names are likely to be older, to be foreign-born, and to have a stronger Jewish identification than other Jews. In fact, the opposite is true. Apparently, name changing was more common among the immigrant generation. American-born Jews, although not identifying in traditional Jewish ways, are comfortable enough in their American status to maintain names which are distinctively Jewish in character. The most important point of our findings, however, is not the direction of the differences which exist between those with DJNs and other Jews, but the fact that the differences are very slight and considerably less than those which are obtained when samples are from Jewish organizational membership lists. Therefore, we believe that a random sample of persons with DJNs is likely to produce a fairly representative sample of American Jews. The low cost of sampling only DJN Jews, particularly for explanatory rather than descriptive studies (i.e., studies which are interested in explaining patterns rather than making accurate estimates of population characteristics), and particularly for studies not concerned with interfaith marriages or immigrant experiences, is likely to outweigh the slight unrepresentativeness which might result. Several caveats might be noted to this conclusion: 1. We have used a national sample and have not looked at possible geographic differences in the characteristics of DJN persons. Given the minimal differences here, it is likely that differences will be minimal in any large-size Jewish community. However, in small Jewish communities, where persons with DJNs are perhaps more likely to be related to each other, and where there are many fewer persons from which to choose, DJN Jews might differ from other Jews to a much greater extent. This needs to be researched further. 2. Sampling Jews with DJNs implies using some existing list of names and addresses such as a phone or house directory. All such lists have limitations which might produce substantial biases (e.g., excluding large numbers of persons with new phone numbers, unlisted

11 SAMPLING BY ETHNIC SURNAMES 257 phone numbers, or those without phones) regardless of who is sampled from them. These inherent biases should be considered when weighing the advantages and disadvantages of sampling from those lists. 3. Sampling from the DJN list presents additional biases: Jewish women married to men who were not born Jewish are not likely to have a Jewish surname and therefore will be underrepresented in DJN samples. Similarly, of course, males who have converted to Judaism will be underrepresented by this sampling method. The names on the DJN list are primarily characteristic of Eastern European Jews. Thus, Sephardic Jews (those from Mediterranean, Asian, and African countries), possibly new Russian immigrants, and many recent Israeli immigrants (the largest new Jewish immigrant group to America) are also likely to be underrepresented by this method. Where there are large settlements of these groups, DJN sampling should be used with caution, perhaps by modifying the name list to include names which are commonly found among these immigrants. 3 The DJN list was developed almost four decades ago. Its full potential as a research tool has yet to be reached. 4 Despite the warnings above, we believe that the benefits of DJN sampling are likely to be substantial and that our research indicates that many of the feared limitations are likely to be slight. The distinctive name method also needs to be developed for use in research on other numerically scarce ethnic groups. For example, there is impressionistic evidence that name lists of Greek-Americans, Chinese-Americans, and Korean-Americans would be even more efficient than the name lists developed for use in research on Jews. There is a great need for developing low-cost sampling procedures particularly for research on American religious denominations and ethnic groups. Further work on name sampling might rapidly advance that development. ' In fact, Ritterband and Cohen (1982) report that in comparing their own version of a DJN list of Jews to the other Jews that were sampled in their New York Jewish population survey, all differences were negligible, except for intermarried Jewish women and Israeli and Russian-born Jews. 4 One ambitious advertising service has collated Jewish names from subscription and organization lists. They now have some 27,000 names and addresses on their computer and can establish the likely proportion of persons with those names who are Jewish within particular locales. At an earlier stage of development, this firm provided researchers in New York with 600 names, which constituted 31 percent of Jewish names in the New York Jewish federation files. Researchers were then able to estimate the number of Jews in an area by weighting the DJNs by this factor (Ritterband and Cohen, 1982).

