!"#$%&'()*+,)!"#(-&./()#0)1/2,34+5)6''45"*+/)7,*-/*/4#+)0"#')/8&)9::;)!#-2%*/4#+)<&+(2(= >"#')?&+&"*/4#+*%)<#'-*"4(#+()/#)/8&)!"#.&(()#0)@A&.#'4+5)7'&"4.*+@ 72/8#"B(C=)<8*"%&()D4"(.8'*+ 1#2".&=)6+/&"+*/4#+*%)E45"*/4#+)F&G4&HI)J#%K)LMI)N#K)OI)1-&.4*%)6((2&=)P8&)N&H)1&.#+, ?&+&"*/4#+)BQ4+/&"I)9::OCI)--K)R:;ST9U !2$%4(8&,)$3=)The Center for Migration Studies of New York, Inc. 1/*$%&)VFW=)http://www.jstor.org/stable/2547154 . 7..

&((&,=)L9X9LXL;9;);R=YU
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at . http://www.jstor.org/action/showPublisher?publisherCode=cmigrations. . Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission. JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

The Center for Migration Studies of New York, Inc. is collaborating with JSTOR to digitize, preserve and extend access to International Migration Review.

http://www.jstor.org

Problems Immigrant Population Generational Process Charles of Hirschman

and

Prospects Adaptation Census: Comparisons From

of from

Studying the ly90

to American"1

the

"Becoming

University of Washington With the loss of the question on parental birthplace in the 1980 and 1990 censuses, there are serious obstacles to current research on immigrant adaption based on the traditional logic of intergenerational progress. The tremendous immigrant streams from more diversity across contemporary than 40 country/region-of-origin groups, however, reinforces the singular of census data for national studies of the post-1965 immigrants importance and their children. A potentially useful research strategy is to examine variations in socioeconomic adaptation by duration of American residence who arrived as children or teenagers. among immigrants Exploratory investigation using this framework reveals a dominant pattern of successful adaptation with greater exposure to American society ("becoming Ameri? can"), but also some mixed patterns that are more consistent with the assimilation model. segmented the U.S. Bureau of the Census dropped the question on the of parents from the decennial census. Although a new question birthplace on ancestry was added, the loss of the parental birthplace question meant that it was no longer possible to directly identify the children of immigrants in census data from 1980 and 1990. Ironically, this loss of critical informa? with the new wave of immigration tion coincided from Latin America and Asia in the 1970s and 1980s. Tracking the progress (or lack of progress) of the children of new immigrants has become a more difficult task without census data on the topic. adequate The hope that survey data, including the Current Population Survey to collect data on the birthplaces which continues of respondents and their parents, might fill this data void is largely illusory. Even the largest national surveys do not contain a sufficient number of cases for any but the major or country-of-origin streams. The heterogeneity of regional immigrant of origin (e.g., Portugal, Laos, El contemporary immigration by country ^his paper was written while the author was a Fellow at the Center for Advanced Study in the Behavioral Sciences. I am grateful for the financial support from the National Science Foundation (SES-9022192) and the extraordinary quality of colleagueship at the Center. In a casual conversation, Frank Furstenberg offered a suggestion that inspired the direction of this paper. Thanks to you, Frank, and to the editor and the anonymous reviewers for their very helpful critiques of the first draft. In 1980, 690 IMR Vol. xxviii, No. 4

Problems

and Prospects

of Studying

Immigrant

Adaptation

691

in immigrant communities Salvador, Colombia, etc.) and the variations in the United States cast serious doubts on studies across places of settlement data. Local studies in particular cities and based upon highly aggregated can provide deeply textured accounts of immigrants and neighborhoods their children in specific settings, but only census data hold the possibility of specific immigrant streams across the country. of comparisons to measure the status of This article examines alternative methods with data from the 1990 Census of Popu? immigrants" "second-generation of that the cross-classification lation. The first method tests the assumption children of ethnic and race data with nativity can identify the native-born from Asia and Latin America. of this the new immigrants Applications method with 1990 census data are limited, however, because most of the of the new immigrants second generation were still children or adolescents in 1990. The second method is to examine the status of immigrants who In many important arrived in the United States as children. ways, these "child immigrants" are probably fairly close to their younger brothers and sisters who were born in the United States. This second method is illustrated with a comparison of the school enroll? ment rates of teenagers and the marriage patterns and the nonmarital fertility of young adults across more than 40 country-of-origin from the populations This preliminary tabulations of the 1990 Census of Population. published analysis shifts the analytical question from intergenerational change to the of "becoming American" (length of exposure to American society). impact there is support for assimilation Although theory that predicts social and economic gains with longer exposure to American society, there is also substan? tial variation by country of origin and the type of socioeconomic outcome.

ANALYSES IN CENSUS The

OF THE DATA

SECOND-GENERATION

HYPOTHESIS

is image of the United States as a "nation of immigrants" popular etched in American consciousness (Fuchs, 1990; Therndeeply political and their strom, 1980). This image has a corollary creed that immigrants descendants should have the same political and economic as prospects natives. Although these beliefs were not universally adhered to, especially in the treatment of racial minorities, the basic ideology has served to the gradual social and economic of successive waves absorption legitimate of immigrants over the last 200 years (Handlin, standard model 1973).The of immigrant is generally framed as an intergenerational progress process - the first - are 1964; Lieberson, (Gordon, 1980). Immigrants generation to achieve status and rarely expected handicapped by their newcomer socioeconomic parity with the native population. Learning a new language, of educational the lack of recognition from a different system, qualifications

692

International

Migration

Review

and hostility toward those with foreign accents and and native prejudice The second generation, cultures are major obstacles for immigrants. the are socialized and educated both in the mainstream children of immigrants, society from public schools and in their ethnic homes and neighborhoods. From the former, the children of immigrants gained the skills necessary for in the American structure, and from the equal participation occupational the immigrant belief that America was a land of latter, many internalized for those with perseverance and determination. opportunity there are still pockets of persistent and poverty, Although inequality overall there has been remarkable socioeconomic across immigrant gains Socioeconomic was reached by generations. parity with native Americans the children of European the middle decades of this during immigrants and Kraly, 1980; Hirschman century (Duncan and Duncan, 1968; Lieberson and Farley, 1985). The story of the struggles of 1990; Neidert 1988, and the success of the second generation is a persistent theme immigrants of the classic works of ethnicity in America (Gordon, 1964; Handlin, 1973; Dinnerstein and Reimers, 1982). This body of research on immigrant generations has been largely shaped by the content and availability of census data. From 1850 onward, a question on country of birth has been included in every decennial census. Data on foreign in every census from 1870 to 1970. These data were collected parentage allowed for the standard comparisons of immigrants, the second generation and everybody else ("native ("native born of foreign or mixed parentage"), born of native parentage"). These generational became the comparisons hallmark of the field and are exemplified by use of the same title, Immigrants and Their Children, for two classic census monographs, written after the 1920 census (Carpenter, 1927) and the 1950 census (Hutchinson, 1956). The generational was limited, however, because it was impos? perspective sible to trace the third generation of immigrants) in (the grandchildren census data. The reference population, "native born of native parentage," included third generation Americans as well as those whose immigrant ancestors arrived in the distant past. The new census data on ancestry but the loss of the question on parental information, provides additional makes it impossible to continue the tradition of research on birthplace based on 1991). It is ironic that scholarship immigrant generations (Farley, census data may reveal less about contemporary and their immigrants children than was possible for earlier waves of immigration. ORTHODOX ADAPTATION AND REVISIONIST THESES OF IMMIGRANT

As suggested in the previous section, the classical hypothesis of immigrant adaptation was of "Americanization" longer residence in the United States led to socioeconomic of differentials with the progress and the narrowing

