Abstract
The post-1960s immigration boom and contemporary demographics have elevated generation-since-immigration as a category that is central to analysts and, more generally, to Americans as they make sense of their place in the world around them. This makes the collection of data on immigrant generations imperative if surveys are to keep up with how the nation’s people think about themselves and each other. A clear portrait of contemporary assimilation, and indeed American progress, depends on possessing the right tools to paint such a portrait. That means that surveys must enable researchers to identify respondents’ generation, particularly the third generation of the post-1965 immigration wave.
Any survey that hopes to reflect the social identities of its respondents has to keep up with the times. And these are complicated times when it comes to identity. The United States, along with other postindustrial societies, is governed by norms and institutions that now allow far greater latitude in how people identify individually (Bean, Brown, and Bachmeier 2015). Indeed, gender, sexual, race, and ethnic identities are arguably more flexible than ever (Brubaker 2016; Saperstein and Penner 2012; Schilt 2010). That dynamism poses significant challenges to surveys that aim to reflect the identities of the people answering them. The massive influx of immigrants in the last four decades only adds to the challenge. As immigrants, their children, and the society that has received them adjust to one another, notions of race, ethnicity, and nation grow in complexity (Jiménez 2017). For survey designers and analysts, this requires hard choices about how to gather and analyze data that tap into group identities.
Gender, race, and ethnicity attract most of the attention regarding changes in identity, categorization, and identification. But large-scale immigration has increased the complexity of a category that is much less subject to individual choice, but nonetheless important for how individuals define themselves and others: generation-since-immigration (hereafter generation), where the first generation consists of immigrants, the second generation the children of immigrants, the third generation the grandchildren of immigrants, and so on. Tectonic shifts in the generational profile of the post-1965 wave of immigrants are already under way in the United States. 1 The shift entails the rise of a third generation whose social, political, and economic fortunes will reveal the extent and kind of assimilation among the descendants of today’s largely non-European immigrants. In addition to being an important temporal marker for analysis of assimilation, generation is also a category relevant to how Americans perceive social insiders and outsiders in everyday life. In immigrant-rich contexts, generation, along with race, class, and gender, can serve as a basis upon which individuals decipher who belongs.
Yet analysts of American society are ill-equipped to make sense of generational distinctions because there is no good way in most surveys to identify a respondent’s generation. And even when generation is ascertainable, it is next to impossible to identify a unique post-1965 third generation. A clear portrait of contemporary assimilation, and indeed American progress, depends on possessing the right tools with which to paint such a picture. That means that surveys must enable researchers to identify respondents’ generations, and particularly the third generation of the post-1965 immigration wave.
The Centrality of Generation to the American Experience
It is incontrovertible that immigration is changing just about every aspect of American life. The numerical significance of the immigrant population relative to the overall population indicates the scale of that change. Today, immigrants make up roughly 13 percent of the population, a rate nearing historic highs (National Academies of Science, Engineering, and Medicine 2015). Immigrants and their children combine to make up one in four individuals in the United States, and in some metro regions, the first and second generations combine to make up roughly half of the population (Jiménez 2017). The effects of immigration on the racial composition of the United States are well known (Frey 2014). Indeed, attention to compositional changes resulting from immigration focus heavily on race and ethnicity, and with good reason. The United States has grown dramatically more diverse in recent decades because of the settlement of immigrants, who come overwhelmingly from Latin America, Asia, and the Caribbean (National Academies of Science, Engineering, and Medicine 2015). Inasmuch as survey classifications of racial groups have become problematic (Prewitt 2013), large government surveys have attempted to keep up with the times by maintaining race and ethnicity 2 as categories of classification, by adopting new racial categories, and, as of 2000, by allowing respondents to check more than one racial category.
When it comes to documenting changes in the generational composition of the United States, surveys, and especially those administered by the government, have attempted but not adequately kept pace. The nation’s generational profile has indeed changed dramatically. Figure 1 uses decennial census data and the Current Population Survey (CPS) to offer a generational profile of the U.S. population since 1970.