12 258 HIMMELFARB, LOAR, AND MOTT Appendix SINGLE-VARIABLE MEASURES OF JEWISH IDENTIFICATION The single variable measures of Jewish identity were derived from the following survey questions: Percent for whom all or most friends are Jewish: Among my friends [none, few, some, most, all or almost all] are Jewish. Percent who have been to Israel: Have you ever been to Israel? Percent who gave to a Jewish charity: In 1969, did [head] make any contributions to any charitable or welfare campaigns? [If yes] Not counting dues, of the total amount contributed by [head] in 1969, about what percent was donated to Jewish charities? Percent who attended Seder last year: In 1969, did you attend a Seder? Percent living in household that observes kosher: Is "kosher" observed in this household? Percent who said Kiddush last Sabbath: Was the Sabbath observed in this household last week? [If yes:] How was the Sabbath observed? Percent who observed Sabbath in some way: Was the Sabbath observed in this household last week? Percent who attend services on all or most weekends: How many times, if any, have you attended Jewish religious services during the past 12 months? [If any:] On what occasions was that? Percent who belong to a synagogue: Are you a member of one or more temples or synagogues? JEWISH IDENTIFICATION SCALES Except for the Devotional and Ideological scales, all other scales of Jewish Identification (JI) used in this study were formed as composite factors. All the items used in a scale and their corresponding factor weights are listed below. The Devotional and Ideological scales were formed as Guttman scales by successively weighting each item with an additional point from the least difficult item to the most difficult item (see the procedure outlined in the SPSS: Statistical Package for the Social Sciences manual). The percentage of respondents "passing" an item is listed next to it. Devotional (Guttman scale) 82% Attended a Seder last year % Lit Chanukah candles 52% Had only matzah in the home on Passover 36% Lit Sabbath candles last week 29% "Kosher" observed in household 20% Recited Kiddush last week 6% Attended Jewish religious services on all or most Saturdays or Sundays during the past 12 months (coefficient of reproducibility =.90; coefficient of scalability =.63) Associational Active in a Jewish club or organization Officer in a Jewish club or organization Active in a temple Fraternal Most friends are Jewish Most neighbors are Jewish Most business associates are Jewish Being Jewish affects choice of where to live

13 SAMPLING BY ETHNIC SURNAMES 259 Ideological (Guttman scale) 89% Agrees that U.S. Jews should do all they can to help Israel 56% Agrees that Six-Day-War raised the status of the Jewish people in the eyes of American non-jews 36% Plan to go to Israel within the next three years 15% Has been to Israel 11% Agrees that U.S. Jews should move to Israel (coefficient of reproducibility =.91, coefficient of scalability =.61) Intellectual Claims to know Jewish culture Claims to know Jewish history Claims to know Jewish religious teachings Claims to know Jewish ethics Esthetic Enjoys Jewish music Enjoys Jewish literature Possesses Jewish art objects Affectional Is happy to be Jewish Agrees that being Jewish means something very definite to him/her Believes in the Jewish religion Feels that it is important that there will always be a Jewish people Would switch from being Jewish to something else if it could be done easily Jewish Identification (JI) Esthetic Intellectual Ideological Affectional Devotional Fraternal Associational References Bock, Geoffrey E "The Jewish schooling of American Jews: a study of noncognitive educational effects." Unpublished Ph.D. dissertation, Harvard University. DellaPergola, Sergio 1977 "The French Jewish population study: progress report and evaluation of research problems." Pp in U. O. Schemlz, P. Glikson, and S. DellaPergola (eds.), Papers in Jewish Demography, Jerusalem: Institute of Contemporary Jewry, The Hebrew University of Jerusalem. Elazar, Daniel 1976 Community and Polity: The Organizational Dynamics of American Jewry. Philadelphia: Jewish Publication Society. Himmelfarb, Harold S "The impact of religious schooling: the effects of Jewish education upon adult religious involvement." Unpublished Ph.D. dissertation, University of Chicago "Measuring religious involvement." Social Forces 53:

14 WO HIMMELFARB, LOAR, AND MOTT Jaret, Charles 1977 "Residential mobility and local Jewish community organization in Chicago." Unpublished Ph.D. dissertation, University of Chicago. Kohs, Samuel C, and W. Blumenthal 1942 "Survey of recreational and cultural needs of the Jewish community." National Jewish Welfare Board, unpublished report, available in the files of the Research Service Bureau, Jewish Federation-Council, Los Angeles, CA). Lazerwitz, Bernard no "Some comments upon the distinctive Jewish names technique." Mimeo, date Bar-Ilan University, Israel (Dept. of Sociology) "The sample design of the national Jewish population survey." New York: Council of Jewish Federations and Welfare Funds "An estimate of a rare population group: fhe U.S. Jewish population." Demography 15 (August): "Jewish-Christian marriages and conversions." Jewish Social Studies 43 (Winter): Lazerwitz, Bernard, and Michael Harrison 1979 "American Jewish denominations: a social and religious profile." American Sociological Review 44: Massarik, Fred 1966 "New approaches to the study of the American Jew." The Jewish Journal of Sociology 8 (December): "National Jewish population study: a new U.S. estimate." American Jewish 75 Yearbook 75: Mayer, Egon 1979 From Suburb to Shtetl: The Jews of Boro Park. Philadelphia: Temple University Press. Ritterband, Paul, and Steven Martin Cohen 1979 "Study design to demographic study of New York's Jews." Xeroxed, Queens College, CUNY (Department of Sociology) "The social characteristics of the New York area Jewish community." Presented at the Annual Meeting of the Association for Jewish Studies, Boston (December). Rosenthal, Eric 1948 "The Jewish population of Chicago, Illinois." Unpublished Ph.D. dissertation, University of Chicago.