10

Data on Immigrants and Immigrant Integration

The study of immigrant integration requires reliable data on the foreign-born and their descendants, with the former providing information on the progress of immigrants with time spent in the United States, and the latter indicating progress toward integration between the first and second generations and beyond. By its very nature, integration is a process that unfolds over time. The pace of integration may be sped up or slowed down by individual characteristics, contexts of reception, or one’s structural position in society, but it always also depends on the duration of exposure to the host country’s culture and society (Alba and Nee, 2003; Portes and Rumbaut, 2014).

Among the most important information to gather to assess integration are data on birthplace/nativity, age and date of arrival, time spent in the United States, and legal status at present and upon entry. The many dimensions of integration (social, economic, political, civic, cultural) require different contextual, family/household, and individual-level data. This chapter focuses on data sources on the foreign-born and second generation, the measurement of the legal status of immigrants, including the undocumented, and challenges to the study of immigrant integration. It closes with recommendations to improve data collection, data access, and ultimately, the understanding of how well immigrants and their offspring are integrating into various arenas.

DATA ON IMMIGRANTS

The primary sources of data on immigrants are administrative data, government surveys, other nongovernmental national and local surveys, and qualitative studies.

Administrative Data

Administrative sources generally come from applications that immigrants file with the Department of Homeland Security (DHS) and/or the U.S. State Department in order to be admitted on either temporary or permanent visas into the United States. These data are produced annually and represent the “flow” of immigrants officially admitted into the United States. The Office of Immigration Statistics (OIS) in the Department of Homeland Security is the administrative home for annual data on all immigrants, both permanent and temporary, to the United States.¹ These aggregate data are useful for national-level questions about the number of documented immigrants arriving each year and the visa types on which they arrive.

Unfortunately the data OIS collects are neither very detailed nor very accessible. DHS collects very limited data on permanent residents as they arrive in the United States, and even less information is compiled upon the entry of temporary migrants. DHS lacks any consistent means of tracking immigrants as they move through different legal statuses. These data offer information about flows and numbers of immigrants in some legal statuses by various demographic characteristics, but are of limited use in assessing integration.

Surveys

For the study of immigrant integration in particular, researchers and policy makers need information beyond administrative data on the inflows of immigrants. That is, to study immigrants’ economic integration, a survey or administrative data also need to capture detailed information on income, social benefits receipt, assets and debt, and so forth, while analysis of civic integration might ask about volunteerism, membership in organizations, and participation in community events. One of the easiest ways to expand the body of knowledge about immigrant integration is to make the collection of key migration variables—place of birth, year of arrival in the United

___________________

¹ The Department of Homeland Security is obviously unable to collect data on immigrants who “enter without inspection” (EWI) by a U.S. Customs officer. However, if and when these undocumented immigrants do apply for a legal status, their information is collected, and their administrative record should include their previous status as EWI.

States and parents’ birthplace—part of most or all of the hundreds of surveys currently conducted to understand U.S. society. In addition, because a standard survey of U.S. residents with a bit over 1,000 sample respondents will, at best, include 100-200 immigrants, data collection should also consider oversampling the foreign-born population.

Government Surveys

The U.S. government also collects data on the characteristics of immigrants through the Census Bureau’s surveys. These data represent the “stock” of the foreign-born and are identified in census and survey data from a question on respondent’s place of birth. This type of question was included on U.S. censuses from 1850 through 2000 and since 2010 has been asked as part of the American Community Survey (ACS), which replaced the census long form. The birthplace question is also included on the Current Population Survey (CPS) and the Survey of Income and Program Participation (SIPP).

Most nationally representative statistics on immigrant integration currently come from the ACS, as its sample size alone among surveys is sufficient to enable statistically significant comparisons of integration outcomes across national origin groups and to consider patterns of integration at the regional, state, and local levels. Each year the ACS conducts around 2.5 million interviews and asks 24 questions about housing and 48 questions on the demographic, social, and economic characteristics of household members. Answers to these questions provide investigators with the major portion of the basic data used to assess immigrant integration.

ACS questions on country of birth, year of arrival, and naturalization enable researchers to examine indicators of immigrant integration by national origin, time since arrival, and citizenship and thus chart how social, economic, and housing characteristics change with time spent in the United States. Housing items of potential interest in assessing integration include kind of housing, age of housing, number of rooms, appliances and services, computer and Internet access, vehicle ownership, heating, utility costs, home value, home ownership, rent payments, tax payments, insurance costs, and mortgage payments. Socioeconomic characteristics of potential interest include educational attainment, foreign language usage, English ability, health insurance coverage, disabilities, marital status, marital history, childbearing, child care arrangements, veteran status, employment, occupation, hours worked, journey to work, and income from various sources.

The March CPS includes a demographic supplement that asks many of the same questions, often in greater detail, along with selected other items; but its sample size of 60,000 yields, on average, information on just 7,800

immigrants, which is sufficient for deriving national-level statistics for large national-origin groups but not for regional, state, and local estimates of smaller national-origin populations. The ACS is used to derive small-area estimates of the number and characteristics of immigrants for purposes of state and local planning, municipal decision making, and public service provision.