U.S. Population by Generation-Since-Immigration, 1970–2010
It is important to note from Figure 1 that the twentieth-century low point for the foreign-born population was 1970, when 87 percent of the American population was third generation or greater (“third-plus” generation). That figure suggests that the paradigmatic American experience of immigration was less of a defining feature of life in 1970 than at any point in the twentieth century. It was also at about that time that the grandchildren of the turn-of-the-century European immigrants were coming of age, and their income, education, intermarriage rates, and neighborhood integration indicated successful assimilation (Alba and Nee 2003). Moreover, immigration took on a nostalgic hue in the collective identities of ethnic groups associated with that wave (Alba 1990; Waters 1990), and the descendants of that immigration wave grafted immigration onto the American national myth of origins (Jacobson 2006). With the U.S. population tilted heavily away from the first and second generations, and with the near complete assimilation of the last large wave of immigrants, the need to distinguish in surveys between different generations did not seem acute. And so 1970 was the last year that the decennial census included a question about parents’ place of birth, from which analysts could distinguish between first, second, and third-plus generations. Starting in 1994, the CPS included a question about parents’ place of birth, but the question has not made its way into more comprehensive government surveys, such as the census or the American Community Survey (ACS).
The large and overwhelmingly non-European immigration in the ensuing decades—the post-1965 immigration—changed the place of immigration in the American experience once again. This time, the source regions of immigrants have been decidedly non-European, with Latin America (especially Mexico), Asia, and the Caribbean sending large numbers. Not only has the most recent wave of immigrants swelled the proportion of the total population that is foreign born to near historic levels, it has also changed the nation’s generational composition. Along with their second-generation children, these immigrants have grown the proportion of the total U.S. population that is either first or second generation. As Figure 1 shows, the share of the U.S. population from those generations nearly doubled between 1970 and 2010. That demographic change underscores the added importance of generation in the contemporary United States as both a temporal marker necessary for tracking patterns of assimilation and as a category that shapes how people make sense of their everyday lives.
Generation: Category of Analysis; Category of Practice
The importance of generation manifests in the way that it operates as both a category of analysis and as a category of practice. The former refers to generation as a category that social scientists use to organize populations for the sake of conducting analysis, usually of the statistical sort. The latter signifies the categories that people use in their everyday life to make sense of the world (Brubaker 2004).
Generation as a category of analysis
The presence of the large immigrant population means that adequately deciphering the contours of any social, political, and economic change entails data on generation. But among scholars, race is a favored category in analyses of differences in social and economic outcomes. For example, economic inequality unambiguously has a racial component. Blacks and Latinos generally sit at the lower end of the socioeconomic distribution, while Asians and whites anchor the high end (Massey 2007). However, examining differences in socioeconomic attainment only by comparing across racial categories misses significant generational differences within these categories (Alba, Jiménez, and Marrow 2014; Jiménez, Fields, and Schachter 2015). Those differences are especially important for groups with long immigration histories. Early research on immigrant populations emphasized generation as a temporal marker of assimilation, noting that with more generations in the United States, immigrant groups tended to resemble the mainstream (Warner and Srole 1945). Implicit assumptions in early assimilation theory that saw assimilation into a white, Protestant, and Anglo-Saxon notion of the American mainstream are problematic (Gordon 1964; Warner and Srole 1945; see Alba and Nee [2003] for a critique). These flaws notwithstanding, an enduring insight from these early theories related to the inevitability of generational change. Scholars ever since have measured assimilation over time using generation as the key temporal marker (National Academies of Science, Engineering, and Medicine 2015).
Mexicans, arguably the most important contemporary immigrant group and one with a long immigration history, illustrate the importance of generation for tracking assimilation patterns. Using historic U.S. decennial census data, Smith (2003, 2006) showed that there are large intergenerational differences among Mexicans in years of schooling and annual income. For example, first-generation Mexican men born between 1900 and 1905 had an average of 3.81 years of school. The second generation, born 25 years later, had an average of 7.88 years of school. The third generation, born 25 years after that, registered an average of 12.61 years of school (Smith 2006, Table 6). These quantitative findings are consistent in kind, if not degree, with trends observed among European and Asian groups whose histories in the United States go back generations (Alba and Nee 2003). Smith could calculate these generational differences because the census, prior to 1980, regularly included a question about parents’ place of birth. 3 Important intergenerational differences also show up in qualitative research on Mexicans. Vasquez (2011) interviewed three generations of Mexican Americans in the same family, showing tremendous variation in mobility patterns, and ethnic consciousness across the three generations.