Problems

and Prospects

of Studying

Immigrant

Adaptation

693

native-born this hypothesis was typically framed as a population. Although of intergenerational social mobility, the same logic can be applied process to the study of intragenerational some Indeed, immigrant adaptation. research has found that duration of residence in the United States leads to of earnings differentials the reduction between immigrants and the nativeborn American population (Chiswick, 1978; Jasso and Rosenzweig, 1990). The standard whether as a inter-or model, intragenerational process, to inform much of current scholarship continues on immigrant progress there has been a questioning in recent years. of this hypothesis although The doubts about the validity of the classical hypothesis are often phased in - the children of terms of declining prospects for the second generation post-1965 immigrants from Asia and Latin America (Reimers, 1992; Daniels, 1990). This revisionist thinking is largely based on the changing character of the American economy over the past two decades and the decline of opportu? nities for upward mobility among immigrants and their children. Poorly paid dead-end jobs in the service sector are not attractive options for the Americaneducated children of immigrants who entered the job market in the 1980s and 1990s (Gans, 1992). Another aspect of the problem is evident in the loss of a considerable share of jobs in the well-paid blue collar sector which was the bulwark of economic advancement for earlier generations of internal and international migrants to cities in the Northeast and Midwest. Another dimension of the revisionist perspective is the declining social in inner city areas where many new immigrants environment have settled. The claim is that contemporary social institutions and the cultural environ? ment place much less priority on social and economic mobility through educational success. The deterioration of public schools, the rise in drug use and violence, and the adversarial culture of many minority youth in inner cities all contribute to an environment with fewer incentives and for second-generation children to get ahead. opportunities immigrant to "American culture" in these conditions Assimilation may be a disadvan? relative to those who remain wedded to the traditional values of the tage and Zhou, 1993). (Portes first-generation immigrant community alternative cannot be conclusively with evaluated perspectives census data. The published tabulations from the 1990 census, an initial opportunity to explore however, provide empirical patterns To set the stage for this aligned with these rather different perspectives. and work, it is useful to begin with a survey of available data on immigrants ethnicity from the 1990 census. THE COMPOSITION IN 1990 OF IMMIGRANT COMMUNITIES These available

There were almost 20 million immigrants counted in the 1990 census. As a fraction of the American population, foreign born was less than 8 percent

694

International

Migration

Review

- a far lower of the total population than that recorded during percentage the early decades of the twentieth century in the peak years of immigration 1994:37). In absolute terms, however, the number (Passel and Edmonston, in 1990 is the highest ever recorded. Moreover, the current of immigrants is different from earlier streams because of its origins wave of immigration from Latin America and Asia. A summary portrait of the 1990 primarily in by country and/or region of origin is presented immigrant population Table 1. The table shows each immigrant nationality group by recency of and pre-1980) as a percentage arrival (1980-1990 of the total immigrant and as a percentage of the total U.S. population. population Of the 19.8 million immigrants in 1990, almost half (44%) arrived during the 1980s. The most dramatic shift in modern immigration (in addition to the rise in numbers) is the shift from Europe to Asia as a major source. Although this trend began during the late 1960s and 1970s, the basic comparison between pre-1980 and post-1980 arrivals shows the differences between con? of immigration. Of the 4 million Euro? temporary and historical composition counted in the 1990 census, over 80 percent had arrived pean immigrants prior to 1980. In contrast, more than half (56%) of the 5 million immigrants from Asia had arrived since 1980. For most Latin American country-of-origin categories, recent migration has been quite substantial, with about 50 percent arriving in the last ten years. For the countries in Central America where civil wars have spawned a refugee exodus (El Salvador and Nicaragua), upwards of 70 percent of immigrants are recent arrivals (since 1980). the country-of-origin vantage point from which to examine is as percentages of all immigrants for the recent past (1980composition 1990) and before (prior to 1980). Among the older pool of immigrants about 30 percent were from Europe, 20 percent from Asia, 43 (pre-1980), and smaller trickles from the former Soviet percent from the Americas, For the last ten years, almost 90 percent of Union, Africa and Oceania. have been from Asia, Latin America and Africa. Except for immigrants Mexico with a 25 percent share of recent immigrants, there are few countries with a dominant share. This tremendous for social range of diversity poses enormous challenges scientists who wish to track the social and economic of the new mobility waves. Even without direct measurement of the second genera? immigrant tion of new immigration, there are several analytical possibilities to explore within the confines of the 1990 census. The other

THE AND

CROSS-CLASSIFICATION NATIVITY

OF ASIAN

AMERICANS

One avenue for tracking the second generation of new immigrants census data is to cross-classify nativity status with other census variables

with that

Problems

and Prospects

of Studying

Immigrant

Adaptation

695

TABLE 1 Foreign-Born PopulationbyPlace of Birth andYearof Entry, United States: 1990 Percent of of Percent of Percent U.S.Population of Born ByYear Entry Country/Region Foreign Total PrePrePrePre(000) 1980-90 1980 1980-90 1980 Total1980-90 1980 Total1980-90 1980 56 100 100 100 7.9 3.5 4.5 Born 19,767 8,663 11,104 44 AllForeign 30 1.6 0.3 1.3 8 82 20 4,017 721 3,296 18 Europe 119 1 1 0.0 0.0 35 84 29 71 0 0.0 France 4 80 89 1 6 0.3 0.0 0.3 712 632 11 Germany 177 1 1 0.1 0.0 154 13 87 0 0.1 23 Greece 110 1 1 0.0 0.0 13 97 12 88 0 0.0 Hungary 170 1 33 81 0 1 0.1 0.0 0.1 Ireland 137 19 581 37 6 94 3 5 0.2 0.0 544 0 0.2 Italy 388 116 70 1 272 30 2 2 0.2 0.0 0.1 Poland 45 165 21 79 1 1 1 0.1 0.0 0.1 210 Portugal 154 486 24 76 3 UnitedKingdom 640 2 4 0.3 0.1 0.2 142 119 16 84 1 0.1 0.0 23 10 0.0 Yugoslavia 334 SovietUnion 132 60 0.1 202 40 2 2 0.1 0.1 2 44 1.1 0.9 Asia 25 32 20 2.0 4,979 2,794 2,185 56 119 102 17 86 14 1 0 0.0 0.0 1 0.0 Cambodia 530 284 46 3 246 54 3 0.1 China 2 0.2 0.1 147 65 1 56 1 1 0.1 0.0 82 44 0.0 Hong Kong 450 250 200 56 44 3 India 2 0.1 2 0.2 0.1 104 1 Iran 106 50 50 1 1 0.1 0.0 210 0.0 153 137 53 47 1 1 0.1 0.1 290 0.1 2 Japan 568 319 3 44 4 Korea 249 56 0.1 2 0.2 0.1 41 76 0.0 Laos 172 131 0 0.1 0.1 12 24 913 448 465 49 51 5 5 4 0.4 0.2 0.2 Philippines 1 160 84 66 34 Taiwan 244 1 0.1 0.1 0.0 2 107 60 44 1 1 Thailand 47 56 0 0.0 0.0 0.0 543 336 38 3 4 Vietnam 207 62 0.1 2 0.2 0.1 53 41 1.5 44 39 3.3 1.7 NorthAmerica 8,124 3,819 4,305 47 745 4 1 83 Canada 124 621 17 6 0.3 0.0 0.2 50 Mexico 19 1.7 0.9 0.9 22 25 4,298 2,145 2,153 50 10 60 9 Caribbean 10 0.8 0.3 0.5 1,938 783 1,155 40 737 188 549 26 4 74 5 0.3 0.1 Cuba 2 0.2 185 163 53 47 Dominican 1 0.1 0.1 0.1 2 2 Rep. 348 1 225 132 93 59 41 1 0.1 0.1 0.0 Haiti 2 334 154 180 46 54 0.1 2 2 2 0.1 0.1 Jamaica 47 69 41 1 1 59 1 0.0 0.0 0.0 Trinidad/Tobago 116 369 67 33 6 America 1,134 765 9 3 0.5 0.3 0.1 Central 465 350 115 75 ElSalvador 4 1 0.2 0.1 0.0 25 2 154 1 226 1 0.1 0.1 0.0 Guatemala 72 68 32 2 71 109 38 65 1 Honduras 35 1 0 0.0 0.0 0.0 169 125 44 74 1 1 0 0.1 0.1 0.0 26 Nicaragua 86 54 37 63 0 0 Panama 0 0.0 0.0 0.0 32 5 497 52 48 6 SouthAmerica 4 0.4 0.2 0.2 1,037 540 146 140 51 1 286 49 1 0.1 0.1 0.1 Colombia 2 143 81 43 57 1 1 62 1 0.1 0.0 0.0 Ecuador 73 48 60 1 1 121 40 0 0.0 0.0 0.0 Guyana 144 87 57 60 40 1 1 1 0.1 0.0 Peru 0.0 364 216 148 59 41 1 0.1 0.1 Africa 0.1 2 2 35 69 34 66 1 0 1 0.0 0.0 0.0 Oceania_104 U.S.Bureau the Census.1993a: of Source: Table1.