Nongovernment Surveys

Two private surveys particularly useful for studying immigrant integration are the New Immigrant Survey (NIS) and the Los Angeles Family and Neighborhood Study (LAFANS), although neither is nationally representative. With the addition of a supplemental sample of 511 immigrant families to the Panel Study of Income Dynamics in 1997/1999, that longitudinal survey is also now used to study integration, although the sample size is still too small for detailed studies and estimates below the national level. A list of all the surveys that collect data on immigration can be found at the UC Berkeley Population Center website (see http://www.popcenter.berkeley.edu/resources/migration_data_sets/data_by_region.php [November 2015]).

Qualitative Data Sources

There are many hundreds of qualitative studies of immigrant integration based on in-depth interviews or participant observation in immigrant communities, families, and institutions such as schools, churches, and workplaces. Most qualitative studies of immigrants in the United States only examine immigrants’ lives in this country, but some span national borders to investigate the transnational lives of immigrants and their families and to compare life in the United States with life in the country of origin (Dreby, 2010; Levitt, 2001).

A growing number of qualitative studies focus on the second generation, and while many of these are based in the traditional gateway cities, qualitative researchers are increasingly examining the dynamics of immigrant integration in newer destinations (Marrow, 2011). The specific topics covered range widely from changing patterns of family life, the development of new identities, and the role of immigrant religious organizations to the difficulties facing undocumented immigrants and their children and the ethnic and racial barriers confronting legal immigrants and their children as well. A welcome development is that some studies have combined qualitative methods with local representative surveys (Kasinitz et al., 2008). Qualitative studies are an important resource, providing valuable insights into the attitudes, values, and beliefs as well as patterns of behavior among immigrants and their children that can provide the basis for further in-

vestigation through nationally representative surveys. One way that they could be improved in the future is if more researchers could deposit their qualitative data in publicly accessible archives, where the data could be used by other researchers to replicate studies or to explore new questions. It is often more difficult to do this with qualitative than with quantitative research because of privacy considerations, but when possible it should be encouraged.

Measuring Legal Status

As discussed throughout this report, legal status affects immigrants and the second generation in myriad ways, both as an outcome of integration and as a determinant of integration. For this reason, it is critical for researchers to have accurate and precise information about the number of people who acquire different legal statuses, especially citizenship and lawful permanent resident (LPR) status, the reasons they do so, the factors that influence acquisition of status, and the length of time it takes. Yet there are few sources of data about the legal status of immigrants that also capture viable information to understand integration. The lack of data stems at least in part from concerns that questions about legal status are sensitive in nature and will therefore yield invalid results or suppress participation (U.S. Government Accountability Office, 2006; Carter-Pokras and Zambrana, 2006). In their analysis of surveys posing questions on legal status, however, Bachmeier, Van Hook, and Bean (2014) found that item nonresponse rates were relatively small, ranging from 0 to 14 percent, and significantly below those typically observed on questions about income. Indeed, they found that after using standard imputation procedures to correct for nonresponse, survey questions on legal status produced estimates of undocumented population composition comparable to those obtained using the widely accepted residual methods employed by the Department of Homeland Security and the Pew Research Center. They recommended that future data collection efforts include questions about legal status to improve models of immigrant integration.

Although most surveys do not collect information about legal status, there are several sources of data on this variable that can aid the study of immigrant integration. These sources are outlined below. In addition, there has been strong interest in identifying immigrants who are living in the United States without legal status because of the disadvantages of this group and their descendants regarding chances for integration. Efforts to collect or estimate data on this population are also discussed below.

Administrative Data Sources on Legal Status

OIS does make available selective demographic information about some legal statuses via a variety of products. The most consistent is the annual Yearbook of Immigration Statistics, “a compendium of tables that provides data on foreign nationals who, during a fiscal year, were granted lawful permanent residence (i.e., admitted as immigrants or became legal permanent residents), were admitted into the United States on a temporary basis (e.g., tourists, students, or workers), applied for asylum or refugee status, or were naturalized.”² The Yearbook offers basic statistical data on naturalized citizens, LPRs, refugees/asylees, nonimmigrants, and enforcement actions, but in general the breadth of these data is rather limited and their collection is limited to a single point in time (upon entry), making them of little use for studying immigrant integration.

Administrative data are most plentiful for LPRs and include annual information on country of origin, state and metropolitan area of U.S. residence, class of admission, age, sex, marital status, and occupation. No data are gathered on education or prior experience in the United States, which are critical factors to consider in any study of integration. For refugees and asylees, published tabulations include only country of origin, age, sex, and marital status. For nonimmigrants (those entering on temporary visas) these tabulations include country of origin, port of entry, category of admission, month of arrival, age, and sex. Data on undocumented migrants who are apprehended and deported include little more than country of origin.

U.S. Citizenship and Immigration Services (USCIS) also produces special reports on the H1B and H2B temporary visas. The H1B report is published annually and includes basic characteristics of workers in that status, but characteristics of H2B visa holders are only published for 2010. The U.S. State Department’s annual Report of the Visa Office provides statistical data on immigrant and nonimmigrant visa issued by consular offices abroad. The U.S. Department of Labor also publishes annual data on labor certifications for various employment-based visas and Permanent Labor Certifications; however, labor certifications do not always translate into applications for visas, so these numbers are limited in their usefulness.