The large post-1965 wave of immigration highlights the continued relevance of generation for tracking assimilation. Intergenerational comparisons are central to research on post-1965 immigration, and especially diverging assessments of second-generation assimilation. There is disagreement about whether the assimilation of the Latino and black post-1965 second generation is characterized by racialization and permanent exclusion (Haller, Portes, and Lynch 2011; Portes and Rumbaut 2001), or by steady, if bumpy, progress (Alba, Kasinitz, and Waters 2011; Kasinitz et al. 2008). The data informing that debate have largely come from three nongovernmental data collection efforts: the Children of Immigrants Longitudinal Survey (CILS), the Immigrant Second Generation in Metropolitan New York (ISGMNY), and the Immigration and Intergenerational Mobility in Metropolitan Los Angeles (IIMMLA) studies. CILS sampled more than 5,000 children of immigrants in the Miami and San Diego metropolitan areas in 1992, when these children were in their early teens, and reinterviewed most of the original sample three years later, and again in 2001–2003. CILS also includes parent surveys and in-depth interviews with select survey respondents. ISGMNY is a survey and interview sample of 1.5- (arrived before the age of 12) and second-generation individuals in the New York metropolitan area. The sample also includes third-plus-generation comparison groups (whites, Puerto Ricans, and African Americans). IIMMLA focused on the mobility trajectories of the young-adult 1.5 and second generations in the Los Angeles metropolitan area. In the case of Mexicans, the sample also includes a substantial number of individuals from the third and later generations. Space constraints prevent me from laying out the findings generated from these data (see, for example, Bean, Brown, and Bachmeier 2015; Kasinitz et al. 2008; Portes and Rumbaut 2001). It suffices to say here that the major findings reveal significant differences along an array of measures between the first and second generations across groups.
As valuable as these data are, they do not capture changes across three generations of the post-1965 immigration, which are crucial to fully understand the extent and kind of contemporary assimilation. Data collection for these studies took place when the second generation of the post-1965 immigration wave was relatively young. But that is no longer the case. The leading edge of today’s second generation has entered middle age, and many are themselves parents to a post-1965 “new third generation” (Jiménez, Park, and Pedroza, forthcoming). These grandchildren of the post-1965 immigrants will write the next chapter in the contemporary American immigrant assimilation story. Julie Park, Juan Pedroza, and I have conducted the first analyses of the new third generation, examining the household circumstances in which they were growing up in 2010 (Jiménez, Park, and Pedroza, forthcoming). We compared the new-third-generation household circumstances to that of a second generation in 1980. Our findings reveal the continued importance of generation as a marker of group change. Because the 1980 U.S. Census lacked a parent place-of-birth question, we had to include in our sample only U.S.-born children (age 19 and under) coresiding with foreign-born parents, age 25 to 54. We identified our 2010 third-generation sample using CPS samples from 2008 to 2013. 4 The parents’ place-of-birth questions on the CPS enabled us to identify a unique third-generation (age 19 and under) coresiding with second-generation parents, age 25 to 54. We measured outcomes at the household level, including differences in parents’ socioeconomic status, having intermarried parents, the presence of extended family, and the ethnic and racial labels that parents give to their children. We found that the Hispanic and Asian new third generation in 2010, compared to the second generation in 1980, has parents who are better educated, earn more, are more likely to be intermarried, and are less likely to be living with extended family. These findings point to an assimilation pattern that resembles patterns among groups with earlier immigration histories.