696

International

Migration

Review

Three census variables identify eth? identify country-of-origin populations. on country of origin: race, Hispanic nic groups with some information origin, and ancestry. This section illustrates the logic of this approach with from the race variable. data on the Asian and Pacific Islander population of this method for identifying The usefulness the second generation is of recent immigration in on the assumption dependent among persons most Asian "race" categories. The measurement of race in the census is a relic of an earlier era when both popular and scientific beliefs assumed that phenotypical variations in were associated with biological capacities. Although this human populations there remain important has long since disappeared, reasons to assumption of the concept. Data classified by race are necessary continue measurement to measure the degree of inequality and segregation that has persisted from In fact, there has been an the days of officially sanctioned discrimination. of "races" in recent censuses as more minority groups lobby for expansion inclusion as recognized and Cresce, 1993). In addi? categories (McKenney tion to the listed categories, in persons can write in additional responses with a sufficient number of cases the "other race" category. All responses are coded and included in the census data file and published tabulations. in the Asian and Pacific Islander population The list of subcategories from the responses to the race question in the 1990 census are listed in the far left column of Table 2. Not all categories mentioned are listed here, only the major ones that were tabulated in the census publication The Asians and Islanders in the United States, 1990 CP-3-5. The groups listed here were Pacific either categories on the census questionnaire or were mentioned as possible examples for filling in the "other race" category. The meaning of these race is not defined in the census beyond the self-identification of the categories who filled in the census questionnaire person (or the household respondent on behalf of other persons in the household). Country of origin from the birthplace question is not necessarily the same as the nationality that is claimed on the race question. For example, persons born in the People's Republic of China, Hong Kong and Taiwan are likely to identify as Chinese on the race question. In addition, there are a significant from (or are the descendants number of persons who immigrated of person who immigrated from) Vietnam, the Philippines, Thailand and Indonesia who identify themselves as Chinese. Similarly, the categories of Asian Indian and Pakistani may include peoples from several countries in South Asia or Southeast Asia with religion as a potentially important determinant of "racial" identity. The slippage in definitions is illustrated with the comparison of the number of immigrants from Asia (4.98 million, see Table 1) and the number of Asians (by race) who are foreign born (4.51 million, see Table 2). Regardless of the exact meaning in Table 2, the overlap categories of the specific Asian and Pacific Islander between Asian origins and immigration

TABLE2 AsianandPacificIslanderPopulation,byNativity, Yearof Entry,andAge,United St Percent TotalEthnic of M (000) _Population_ Population Native Total Population Bom by Population Race TotalU.S.Population (millions) 248,710 228,943 Non-Asian Pacific and Islander 241,483 226,274 AsianandPacific Islander 2,668 7,227 Asian 6,876 2,363 506 Chinese 1,649 506 1,420 Filipino 585 866 Japanese 193 786 AsianIndian 797 218 Korean 119 693 Vietnamese 31 149 Cambodian 94 32 Hmong 147 31 Laotian 91 22 Thai 30 5 Indonesian 19 82 Pakistani 351 305 Pacific Islander 284 256 Polynesian 206 203 Hawaiian 45 58 Samoan 7 17 Tongan 45 55 Micronesian 48 42 Guamanian Melanesian _2_ _I_ Table1. Source: Bureau the Census, U.S. of 1993b,

Bom Born Native Native Foreign Foreign 1980-90_Pre-1980 Born 1980-90_Pre-1980 Born 1 4 32.5 3 92 8,664 11,104 4 32.7 3 94 6,042 9,167 15.6 27 37 37 2,662 1,937 14.7 28 38 34 2,597 1,916 16.3 30 39 31 493 649 33 14.3 36 32 448 465 15 35.3 18 68 153 127 8.8 32 44 25 346 248 9.0 41 32 27 327 252 6.7 26 17 42 181 293 9 4.7 70 21 14 104 16 5.2 50 34 47 15 5.4 16 63 21 93 24 11.7 44 32 24 40 29 30 12.0 17 53 9 16 6.7 26 51 23 21 42 6 23.5 7 87 21 25 5 23.6 5 90 13 15 0 26.1 0 99 1 1 14 17.2 9 78 8 5 7.4 29 41 29 5 5 5 23.9 13 3 7 4 24.9 8 4 2 6.6 _22_ _2Q_ _4_ _2_

698

International

Migration

Review

some information on immigrant close enough to provide is probably in the 1990 census report, CP-3-5, The published tabulations generations. characteristics of the Asian and show a wide variety of social and economic status. Table 2 presents the absolute Pacific Islander populations by nativity the percentage numbers (in thousands), by nativity status composition born, foreign born who arrived after 1980, and foreign born who (native arrived prior to 1980), the median age of these three groups, and the absolute numbers in the age group 15-24. Although most native-born Asian are probably the children of immigrants, all but a few were still Americans too young in 1990 to provide a base for research on the socioeconomic of the second There are, however, a significant generation. adaptation number of immigrants who arrived as children and are probably close for the second generation. proxies - almost 3 By "race," there were 7.2 million Asian and Pacific Islanders of the total U.S. population - counted in the 1990 census. Aside from percent the fact that Asians and the Pacific Islander populations share homelands across of the two groups the Pacific Ocean from the U.S. mainland, the combination of Pacific Islanders are makes very little sense. The distinctive characteristics obscured in the composite category in which they are less than 5 percent. most Pacific Islanders are not immigrants, but the indigenous Moreover, of American territories. The largest group, Hawaiians (206,000 of the peoples 351,000 Pacific Islanders), are largely resident in the state of Hawaii. Under the Asian category, there is tremendous diversity: 13 populations are listed separately in the here, but many smaller groups are included overall category. There were six major Asian populations with more than a half million persons in the United States in 1990: Chinese (1.6 million), Asian Indian (.8 million), (1.4 million), (.9 million), Filipino Japanese Korean (.8 million) and Vietnamese (.7 million). Within the Pacific Islander into the origin categories of Polynesian, Micronesian, category (subdivided and Melanesian), there is also considerable diversity. The major value of Table 2 lies in the comparison of the numbers of and foreign-born native-born persons of Asian and Pacific Islander origin. Almost all of the Pacific Islander population are native born: 99 percent of 78 percent of the Samoans, and 88 percent of the Guama? the Hawaiians, nians. These groups are classified as native born because they are indigenous peoples of American territories. Within the Asian American population, only about one-third of the total is native born. The Japanese-American population contains a native-born majority, but among the other groups, the percentage of native born ranges from a high of 36 percent for Filipino Americans to a low of 17 percent for those of Vietnamese and Indonesian origins. What can census data tell us about the second-generation Asian Ameri? cans? First, it seems that almost all native-born Asian Americans, with the of the Japanese-American and (and some Chinese exception population

