1. What are the political consequences for growing inequality?

Educational Researcher, Vol. 48 No. 7, pp. 407 –420 DOI: 10.3102/0013189X19860814

Article reuse guidelines: sagepub.com/journals-permissions © 2019 AERA. http://er.aera.net

OCTOBER 2019 407

Racially segregating students of color in certain schools—separated from White or middle-class peers—often cor-responds with unfair financing of schools, regressive allocation of quality teachers, and culturally limited curricula (e.g., Noguera, Pierce, & Ahram, 2015; Vasquez Heilig, Brown, & Brown, 2012). But little is known about whether the racial or economic segregation of young Latino children—as they enter kindergarten—is easing or growing worse.

We do know that Latino children have enjoyed gains in early learning and social development during their preschool years since the 1990s (Bassok, Gibbs, & Latham, 2018; Reardon & Portilla, 2016). Yet the segregated settings that beset many Latino children as they enter school likely constrain this progress (Owens, 2018). An earlier generation of research detailed how Black children benefit from integrated schools (for review, see Cook, 1984). Yet little is known empirically about recent trends in levels of racial and economic segregation that confront Latino children at entry to elementary school.

We contribute to the segregation literature by focusing on schools that young Latino children enter, matching universe data for the nation’s schools and districts with family and child-level data, 1998 to 2010. This allows us to (a) observe trends in the isolation of Latino children within certain schools in school

districts nationwide, (b) disaggregate trends for children of immigrant and native-born parents, and (c) learn how school ver- sus neighborhood contexts are changing for Latino children. We also ask whether patterns of racial and economic segregation may move independently over time, as many Latino families spread across the nation, some enjoying upward mobility (Dondero & Muller, 2012; Reardon & Bischoff, 2011).

Overall, we find intensifying segregation of Latino children from White peers among schools in districts that enroll at least 10% Latino pupils; this set against already high levels of racial isolation. Nor have the nation’s 10 districts serving the poorest concentrations of students shown discernible progress toward integrating Latino students among their constituent schools. Children from low-income families—regardless of race or Latino ethnicity—did increasingly attend school with middle-class peers over the 1998 to 2010 period.

Our more textured child and family-level data verify that Latino kindergartners experienced declining exposure to White peers during the period. Those of foreign-born mothers entered

860814EDRXXX10.3102/0013189X19860814Educational ResearcherEducational Researcher research-article2019

1University of California, Berkeley, CA 2University of Maryland, College Park, MD 3University of California, Irvine, CA

Worsening School Segregation for Latino Children? Bruce Fuller1, Yoonjeon Kim1 , Claudia Galindo2, Shruti Bathia1, Margaret Bridges1, Greg J. Duncan3, and Isabel García Valdivia1

A half century of research details how segregating racial groups in separate schools corresponds with disparities in funding and quality teachers and culturally narrow curricula. But we know little about whether young Latino children have entered less or more segregated elementary schools over the past generation. This article details the growing share of Latino children from low-income families populating schools, 1998 to 2010. Latinos became more segregated within districts enrolling at least 10% Latino pupils nationwide, including large urban districts. Exposure of poor students (of any race) to middle-class peers improved nationwide. This appears to stem in part from rising educational attainment of adults in economically integrated communities populated by Latinos. Children of native-born Latina mothers benefit more from economic integration than those of immigrant mothers, who remain isolated in separate schools. We discuss implications for local educators and policy makers and suggest future research to illuminate where and how certain districts have advanced integration.

Keywords: early childhood; early learning; educational policy; equity; immigration/immigrants; Latino children;

longitudinal studies; regression analyses; segregation

FEATURE ARTICLES

408 EDUCATIONAL RESEARCHER

even more sharply segregated schools, compared with Latino peers of native-born mothers. At the same time, Latino children more often resided in neighborhoods with higher educational attainment, and parents’ schooling predicted children’s entry to economically integrated schools.

We discuss implications for local educators and policy mak- ers, then sketch future research priorities. Trends in neighborhood segregation condition school segregation and improved exposure to middle-class peers. But little is known on how education lead- ers and policy makers might advance gains at the community level for Latino children, progressively financing schools and fairly allocating high-quality teachers. This article does not revisit the consequences of deep-seated segregation on learning. We do unpack local variation in the school and neighborhood contexts of Latino children and how these settings are changing over time.

Integrating Latino Children—Progress or Regress?

Schools can act to ease or harden disparities experienced by chil- dren of Latino or low-income families. Yet as educators and policy makers attempt to buoy Latino children—seeking to advance their early growth and learning—little is known about whether they are entering less or more segregated schools and how trends may differ for the offspring of immigrant versus later-generation parents. The segregated facets of neighborhoods, especially housing patterns, contribute to the extent to which Latino children attend schools with White or middle-class peers. This article focuses on trends in school segregation, while also reporting on relevant features of neighborhoods in which Latino children are raised.

Tracing school and residential segregation has become more complicated as many Latino families exit urban enclaves (partly fostered by rising educational attainment), bound for what demographers term new destinations—suburbs, exurbs, or rural areas—still mainly White in their demographic composition, often hosting schools that are ill-equipped to serve diversifying children and families.

“Migration [within the United States] ostensibly permits resi- dents to overcome place-linked disadvantages,” as Tienda and Fuentes argue (2014, p. 505). But “Institutional barriers, eco- nomic conditions, and housing constraints limit the realization of social mobility.” And even when Latino parents search out materially better-off neighborhoods, their children may not enter more integrated schools. Housing patterns and local dis- trict policies (or inaction) may limit the likelihood of finding integrated schools.

Overall, it is not clear whether school segregation for young Latino children is increasing or decreasing nationwide. Crosnoe (2005) found that Mexican-heritage children sorted into schools with greater concentrations of poor and non-White stu- dents relative to White peers, drawing on a nationally represen- tative sample. Orfield and Lee (2007) reported that 60% of all Latino students (K–12) attended high-poverty schools nation- wide (at least half in poverty), as did less than one-fifth of all White pupils (18%).

Tracking student composition in 350 metro areas, Stroub and Richards (2013) found that after peaking in the late 1980s, mean levels of racial segregation among schools within districts had fallen modestly by 2009. They also found modest gains in the integration of Latino students across K–12, in part mirroring family movement to once predominantly White suburbs (con- sistent with Fiel, 2013; Iceland & Sharp, 2013). Racial integra- tion among non-White groups has improved modestly as well, but not necessarily across economic classes (Reardon & Owens, 2014). That is, the capacity of families to move into middle-class neighborhoods (independent of race or ethnicity) may condition the likelihood of selecting an integrated school.

Turning to residential segregation, Iceland and Sharp (2013) found that the average Latino resident was less likely to see a White neighbor in 2010, compared with 1980, after examin- ing segregation among communities situated in 366 metro areas. We know that Latinos displayed slightly rising levels of residential segregation from Whites, 1970 to 2010, across 287 metro areas (Alba & Foner, 2015). And aging suburbs, although better-off economically compared with immigrant enclaves, remain quite segregated racially (Lichter, Parisi, Taquino, & Grice, 2010).

A majority of Latinos by 2010 resided in suburban parts of metro areas for the first time nationwide, in part stemming from migration to new destinations in the Midwest and South (Hirschman & Massey, 2008). Fully one-third of Mexican immi- grants to the United States, between 1995 and 2000, settled into nontraditional states (Lichter et al., 2010). Many more left tradi- tional immigrant gateways, heading out to diversifying suburbs.

These findings prompt the question of how residential segre- gation conditions the isolation of Latino children in certain schools. Both family selection and place-based factors appear to determine when segregation among schools diminishes or grows worse. As certain Latino families sort into new destinations, for instance, this alters the ethnic and economic mix of district enrollments. Better educated Latino parents may seek middle- class neighborhoods, while still enrolling their children in pre- dominantly Latino schools—achieving greater exposure to middling, but not necessarily White, peers (Reeves & Busette, 2018).

Incumbent residents of neighborhoods also shape evolving lev- els of segregation via higher fertility rates among some groups, along with the out-migration of White families. Part of the “reseg- regation” of schools stems, not so much from school policy, but from higher birth rates among Latina mothers, relative to other groups residing in the same neighborhoods (Logan, 2004).