But there may also be important differences. We also analyzed how intermarried parents labeled their children. We found that third-generation children growing up in intermarried households with at least one Hispanic parent were more likely to receive a Hispanic label compared to a second generation in the same kind of household in 1980. The same was true for Asians: third-generation children growing up in an intermarried household that included at least one Asian parent were more often given an Asian label compared to second-generation children with intermarried parents in 1980. It is important to note that, starting in 2000, the census and the CPS gave respondents the option of checking multiple racial categories. And so while intermarried parents in 1980 were forced to select a single racial category for their children, parents in 2010 could select more than one racial category. The census and CPS treat the Hispanic category as an ethnic category, distinctive from race. Thus, parents select whether their children are Hispanic and their children’s race in response to separate questions. It is possible that parents could have used the Hispanic and race categories to communicate their child’s mixed ancestry, but it is impossible to know their intension from their survey responses. These caveats aside, parents continue to identify their children with the categories that reflect the origins of the immigrant generation.
The larger point from analysis of the new third generation is that among the post-1965 immigration wave, generation is an analytical category that tracks important changes within ethnic and racial groups. Moreover, generation is especially important right now because the rise of the new third generation is the best opportunity that social scientists have for gaining a fuller picture of contemporary assimilation.
Generation as a category of practice
Generation is more than a category central to analyses of group change. Generation is also a category of practice that structures how individuals categorize themselves and others in everyday life. Intragroup identity processes centrally entail drawing distinctions among group members along generational lines. My own ethnographic research on later-generation Mexican Americans—those whose families came to the United States before 1940—in Garden City, Kansas, and Santa Maria, California, shows how Mexican Americans with deep generational roots in the United States and more recently arrived immigrants and their children distinguish each other based on generation (Jiménez 2010). The generational distinctions were animated by characteristics and behaviors that later-generation Mexican American respondents believed were associated with first- and second-generation Mexicans, and not people like themselves, whose families had “been here forever” (Jiménez 2010). Later-generation respondents saw generational differences in the degree of knowledge about Mexico, the ability to speak Spanish, the kinds of food that people eat, and tastes for popular culture. Mexican Americans in Garden City—where there have been two very distinctive waves of Mexican immigration—saw these generational differences in clearer terms than their counterparts in Santa Maria, where immigration has been more continuous. But the more relevant difference to everyday life had to do with the number of generations that individuals’ families had been in the United States. Similar findings appear in other ethnographic research on generations of Mexican Americans (Macias 2006; Ochoa 2004; Vasquez 2011), Asian Americans (Tsuda 2016; Tuan 1998), and Polish Americans (Erdmans 1998).
In addition to marking differences within ethnic groups, generation is also a category of practice pertinent to intergroup relations. The relevance of generation is particularly prominent in areas where immigrants have settled in large numbers. My ethnographic research among “established individuals”—those who are U.S.-born of U.S.-born parents—in three Silicon Valley communities illustrates the importance of generation as a marker of everyday difference that cuts through racial group distinctions. One of the communities that I studied, Cupertino, is a high-skilled immigrant gateway with large South-Asian Indian and Chinese first and second generations. The established population was largely white, but also included later-generation Chinese and Japanese Americans. Interviews with the later-generation individuals there showed that the number of generations that someone’s family has been in United States was a basis for perceptions of difference and similarity. For example, respondents reported that there were few differences between established Asian American families and white families precisely because both had deep generational roots in the United States. But respondents noted significant differences between established Asian American families and those that were headed by foreign-born parents. When asked what differentiated these families, respondents said that established Chinese and Japanese Americans were more “whitewashed,” which they defined by weak observance of ethnic traditions, a more relaxed approach to schooling, and greater sociability. Similar findings emerged in the other Silicon Valley locations that I studied: East Palo Alto and Berryessa (a neighborhood of San Jose) (Jiménez 2016). In all three of the locations, generational differences within racial groups were often just as important as markers of group distinctions as racial identity (Jiménez 2017). 5
Measuring Generations
That generation is both a category of analysis and a category of practice implies an imperative to track generation in surveys. Accomplishing that aim is not simple. The inclusion of a parents’ place-of-birth question in the CPS has allowed researchers to identify first, second, and third-plus generations. For researchers interested in post-1965 immigrant assimilation, that generational breakdown has only limited adequacy. It has allowed the documentation of progress between the adult first- and second-generations of this immigration wave, and across three-plus generations of groups with older immigration histories.