Problems

and Prospects

of Studying

Immigrant

Adaptation

699

are the children of immigrants.2 This is a simple Filipino Americans), - there has been too little time product of the recency of Asian immigration to be formed. Some indirect evidence for this claim for a third generation share of immigration can be read from the substantial during the 1980s is given in (relative to the period prior to 1980). Much stronger evidence the third panel of Table 2, which shows the median age of each group by with half of the nativity status. The median age is the fiftieth percentile, younger than the median age. population The foreign-born Asian populations have "normal" age structures with median ages in the 20s and 30s. In contrast, most of the native-born Asian are "child populations" with very young median ages - often populations below 10. For these populations, we can safely assume that almost all of the native born are the children of immigrants (and still children at the time of the census). The native-born is, of course, the major Japanese population and there are probably nonnegligible fractions of the Chinese exception, and Filipino populations that are third generation or higher. The conclusion that most native-born are second gen? Asian Americans eration in 1990 does not mean that census data can be used for intergen? erational comparisons on most social and economic variables. The relative - which allows us to youth of the native-born samples safely assume their status - also means that relatively few have reached the young generational adult years where completed schooling, family formation and employment can be analyzed. The underlying problem is that there are relatively small of native-born numbers adults in many Asian-American to populations of second-generation The census in the year analyze processes adaptation. 2000 will contain much larger samples of older native-born Asian Ameri? cans, but there will also be less certainty that all are second generation. To illustrate the potential for research on second-generation Asian Ameri? cans, the last two columns in Table 2 show the actual numbers (in thousands) of the native born of each nationality group from age 15 to 24 with the in the same age group. The age group for immigrants comparable figure 15-24 embodies the critical years of the school to work transition and of the life years of early family formation potential important segments course to monitor. Any study of the second-generation hypothesis among Asian Americans will have to contend with the very small numbers of native born for most specific nationality groups at this stage of the life course. The largest numbers of the native born for most groups are actually in the 0-4 - too and 5-9 age categories One young to show much differentiation. 2There is a substantial literature on the descendants of the Japanese immigrants who immi? grated to the U.S. mainland and Hawaii around the turn of the century (Kitano, 1976; Montero, 1980; Bonacich and Modell, 1980). Most Japanese Americans are third- or fourthgeneration Americans. There is an even earlier history of large-scale Chinese immigration in the late nineteenth centuiy, but the large influx of contemporary Chinese immigration has tipped the balance toward a majority foreign-born population.

700

International

Migration

Review

but analytical possibility is to aggregate across country-of-origin categories, this strategy may sacrifice a potentially variable. important explanatory Another possibility is to focus on the significant number of Asian-Ameri? who are "almost native born" - persons who immigrated can immigrants as children. The last column shows the number of immigrants age 15-24 in 1990. This sample is typically larger than the native born for the same age On many dimen? group (except for the Japanese-American population). who arrive as children even as teenagers) are sions, the immigrants (or closer to their American-born than they are to their probably siblings can be parents who arrived as adults. Analyses based on this population used to explore the impact of "Americanization" on (duration of residence) adaptation processes.

THE AND

CROSS-CLASSIFICATION NATIVITY

OF HISPANIC

AMERICANS

The same exercise shown for Asian Americans in Table 2 is repeated in Table 3 for the 1990 census population of Hispanic Americans. The 1990 census question on Hispanic origin read: "Is this person of Spanish/Hispanic origin?" If the response was yes, the listed choices included: Mexican/Mexican AmeriPuerto Rican, Cuban, and Other Spanish/Hispanic. For the can/Chicano, "other" category, respondents were encouraged to write in specific nationality and several examples were provided as possible answers. Table 3, based groups, on the published tabulations in the 1990 census report, Persons of Hispanic Origin in the United States, CP-3-3, includes the three listed groups (Mexicans, Puerto Ricans and Cubans), one Caribbean population six Central (Dominicans), American populations, nine South American populations, and three general categories of Spaniard, Spanish and Spanish American. With the exclusion of the non-Hispanic groups from the Caribbean and South America, the list of national-origin in Table 3 is somewhat populations different from the country-of-birth of the foreign-born categories population in Table 1. Moreover, the number of foreign-born with Latin persons American heritage (Hispanic identity) from a specific country (Table 3) may differ from the number of persons reported to have been born in the same differ because (Table 1). These numbers country country of birth and national heritage are not always the same and because non-Hispanics from Latin America are excluded from Table 3. For example, 86,000 are reported to be born in Panama by foreign birth (Table 1), but only 62,000 persons of Hispanic origin from Panama are foreign born (Table 3). Analysts need to be very cautious and aware of these nuances of census measurement. The in census data may be somewhat arbitrary, but they reflect the distinctions complex conceptual problems of measuring ethnicity.

TABLE 3 Hispanic PopulationbyNativity,Yearof Entry,andAge,United States; 19 of Percent TotalEthnic (in _Population Population millions) Born Born Native Native Foreign Total Foreign NativeBorn 1980-90 Pre-1980 Bom 1980-90_Pre-1980_Rom by Origin Population Population Hispanic 32.5 92 TotalU.S.Population 11,104 8,664 248,710 228,943 33.3 95 2 7,237 4,689 226,810 214,884 Origin Non-Hispanic 18 19.6 18 64 NorthAmerican 3,867 14,058 3,975 21,900 17 17 18.3 67 Mexican 2,212 2,248 13,393 8,933 1 1 25.4 99 17 16 Puerto Rican 2,619 2,652 53 15.9 19 559 28 195 299 Cuban 1,053 18.9 32 22 46 OtherHispanic 1,078 1,516 4,802 2,207 33 9.6 38 171 29 196 153 520 Dominican 55 24 8.4 21 316 730 278 Central American 1,324 40 14.3 30 32 17 23 57 18 CostaRican 55 25 8.2 67 20 149 53 269 Guatemalan 9.6 25 52 33 23 68 131 30 Honduran 10.4 60 21 19 42 38 122 203 Nicaraguan 39 16.3 33 28 36 26 30 92 Panamanian 61 20 6.0 19 113 346 565 106 Salvadoran 37 38 12.3 379 25 397 260 SouthAmerican 1,036 47 31 16.9 47 23 31 101 23 Argentinean 47 29 12.2 11 24 9 18 38 Bolivian 11.4 30 42 26 29 18 69 21 Chilean 11.6 36 39 136 26 146 379 98 Colombian 42 12.5 32 80 26 191 50 62 Ecuadorian 6.4 43 29 29 3 2 7 2 Paraguayan 11.5 47 29 51 23 41 175 Peruvian 45 13.9 36 18 10 4 22 Uruguayan 13.8 50 23 11 27 13 48 24 Venezuelan 6 12 28.6 60 30 519 429 Spaniard 9 7 27.4 84 39 30 445 376 Spanish 26.8 92 American _L_ _22_ Spanish of U.S. Bureau the Census,1993c.Table1. Source:

