A Brief Review of Existing Research
    While most academic studies of immigrants and education have been ethnographic, there have been several important quantitative studies of immigrants and educational attainment, including Betts and Loftstrom (1998) and Vernez and Abrahamse (1996).1   Two important findings emerge from this literature.  First, immigrant children are at least as likely as native-born children to enroll in school, and, second, the educational attainment of immigrants is, in many respects, comparable to that of the native born. In addition, Betts and Lofstrom (1998) find some evidence that the success of the immigrants comes at the direct or indirect expense of the educational attainment of the native born. Put simply, immigrants might crowd out the native born in competition for educational resources or opportunities.
    The research on educational resources and immigrants has focused on the costs associated with limited English proficiency. Duncombe and Yinger (1997) and Downes and Pogue (1994), for example, find that increasing the representation of LEP students increases district costs.  Although, as discussed below, additional resources may be available for schools and districts serving LEP students, these resources may be insufficient to cover the additional costs, implying a decrease in resources available for educational programs.2
    Rivera-Batiz (1996) examines the impact of immigrants on schools per se, using New York City school level data to examine the determinants of passing rates on reading and math exams.3 He finds that the proportion of recent immigrants in a school has a positive impact upon outcomes while the proportion of LEP students has a negative impact. Our analysis builds upon this research, using richer data and investigates resources as well as performance.

Policy Context
    As detailed in Gershberg (2000), the number and proportion of immigrant and LEP students in New York City has grown since the 1980s, contributing significantly to school overcrowding in some neighborhoods and creating a public perception that the school system has “an immigrant problem” that the school system is poorly equipped to handle.
    Students in New York City are categorized as LEP eligible if the first language spoke in their home is not English and if they score below the 40th percentile on a test of English language skills.  LEP eligible students must, then, enroll in one of two kinds of programs either freestanding English as a Second Language (ESL) or Bilingual Education. ESL programs provide one to two pullout classes per day of training while subject courses are taught in English. Bilingual programs provide ESL training and subject classes taught in the students’ native languages. Bilingual programs may not be available at every school -- schools must provide bilingual education only if they have twenty or more students in the same grade speaking the same language.
Recent work by the New York City Board of Education (NYCBOE (2000)) found that the academic success of LEP students – measured, primarily, by their exit from ESL/Bilingual programs – depends critically upon the grade at which they enter the New York City public schools.4  Those entering in elementary school, especially kindergarten and first grade, do the best, followed by those entering in high school.  The implication is that the immigrant experience and the needs of immigrants differ significantly between elementary, middle and high school and our empirical investigation should treat these separately.5
Nearly all public policy aimed at immigrant students in New York City and State relates to teaching English and/or bilingual education. As an example, approximately $81 million in state aid was provided to fund ESL and bilingual education programs in 1996-97 Federal aid for assisting in the education of LEP students was approximately $23.5 million.  In contrast, there is a small Federal program, the Emergency Immigrant Education Program (EIEP), aimed at immigrants, per se, however, at approximately $5 million in funding in 1996-97, the EIEP is too small to have a great impact on educational resources. (Gershberg, 2000)
    New York City itself has little in the way of an articulated policy toward educating immigrants. There are, by now, seven “Newcomer” schools, which concentrate on teaching only new immigrants.  Interestingly, these have not arisen out of any organized City or State policy, but rather out of various “grassroots” efforts to create opportunities and appropriate educational programs for new immigrants.

Why Try to Differentiate Between Recent Immigrants and LEP?
    There is growing popular concern that we have “an immigrant problem” in U.S. public schools, and by extension in New York City’s schools as well. One piece of social evidence is the tone of the debate around the campaigns to end bilingual education led by silicon valley financier Ronald Unz. Much of the reliable research indicates, however, that the real problem is not so much an immigrant problem, but what Ruiz-de-Velasco and Fix (2000) call Long-Term LEPs. Zehr (2001c) describes long-term LEPs as “youths who have learned to function socially in an English-speaking environment but keep the LEP label for years because they can’t read or write well in English.” Data and trends reported in NYCBOE (2000) suggest strongly that the same is true for New York City. If the school system is concerned with reducing the number of Long-Term LEPs in the future, it should be concerned with three basic potential current “sources”: (1) current recent immigrants, (2) current non-recent immigrants; and (3) non-immigrant LEP students. This last group would be made up mostly of children of immigrants and Puerto Ricans.  While our empirical work presented below does not provide any immediate answers to this obviously important and difficult problem, we do provide insight that should prove beneficial in future research. First of all, we provide insight into the educational  environment of, and exposure to educational resources experienced by, recent immigrants. Our findings also suggest that programs aimed at recent immigrants should be different than those for non-recent immigrant LEP students, and different as well for students arriving and entering different levels of schooling. The first years after arrival are critical for recent immigrants. As Carola Suarez-Orosco, notes: “There are energies we could harness as a society, but we’re not. Kids come in with energy and quickly lose hope.”6 (Zehr, 2001c) We expect that the research presented here will be of use to policymakers charged with designing programs to reduce the Long-Term LEP population.