Yet we still know little about whether young Latino children are entering schools in which exposure to White or middle-class peers is improving or diminishing and how these trends vary for the offspring of immigrant versus native-born Latino parents. Nor do we know whether children from low-income children (independent of their race or ethnicity) interact more or less with middle-class peers over time. This conditions the experience of Latino children from poor households. We examine these trends over the 1998 to 2010 period, an era that witnessed unstable policies toward English learners, resurging hostility toward immigrants, and recovery from the Great Recession.

OCTOBER 2019 409

Mapping Segregation—Variation Across Geographies and Social Classes Children of Latino heritage, of course, make up a rising share of school enrollments across the nation. Over one-fifth of all kin- dergartners were of Latino heritage in 17 states by 2012 (Stepler & Lopez, 2016). As Latino parents radiate out from urban enclaves, some do settle in less segregated neighborhoods.1 This suggests that levels of residential integration may be improving in many neighborhoods, or at least variability grows wider, com- pared with old Latino enclaves in urban centers. This may condi- tion trends in school segregation.

To illustrate local variation in residential segregation, Figure 1 displays the share of population, White, for Los Angeles census tracts in which at least one-fourth of all residents were Latino (based on kindergarteners in the federal Early Childhood Longitudinal Study [ECLS], 2010). We see a slight presence of White neighbors in predominantly Latino tracts downtown (near the “101” freeway label), with greater integration east and northwest of the central city.

The L.A. pattern of geographic dispersal backs an optimistic theoretical position known as spatial assimilation, claiming that immigrant groups will assimilate into the middle class as they enjoy gains in education and job status over time (Alba, Kasinitz, & Waters, 2011). These advances by individuals accumulate, according to this account, to lift the economic status and racial

integration of neighborhoods, increasingly populated by second- and third-generation descendants of immigrant settlers.

This theoretical lens suggests that the isolation of low-status children in particular schools will ease as residential segregation lessens, as experienced with White European immigrants a cen- tury ago. But whether this allegedly natural drift toward assimi- lation will be enjoyed by contemporary Latino immigrants, especially when racialized and stigmatized in many communi- ties, remains unknown. Nor is it clear that Latinos entering racial or economically integrated neighborhoods will sort into integrated schools.

Other scholars counter with place stratification theory, argu- ing that class and racial markers continue to leave Latinos in subordinate and isolated positions, even as they move to promis- ing economic destinations in which to raise children. “Economic mobility is no guarantee of residential integration,” as Lichter, Parisi, and Taquino (2015, p. 36) argue.2 This suggests that racialized markers will segregate Latinos into separate schools, even when the larger geographic unit (metro area or school dis- trict) is becoming more diverse. One troubling case occurs when immigrant Latinos move into new destinations, where incum- bent residents remain hostile or school districts are ill prepared, often assigning newcomers to highly segregated schools.

This theoretical position highlights the pivotal role of educa- tion leaders, as they allocate Latino students and scarce resources

FIGURE 1. Varying levels of Latino segregation for sampled Los Angeles census tracts, 2010. Source. U.S. Census Bureau (2011) and National Center for Education Statistics (2001). Cartography by GreenInfo Network, www.greeninfo.org.

410 EDUCATIONAL RESEARCHER

to particular schools. That is, racial status and local histories must be explicitly taken into account if school segregation is to be effectively addressed. And conscious school interventions are more likely to make a difference under this theoretical view- point: Levels of school segregation are not necessarily driven by residential or housing patterns alone.

Ecological theory may further explain variation in levels of segregation observed among schools within districts. Urban ecologists take a broader look, studying how the movement of people, jobs, and housing between urban centers and suburbs comes to racially segment groups among differing schools.3 Patterns of residential segregation, housing prices, and the city’s racialized political economy all shape which families populate differently situated school districts (Reardon & Owens, 2014).

Differential fertility rates for Latinas, tied to maternal educa- tion levels, then alter the complexion of districts and school attendance zones over time. Legal histories and district policies shape efforts to ease or sustain the segregation of non-White stu- dents in separate schools as well (Henig, Hula, Orr, & Pedescleaux, 2001). Whereas the urban ecology frame stems from older structural views of the city’s political economy, the school institution may act with some autonomy to ease the isola- tion of particular students in certain schools.

The Latino case becomes more complicated by the wide vari- ation in the class position of this growing population, differing between the offspring of immigrant versus later-generation Latinos. Rising school attainment, for instance, spurs exit from urban enclaves and upward mobility for some Latino parents (Bean, Brown, & Bachmeier, 2015).

We know that residential segregation is higher in locales host- ing greater shares of foreign-born, rather than native-born, Latinos (Iceland & Nelson, 2008). In contrast, native-born Latinos with children tend to live in higher-income areas and closer proximity to Whites (Fuller, Bein, Kim, & Rabe-Hesketh, 2015; South, Crowder, & Chavez, 2005).4 Markers of social class, especially the nativity of parents, may interact with neigh- borhood attributes to shape children’s attainment (Brazil, 2016; Owens, 2010).

But it is unknown whether young Latino children have entered more or less racially segregated schools in recent decades. Recent findings show that American society overall is becoming more economically segregated, as affluent Americans increasingly reside in exclusive enclaves (Gibson-Davis & Percheski, 2018; Reardon & Bischoff, 2011). Yet for the wide middle class, we have little understanding of whether Latino children are gaining greater exposure to middle-class children (economically inte- grated schools) and how these trends may vary between immi- grant and later-generation Latino families.

Contexts of School Reception As Latino families spread to new areas, little is known about the kinds of schools they enter. Dondero and Muller (2012) found fewer bilingual teachers and less access to advanced courses in new-destination schools, compared with schools in older Latino enclaves. Focusing on the 30 largest “new settlement areas,” Fry (2011) found that Latino students in K–12 enjoyed greater

exposure to White and more affluent peers, compared with enclaves. Other work finds that “the traditional sites … of Latino growth – where the knowledge and resources to turn around the problem are potentially in greater abundance – are no longer the sites of greatest growth” (Gándara & Mordechay, 2017, p. 151).

Place certainly matters for how schools receive immigrant and later-generation Latino children, manifest by teachers with varying cultural competence, those who speak Spanish or reach out to parents (Perreira, Fuligni & Potochnick, 2010). How dis- tricts allocate resources among schools often stems from segre- gated housing and schools, along with local educators’ varying commitment to racial or economic integration (Orfield & Lee, 2007; Vasquez Heilig, Khalifa, & Tillman, 2014).

Such place-based policies can hold long-term consequences: Mexican American females raised in Texas, for instance, display lower school attainment and higher fertility, compared with peers in California (Van Hook, Bean, Bachmeier, & Tucker, 2014). And we know that Black-White achievement gaps are wider in highly segregated districts (Owens, 2017).

Research Questions and Analytic Strategy

In sum, evidence remains mixed on whether Latino students attend schools that offer rising or diminishing exposure to White or mid- dle-class peers. Nor do know whether children from poor house- holds (independent of race or Latino ethnicity) attend schools offering greater interaction with middle-class peers over time. Even less is known about segregation trends for young Latino children as they enter kindergarten, along with the differing experiences of immigrant and native-born offspring. And as diverse Latino fami- lies spread out to diverse suburbs and exurbs, have they benefited from more racially or economically integrated schools?

This article informs these empirical gaps by asking whether Latino children’s exposure to White peers in elementary school increased or declined between 1998 and 2010. We also describe whether exposure of poor to nonpoor children (regardless of race or ethnicity) changed in districts enrolling significant shares of Latino pupils. We report standard measures of racial and eco- nomic segregation among schools within the nation’s school dis- tricts. Then, we draw on representative cohorts of individual kindergartners over the same period to replicate these patterns and, going deeper, to examine differences in school and neigh- borhood contexts.

Racial segregation among schools within districts— RQ-1A: Did the racial segregation of Latino children within

isolated elementary schools increase or decline in the nation’s school districts, 1998 to 2010?