702

International

Migration

Review

in 1990, the Mexican-origin Americans Of the 22 million Hispanic over 13 million (over 60%), with the other major comprises population groups being: Puerto Ricans (2.7 million), Cubans (1 million), Dominicans (1 million), (1.3 million), South Americans (.5 million), Central Americans and the generic categories of Spaniard, Spanish and Spanish American (1.1 is much more likely to be million). Overall, the Hispanic-origin population native born than the Asian and Pacific Islander population (64% to 34%), of Hispanic variation across the spectrum but there is enormous popula? tions. Almost all (99%) of the Puerto Rican population is classified as native are classified as native born; all those born on the island or the mainland of Spaniards born. The birthplace (83% native born) is very composition similar to those who label themselves (84% native born) and Spanish largely repre? Spanish American (99% native born). These three categories Southwest that predate sent persons with historical roots in the American into the United States in the mid-nineteenth the region's incorporation century. two-thirds of the Mexican-origin are native born, Although population is so large that the foreign-born the overall population Mexican population (4.5 million) swamps the size of all other groups in Table 3. The number of Mexican Americans is almost three-fifths of all foreign-born foreign-born and over one-fifth of all immigrants. All other Hispanic popula? Hispanics tions have only a minority (20-30% in most cases) native born. All of these from recent immigration groups gained substantially during the 1980s, especially those from Central America and some groups from South Amer? ica. How well does the native-born Hispanic population represent the second It depends on the group. Almost all of the Spaniard/Span? generation? and the majority of the Mexican-Ameri? ish/Spanish-American populations have been in the United States for many generations, can population and inferences about generational from comparisons between the native change born and the foreign born would not be warranted. For the Puerto Rican data on nativity really do not speak to the question of birthplace population, in the island or the mainland. For most of the other Hispanic groups from the Caribbean Central America and South America, the (Dominicans), of the native born are probably second-generation Americans. This majority seems evident from the significant share of recent (post-1980) immigration and the median age of the native-born of each group (less than segment age 15 in most cases and less than age 10 for several groups). in Table 3 show the numbers of native-born and 15 and 24. As with Table 2, these between foreign-born Hispanics age figures are meant to illustrate the potential samples for an analysis of early socioeconomic attainment and family formation. Except for the groups for last two columns The

Problems

and Prospects

of Studying

Immigrant

Adaptation

703

the second generation which the native born do not represent (Mexicans, the numbers of Puerto Ricans and Spaniards/Spanish/Spanish Americans), native born are very small (because the native-born are few and populations also very young). In general, the numbers of young immigrants (many of whom arrived as children) represent much more sizable samples for analysis of the impact of exposure to American society on immigrant adaptation. THE CONCEPT AMERICAN" AND MEASUREMENT OF "BECOMING

and the socioeconomic The concepts of immigrant life course generations The standard image of immigrants have not been well integrated. (and the in most immigration conventional research) is of young adults assumption in their country of origin. The parallel who were socialized and educated is that they are the first members of image of the children of immigrants in American their community to be educated and socialized institutions. The problem with this logic is that a significant fraction of historical and has been of families who brought children with contemporary immigration - often labeled the 1.5 - were them. These child immigrants generation but they were largely educated classified as immigrants, and socialized in the United States. Because census tabulations sufficient rarely provided on year of immigration, information most analyses relied on the standard An unexplored is whether countrygenerational hypothesis comparisons. differences stock populations of-origin among foreign may reflect the age of immigrant streams. composition The critical variable of timing is not simply the duration of residence in the United States, but also the age at arrival. The role and potential of the home, neighborhood, influence schools, peers, and the mass media are sharply differentiated the ages of children and teenagers. by Immigrants who learned English in their teenage years will often retain an accent, while their younger who siblings will speak like "natives." Youthful immigrants in American are entirely educated schools may face fewer problems of than those who enter at older ages, but the youngest immigrants adjustment to peer pressures that are at odds with may also be more susceptible influences from the home. The measurement of age at arrival and years of in the United States provide benchmarks duration to test the impact of American" or exposure to American "becoming society. This framework does not replace the idea of immigrant but enriches it to take generations, into account the variation in age composition of immigrants. A recently released 1990 census report, The Foreign-Born Population in the United States, 1990 CP-3-1, tabulates a variety of social characteristics of States (before 1980, immigrants by the year of arrival in the United The cross-classification 1980-1981, 1982-1984, 1985-1986, 1987-1990).

704

International

Migration

Review

of current age (in 1990) and year of arrival also identifies age at immigra? tion. Among the important socioeconomic variables in these tabulations are at the older school enrollment and children ever born. School enrollment and 18-19) is a critical indicator for graduation from teenage years (15-17 school and the transition to college. The data allow for the comparison high of current enrollment rates of teenage who arrived ten (or immigrants more) years earlier (almost native born) with those who arrived less than three years before the census. A comparable the marital status analysis can be conducted comparing and early fertility of women immigrants, age 15-24, in 1990. The data allow for the measurement of nonmarital fertility, an indicator of nonconformity This variable is particularly important to middle class aspirations. to exam? assimilation ine in light of the segmented thesis that greater exposure (and culture among some immigrant to American assimilation) groups may be attainment. associated with mixed prospects for socioeconomic The classical assimilation model would imply that teenage immigrants who have been in the United States longer are more likely to be enrolled in high school and to have lower nonmarital fertility than those who just arrived. The assimilation model questions this expectation. If greater exposure segmented to American society has primarily been to inner city environments that are not conducive for upward social mobility (educationally or economically), then children with longer U.S. residence (and a younger age at arrival) immigrant may not be doing better than recent arrivals. A satisfactory test of these competing would require data on hypotheses the socioeconomic status of immigrant families and the character of their residential locations. Some aspects of these data are available in the Public Use Microdata Sample (PUMS) census files. In this preliminary investiga? between national origins, duration of residence, and tion, the associations in the United States are explored. adaptation Descriptive patterns across more than 40 country-of-origin should provide immigrant populations some important initial insights on the process. PATTERNS OF EDUCATIONAL

ENROLLMENT

Table 4 shows the percentage of immigrant teenagers enrolled in school for two age groups (15-17 and 18-19) by year of immigration (five categories from pre-1980 to 1987-1990). The younger reflects age group, 15-17, - the likelihood in high school (or its complement enrollment of dropping out before graduation). at the older age group, 18-19, Higher enrollments are a mixture of both high school and college and should attendance, to post-secondary provide a fairly good indicator of continuation schooling. All country/region-of-origin groups are included in Table 4. In most cases, the numbers of teenagers in each age group by year of entry and country

Problems

and Prospects

of Studying

Immigrant

Adaptation

705

in the hundreds or thousands, but in a few cells the of origin numbered of observations below 100. Consequently, attention number is dropped focused on the broad patterns, not on individual cells.3 is one of higher educational The overall pattern for all immigrants ratios with duration of residence enrollment (or younger age at arrival) in for the most the United States. The time series shows a modest disadvantage recent arrivals - those who arrived in the last five years - but little variation For the high school age group (15-17), among those with longer residence. hover around 90 percent, dropping to the 80 to 70 percent enrollments for recent arrivals in some groups. For the sample of older teenagers range were (age 18-19), about 70 percent of earlier (1984 or before) immigrants enrolled, but only about 50 percent of the recent immigrants. A close examination of the country/region reveals two dis? populations The first is high and constant tinct patterns. of enrollment, regardless in the United States. The second pattern is an initial duration of residence - indicated - for the most recent arrivals handicap by lower enrollment followed by higher enrollment for those with five or more years of residence. The first pattern of consistent high enrollments (90% and higher) is evident for the younger age group (age 15-17) among immigrants from almost all Asian countries, some European countries, the former Soviet Union, several Caribbean and South American countries, Canada, and Africa. The same lower rates) is evident for a smaller number of (at somewhat pattern in the older age group for many, but not all, of the immigrant populations For example, same countries. among African immigrants, high school rates are always above 90 percent and college enrollment enrollment rates (age 18-19) are always 80 percent or higher. The pattern of rising enrollment with longer residence in the United States is exemplified for whom high school by the case of Mexican immigrants enrollment (age 15-17) rose from 52 percent for recent arrivals to 88 percent for those who arrived prior to 1980 (college-age enrollments rose from 22% to 57% across the same categories). This pattern, which is evident for many other countries, especially for immigrants from Central America, provides support for the thesis of immigrant adaptation. Time in the system and/or socialization at a younger age leads to increasing opportunities or incentives for continued For the older group (age 18-19), the most recent arrivals are less enrollment. likely to be enrolled than those who arrived some years earlier. Perhaps those who arrived at younger ages were more likely to have graduated from a U.S. high school and were able to continue to college. There are also some mixed patterns that are consistent with the possibility to American culture leads to poorer prospects that greater exposure for 3The published tabulations did not permit the computation of enrollment ratios separately for those age 15-17 and those age 18-19. I am very grateful to Dr. Jorge del Pinal, Chief of the Ethnic and Hispanic Statistics Branch of the Bureau of the Census, for his kind assistance in producing this table from the summary file SSTF1.