But there are at least two reasons why the CPS data are inadequate for tracking trends in immigrant integration beyond the second generation. The first is that the CPS sample is too small for the regional, metropolitan, and small-area data necessary to explain intergenerational change. When so many outcomes are linked to the places in which individuals live (Sampson 2012; Sharkey 2013), understanding intergenerational change requires data on the neighborhoods in which people grow up and the neighborhoods that they live in as adults. A second issue, and one stemming from the small sample size, is that the CPS does not enable analysts to make group distinctions beyond the largest ethnic groups. An important theme in assimilation research past and present is that assimilation outcomes between ethnic groups differ widely. Some groups, like Vietnamese, Hmong, Salvadorans, and Dominicans are too small to show up in the CPS prominently enough to permit statistical analysis. And yet these groups are important to a full understanding of American assimilation. A third issue is that the CPS forces researchers to lump all individuals who are U.S.-born of U.S.-born parents into a “third-plus-generation” category, without the ability to identify a unique third generation separate from those who are fourth generation and beyond. The need to identify a unique third generation comes from the demographic rise of the new third generation (Jiménez, Park, and Pedroza, forthcoming), whose outcomes will offer a fuller picture of the post-1965 immigrant integration than the one that currently relies on analyses of the first and second generations only (Alba, Kasinitz, and Waters 2011; Haller, Portes, and Lynch 2011; Kasinitz et al. 2008; Portes and Rumbaut 2001). As it stands, the only way to identify the new third generation is by selecting households from the CPS with second-generation parents who are coresiding with U.S.-born children (Jiménez, Park, and Pedroza, forthcoming). That selection strategy restricts analyses to children, leaving analysts to infer from their household circumstances what adulthood has in store. When this new third generation enters adulthood and moves out of their parents’ homes, they will become part of the CPS’s third-plus generation and thus be indistinguishable in the data from individuals who descend from earlier waves of immigrants of the same ethnoracial origins. The ability to adequately track the multi-intergenerational change of the post-1965 immigration will thus be lost.
What about nongovernment surveys? Surveys aimed at understanding the second generation have done a great deal to advance research on assimilation. However, these surveys are costly and tend to focus on specific regions, limiting valuable comparisons across regions.
Parents’ and grandparents’ place of birth
What can be done? At the very least, governmental surveys—ideally the ACS—would return to the old census practice (and current CPS practice) of asking respondents about parents’ place of birth. Including that question would be a step in the right direction, but it would still not enable analysts to identify a unique third generation separate from a fourth-plus generation. In addition to a parents’ place-of-birth question, including a grandparents’ place-of-birth question would allow for such a distinction. I recognize that making already long surveys such as the ACS longer can hurt response and completion rates, and some respondents may not know where their grandparents were born. An intermediary step between excluding a grandparents’ place-of-birth question and the current slate of questions would be to ask a simple yes-no question: Were any of your grandparents born outside of the United States? The answer to that short question would allow researchers to identify a unique third generation (the U.S. born with U.S.-born parents who answer “yes” to the grandparents’ place-of-birth question would constitute the third generation), and thus yield a more precise view of assimilation among contemporary immigration waves.
At the heart of scholarly and public debates about immigration is whether today’s immigrants are assimilating. It is an emotionally charged debate, and one that is more productively hashed out with good data. Contemporary immigration has elevated generation as a central category of analysis and of practice in American life, making the collection of data on generation an imperative if surveys, both government and private, are to keep up with how Americans think about themselves and each other. A better capture of generation data on surveys does more than keep up with the times. It would allow researchers to offer a fuller understanding of the times we live in.
Footnotes
Notes
Tomás R. Jiménez is an associate professor of sociology and comparative studies in race and ethnicity at Stanford University. He is author of The Other Side of Assimilation: How Immigrants Change American Life (University of California Press 2010).