The U.S. Department of Justice publishes detailed statistics on cases the Executive Office for Immigration Review (EOIR) adjudicates each year in its Statistics Yearbooks, which provides some detailed information on grants of asylum. Other government agencies, including the Congressional Research Service, Congressional Budget Office, and Government Accountability Office also produce periodic reports on various legal statuses, but it

___________________

² See http://www.dhs.gov/yearbook-immigration-statistics [November 2015].

is unclear whether they use publicly available data from the sources listed above or nonpublic data only available within the federal government.

During the course of its work, the panel made a formal request to USCIS (one of the study’s sponsors) for additional data on trends in naturalization rates and characteristics of naturalized citizens, estimates of eligible-to-naturalize populations and their characteristics, and information about assistance and fee waivers for applications for naturalization from USCIS. Some of these data were previously published in the Yearbook but no longer appear in the public data sources. Although USCIS did provide the panel with data on derivative naturalizations, none of the other data requests were met.

Survey Data on Legal Status

The only nationally representative survey that asks directly about legal status is the SIPP (Hall et al., 2010). The SIPP consists of a series of monthly panel surveys that began in 1984, with each panel lasting from 2.5 to 4 years. The most recent panel was fielded in 2014 and like prior surveys was a multistage stratified sample of the civilian, noninstitutionalized population. The SIPP is designed to compile detailed data on labor force status, program participation, and income to measure the economic situation of U.S. residents (U.S. Census Bureau, 2015). The study is designed as a continuous series of national panels, with sample size ranging from approximately 14,000 to 52,000 interviewed households.

For each person born abroad, the SIPP questionnaire asks “when [NAME] moved to the United States to live, what was his or her immigration status?” The response categories include immediate relative or family-sponsored permanent resident; employment-based permanent resident; other permanent resident; granted refugee status or granted asylum; nonimmigrant, such as a diplomatic, student, business, or tourist visa; and other. Respondents are then asked whether their status has changed to legal permanent residence. Undocumented status is estimated as a residual category inferred by process of elimination. Posing these questions produces very modest levels of nonresponse and has no significant effect on responses to subsequent questions or to survey follow-up rates. The questions appear to yield estimates of population characteristics comparable to the indirect methods described above (Bachmeier et al., 2014). The SIPP provides detailed data on labor force participation, income, assets, education, entitlement usage, and health insurance coverage and, given its large nationally representative sample, represents a potentially rich though underutilized source of information on the consequences of unauthorized status for immigrants and their families.

Although the SIPP is the only nationally representative survey of the

entire U.S. population to contain questions on legal status, these questions have been asked on a variety of other surveys of specific populations or regions. Below, the panel reviews some of the most useful publicly available surveys, but the surveys we describe are not a comprehensive list.

LAFANS, a representative sample of Los Angeles County fielded between 2000 and 2002, also contained a module on legal status that has been used to infer a lack of documentation after eliminating a series of legal categories (see Prentice et al., 2005; Bachmeier et al., 2014). A series of five questions asks first whether the respondent is a U.S. citizen and if the answer is “no” goes on to ask in serial order whether the respondent has permanent residence; asylum, refugee status, or temporary protected status; a tourist visa, work visa, or other document permitting one to stay in the United States; and if the later document is still valid. Although LAFANS is representative for the area covered by its frame, it is not nationally representative; the relative size and composition of the undocumented population derived for Los Angeles County lines up closely with indirect estimates for that county using the residual approach (described below), at least with respect to age, sex, and national origin (Bachmeier et al., 2014). Given the wealth of data contained in LAFANS on households, adult and child members, neighborhoods, and schools, it offers a promising source for studying the consequences of unauthorized status in the nation’s largest single undocumented population.

Another source of data on legal status is the NIS, which began with a representative pilot survey of 1,134 legal immigrants who acquired legal residence documents during July and August of 1996 and were followed for the next year. The pilot survey compiled detailed information about migratory experiences in the United States prior to admission as LPRs, enabling identification of those with prior undocumented status. Using these data, Massey and Malone (2002) estimated that 32 percent of all “new” LPRs had previously been undocumented migrants, roughly the same figure estimated by Jasso et al. (2008), although among groups such as Mexicans and Central Americans the figure was much higher.

The pilot survey was followed in 2003 with a probability sample of the entire cohort of LPRs who acquired their residence documents between May and November of 2003, including interviews with 8,573 adults, 810 parents of sampled child immigrants, and 4,915 spouses. Follow-up interviews were conducted with this cohort between June 2007 and December 2009, yielding data on 3,902 adults, 392 children, and 1,771 spouses. Although the full NIS did not include the detailed module on pre-arrival experiences administered in the pilot, it did identify previously unauthorized immigrants who were legalized under the Immigration Reform and Control Act’s (IRCA’s) registry provisions, suspension of deportation, cancellation of removal, and special legalization programs, again comprising

around 32 percent of all immigrants. Both sources of data allow researchers to study the effects of legal status on integration (e.g., Weeraratne and Massey, 2014).