706

International

Migration

Review

TABLE4 Percent Enrolled in School of Foreign-Born Youth, byAgeandYearof Entry, United States; 1990 Persons 15-17 in 1990 Persons 18-19 in 1990 Age Age of Entry_Year of Entry_ _Year PrePre1985-86 1982-84 1980-81 1980 1987-90 1985-86 1982-84 1980-81 1980 _1987-90 87 75 91 91 48 57 70 AllForeign 71 Born 92 72 94 95 88 73 81 92 95 74 77 68 Europe 94 83 100 81 84 100 90 88 67 France 73 96 93 91 95 84 67 77 86 61 72 Germany 81 94 86 74 100 83 81 Greece 82 82 68 97 75 94 100 65 87 88 82 90 52 Hungary 90 87 100 81 77 39 61 69 39 Ireland 66 84 87 81 67 100 55 57 92 76 61 Italy 88 98 97 84 71 97 80 80 Poland 83 70 83 87 90 89 64 82 48 64 54 62 Portugal 97 94 97 98 79 UnitedKingdom 91 77 67 82 78 79 96 97 96 100 55 91 36 68 68 Yugoslavia 86 100 93 63 97 77 84 84 83 SovietUnion 92 91 95 95 96 96 79 86 85 86 Asia 88 94 74 91 91 95 79 83 92 84 83 Cambodia 93 93 94 96 95 94 86 90 93 China 92 96 98 96 98 95 88 98 95 92 94 Hong Kong 95 95 97 90 98 88 India 92 82 94 92 97 91 97 96 97 84 Iran 89 80 87 82 95 95 100 95 96 85 92 96 92 76 Japan 93 95 97 96 86 87 89 Korea 92 88 90 95 89 91 94 95 68 80 80 Laos 76 80 89 95 95 95 95 56 79 80 85 82 Philippines 95 98 96 97 96 90 94 Taiwan 99 92 95 93 91 97 99 98 89 79 Thailand 92 79 72 91 96 97 77 96 96 88 83 Vietnam 89 91 63 80 87 88 31 89 56 NorthAmerica 59 42 61 91 97 96 95 Canada 97 86 86 79 83 76 70 81 52 86 30 43 53 Mexico 22 57 88 Caribbean 92 67 74 92 92 69 73 72 83 90 91 75 67 75 Cuba 66 67 92 67 94 88 92 93 57 69 Dominican 65 61 62 Rep. 85 91 93 94 89 Haiti 81 83 74 92 77 82 93 93 89 95 70 75 92 76 73 76 Jamaica 89 90 87 63 92 56 85 53 80 Trinidad/Tobago 91 89 90 America 74 44 58 65 Central 92 92 65 62 86 91 89 37 El Salvador 91 51 61 60 66 88 38 51 58 Guatemala 92 92 92 48 67 90 94 91 48 60 60 Honduras 92 72 55 93 94 57 95 93 74 87 67 72 Nicaragua 100 79 100 83 69 80 76 68 Panama 61 94 91 94 94 64 74 75 77 SoudiAmerica 72 96 94 91 65 92 67 66 Colombia 72 72 85 94 58 73 94 66 78 Ecuador 92 74 90 89 94 59 93 75 74 78 79 Guyana 93 96 98 96 66 64 Peru 69 82 82 94 95 94 83 83 Africa 92 84 80 87 93 97 96 96 74 63 61 59 73 Oceania_ The in Source: SSTF1: Foreign-Born Population die UnitedStates,1990.

Problems

and Prospects

of Studying

Immigrant

Adaptation

707

immi? mobility. For example, younger (age 15-17) Caribbean do not always have higher Dominicans, Haitians, etc.) grants (Cubans, enrollments educational among the groups that have been here the longest. are more likely This is an important finding because Caribbean immigrants with an African-American to be black and to live in neighborhoods majority. Note that this pattern does not hold for the older (age 18-19) Caribbean The other groups with inconsistent immigrants. patterns of educational enrollment are some European at the immigrant groups, particularly The interpretation would be that greater exposure to college-age range. has dulled the ambitions American for higher education which society remain strongest among the newest immigrants. educational There are many threats to all of the preceding The census interpretations. is inclusive of all residents whether temporary enumeration or permanent and whatever their reasons for being in the United States. For some nationality groups, including many from Asia, foreign students are included with permanent For other groups, temporary laborers (per? immigrants. and temporary are included (Central America) refugees haps Mexicans) with immigrants. census data allow for differentiation of citizens Although there is no measure of visa status. Moreover, cross-sectional and noncitizen, data do not permit independent of selectivity of immigration assessment Small cell sizes, during particular years of arrival and duration of residence. especially for many of the European groups, also reinforces caution. In spite of these measurement problems, my reading of Table 4 suggests that there is modest evidence for the thesis of (but not overwhelming) to American immigrant Longer exposure adaptation. society (or younger for age at arrival) is generally a positive factor in obtaining more education Some groups, especially those from Asia and Africa, have very teenagers. enrollment Caribbean high educational regardless of duration of residence. resemble Asians for recent arrivals (very high enrollment ratios), but groups also show patterns of lowered education for those with longest Ameri? they can residence, as do some Central American in the older age immigrants range.

EARLY

MARRIAGE

AND

NONMARITAL

FERTILITY

Table 5 shows two panels with 1990 census data on the percent married of women, young women, age 15-24, and mean fertility among unmarried 15-24, for all immigrant populations (the same list as in Table 4) by year age of entry into the United States. The question this analysis is the guiding of additional of American residence on "adaptation" years impact during the formative and young adulthood, but the expecta? ages of adolescence tions are a bit different than in the previous table. A younger age at is considered to be "traditional" behavior, and greater exposure marriage