RQ-1B: Did the segregation of students of low-income fami- lies from middle-class peers regardless of race or ethnicity in isolated elementary schools (economic segregation) increase or decline, 1998 to 2010?

Racial segregation for Latino kindergartners and subgroups and explaining variation—

RQ-2A: Did the nation’s average Latino kindergartner enter an elementary school with rising or declining concentra- tions of Latino peers (racial segregation)?

OCTOBER 2019 411

RQ-2B: Did the nation’s average Latino kindergartner enter an elementary school with rising or declining concentra- tions of peers from low-income families (economic segregation)?

RQ-2C: Among Latino kindergartners from economically diverse families, what markers of social class predict higher or lower levels of racial segregation?

Method

Our analysis builds from two tandem data sets. First, we assem- bled enrollment data for all public elementary schools and dis- tricts nationwide, drawn from the Common Core of Data (CCD) in 1998 and 2010, compiled by the National Center for Education Statistics (NCES, 2001).5 This allowed us to con- struct multiple measures of racial and economic segregation among schools within districts. Then, we replicated observed patterns for Latino kindergartners at school entry, along with disaggregating trends for subgroups of Latino children, drawing on nationally representative samples of kindergartners in the same years, 1998 and 2010, from the Early Childhood Longitudinal Study (ECLS-K; NCES, 2001).6

The CCD in 1998 includes enrollment data broken down by ethnicity and eligibility for free or reduced-price meals (FRPM) for a universe count of 50,529 elementary schools nationwide situated in 13,215 school districts. The corre- sponding counts in 2010 were 53,636 and 13,849, respectively. The ECLS-K data are drawn from national probability samples of 7,219 children in 1998 (nested in 442 districts) and 8,627 (in 419 districts) for 2010, after losing cases (a) with missing nativity data on the mother, (b) absent a match between school or family to the respective census tract, or (c) when all covari- ates necessary for regression estimates were not available for the final analysis (NCES, 2018; U.S. Census Bureau, 2011). We weighted all data using the sampling weights provided in the ECLS-K data set.

Measures

Latino children, families, and subgroups. The CCD reports enrollment counts of elementary school children identified by local education authorities as Latino. The ECLS-K data goes deeper, based on field interviews of the household respondent, usually the mother. Each self-identified as of Latino origin or another ethnic heritage and whether they were native or foreign- born. We only draw on data collected by field staff when the child was attending kindergarten. Subsequent interviews with mothers include questions about country of origin, but this information was not utilized in the present article.7

District and school-level segregation. Multiple indicators of segregation—the extent to which one group disproportionately resides in certain units (schools) situated within a larger geographic or institutional unit (districts)—are commonly used in the demog- raphy and immigration literatures (Reardon & Owens, 2014). We describe the conceptual justification for each measure.

For each of the nation’s districts that host at least one elemen- tary school, we calculated the two-group interaction (exposure) index, interpretable as the probability that a randomly drawn Latino student shares a school with a White peer (Massey & Denton, 1988; Owens, 2017). This measure indicates the extent to which Latinos are exposed to Whites in elementary schools. This index ranges higher—indicating strong integration—when Latino students are evenly distributed among schools within the district, relative to the distribution of White peers. (The formula for calculating each index appears in the appendix.)

Given that Latino children may attend schools populated by multiple ethnic groups, not limited to Whites, we also calculated entropy for each school, measuring the evenness of the represen- tation of two or more groups (i.e., Black and Asian-heritage chil- dren). Finally, we include the dissimilarity index (D), the absolute value of what percentage of White students would have to exit a school to reach parity with the Latino share. Shares of pupils enrolled who are Latino, White, or FRPM-eligible are reported. Analyses were conducted for all the nation’s elementary schools and host districts, then separately for districts enrolling at least 10% Latino children.

We use the term “middle class” to describe kindergarteners whose family income exceeded the eligibility threshold for FRPM. Change in the percentage of students qualifying for FRPM does not always track against child poverty rates, one reason that we report the share of enrollment who are English learners, along with neighborhood attributes for sampled kindergartners drawn from the ECLS-K data (Hoffman, 2012; NCES, 2018).8

Neighborhoods. Given interest in how school segregation varies among types of neighborhoods, we matched each kindergartner to her or his census tract of residence. This allows us to report from the ECLS-K data median household income, poverty rates, and educational attainment of resident adults in 1998 and 2010, as key indicators of the child’s social context.9 We also determined whether the family resided in a new destination or traditional urban enclave, utilizing Tran and Valdez’s (2015) procedure.

Markers of family social class. We report on attributes of kinder- gartners and households, focusing on markers of class that help to explain variation in segregation among schools. These factors include household income (adjusted to 2018 dollars); maternal education as dichotomous indicators of less than high school, diploma, some college, or bachelor’s degree or more; non-Eng- lish home language; and female-headed household or not. After reporting descriptive trends, 1998 to 2010, we estimate the extent to which the class position of Latino families contributes to the level of school segregation experienced by their kindergartner.10

Findings

Shifting Segregation Levels for Latino Children (RQ1)? We first replicate earlier work showing that the nation’s schools serve rising shares of Latino pupils and students from low- income families regardless of race or ethnicity, over the 1998 to 2010 period. Latino students experienced declining exposure to

412 EDUCATIONAL RESEARCHER

White peers over the period in districts with at least 10% Latino enrollment. At the same time, students from low-income house- holds (regardless of race) were more likely to attend school with a middle-class peer.

We see in column 1 (Table 1) that Latino children’s expo- sure to White peers across all school districts remained con- stant on average (row A, index score, .61), 1998-2010; whereas the likelihood that poor children were exposed to middle-class peers increased markedly (.33 to .45, p < .001, about two- fifths SD).

Turning to districts with enrollments at least 10% Latino (row B), we see the interaction exposure index declining from .50 to .47 (p < .001) by a small level of magnitude (.11 SD). We again see a rising likelihood of poor students being exposed to middle-class students among schools (regardless of race); the index rising from .31 to .38 (p < .001, .29 SD).

Note that many more districts enrolled at least 10% Latino children in 2010 (4,277), compared with the count in 1998 (2,321). Trends did not appreciably change when comparing 1998 and 2010 levels only for the original 2,321 districts. This constant set of districts also displayed declining interaction between Latino and White children, along with improving inter- action between children of poor and those of middle-class fami- lies (again, regardless of race or ethnicity). The share of students enrolled, FRPM eligible, climbed from 38% to 61% over the

period. This was partly due to liberalized eligibility for free and reduced-price meals at school.

We calculated segregation indices for the nation’s 10 poorest districts (based on FRPM shares) enrolling more than 50,000 pupils and at least 10% Latino, as reported in row C.11 Overall, we see considerably lower Latino-White interaction scores, indi- cating more severe segregation of Latino children in separate schools, relative to national averages. Schools in these 10 districts enrolled increasing shares of Latino students, whereas the per- centage of students, White, declined over the period.

Placing these index values in context, large urban districts tend to display quite high levels of racial segregation, such as Los Angeles or San Antonio, where the interaction index dropped below .10 in 2010. On the other hand, certain subdistricts of New York City schools display stronger integration of Latinos, above .35 on the index.

Patterns are quite similar when moving to the districts from which ECLS-K kindergartners were sampled (rows D and E), except that the likelihood of Latino children’s exposure to White peers fell from .65 to .51 over the period (p < .001, .40 SD). It may be that Latino families with kindergarten-age children reside in more segregated neighborhoods relative to families with older children attending elementary school. Again, we see rising expo- sure of poor to middle-class children, a gain of moderate magni- tude among the ECLS-K school districts (p < .001, .35 SD).