Another source of data on formerly unauthorized migrants is the Legalized Population Survey, a sample of 6,193 undocumented immigrants who sought LPR status through IRCA and were interviewed in 1989 with a follow-up survey of 4,012 respondents in 1992. The survey collected information on the labor market characteristics of those who gained legal status at three times: just prior to legalization while still undocumented, at legalization, and 5 years following legalization.

Although primarily designed to assess the progress of second generation immigrants, the Immigration and Intergenerational Mobility in Metropolitan Los Angeles Study (IIMMLA) included questions on legal status. The presence of undocumented parents and children in the Los Angeles study enabled work on the effects of legal status on the integration of undocumented migrants and their citizen children (see Bean et al., 2015).

Three specialized population surveys that also contain questions on legal status are the Latino National Survey, the National Agricultural Workers Survey, and the National Asian American Survey. The Latino National Survey is a representative national survey of some 8,634 Latino residents of the United States implemented in 2005-2006. Unfortunately, only a minority of respondents to the Latino National Survey are foreign-born, drastically reducing the sample size of immigrants and rendering it too small for all but national estimates on large groups such as Mexicans.³

The National Agricultural Workers Survey was an employment-based, random survey of U.S. crop workers. It collected demographic, employment, and health data in face-to-face interviews with over 56,000 workers in 1988-1989 and ascertained which respondents were undocumented and which had applied for legalization under IRCA’s Special Agricultural Worker Program.

The National Asian American Survey was a telephone survey of 5,159 self-identified Asian or Asian American residents of the United States fielded in 2008. It included a direct question about current legal status but 93 percent of respondents reported either being U.S. citizens or legal visa holders, with just 5.2 percent saying they didn’t know or refusing to answer, thus potentially identifying them as undocumented.

Finally, both the Mexican Migration Project (Durand and Massey, 2004) and the Latin American Migration Project (Donato et al., 2010) ask direct questions on legal status from representative samples of immigrants

___________________

³ The sample size for the LNS is still the largest publicly available sample of the Latino foreign-born population, and some of the state subsamples should allow for national-origin level analysis of the foreign-born—e.g., California, Florida, and Texas.

from specific sending communities who are interviewed mostly in their countries of origin. Although neither dataset is necessarily representative of migrants from Mexico or other Latin American nations, systematic comparisons with objective sources suggest they cover their respective immigrant populations well (Massey and Zenteno, 2000; Massey and Capoferro, 2004), and both surveys have been used extensively to study the effect of legal status on labor market integration.

The above surveys all follow the best practice of making data publicly available in publicly accessible archives. All surveys of immigrants and the second generation should do this.

Qualitative Data on Legal Status

A variety of qualitative studies have examined the effects of legal status on immigrant integration, and many of these are cited in Chapter 3. Although qualitative datasets are generally too small to make broad generalizations about the effects of legal status on immigrant integration, they provide important guidance for how legal status might be measured and what types of social variables intersect with legal status to help or hinder immigrants’ integration prospects. Qualitative studies can also help researchers assess the effects of legal status in particular local contexts and for smaller groups of immigrants.

Sources of Data on Undocumented Immigrants

Standard residual methods for measuring the undocumented have been perfected in recent decades to derive indirect estimates of the number, location, and basic demographic characteristics of undocumented migrants using the Decennial Census and the CPS and ACS surveys. Indirect estimates of the size and characteristics of the undocumented population are regularly published by OIS,⁴ the Pew Research Center,⁵ and the Center for Migration Studies.⁶ The Migration Policy Institute has also published estimates for the undocumented at the national and select state and county level, including estimates of the population eligible for Deferred Action for Childhood Arrivals (DACA) and Deferred Action for Parental Accountability (DAPA, also called Deferred Action for Parents of Americans and Lawful Permanent Residents).

In order to derive an independent benchmark estimate of the undocu-

___________________

⁴ See http://www.dhs.gov/immigration-statistics-publications [November 2015].

⁵ See http://www.pewhispanic.org/category/publications/ [November 2015].

⁶ See http://cmsny.org/researchprojects/democratizingdata/us/unauthorizedtables/ [November 2015].

mented population, the U.S. Government Accountability Office (2009) has recommended the application of a “two-card method” in conjunction with the CPS (see U.S. Government Accountability Office, 2009, for details). This method seeks to overcome concerns about asking about legal status because those without documents only admit to being in a category that contains undocumented migrants among several other legal statuses, making it impossible to identify any single person as being undocumented. Although this design feature guarantees privacy, it also makes it impossible to identify the characteristics of the undocumented population. As a result, the two-card method has been little used in research on immigrant integration to date.

In summary, the inability to identify legal status among the foreign-born enumerated in the Decennial Census or in the ACS and CPS leaves a major determinant of immigrant integration unmeasured, thus potentially biasing models estimated to predict how integration varies with time spent in the United States (Massey and Bartley, 2005). Among Mexican immigrants, in particular, one cannot know whether a relatively slow pace of socioeconomic integration is a general feature of the population or is simply the average of a fast pace of integration among legal immigrants and citizens and a slow pace among undocumented immigrants. Moreover, to the extent that the prevalence of undocumented migrants varies across national origins, differences in integration will be confounded with differences in the rate of legal documentation.