708

International

Migration

Review

On the other to American society would lead to marital postponement. to American hand, nonmarital fertility would be a sign of acculturation social mobility. culture of youth that are least likely to experience upward the cumulative Nonmarital fertility is measured by subtracting fertility women, age 15-24, from the cumula? (children ever born) of ever-married tive fertility of all women, age 15-24. The nonmarital fertility rate (mean nonmarital CEB) is computed by dividing the cumulative fertility of nevermarried women, age 15-24, by the number of never-married women, age 15-24, in 1990. For the total foreign-born there is modest support for the population, that greater exposure to American society (earlier immigration) hypothesis About one-third of young women leads to a greater marital postponement. who are recent immigrants (after 1985) are married, but only about onefourth of young women who immigrated earlier were married in 1990. It is law, not differential possible that selectivity based on immigration exposure to American (or socialization) society, explains this pattern. For example, to an American resident) may be correlated with being married (especially for admission to the United States for young immigrant women eligibility in their late teens or early twenties. For young women of the same age who arrived in the United States as teenagers or earlier, the only link between the timing of immigration and the timing of marriage would be the from that of the country of origin to the United of acculturation hypothesis in the relationship States. Variations across national-origin populations suggest, however, that there is more than selectivity at work. The hypothesized (more U.S. exposure leads to more mari? relationship tal postponement) is evident for immigrants from some, but not all, coun? tries in Europe, Latin America and Asia. In particular, the association is for immigrants from India, the Philippines, Cambo? strong and consistent dia, Mexico, Peru and Africa. For many other countries, Vietnam, Thailand and Taiwan in Asia and for most of the Caribbean and Central American there is no clear relationship. countries, in levels of age at marriage are also major differences across In general, countries. from Asia, some Caribbean sending immigrants countries (Haiti and Jamaica) and Africa seem to be much more likely to than immigrants from Europe, Mexico and Central postpone marriage America. Whether these national variations as class might be explained as opposed to some cultural preferences, cannot be addressed differences, within the limits of published data. data on nonmarital fertility also reveal more variation by national than by duration of residence in the United States. In general, origin nonmarital from Europe and Asia fertility is much lower for immigrants than for immigrants from Mexico, the Caribbean and Central America. Levels of nonmarital from South America and Africa fertility for immigrants The There

Problems

and Prospects

of Studying

Immigrant

Adaptation

709

TABLE5 and Percent Married Non-MaritalFertility of Foreign-Born Women, 15-24, byYearof Entry,United States: 1990_ _Age MeanCEBper Non-Married Percent Married Women, Woman, Age 15-24 Age 15-24 Year Entry of Year Entry of PrePre1987-90 1985-86 1982-84 1980-81 1980 1987-90 1985-86 1982-84 1980-81 1980 0.18 33 35 0.15 0.15 0.12 AllForeign 26 24 0.12 28 Born 0.03 33 17 27 0.02 0.04 0.05 0.05 40 22 Europe 0.00 0.00 17 35 11 20 28 0.04 0.10 0.10 France 39 51 58 0.05 0.06 0.08 0.08 0.10 23 42 Germany 0.00 0.11 0.00 0.00 0.02 39 59 59 24 26 Greece 38 19 0.00 0.00 0.00 47 6 32 0.00 0.08 Hungary 15 0.03 20 0.01 0.06 0.00 0.04 Ireland 20 27 25 31 0.05 40 22 26 0.02 0.05 0.00 0.04 20 Italy 17 31 36 23 0.03 0.01 0.03 0.03 0.02 Poland 20 0.09 41 47 26 28 0.01 0.03 0.03 0.04 25 Portugal 35 10 0.05 0.07 0.06 UnitedKingdom 39 15 21 0.02 0.05 37 0.05 0.00 55 66 41 0.00 0.00 0:03 27 Yugoslavia 0.01 0.00 16 22 22 0.00 0.01 0.02 SovietUnion 26 20 14 17 16 0.03 0.05 Asia 0.05 0.06 0.03 27 21 18 11 0.14 0.14 0.07 Cambodia 25 23 0.25 0.24 21 7 7 16 0.01 0.04 0.01 0.01 0.03 China 26 12 10 9 6 0.00 0.00 0.03 10 9 0.03 0.00 Hong Kong India 49 31 17 0.02 0.01 0.01 0.01 12 0.02 28 30 10 15 Iran 17 0.02 0.05 0.00 0.01 21 0.02 10 10 17 0.01 0.00 0.01 20 0.06 0.10 Japan 9 9 0.01 0.01 16 0.01 Korea 28 0.02 0.02 30 30 0.14 50 0.16 0.19 Laos 29 0.21 0.08 29 18 16 0.05 0.05 0.05 0.05 0.06 28 24 20 Philippines 6 6 0.00 11 5 0.01 0.01 0.01 Taiwan 4 0.02 16 13 0.01 14 26 0.00 0.04 0.07 0.07 Thailand 32 11 13 13 0.04 Vietnam 15 12 0.12 0.06 0.04 0.04 37 41 37 30 32 0.27 NorthAmerica 0.20 0.28 0.24 0.21 17 0.03 19 22 0.06 0.04 0.03 0.04 Canada 29 29 39 44 50 0.19 0.31 0.33 48 32 Mexico 0.28 0.22 19 18 20 0.15 0.19 0.19 0.16 0.21 Caribbean 26 22 45 0.04 0.03 Cuba 28 0.04 0.08 0.08 24 29 27 30 0.17 21 32 0.38 0.32 Dominican 23 0.22 0.24 Rep. 29 Haiti 18 16 16 0.09 0.15 12 18 0.22 0.22 0.15 14 10 10 13 0.19 12 0.19 0.17 0.24 0.20 Jamaica 0.14 18 0.30 0.17 0.21 27 25 25 0.16 Trinidad/Tobago 19 33 America 34 27 28 0.35 Central 27 0.28 0.29 0.26 0.19 30 30 0.33 0.41 ElSalvador 36 34 0.34 24 0.29 0.22 31 0.33 0.43 36 23 29 Guatemala 32 0.27 0.27 0.23 36 0.30 0.33 Honduras 31 30 26 29 0.29 0.21 0.24 21 27 0.16 0.13 0.10 28 27 0.21 0.21 32 Nicaragua 35 15 0.08 Panama 27 0.15 0.10 0.21 27 27 0.28 30 0.08 0.09 22 21 0.10 0.10 SouthAmerica 32 27 0.12 34 35 28 0.12 0.17 0.11 Colombia 27 24 0.24 0.12 0.11 33 34 28 0.09 0.11 34 0.08 0.15 Ecuador 28 16 14 0.07 19 18 0.08 0.15 0.09 0.14 21 Guyana 0.07 34 30 17 0.07 Peru 23 22 0.06 0.13 0.09 14 15 11 0.07 0.07 0.05 Africa 28 22 0.07 0.08 0.03 0.31 34 26 24 0.04 0.04 0.09 28 32 Oceania_ Source: Bureau the Census,1990a.Table U.S. of 1.

710

International

Migration

Review

to these regional patterns, however. There are exceptions are intermediate. from Asia (Cambodia, Laos and, to a lesser extent, The refugee populations had higher levels of nonmarital to Vietnam) fertility that are comparable the Caribbean pattern, but not as high as Central American levels. for the alternative thesis that greater Again, there is some support lead to higher rates of nonmarital to American society might exposure The patterns among the Caribbean origin populations. fertility, especially across year-of-arrival are, however, not monotonic groupings. Although the evidence is a bit shaky, nonmarital fertility appears to be a bit lower among the longer duration categories (younger age at arrival) within the Central American populations and the Southeast Asian refugee populations. Might this be a sign of immigrant to the "successful path" adaptation adjustment for upward mobility for immigrants?

CONCLUSIONS The incorporation of the new wave of post-1965 and their immigrants children is one of the most important challenges facing American society. The importance of the topic is reflected in the great debates over immigra? tion policy and the several national commissions over the last two decades that have been charged with reviewing the progress of immigrants and their and Passel, impact on American society (Briggs, 1984; Bean, Edmonston of Labor, 1989). 1990; U.S. Department What has not happened, however, has been any comparable national effort to collect new data that might allow for serious scientific research on the topic. Collection of immigration and emigration statistics is so poor as to be a national scandal (Levine, Hill and Warren, 1985). There have not been innovative efforts to collect survey data on a scale comparable to those launched to study other national issues such as aging and retirement, youth employment, crime or drug use. This leaves the census as the major, and often the victimization, only, source of data for the study of contemporary immigration. sent 100,000 or more During the 1980s, more than twenty countries of immigrant and immigrants to the United States. The size and composition in Los Angeles, Miami and New York are very different ethnic communities from each other and from almost every other metropolitan area in the country. With such great diversity and weak data resources, about the knowledge economic and cultural mobility of the post-1965 wave of immigrants social, often seems to rest more on impressionistic information than on solid evidence. For immigration researchers and policymakers, the census ap? season. The pears like the tired old horse to farmers at spring plowing census, like the work horse, will do the basics, but it could be done much better if modern machinery were available for the task.