Table 1 Change in Segregation for the Nation’s School Districts and Elementary Schools and for ECLS-K Sample of

Units Serving Kindergartners, 1998 to 2010

District Segregation School Segregation

N

Latino Exposure to

Whites

FRPM Exposure to Non-FRPM

N

Percent Enrolled Latino

Percent Enrolled White

Percent Enrolled FRPM Dissimilarity

Racial Entropy

[1] [2] [3] [4] [5] [6] [7]

Universe counts nationwide A. All districts 1998 13,215 .61 (.38) .33 (.33) 50,529 .14 (.25) .64 (.35) .28 (.30) .65 (.33) .38 (.24) 2010 13,849 .61 (.37) .45 (.25) 53,636 .22 (.28) .53 (.37) .52 (.29) .56 (.32) .46 (.23) B. All districts with enrollment >10% Latino 1998 2,321 .50 (.26) .31 (.27) 13,793 .44 (.30) .36 (.28) .38 (.34) .46 (.29) .58 (.19) 2010 4,277 .47 (.27) .38 (.22) 25,568 .43 (.29) .35 (.27) .61 (.27) .45 (.28) .59 (.20) C. Nation’s 10 poorest districts (FRPM) >10% Latino enrollment 1998 10 .07 (.05) .16 (.04) 1,097 .58 (.31) .11 (.17) .76 (.23) .52 (.32) .65 (.36) 2010 10 .05 (.35) .25 (.10) 1,283 .66 (.30) .08 (.16) .68 (.25) .63 (.31) .59 (.35) Sampled units hosting ECLS-K kindergartners D. All sampled districts 1998 442 .65 (.41) .37 (.32) 1,297 .18 (.26) .51 (.37) .37 (.34) .54 (.34) .40 (.25) 2010 419 .51 (.32) .48 (.25) 1,086 .27 (.28) .44 (.33) .54 (.29) .49 (.31) .50 (.23) E. All sampled districts with enrollment > 10% Latino 1998 148 .37 (.39) .34 (.26) 503 .43 (.28) .32 (.27) .47 (.36) .44 (.27) .56 (.20) 2010 239 .36 (.25) .39 (.21) 668 .42 (.27) .33 (.26) .59 (.27) .42 (.27) .59 (.20)

Note. FRPM = free or reduced-price meals; ECLS-K = Early Childhood Longitudinal Study.

OCTOBER 2019 413

Turning to school-level indicators of segregation, we see a jump in the share of elementary students nationwide of Latino heritage, rising from 14% in 1998 to 22% in 2010, on average (row A), but no appreciable change for districts with 10% Latino enrollment (row B). The rise in FRPM pupils is larger, climbing from 38% to 61% among schools in districts with 10% Latino children or more.

The dissimilarity index fell about one-fourth SD, and entropy climbed about one-third SD—both measures indicating a more even racial distribution of students within the nation’s average elementary school. This is consistent with the rise in integration among non-White populations among neighborhoods reported above (Reardon & Owens, 2014). Yet the gains we observed are not apparent for elementary schools situated in districts with at least 10% Latino enrollment. Patterns are quite similar for schools from which ECLS-K children were sampled (rows D and E).

Differences for Latino Subgroups and Neighborhoods (RQ2) Next, we describe school contexts experienced by individual kin- dergartners, focusing on Latino subgroups, drawing on the ECLS-K data. Table 2 reports changing indicators of Latino chil- dren’s exposure to White or middle-class peers. We see that the percentage of students, White, for the mean Latino kindergartner declined from 37% to 30% between 1998 and 2010 (p < .05).

We observe a corresponding rise in the share of students identi- fied as English learners, climbing from 23% to 33% (p < .01). This confirms that rising concentrations of children from poor families are not merely artifacts of liberalized FRPM eligibility. Entropy scores increased over the period for the mean kindergart- ner, suggesting a more even distribution among multiple ethnic groups.

The mean Latino kindergartner experienced deteriorating neighborhood conditions in terms of falling household income and rising poverty. Median income fell from $56,291 in 1998 to $52,348 in 2010 (current dollars), likely due to the post-2007 recession. Yet educational attainment rose for adults in the mean Latino child’s tract: 42% of all adults reporting some college in 1998, rising to 47% in 2010. Educational aspirations may oper- ate somewhat independently of economic shocks, a hopeful finding for educators and policy makers who endeavor to com- bat segregation.

Table 3 reveals sharply differing contexts for Latino children, depending on maternal nativity. The share of peers, White, dropped from 47% in 1998 to 37% in 2010 for the mean Latino kindergartner with a native-born mother (p < .05). These shares equaled 31% and 25%, respectively, for the corresponding child with a foreign-born mother. Both subgroups of children entered schools with rising shares of English learners: climbing from 14% to 23% for native-born, and 30% to 39% for foreign-born over the period (p < .05 in both cases). Children of foreign-born

Table 2 Descriptive Statistics for Census Tracts and Schools in Which Sampled Families Reside, Split by Ethnicity, for

1998 and 2010 Cohorts (Means and SDs Reported)

1998 2010

All White Latino All White Latino

Schools Ethnicity and segregation Percent enrollment, White 0.68

(0.31) 0.80

(0.21) 0.37

(0.32) 0.59

(0.31) 0.73

(0.22) 0.30

(0.29) Percent enrollment, Latino 0.13

(0.22) 0.07

(0.11) 0.46

(0.32) 0.19

(0.25) 0.11

(0.14) 0.53

(0.33) Dissimilarity 0.67

(0.30) 0.76

(0.25) 0.53

(0.32) 0.58

(0.30) 0.66

(0.27) 0.56

(0.31) Entropy 0.38

(0.23) 0.33

(0.21) 0.48

(0.22) 0.46

(0.23) 0.44

(0.22) 0.48

(0.24) Social class Percent enrollment, eligible for subsidized meals 0.28

(0.30) 0.21

(0.24) 0.43

(0.35) 0.43

(0.30) 0.35

(0.26) 0.59

(0.33) Language Percent enrollment, English learners 0.06

(0.15) 0.02

(0.07) 0.23

(0.27) 0.12

(0.21) 0.06

(0.13) 0.33

(0.29) Tracts Median household income ($) 65,887

(27,346) 70,465

(27,060) 56,291

(24,739) 61,760

(27,594) 66,795

(27,608) 52,348

(22,398) Percent population in poverty 11.7

(10.08) 8.78

(6.88) 17.83

(12.19) 14.38

(11.68) 11.24 (8.98)

20.44 (13.86)

Percent population with some college or more 51.3 (17.96)

54.28 (16.84)

41.82 (19.12)

57.29 (17.37)

60.95 (15.59)

47.23 (18.68)

N of matched families 7,219 4,957 964 8,627 5,372 1,468

414 EDUCATIONAL RESEARCHER

mothers were more highly concentrated in urban enclaves and new destinations, relative to peers with native-born mothers residing in all other tracts.

Median household income held steady for the average Latino kindergartner with a native-born mother but fell sharply for the mean peer with a foreign-born mother (falling by $7,721, p < .05). School attainment climbed, most markedly in neighbor- hoods of foreign-born mothers, even in the recession’s wake (Table 4).

Which Latino Children Sort Into Segregated Schools? Given the wide diversity of Latino families, we asked whether certain markers of social class hold explanatory power in account- ing for variation in levels of school segregation. Results appear in Table 5, when estimating three measures of segregation, regress- ing on social-class features of Latino families. Column 1 shows

that home language is a major driver of the share of the average kindergartner’s peers who are Latino. This percentage is 17 points higher for Latino kindergartners of Spanish-speaking par- ents, compared with those from English-speaking homes. Family income, a second marker of class position, is negatively related to entering a school with higher concentrations of Latino peers.