DATA ON THE SECOND GENERATION

As discussed throughout the report, family and household circumstances, as well as the general climate toward immigrants and the policy environment that immigrants enter, are critical in determining patterns and processes of social and economic integration for the second generation and beyond. In addition to the parents’ own legal status, age and date of arrival, and time spent in the United States, other key variables affecting the integration of the second generation include whether both parents were foreign-born; what language is spoken at home; household socioeconomic and demographic composition; and general indicators of parental health, education, occupation, and income. In addition, more distal variables such as the general policy environment toward immigrants (welcoming or restrictive), neighborhood characteristics, the types of schools attended; and the availability and quality of English-as-a-Second-Language (ESL) programs and other social and health services affect immigrant integration and should be measured with respect to immigrant-descendent generation.

In general, resources accessible to children within the household while growing up can be expected to play an outsized role in determining the

nature and extent of their later integration into American society. For older children of immigrants who no longer live at home, of course, a specific question on the birthplace of parents is required to identify members of the second generation but rarely is much additional information gathered about the parents or the family in which they came of age. Despite this information gap, circumstances in the family of origin are nonetheless critical to understanding current trajectories of integration among adult members of the second generation, underscoring the need for longitudinal data in studies of immigrant integration, especially in the second generation.

The second generation may be identified in one of two ways. Minor children of immigrants are easily identified as long as they remain in the household of their immigrant parents, who are themselves identified from the birthplace question. The adult children of immigrants, however, must be identified using a separate question on the birthplace of parents: a question that was asked on every Decennial Census from 1870 to 1970 but was eliminated on the 1980, 1990, and 2000 census forms and was not included on the ACS in 2010. Since 1996 a parental birthplace item has been asked in the March supplement to the CPS, but the small sample size makes it difficult to create reliable estimates for most second generation immigrant populations (for one potential method, see Ramakrishnan, 2005), especially at the state and local level. There are other limitations to the CPS that limit its usefulness for substate-level analysis: more than a third of county level identifiers are not available in the public release of the CPS due to concerns about privacy; other data might be available only through a handful of restricted data centers

At present there is no reliable source of information on adult second generation immigrants based on a large, nationally representative sample. As noted in Chapter 6, because the U.S. Census Bureau data relies on self-identification of race and Hispanic origin and because identity is related to socioeconomic status, the identification of the second and especially the third and higher generations may be increasingly inaccurate and may introduce systematic errors in measurements of intergenerational mobility.

To fill the gap, private organizations led by the Russell Sage Foundation have funded a series of specialized surveys of second generation immigrants in San Diego and Miami (Portes and Rumbaut, 2014), New York (Kasinitz et al., 2008), and Los Angeles (Brown et al., 2011). Telles and Ortiz (2008) used a survey of Mexican Americans in California and Texas, originally conducted in 1965, and then relocated the original respondents and their descendants. They demonstrated that having information on biological generations—tracing great-grandfathers, grandfathers, fathers, and sons—yielded a different trajectory of integration than measuring generation as time since immigration and examining cross-sectional differences among individuals of different immigrant generations but similar age cohorts. In

addition, Grusky and colleagues (2015) recently recommended the creation of an “American Opportunity Study” to develop the capacity to link records across the Decennial Census, the ACS, and administrative records (see Box 10-1). These linkages would significantly enhance researchers’ ability to monitor social mobility across generations, a key component in the measurement of immigrant integration. Overall, the lack of a parental birthplace on the ACS and its absence from the 1980-2000 census long forms constitutes a huge gap in the nation’s statistical system and is the largest single barrier to studying the intergenerational integration of immigrants (Massey, 2010). As the third generation grows in size, the lack of a question on grandparents’ place of birth also means that researchers are unable to trace intergenerational integration as it advances beyond the

children of immigrants. During the greatest period of mass immigration since the early 20th century, when the population of immigrants rose from 14 to 40 million and the second generation proliferated, the nation has lacked a reliable means of assessing the progress and characteristics of the children of immigrants.

CHALLENGES TO THE STUDY OF IMMIGRANT INTEGRATION

Analyzing the progress of immigrants as individuals presents many challenges despite the existence of large, nationally representative sources of data on the foreign-born. In addition to the lack of data on legal status and the absence of a large nationwide sample of the second generation, additional challenges include the ambiguity in defining duration of U.S. experience among immigrants who undertake multiple trips in and out of the country, the difficulty of identifying the intention to settle in the United States and when settlement occurs, the relatively small share of immigrants in most general-population samples, and lack of data on contexts of reception.