Problems

and Prospects

of Studying

Immigrant

Adaptation

711

of studying immigrant The classical model (and method) adaption by from census data was dealt a mortal blow comparisons intergenerational with the loss of the census question on the nativity of parents in the 1980 This article has explored and 1990 censuses. for using other prospects features of 1990 census data to investigate the adaptation of immigrants One potential method is to consider and their children. the native-born members of certain race and ethnic communities as equivalent to the This assumption second generation. seems to be justified for most Asian and for some groups from the nationality groups (with a few exceptions) Caribbean and Central America. The only problem is that most of the of these new immigrant second generation groups were still children (below age 15) in 1990. The numbers of each national-origin population in the age categories of greatest interest (e.g., age 15-24) typically number For the larger groups (Filipinos and Chinese), a few thousand. the assump? tion that the native born represent the second generation is more open to question. An alternative research strategy is to examine the impact of duration of on the socioeconomic residence (or age at immigration) outcomes of immi? grants. If this technique is applied to teenagers or young adults, duration of residence can span the period from childhood to the years immediately before the census. Although cloud the problems of the selectivity of immigration the impact of "becoming American" (meaning picture, longer exposure to American society and arriving at a younger age) may have a positive, negative, or neutral influence on socioeconomic outcomes. second half of this article explored this question, in a preliminary fashion, with data from the published tabulations of the 1990 census for two outcomes: educational enrollment and early marriage and nonmarital fertility. Although the results were mixed, the most common pattern seems to be that greater exposure to American society during childhood is positive. There were hints of a different pattern, however. In a few instances, especially for Caribbean those with longest U.S. residence had slightly lower high school immigrants, enrollment rates and slighdy higher levels of nonmarital childbearing. Caribbeans are one of the most likely of immigrant populations to be closest to inner It may be that the classical and revisionist hypotheses city minority populations. of immigrant adaptation describe different aspects of important contemporary trends. These variations may become clearer when the analysis is broadened to include the differential economic status of immigrant families and their communities of settlement. The

REFERENCES Bean, F. D., B. Edmonston, and J. S. Passel, eds. 1990 Undocumented Migration to the UnitedStates.Washington, DC: Urban Institute.

712

International

Migration

Review

Bonacich, E., and J. Modell Basis of EthnicSolidarity: Small Businessin theJapaneseAmericanCommu? 1980 The Economic nity. Berkeley: University of California Press. Briggs, V. 1984 Policyand theAmericanLaborForce.Baltimore, MD: The Johns Hopkins Immigration University Press. Carpenter, N. 1927 Immigrantsand Their Children,1920: A StudyBased on CensusStatisticsRelative to the Foreign-Bornand the Native Whiteof Foreign or Mixed Parentage.Washington, DC: Government Printing Office. Reprinted in 1960 by Arno Press, New York. Chiswick, B. R. 1978 "The Effect of Americanization on the Earnings of Foreign-Born Men,"Journal of PoliticalEconomy,86:897-921. Daniels, R. A 1990 Comingto America: Historyof Immigrationand Ethnicityin AmericanLife. New York: Harper Collins. Dinnerstein, L., and D. M. Reimers A EthnicAmericans: Historyof Immigration Assimilation.Cambridge, MA: Harper and 1982 and Row. Duncan, B., and O. D. Duncan 1968 "Minorities and the Process of Stratification," American Sociological Review, 33:356-364. Farley, R. 1991 "The New Census Question about Ancestry: What Did It Tell Us?", Demography, 28:411-430. Fuchs, L. H. 1990 The AmericanKaleidoscope: Race, Ethnicityand the Civic Culture.Hanover, NH: Uni? versity Press of New England. Gans, H. 1992 "Second-Generation Decline: Scenarios for the Economic and Ethnic Futures of the Post-1965 American Immigrants," Ethnicand Racial Studies, 15:173-192. Gordon, M. 1964 Assimilationin AmericanLife. New York: Oxford University Press. Handlin, O. 1973 The Uprooted:The Epic Story of the GreatMigrations That Made the American People. Second Edition. Boston: Little, Brown. (1951) Hirschman, C, and E. P. Kraly 1990 "Immigrants, Minorities, and Earnings in the United States: 1950," International MigrationReview, 24:4-33. "Racial and Ethnic Inequality in the United States: 1940 and 1950," Ethnicand Racial Studies, 11:332-365. Hutchinson, E. P. 1956 Immigrantsand Their Children,1850-1950. New York:John Wiley and Sons, Inc. Jasso, G., and M. R. Rosenzweig 1990 TheNew Chosen in New York:RussellSage Foundation. People: Immigrants theUnitedStates. Kitano, H. H. L. 1976 Second Edition. Englewood Cliffs, JapaneseAmericans:The Evolutionof a Subculture. NJ: Prentice-Hall. Levine, D. B., K. Hill, and R. Warren, eds. 1985 A ImmigrationStatistics: Storyof Neglect. Washington, DC: National Academy Press. Lieberson, S. 1980 A Piece of the Pie: Blacks and WhiteImmigrantsSince 1880. Berkeley: University of California Press. 1988

Problems

and Prospects

of Studying

Immigrant

Adaptation

713

McKenney, N. R., and A. R. Cresce 1993 "Measurement of Ethnicity in the United States: Experiences of the U.S. Bureau of the Census." In Challengesof Measuringan Ethnic World: Science,Politics,and Reality: on Proceedings theJoint Canada-UnitedStatesConference the Measurement Ethnicity, of of April 1-3, 1992. Washington, DC: Government Printing Office. Pp. 173-221. Montero, D. 1980 Japanese Americans: Changing Patterns of Ethnic Affiliation over Three Generations. Boulder, CO: Westview Press. Neidert, L. J., and R. Farley 1985 "Assimilation in the United States: An Analysis of Ethnic and Generation Differ? ences in Status and Achievement," AmericanSociologicalReview, 50:840-850. Passel, J. S., and B. Edmonston 1994 "Immigration and Race: Recent Trends in Immigration to the United States." In Immigrationand Ethnicity:The Integrationof AmericasNewestArrivals. Ed. B. Edmon? ston andj. S. Passel. Washington, DC: The Urban Institute Press. Pp. 31-71. Portes, A., and R. G. Rumbaut A 1990 America: Portrait.Berkeley: University of California Press. Immigrant Portes, A., and M. Zhou 1993 "The New Second Generation: Segmented Assimilation and Its Variants,"Annals of theAmerican Academy Politicaland Social Sciences,530:74-96. Nov. of Reimers, D. M. Still the GoldenDoor: The Third WorldComesto America.Second Edition. New York: 1992 Columbia University Press. Thernstrom, S., ed. HarvardEncyclopedia AmericanEthnicGroups.Cambridge, MA: Harvard University 1980 of Press. U.S. Bureau of the Census Census of Population: The Foreign-BornPopulationin the United States, 1990 CP-3-1. 1993a Washington, DC: Government Printing Office. 1993b 1993c 1990 Censusof Population:The Asians and PacificIslandersin the United States, 1990 CP-3-5. Washington, DC: Government Printing Office.

1990 CensusofPopidation:Personsof Hispanic Origin in the UnitedStates,1990 CP-3-3. Washington, DC: Government Printing Office. U.S. Department of Labor, Bureau of International Labor Affairs 1989 The Effectsof Immigrationon the U.S. LaborEconomyand LaborMarket. Immigration Policy and Research, Report No. 1. Washington, DC: U.S. Department of Labor.