Robust results also appear when estimating the share of enrollment, FRPM eligible (column 2). Beyond household income and home language effects, Latino children whose par- ents completed some college entered elementary schools with lower shares of poor peers. This helps to explain why integration of children from poor and middle-class homes improved over the period. We earlier saw that the average Latino child resided in a neighborhood with higher educational attainment in 2010, relative to 1998. The regression finding now shows that Latino children with better educated parents entered schools with lower shares of FRPM peers. To the extent that low-income Latino

Table 3 Change Across Cohorts, 1998 to 2010, for Tracts and Schools in Which Sampled Native or

Foreign-Born Latina Mothers Reside (Means and SDs Reported)

1998 2010 Change Across Cohorta

Variables Native-Born

Latino Foreign-Born

Latino Native-Born

Latino Foreign-Born

Latino Native-Born

Latino Foreign-Born

Latino

Schools Ethnicity and segregation Percent enrollment, White 0.47

(0.31) 0.31

(0.30) 0.37

(0.30) 0.25

(0.27) –0.10* –1.98

–0.06 –1.48

Percent enrollment, Latino 0.37 (0.30)

0.53 (0.32)

0.45 (0.33)

0.58 (0.31)

0.08 1.52

0.05 1.06

Dissimilarity 0.49 (0.31)

0.55 (0.32)

0.55 (0.30)

0.56 (0.32)

0.06 0.77

0.01 0.22

Entropy 0.49 (0.20)

0.47 (0.24)

0.50 (0.24)

0.47 (0.24)

0.01 0.16

0.00 –0.02

Social class Percent enrollment, eligible for subsidized

meals 0.35

(0.30) 0.49

(0.36) 0.49

(0.33) 0.66

(0.31) 0.14* 2.55

0.17* 2.5

Language Percent enrollment, English learners 0.14

(0.22) 0.30

(0.29) 0.23

(0.26) 0.39

(0.29) 0.09* 2.02

0.09* 2.04

Tracts Median household income ($) 58,677

(26,153) 54,558

(23,531) 57,006

(24,490) 49,581

(20,576) –1,671 –0.47

–4,977† –1.74

Percent population in poverty 15.82 (11.48)

19.29 (12.49)

17.95 (13.25)

21.92 (14.01)

2.13 0.99

2.63 1.23

Percent population with some college or more 46.31 (18.78)

38.55 (18.72)

51.61 (19.37)

44.62 (17.77)

5.30† 1.8

6.07* 2.47

Immigrant destinationsb

Traditional destinations 0.31 (0.46)

0.43 (0.50)

0.38 (0.49)

0.43 (0.50)

F(1.55, 834.95) F(1.61, 771.55)

New destinations 0.29 (0.45)

0.41 (0.49)

0.25 (0.44)

0.37 (0.48)

=0.53 =0.52

N of matched families 411 553 564 904

aCross-cohort difference in first rows and t-statistics in second rows. Significant differences in t-tests for independent samples, cross cohort. bTest for independence between immigrant destination and cohort is conducted. The Pearson χ2 statistic is corrected for survey design and converted into an F statistic. †p < .10. *p < .05.

OCTOBER 2019 415

families migrate into middle-class communities, this helps to explain improving economic integration.

Discussion

A half century of research details the corrosive effects of segregat- ing certain groups in separate schools, distant from White or middle-class peers (e.g., Cook, 1984; Johnson, 2019; Reardon & Owens, 2014). Our results verify how levels of racial and eco- nomic segregation have been slow to move in recent decades for young Latino children, with notable gains in exposure to mid- dle-class (often fellow Latino) peers. At the moment these chil- dren enter school, most face highly segregated settings, especially the offspring of immigrant parents.

Equity-minded reformers often assume they can tinker with the curriculum, the niceties of testing, or performance standards. Yet racial segregation, as a deep-seated structural constraint, mir- rors disparities in school funding, access to quality teachers, and monistic forms of knowledge that remain insensitive to cultural variety (Vasquez Heilig, Khalifa, & Tillman, 2014). Our results confirm this deeply institutionalized and racially arranged social order.

A pair of contextual dynamics remain key: The share of ele- mentary pupils of Latino heritage continues to grow in many districts, both in cities and out in diversifying suburbs. This means that rising percentages of children from low-income fami- lies populate the nation’s schools. Then, we find that Latino

students became more racially segregated over the period—less likely to interact with White peers in the same schools—for dis- tricts enrolling at least 10% Latino pupils. The nation’s 10 poor- est districts, enrolling at least 50,000 students, already quite racially segregated in 1998, backslid even further by 2010.

The textured child and family-level data show that local gains in economic integration may be driven in part by rising educa- tional attainment in neighborhoods that are increasingly settled by young Latino families. Even as many neighborhoods popu- lated by children of immigrants felt a sizeable decline in mean household income during the recession, school attainment of resident adults continued to climb. And our regression accounts of variation in school segregation show that Latino kindergart- ners enter less economically segregated schools when their par- ents are better educated. We also saw how elementary schools displayed a more even distribution among multiple racial groups (entropy) over time. This widening pluralistic blend of popula- tions may advance economic integration as well.

A less optimistic hypothesis stems from the fact that some middle-class families in the recession’s wake, at times losing their homes, fell into poorer neighborhoods. The net worth of Latino households fell from $23,600 to $13,700 (42%) between 2007 and 2013, compared with the decline for Whites, from $192,500 to $141,900 (26%; Kochhar & Fry, 2014). Rising economic integration may have occurred ironically via the downward mobility of Latino families. Future research is required to learn whether upward mobility or suburban migration of second and

Table 4 Descriptive Statistics at the Mother and Family Level by Ethnicity and Nativity Split by Cohort,

1998–2010 (Means and SDs Reported)

1998 2010 Change 1998 to

2010a

Variable All White

Native- Born

Latino

Foreign- Born

Latino All White

Native- Born

Latino

Foreign- Born

Latino

Native- Born

Latino

Foreign- Born

Latino

Social class Yearly family income ($) 82,263

(77,738) 93,948

(81,687) 66,855

(51,201) 45,425

(54,309) 81,211

(68,618) 95,812

(69,165) 65,403

(60,450) 37,704

(41,236) –1,452

–0.3 –7,721* –2.13

Maternal education Less than high school 0.11

(0.32) 0.06

(0.23) 0.18

(0.38) 0.47

(0.50) 0.11

(0.32) 0.04

(0.21) 0.15

(0.36) 0.46

(0.50) –0.03 –0.64

–0.01 –0.21

High school diploma 0.31 (0.46)

0.29 (0.46)

0.38 (0.49)

0.29 (0.45)

0.20 (0.40)

0.17 (0.37)

0.27 (0.45)

0.29 (0.46)

–0.11** –3.46

0.00 0.18

Some college 0.33 (0.47)

0.35 (0.48)

0.34 (0.47)

0.17 (0.38)

0.33 (0.47)

0.34 (0.47)

0.38 (0.48)

0.16 (0.36)

0.04 0.98

–0.01 –0.43

Four-year degree or more 0.25 (0.43)

0.30 (0.46)

0.10 (0.29)

0.08 (0.26)

0.36 (0.48)

0.44 (0.50)

0.20 (0.40)

0.09 (0.29)

0.10** 3.39

0.01 0.73

Language Non-English home

language 0.09

(0.29) 0.01

(0.10) 0.12

(0.33) 0.76

(0.43) 0.13

(0.34) 0.01

(0.11) 0.10

(0.30) 0.78

(0.42) –0.02 –0.89

0.02 0.67

Household structures and activity Single-parent family 0.17

(0.38) 0.12

(0.33) 0.19

(0.40) 0.16

(0.37) 0.19

(0.40) 0.14

(0.35) 0.23

(0.42) 0.18

(0.38) 0.04 1.11

0.02 0.66

aCross-cohort difference in first rows and t-statistics in second rows. Significant differences in t-tests for independent samples, cross cohort. *p < .05. **p < .01.

416 EDUCATIONAL RESEARCHER

later-generation Latinos improves the economic integration of young children.

It is encouraging that class-based integration widened for elementary students, despite the recession, although this gain was not observed in the nation’s 10 largest and poor districts. The mean Latino child resided in a tract with higher educational attainment in 2010, compared with the corresponding child in 1998. One explanation is that many Latino families exited tradi- tional urban enclaves over the period. Other studies reveal how attainment levels climbed for adult Latinos, independent of resi- dential movement, especially for young women.12 The educa- tional aspirations of many Latinos seem quite resilient, even when residing in economically fragile communities.

Our findings mesh with other work showing a significant postrecession recovery for native-born Latinos, along with sus- tained upward mobility when compared with first-generation Latinos (Tran & Valdez, 2015). Median income climbed for Latino households nationwide from $46,046 in 2006, on the eve of recession, to $50,486 in 2017 (9.6% gain in 2017 dollars).