Ambiguity in Duration and Intent to Settle

Beyond legal status, a key variable in all models of immigrant integration is an indicator of time spent in the United States. The 2000 Decennial Census long form and the 2010 ACS assessed time in the United States with the question “when did this person come to live in the United States?” The 1970 through 1990 Census long forms asked foreign-born person when they came to the United States “to stay.” Most researchers simply subtract the date of entry (to live or to stay) from the date of the survey to estimate total time spent in the United States. However, recent research based on retrospective longitudinal data indicates that the vast majority of persons entering the United States as new LPRs are not arriving for the first time. For example, among the 2003 cohort of new legal resident aliens surveyed upon entry by the NIS, 66 percent had prior experience in the United States through a variety of pathways, and total time accumulated in the United States varied widely by legal status (Massey and Malone, 2002). The question itself tacitly assumes that foreign-born persons have indeed made a decision to settle in the United States, but in fact some LPRs and naturalized citizens use their residence documents as a convenience to come and go (Massey et al., 1987; Massey and Akresh, 2006). Even among those who intend to live out their lives in the United States, answering the question requires a judgment about when exactly it was that they came with the intention of settling (Massey and Malone, 2002).

Thus measures of time spent in the United States derived from govern-

ment censuses and surveys may contain significant measurement error and considerable but unknown potential for bias across national-origin groups. Although total time accumulated in the United States and the total number of trips during this time are probably most relevant in determining integration outcomes, these quantities are generally not available from standard data sources. In sum, models of immigrant adaptation and integration estimated from the Decennial Census, ACS, or CPS are likely to be strongly affected by omitted variable bias (owing to the lack of information on legal status and parental birthplace) and measurement error (because of unreliability in accurately capturing time spent in the United States).

Small Sample Size

Moving beyond these standard sources of national data, the possibilities for studying immigrant integration are even more restricted given the relatively small size of the foreign-born population. The foreign-born presently constitute around 13 percent of the U.S. population, so that a representative national sample of 2,000 people would yield just 260 immigrants distributed across a wide variety of divergent groups. Unless non-Mexican Latinos and Asians or immigrants themselves are oversampled, standard surveys are unlikely to produce sufficient numbers of immigrants for meaningful analysis. There are some exceptions, including the National Health Interview Survey, the Panel Study of Income Dynamics, and the National Health and Nutrition Examination Survey. Other surveys target those of Latino and Asian origin and thus include a large number of immigrants by design (e.g., the Latino National Survey and Asian American National Survey). But even when surveys contain larger subsamples of immigrants, very few ask a question on parental birthplace, precluding studies of intergenerational integration.

Lack of Longitudinal Data

Another limitation is the lack of information about immigrants and their descendants in longitudinal data sources. One of the best ways to capture change over time, as is implicit in an idea such as “integration,” is through longitudinal studies that reinterview the same respondents at multiple time points. This helps researchers identify whether economic, health, or civic integration happens gradually and steadily over time or if there are key inflection points in the process. One problem for cross-sectional surveys is that researchers face a much harder time identifying the temporal order of events that might point to potential mechanisms of change. For example, while the prior census enumerations captured whether a foreign-born resident was a U.S. citizen at the time of enumeration, as well as whether he or

she was married, had a job, owned a home and so forth, one does not know whether marriage, employment, and homeownership occurred before or after acquiring citizenship. Recent changes to the ACS that ask respondents the year in which they naturalized will help to specify the determinants of citizenship, but longitudinal studies are a gold standard for understanding mechanisms of integration.

A model for how the United States could provide longitudinal data based on administrative records linking visa status and mode of entry to subsequent income is the Canadian government’s Longitudinal Immigrant Database.⁷ This database links taxation records with immigration records over time, allowing researchers to examine labor force participation, earnings, internal migration, and income over time by class of admission and other attributes such as language ability (Hiebert, 2009). The availability of these data allows researchers to provide evidence on the long-term integration of people admitted under different visa types. This type of database would not only allow U.S. researchers to more accurately describe integration over time, it would provide lawmakers with direct information about the effects of U.S. immigration law such as the relative costs and benefits over time of admitting different types of skilled workers.

Contexts of Reception

Even though there are difficulties in measuring changes to American society, the panel defined integration as a two-way process. To understand that process in both directions, data are needed not only on immigrants—their background, legal status, attitudes, economic condition, health status, and so forth—but also on the communities in which they live. In studies that include a detailed geographic identifier, researchers can embed the information about an individual in aggregate data on the individual’s neighborhood, town, or metropolitan area, usually using Decennial Census data. However, such place-based data are often incomplete for some of the integration outcomes highlighted in this report. The research community needs better data, for example, on the civic or institutional infrastructure within immigrant communities—for instance, the number of religious institutions offering non-English services, the number of community-based social service agencies, the number of refugee resettlement organization, or the number of legal aid clinics, especially those able to process immigration paperwork. A few studies have begun to evaluate civic infrastructure using, for example, data on 501(c)(3) nonprofit organizations distributed by the National Center for Charitable Statistics, but these databases do not have

___________________

⁷ See http://www23.statcan.gc.ca/imdb/p2SV.pl?Function=getSurvey&SDDS=5057 [November 2015].

consistent information on whether the organizations are focused on serving immigrant communities(see also de Leon and Roach, 2013).