This recovery for Whites moved from $63,892 to $68,145 (6.7% gain; U.S. Census Bureau, 2018b). Still, lower and unsta- ble incomes have persisted for foreign-born Latino parents (Gennetian, Rodrigues, Hill, & Morris, 2015).

Our fine-grain data on kindergartners detail a similar trend of increasing segregation as Latino kindergartners enter school. This remains troubling, as governments invest more heavily in preschool, aiming to narrow disparities in school readiness. The average Latino kindergartner entered a school with enrollments, 37% White in 1998, falling to 30% by 2010; the share of enroll- ment, Latino, climbed from 46% to 53%. The percentage of FRPM-eligible peers jumped from 43% to 59%, climbing more than the increment tied to liberalized program eligibility. Future research should build alternative measures of economic integra- tion, as the federal definition of FRPM-eligible continues to shift (Domina et al., 2018).

A pressing question remains of whether and how policy mak- ers and education leaders can effectively reduce the segregation of Latino students, independent of demographic and economic

Table 5 Estimating Child Composition, Racial and Class Exposure in School From Family Socioeconomic Status (SES)

Covariates for Latinos Families With Cross-Cohort Interactions (Unstandardized Coefficients and SEs Reported)

Variable Percent, Latino Percent, Free or Reduced-Price

Meal (FRPM) Eligible Percent, English Learners

Yearly family incomea –.0011*** (.0002)

–.0013*** (.0002)

–.0006*** (.0002)

Maternal education Some college –.0595

(.0329) –.0804** (.0303)

–.0307 (.0285)

Four-year degree or more –.0884 (.0629)

–.094* (.0404)

–.0452 (.0444)

Non-English home language .1754*** (.0276)

.131*** (.0320)

.1644*** (.0263)

Single-parent .0069 (.0293)

–.0174 (.0373)

.0309 (.0288)

Year .065 (.0389)

.1856*** (.0381)

.1084** (.0390)

Interaction with year Yearly family income .0003

(.0003) –.0001 (.0003)

–.0001 (.0002)

Maternal education Some college .0304

(.0409) –.0105 (.0371)

–.0205 (.0359)

Four-year degree or more .0078 (.0748)

–.1233* (.0521)

–.0547 (.0530)

Non-English home language –.0468 (.0351)

–.0411 (.0378)

–.0158 (.0341)

Single-parent .0571 (.0370)

.0595 (.0423)

.0228 (.0360)

Constant .4151*** (.0280)

.4191*** (.0303)

.1674*** (.0316)

F for model F (11, 723) = 15.91 F (11, 723) = 24.48 F (11, 723) = 17.40 R 2 .15 .23 .18

aFamily income was centered around $36,720, which is close to the median of family income in year 1998 and 2010, and divided by $1,000. Thus, a unit increase corresponds to $1,000 increase in family income. *p < .05. **p < .01. ***p < .001.

OCTOBER 2019 417

forces. These state and local actors face stiff headwinds: the steady growth of Latino populations, including poor and immi- grant families. And many Latino parents continue to exit tradi- tional urban enclaves, migrating to aging suburbs and exurbs, where the reception offered by educators and fellow residents ranges from ill-prepared to downright hostile.

At the same time, an expanding Latino middle class is mov- ing to economically better-off neighborhoods and schools that are more integrated, at least along social-class lines. Future research should further untangle such local variations in the Latino experience and the extent to which racialized markers continue to segregate Latino children in certain schools. The interplay of demographic trends and educational policies—the incursion of market-oriented reforms, for instance—offers another ripe area of study (Fiel, 2015).

How do these trends inform contemporary policy efforts and district-level strategies for lessening the corrosive effects of segre- gation? First, the independence of economic integration vis-à-vis racial integration offers encouraging news for Latino families in some locales. We must learn more, however, as to whether mid- dling Latino parents and children, in turn, benefit from higher quality schools as they enter less segregated settings. Second, the fact that adult educational levels in predominantly Latino neigh- borhoods continue to climb—despite the recession’s onset— offers encouraging news as well.

Third, one hopes that upward mobility and school integration benefit Latino families and young children when the economy grows. Our findings confirm how children of native-born parents do enjoy more integrated schools and materially better-off neigh- borhood, compared with immigrant peers. But under what eco- nomic or schooling conditions do young offspring of foreign-born mothers share in these gains? And education leaders—aiming to serve increasingly diverse Latino children—must discern how the conditions of immigrant versus later-generation families can dif- fer dramatically, even when attending the same school.

Some education leaders have persevered in their efforts to racially integrate students among schools within their districts, for instance by expanding magnet or dual-language schools (Riel, Parcel, Mickelson, & Smith, 2018). District-managed choice efforts seek to balance parental preferences with the com- mon cause of racial integration, as in Cambridge or San Francisco. Other district leaders, when direct integration efforts fail, at least attempt to equalize the allocation of dollars or qual- ity teachers among racially segregated schools (Johnson, 2019; Schwartz, Rubenstein, & Stiefel, 2009; Fuller & Lee, 2018).

Still, our findings highlight how differing fertility rates among groups, out-migration from old urban enclaves, and long-term progress in educational attainment will shape whether Latino parents find better integrated schools in their neighbor- hoods. Local educators and policy makers must candidly con- front these deep-seated structural forces that shape varying degrees of segregation, then look for institutional openings to better integrate Latino children, widening their opportunities along lines of race and social class.

ORCID ID

Yoonjeon Kim https://orcid.org/0000-0002-1214-3549

NOTES

This article stems from the Latino Contexts and Early Development Project, originally funded by the Packard and Spencer foundations, along with Berkeley’s Institute of Human Development. Mr. Fuller coordinated this study, and Ms. Kim led the quantitative analysis. Ms. Galindo codirects the project. Team members are listed alphabetically. Heartfelt thanks to Ann Owens and Sean Reardon for their guidance and to Elise Castillo, who compiled much of our data, along with Gail Mulligan at the National Center for Education Statistics for her gra- cious fielding of endless questions.

1Iceland and Nelson (2008) replicated Massey’s finding, while showing how Puerto Rican and Black Latinos have not shared the upward mobility or integrated settings more commonly achieved by later-generation Latinos.

2A third position accents how Latinos increasingly raise their chil- dren in multiethnic suburbs or gentrifying urban centers with Asian, Black, and White neighbors (e.g., Bean et al., 2015).

3The Los Angeles School of suburban dispersal of families pio- neered much of this work, in contrast to the Chicago School (e.g., Soja, 2014). More recently, urban-ecological thinking has been advanced by students of Baltimore (Grove, Cadenasso, Pickett, Machlis, & Burch, 2015).

4White and Sassler (2000) earlier showed how the class position of Latino families, not surprisingly, predicts the economic and demo- graphic features of the neighborhoods in which they settle. Younger Latino families appear more likely to migrate to new destinations, often finding low-wage jobs (Donato, Tolbert, Nucci, & Kawano 2008).

5NCES defines elementary (or primary) schools as starting at pre-K or kindergarten and not having grades beyond eighth. Districts solely hosting charter schools and high school districts were excluded.

6Means are nationally representative for all and Latino children when properly weighted. A nearly identical protocol was followed to collect data from families and schools for the tandem cohorts (G. Mulligan, National Center for Education Statistics, personal commu- nication providing guidance on representativeness of Latino subsam- ples in ECLS-K for 1998 and 2010 cohorts, 2015; Rock & Pollack, 2002).

7Sampled ECLS-K parents were not asked about nativity until the first-grade home visit, resulting in lost cases relative to the kindergarten sample.

8The family eligibility cut-point, 185% of the federal poverty line, remained constant over the period. Yet Washington moved to liberal- ize eligibility and the ease of certifying families (Food and Nutrition Service, 2009; Hoffman, 2012). Between 1998 and 2004, the share of children deemed eligible for FRPM climbed 3%, while the child poverty rate declined a like amount. FRPM eligibility rose from 38% to 48% nationwide, 2000–2010 (NCES, 2018).

9A linguistically isolated household hosts no member, 14 years or older, who speaks English fluently, as defined by the census.

10We estimated whether the mother’s Latino ethnicity, as racialized marker, further contributed to the intensity of segregated schools, after propensity-matching of Latino and White families on social-class mark- ers. Results available from the authors.