Localities also vary in the degree to which they welcome or develop initiatives to facilitate immigrant integration (see White House Task Force on New Americans, 2015). Academics regularly compare the policy frameworks for immigrant integration across countries. One of the most developed is the MIPEX project,⁸ which has 144 policy indicators across 38 countries over 8 years to track changes in policy and variation between countries. Other indicators, such as the Multiculturalism Policy Index⁹ and the Citizenship Observatory¹⁰ have generated cross-national policy comparisons for more focused topics such as diversity and citizenship policy. Using these indices, researchers can evaluate whether policy differences across countries tend to correlate with better or more problematic integration outcomes, such as naturalization or economic outcomes.

It would be valuable to extend this contextual approach to localities within the United States. A few studies have tried to see whether states or municipalities with harsher anti-immigrant ordinances impede immigrants’ integration or, conversely, whether outreach policies have real effects on integration (see Mollenkopf and Pastor, 2013; Flores, 2014).

RECOMMENDATIONS

The most serious current gap in the U.S. statistical system on immigration is the lack of a question on parental birthplace for a large representative sample of the U.S. population. A question on the birthplace of parents would enable clear identification of second generation immigrants of all ages and enable researchers to assess their social, economic, and cultural integration not only within the United States as a whole but across various national origins in different regions, states, and metropolitan areas. Although it remains important to assess the relative progress of foreign-born immigrants as they adapt to life in the United States, the more critical issue for the nation is the progress of their U.S.-born children, since they are native citizens who in a very real way represent America’s demographic future as they inevitably produce more citizens.

With millions of long-term U.S. residents lacking legal documentation and millions more children growing up in a household with one or more

___________________

⁸ See http://www.mipex.eu/what-is-mipex [November 2015].

⁹ See http://www.queensu.ca/mcp/immigrant.html [November 2015].

¹⁰ See http://eudo-citizenship.eu/ [November 2015].

unauthorized parents, undocumented status clearly emerges as a major constraint on socioeconomic mobility among immigrants and thus a key determinant of the prospects for their children as well. A question by which respondents select among various well-defined legal statuses at entry or at present, leaving those in undocumented status to be identified by process of elimination, now appears on the SIPP and the LAFANS. The question provides important information on the status of immigrants. If the indirect estimation question functions well in the CPS, it could also be added to the ACS.

As indicated by the New Immigrant Survey, the Mexican Migration Project, and other surveys, legal status is a dynamic variable that changes over time as immigrants move from temporary to permanent legal statuses or between unauthorized and authorized circumstances. The attainment of legal status and eventual citizenship are likely to be crucial steps in the process of economic and social integration, yet researchers presently lack the means to model them.

Just as the 1986 Immigration Reform and Control Act mandated a survey of the immigrants who legalized, any future legislation to address the legalization of millions of undocumented immigrants should do the same. A legalization program creates a targeted opportunity to learn more about the population of immigrants living in the United States without legal status. Data collection and analysis of the legalized population—how they entered the United States, where they fit into the labor market, demographic characteristics, family composition, use of social services, migration behavior and origins—in the 1986 program illuminated the behavior of a population for which there previously was little systematic information. Learning more

about today’s unauthorized population will help provide insights into an otherwise elusive population and assist in creating new policies to address undocumented immigration and immigrants.

A cost-effective way to improve data accuracy and aid research on intergenerational changes in immigrant integration is to link administrative data with Decennial Census and survey data. Records from the Survey of Income and Program Participation (SIPP) could link to respondents and their children and parents, to enable investigation into intergenerational and intragenerational issues, including but not limited to mobility, long-run outcomes of early life circumstances, and intergenerational effects for immigrants and their descendants. Matched individual-level records from Decennial Censuses (and the ACS) with income data from Internal Revenue Service and the Social Security Administration would allow for longitudinal studies of the socioeconomic progress of immigrants in American society and allow for the measurement of both intracohort change and intercohort (for cohorts based on time of arrival in the United States) change for successive waves of immigrants. Matched Census and USCIS records would allow for in-depth studies of pathways to legalization and also the impact of legal status on socioeconomic outcomes of individuals and their children.

USCIS and other federal government agencies produce a range of administrative data about immigrants, including flows of new arrivals by visa status and data on newly naturalized U.S. citizens. However, the published data are aggregated with only some very basic cross-tabulations. It is impossible to use these data for more fine-grained analysis, for example, to

compare whether women from specific countries over the age of 50 living in areas of the United States with more legal aid support in their native language are more likely to naturalize than similar female migrants living elsewhere. To support such finer grained analyses, researchers need micro-level data on individuals and the ability to link such data with additional information, such as aggregate data on localities.

Many common data sources for social science research lack large enough samples of Hispanics, Asians, or immigrants to effectively use in the study of immigrant integration. In recent years, however, a number of nationally representative studies have added oversamples of Hispanics and Asians or immigrants, including the National Health Interview Survey, the Panel Study of Income Dynamics, and the National Health and Nutrition Examination Survey. These surveys could provide a model for oversampling key populations.

Contexts of reception are critical for immigrant’s integration prospects, not just on the national or state scale but also on the local and neighborhood scale. Researchers therefore need access to small-area data. In addition, researchers need these data in a timely manner if they are going to capture processes that are unfolding continuously.