11These 10 include large urban districts such as Dallas, Houston, Los Angeles, Miami, Milwaukee, and San Diego.

12The share of Latina women, age 25–29, who attained a 4-year college degree nationwide climbed from 66% to 74% between 1998 and 2010. These percentages equaled 60% to 66% for Latino males (U.S. Census Bureau, 2018a).

13For tract and neighborhood descriptives, sample weights C1PW0 (1998) and W1P0 (2010) were used; and for school descrip- tives, weights C1CPTW0 (1998) and W1T0 (2010) were applied. We

418 EDUCATIONAL RESEARCHER

use stratum and PSU identifiers that correspond to sample weights to compute variance estimates based on Taylor series methods.

14Since families in a school possess the same value on the depen- dent variable, error terms may be correlated. To correct for this, we ran models with cluster-robust standard errors (Crowder & South, 2008). Because Stata does not provide options for specifying a stratification variable under robust-cluster options, we specify sample weights using the pweight option.

REFERENCES

Alba, R., & Foner, N. (2015). Strangers no more: Immigration and the challenges of integration in North America and Western Europe. Princeton, NJ: Princeton University Press.

Alba, R., Kasinitz, P., & Waters, M. (2011). The kids are (mostly) alright: Second generation assimilation. Social Forces, 89, 763–773.

Bassok, D., Gibbs, C., & Latham, S. (2018). Preschool and chil- dren’s outcomes in elementary school: Have patterns changed nationwide between 1998 and 2010? Child Development, Online. Retrieved from https://onlinelibrary.wiley.com/doi/abs/10.1111/ cdev.13067

Bean, F., Brown, S., & Bachmeier, J. (2015). Parents without papers: The progress and pitfalls of Mexican American integration. New York: Russell Sage Foundation.

Brazil, N. (2016). The effects of social context on youth outcomes: Studying neighborhoods and schools simultaneously. Teachers College Record, 118, 1–30.

Cook, T. (1984). What have Black children gained academically from school integration? In T. Cook, D. Armor, R. Crain, & N. Miller, et al. (Eds.), School desegregation and Black achievement (pp. 6–42). Washington, DC: National Institute of Education.

Crosnoe, R. (2005). Double disadvantage or signs of resilience? The elementary school contexts of children from Mexican immigrant families. American Educational Research Journal, 42, 269–303.

Crowder, K., & South, S. (2008). Spatial dynamics of white flight: The effects of local and extralocal racial conditions on neighborhood out-migration. American Sociological Review, 73, 792–812.

Domina, T., Pharris-Clurej, N., Penner, A., Penner, E., Brummet, Q., Porter, S., & Sanabria, T. (2018). Is free and reduced-price lunch a valid measure of educational disadvantage? Educational Researcher, 47, 539–555.

Donato, K., Tolbert, C., Nucci, A., & Kawano, Y. (2008). Changing faces, changing places: The emergence of new non-metropolitan immigrant gateways. In New faces in new places: The changing geog- raphy of American immigration (pp. 75–98). New York: Russell Sage Foundation.

Dondero, M., & Muller, C. (2012). Social stratification in new and established Latino destinations. Social Forces, 91, 477–502.

Fiel, J. (2013). Decomposing social resegregation: Social closure, racial imbalance, and racial isolation. American Sociological Review, 78, 828–848.

Fiel, J. (2015). Closing ranks: Closure, status competition, and school segregation. American Journal of Sociology, 121, 126–170.

Food and Nutrition Service. (2009). Direct certification in the national school lunch program: State progress in implementation. Washington, DC: Author.

Fry, R. (2011). The Hispanic diaspora and the public schools: Educating Hispanics. In D. Leal & S. Trejo (Eds.), Latinos and the economy: Integration and impact in schools, labor markets, and beyond (pp. 15–36). New York: Springer.

Fuller, B., Bein, E., Kim, Y., & Rabe-Hesketh, S. (2015). Differing cog- nitive trajectories of Mexican American toddlers: The role of class,

nativity, and maternal practices. Hispanic Journal of Behavioral Sciences, 37, 139–169.

Fuller, B., & Lee, J. (2018). Progress in Los Angeles schools: Rigor, fairness, and engaging families. Los Angeles: United Way.

Gándara, P., & Mordechay, K. (2017). Demographic change and the new (and not so new) challenges for Latino education. Educational Forum, 81, 148–159.

Gennetian, L., Rodrigues, C., Hill, H., & Morris, P. (2015). Low and stable income: Comparisons among Hispanic children, from 2004 through the period following the Great Recession. Washington, DC: Child Trends.

Gibson-Davis, C., & Percheski, C. (2018). Children and the elderly: Wealth inequality among America’s dependents. Demography, 55, 1009–1032.

Grove, J., Cadenasso, M., Pickett, S., Machlis, G., & Burch, W., Jr. (2015). The Baltimore school of urban ecology: Space, scale, and time for the study of cities. New Haven, CT: Yale University Press.

Henig, J., Hula, R., Orr, M., & Pedescleaux, D. (2001). The color of school reform: Race, politics, and the challenge of urban education. Princeton, NJ: Princeton University Press.

Hirschman, C., & Massey, D. (2008). Places and peoples: The new American mosaic. New York: Russell Sage Foundation.

Hoffman, L. (2012). Free and reduced-price lunch eligibility data in ED-Facts: A white paper on current status and potential changes. Washington, DC: U.S. Department of Education, Office of Planning, Evaluation and Policy Development.

Iceland, J., & Nelson, K. (2008). Hispanic segregation in metro America: Exploring the multiple forms of spatial assimilation. American Sociological Review, 73, 741–765.

Iceland, J., & Sharp, G. (2013). White residential segregation in U.S. metropolitan areas: Conceptual issues, patterns, and trends from the U.S. census, 1980 to 2010. Population Research & Policy Review, 32. doi:10.1007/s11113-013-9277-6.

Johnson, R. (2019). Children of the dream: Why school integration works. New York: Basic Books.

Kochher, R., & Fry, R. (2014). Wealth inequality has widened along racial, ethnic lines since end of Great Recession. Washington, DC: Pew Research Center.

Lichter, D., Parisi, D., & Taquino, M. (2015). Spatial assimilation in U.S. cities and communities? Emerging patterns of Hispanic segre- gation from Blacks and Whites. Annals of the American Academy of Political and Social Sciences, 660, 36–56.

Lichter, D., Parisi, D., Taquino, M., & Grice, S. (2010). Residential segregation in new Hispanic destinations: Cities, suburbs, rural communities compared. Social Science Research, 39, 215–230.

Logan, J. (2004). Resegregation in American public schools? Not in the 1990s. Albany, NY: Lewis Mumford Center.

Massey, D., & Denton, N. (1988). Suburbanization and segregation in U.S. metropolitan areas. American Journal of Sociology, 94, 592–626.

National Center for Education Statistics (NCES). (2001). ECLS-K base year public-use data file and user’s manual (2001-029). Washington, DC: United States Department of Education.

National Center for Education Statistics (NCES). (2018). Table 204.10. Number and percentage of public school students eligible for free and reduced-price lunch by state: Selected years, 2000-01 through 2013-14. Washington, DC: Author. Retrieved from https://nces. ed.gov/programs/digest/d17/tables/dt17_216.60.asp

Noguera, P., Pierce, J., & Ahram, R. (Eds.). (2015). Race, equity, and education: Sixty years from Brown. New York: Springer.

Orfield, G., & Lee, C. (2007). Historic reversals, accelerating resegrega- tion, and the need for new integration strategies. Los Angeles: UCLA Civil Rights Project.

OCTOBER 2019 419

Owens, A. (2010). Neighborhoods and schools as competing and rein- forcing contexts for educational attainment. Sociology of Education, 83, 287–311.