Population surveys continue to be an important source of health status and use of health services among immigrants and the U.S.-born. Unlike administrative and clinical data, population surveys are able to reach people who do not seek preventive or treatment services. Since there is a sizable portion of people who either do not access health care or do so irregularly, surveys provide data on people who are less likely to receive medical attention. Moreover, surveys are able to include more dimensions of social life that are important for understanding health status and health care (e.g., socioeconomic status, English language proficiency, country of origin, length of time in the United States, age of immigration).

REFERENCES

Hall, M., Greenman, E., and Farkas, G. (2010). Legal status and wage disparities for Mexican immigrants. Social Forces, 89, 491-514.

Hiebert, D. (2009). The Economic Integration of Immigrants in Metro Vancouver. Working Paper Series No. 09-08. Vancouver: Metropolis British Columbia Centre for Excellence for Research on Immigration and Diversity.

Jasso, G., Massey, D.S., Rosenzweig, M., and Smith, J.P. (2008). From illegal to legal: Estimating previous illegal experience among new legal immigrants to the United States. International Migration Review, 42, 803-843.

Kasinitz, P., Waters, M.C., Mollenkopf, J.H., and Holdaway, J. (2008). Inheriting the City: The Children of Immigrants Come of Age. New York: Russell Sage Foundation.

Levitt, P. (2001). The Transnational Villagers. Berkeley: University of California Press.

Marrow, H. (2011). New Destination Dreaming: Immigration, Race, and Legal Status in the Rural American South. Stanford: Stanford University Press.

Massey, D.S. (2010). Immigration statistics for the 21st century. The ANNALS of the American Academy of Political and Social Science, 631, 124-140.

Massey, D.S., and Akresh, I.R. (2006). Immigrant intentions and mobility in a global economy: The attitudes and behavior of recently arrived U.S. immigrants. Social Science Quarterly, 88, 30-47.

Massey, D.S., and Bartley, K. (2005). The changing legal status distribution of immigrants: A caution. International Migration Review, 34, 469-484.

Massey, D.S., and Capoferro, C. (2004). Measuring undocumented migration. International Migration Review, 38, 1075-1102.

Massey, D.S., and Malone, N. J. (2002). Pathways to legalization. Population Research and Policy Review, 21, 473-504.

Massey, D.S., and Zenteno, R. (2000). A validation of the Ethnosurvey: The case of Mexico-U.S. migration. International Migration Review, 34, 765-792.

Massey, D.S., Alarcón, R., Durand, J., and González, H. (1987). Return to Aztlan: The Social Process of International Migration from Western Mexico. Oakland: University of California Press.

Mollenkopf, J., and Pastor, M. (2013). Struggling over Strangers or Receiving with Resilience? The Metropolitics of Immigrant Integration. Working Paper. Available: http://brr.berkeley.edu/wp-content/uploads/2013/05/Mollenkopf-Pastor-struggling-strangers.pdf [October 2015].

National Research Council. (1985). Immigration Statistics: A Story of Neglect. Panel on Immigration Statistics. D.P. Levine, K. Hill, and R. Warren (Eds.). Committee on National Statistics, Commission on Behavioral and Social Sciences and Education. Washington, DC: National Academy Press.

National Research Council. (2013). Developing New National Data on Social Mobility: A Workshop Summary. A. Smith, Rapporteur. Committee on Population and Committee on National Statistics, Division of Behavioral and Social Sciences and Education. Washington, DC: The National Academies Press.

Portes, A., and Rumbaut, R.G. (2014). Immigrant America: A Portrait, Updated and Expanded. Oakland: University of California Press.

Prentice, J.C., Pebley, A.R., and Sastry, N. (2005). Immigration status and health insurance coverage: Who gains? Who loses? American Journal of Public Health, 95(1), 109-116,

Ramakrishnan, S.K. (2005). Democracy in Immigrant America: Changing Demographics and Political Participation. Stanford, CA: Stanford University Press

U.S. Census Bureau. (2015). Survey of Income and Program Participation. Available: http://www.census.gov/sipp/ [October 2015].

U.S. Government Accountability Office. (2006). Estimating the Undocumented Population: A “Grouped Answers” Approach to Surveying Foreign-Born Respondents. GAO-06-775. Report to the Subcommittee on Terrorism, Technology and Homeland Security, Committee on the Judiciary, U.S. Senate, Washington, DC.

U.S. Government Accountability Office. (2009). Estimating Irregular Migration in a Survey: The “Two-Card Follow-Up” Method. Working Paper No. ESA/P/WP.208. New York: United Nations, Department of Economic and Social Affairs, Population Division.

Weeraratne, B., and Massey, D.S. (2014). Does Past Unauthorized Immigrant Status Result in a Wage Penalty for Legalized Immigrants? Presented at the American Economic Association Annual Meeting, January 4, Philadelphia, PA.

White House Task Force on New Americans. (2015). Strengthening Communities by Welcoming All Residents: A Federal Strategic Action Plan on Immigrant & Refugee Integration. Available: https://www.whitehouse.gov/sites/default/files/docs/final_tf_newamericans_report_4-14-15_clean.pdf [October 2015].

This page intentionally left blank.