Owens, A. (2017). Racial residential segregation of school-age children and adults: The role of schooling as a segregating force. Russell Sage Journal of the Social Sciences, 3. Retrieved from https://www . rsfjournal.org/doi/full/10.7758/RSF.2017.3.2.03

Owens, A. (2018). Income segregation between school dis- tricts and inequality in students’ achievement. Sociology of Education, 91. Retrieved from http://journals.sagepub.com/doi/ pdf/10.1177/0038040717741180

Perreira, K., Fuligni, A., & Potochnick, S. (2010). Fitting in: The roles of social acceptance and discrimination in shaping the academic motivations of Latino youth and in the U.S. southeast. Journal of Social Issues, 66, 131–153.

Reardon, S., & Bischoff (2011). Income inequality and income segre- gation. American Journal of Sociology, 116, 1092–1153.

Reardon, S., & Owens, A. (2014). 60 years after Brown: Trends and consequences of school segregation. Annual Review of Sociology, 40, 199–218.

Reardon, S., & Portilla, X. (2016). Recent trends in income, racial, and ethnic school readiness gaps in kindergarten entry. AERA Open, 2. Retrieved from http://journals.sagepub.com/doi/full/10.1177/ 2332858416657343

Reeves, R., & Busette, C. (2018). The middle class is becoming race- plural, just like the rest of America. Washington, DC: Brookings Institution. Retrieved from https://www.brookings.edu/podcast- episode/the-changing-identity-of-americas-middle-class/

Riel, R., Parcel, T., Mickelson, R., & Smith, S. (2018). Do magnet and charter schools exacerbate or ameliorate inequality? Sociological Compass, 12.

Rock, D., & Pollack, J. (2002). Early Childhood Longitudinal Study- Kindergarten Class of 1998–99: Psychometric report for kindergarten through first grade (NCES 2002–05). Washington, DC: National Center for Education Statistics.

Schwartz, A., Rubenstein, R., & Stiefel, L. (2009). Why do some schools get more and others less? An examination of school-level funding in New York City. New York: Institute for Education and Social Policy.

Soja, E. (2014). My Los Angeles: From urban restructuring to regional urbanization. Los Angeles: University of California Press.

South, S., Crowder, K., & Chavez, E. (2005). Exiting and entering high-poverty neighborhoods: Latinos, Blacks and Anglos com- pared. Social Forces, 84, 873–900.

Stepler, R., & Lopez, H. (2016). U.S. Latino population growth and dis- persion has slowed since the onset of the Great Recession. Washington, DC: Pew Research Center.

Stroub, K., & Richards, M. (2013). From resegregation to reintegra- tion: Trends in the racial/ethnic segregation of metropolitan public schools, 1993–2009. American Educational Research Journal, 50, 497–531.

Tienda, M., & Fuentes, N. (2014). Hispanics in metropolitan America: New realities and old debates. Annual Review of Sociology. Retrieved from doi:10.1146/annurev-soc-070913-043315

Tran, V., & Valdez, N. (2015). Second-generation decline or advan- tage? Latino assimilation in the aftermath of the Great Recession. International Migration Review, 50, 1–36.

U.S. Census Bureau. (2011). American Community Survey and the Decennial Census of Population and Housing, 2010. Washington, DC: Author. Retrieved from http://www.census.gov/acs/www/

U.S. Census Bureau. (2018a). Current Population Survey historical time series tables, 2017. Retrieved from https://www.census.gov/data/

tables/time-series/demo/educational-attainment/cps-historical- time-series.html

U.S. Census Bureau. (2018b). Current Population Survey income trends. Table A1. Households be total money income, race, and Hispanic ori- gin of householder, 1967-2017. Retrieved from https://www.census. gov/data/tables/2018/demo/income-poverty/p60-263.html

Van Hook, J., Bean, F., Bachmeier, J., & Tucker, C. (2014). Recent trends in coverage of the Mexican-born population of the United States. Demography, 51, 699–726.

Vasquez Heilig, J., Brown, K., & Brown, A. (2012). The illusion of inclusion: A critical race theory analysis of race and standards. Harvard Educational Review, 82, 403–424.

Vasquez Heilig, J., Khalifa, M., & Tillman, L. C. (2014). High-stakes reforms and urban education. In H. R. Milner & K. Lomotey (Eds.), Handbook of urban education (pp. 523–537). New York, NY: Routledge.

AUTHORS

BRUCE FULLER is a professor of education and public policy at the Graduate School of Education, University of California, Berkeley, CA 94720; [email protected]. He studies institutions, often the social and political organization of schooling and processes that contribute to the early learning and socialization of Latino children.

YOONJEON KIM is a postdoctoral scholar in the Graduate School of Education at the University of California, Berkeley, Center for the Study of Child Care Employment, University of California, Berkeley, CA 94704; [email protected]. Her research examines how macro- institutional factors shape the social organization of schools and learn- ing in the United States and cross-nationally.

CLAUDIA GALINDO is an associate professor of educational policy in the College of Education at the University of Maryland, College Park, MD 20742; [email protected]. Her work focuses on the school experiences and achievement of Latino students, and the role of family and community supports.

SHRUTI BATHIA is a PhD student in social science methods at the Graduate School of Education, University of California, Berkeley, CA 94720; [email protected]. Her research centers on applica- tions of psychometrics and statistics to study human behavior.

MARGARET BRIDGES is a research scientist at the University of California, Berkeley Institute of Human Development, Berkeley, CA 94720; [email protected]. Her work focuses on how families and preschool contribute to young children’s social-emotional and cognitive development.

GREG J. DUNCAN is a distinguished professor in the School of Education at the University of California, Irvine, CA 92697; gduncan @uci.edu. His research focuses on economic mobility and the effects of poverty on children’s development.

ISABEL GARCÍA VALDIVIA is a PhD candidate in sociology at the University of California, Berkeley, Department of Sociology, Berkeley, CA 94720; [email protected]. Her research centers on migra- tion and inequalities faced by Latinx families, including education and legal status.

Manuscript received July 10, 2018 Revisions received October 22, 2018; January 8, 2019;

April 28, 2019; May 31, 2019 Accepted June 6, 2019

420 EDUCATIONAL RESEARCHER

Methods Appendix

Details on Segregation Indices Interaction index. The interaction (or exposure) index (e.g., Latino children’s exposure to Whites) is specified as (Massey & Denton, 1988; Owens, 2017; Reardon & Bischoff, 2011)

x y *

i=1

n P =

x X

y s

i i

i

⎡ ⎣⎢

⎤ ⎦⎥ ⎡

⎣ ⎢

⎤

⎦ ⎥∑ ,

where xi, yi, and si are the counts of Latinos, Whites, and the total student enrollment in elementary school i, respectively. X is the total elementary school enrollment of Latinos in the school district. The interaction index can be interpreted as the probability that a randomly drawn Latino child goes to the same school with a White child. Similarly, we calculated free or reduced-price meal (FRPM)–eligible children’s expo- sure to non-FRPM peers.

Racial diversity or entropy index. A school’s entropy (E) is specified as

E Q log Qrr

R

r

R= =1

1∑ ,

where R is the number of racial and ethnic groups in each school. Qr is the proportion of racial or ethnic group r. School entropy, which represents the extent of even distribution among the

groups within a school, varies from 0 (when the school contains only a single group) to 1 (when the racial groups in the school are evenly proportioned).

Estimating Levels of School Segregation for Latino Kindergartners All descriptive statistics are estimated using Stata 13 svy com- mands specifying stratification, sampling units, and survey weights.13 After descriptively breaking down differences in the segregation indices by parental nativity, home language, and other markers of social class, we estimate their independent con- tribution via ordinary least squares (OLS) regression. We interact a child-cohort dummy with social-class markers to see if their contributions are increasing over time, in the form

Y = + Year + X + X Year +j 0 1 ij pij 2

p

pij pij 2

p

pij ij ijβ β β γ ε∑ ∑ .

Here, Yj is the attribute of school j. Dummy Yearij indicates whether a child i in context j was sampled in 2010. X pij are family-level attributes of child i in school j. Therefore, β0 represents the mean proportion of students, Latino or in poverty (FRPM) in 1998, and β1 represents the change, 1998–2010, when all family background variables are equal to zero. βpij indicates the slopes of family variables in 1998, and γpij indicate the change in the effects of class markers on school ethnic and class composition (segregation) between 1998 and 2010.14