Data and Measures
    This study uses school-level data from the New York City Board of Education’s Annual School Reports (ASR) for 1996-1997 and 1997-1998 and School Based Expenditure Reports (SBER) for 1997-1998.7  The ASRs provide information on the test scores and demographic characteristics of students, as well as teacher characteristics. The SBERs contribute expenditure data, pupil-teacher ratio, and the percentage of students in part- and full-time special education. Variables capturing the interaction between socioeconomic and demographic characteristics were calculated based upon a student level data file provided by the NYCBOE’s Division of Assessment and Accountability only for elementary and middle schools.
While there are over 1100 public schools in New York City, several schools were excluded due to missing or incomplete data. The resulting sample includes 1,097 schools, and more than a million students.8 The final sample contains 691 elementary schools, 233 middle schools and 173 high schools.9
    School-level performance was captured by a test in reading proficiency (CTB) and in math proficiency (CAT).  In 1997-1998, average Normal Curve Equivalents (NCE) were reported for each school, however, in 1996-1997 only the percentage of students performing above the 50th percentile (based on a national sample) were reported.  We use data on test performance for the fifth grade for 1997-98 and the ‘lagged’ value of fourth grade performance (performance on the fourth grade test in 1996-97 for the same school) in our elementary school analyses. Our middle school analyses use eighth grade tests for 1997-98 and the lagged value of seventh grade tests (performance on the seventh grade test in 1996-97 for the same school).10  The absence of consistent performance data precludes a high school performance analysis.
Demographic data include the percent of immigrants who arrived in the United States within the past three years (recent immigrants), the percentage of students who are female, eligible for free lunch, limited English proficient (LEP), black, Hispanic or Asian.  Interaction variables include a breakdown of recent immigrants by race, limited English proficiency, and poverty, a breakdown of the ‘poor’ population (free lunch eligible) by race, and a breakdown of the LEP population by race and poverty.11  We use three resource measures, expenditure per pupil, pupil-teacher ratio and teacher education, the percent of teachers with a Masters degree.12
    Notice that our data describe only the population of recent immigrants - and not the population of students who are not native-born nor the ‘second generation’ population - the children of immigrants.  Thus, we will also analyze the LEP population in an effort to capture the larger group.  Unfortunately, this group misses the non-native born who are not LEP eligible – including, for example, Caribbean students – an important oversight.

Methods and Results
Statistical Portraits of Schools Attended by Recent Immigrants and LEP Students

    Enrollment-weighted means provide a portrait of the school attended by the ‘average student’ – or the average student of some particular group, such as immigrant students. Enrollment-weighted means differ from unweighted means due to differences in the characteristics of schools that are correlated with enrollment. Weighted means will, in turn, differ from one another to the extent that the distribution of immigrant (LEP) students differs from the distribution of pupils overall.  These statistics allow us to examine the extent to which the immigrant experience differs from the experience of the typical students.
We also use two conventional measures of segregation and racial composition – dissimilarity indices and exposure indices. Dissimilarity indices measure the percentage of all immigrants (or other group) who would have to change schools in order for the group to be evenly distributed across schools. The dissimilarity index is calculated as
D = 100*  xi /xi - yi /yi   /2
where xi  represents the number of immigrants in school i and yi represents the number of non-immigrants in school i.  D ranges from a low of zero, when immigrants and non-immigrants are distributed identically, to a high of 100, when immigrants are completely segregated – that is, there are no schools that include both immigrants and non-immigrants. For comparison purposes, we also calculate dissimilarity indices for LEP students and other demographic groups.
Exposure indices measure the degree of contact between immigrants and students in other socioeconomic groups.  The exposure of immigrants to students of type Y is calculated as:
EXY =  xi (yi/ti)/ xi =   [ (xi /xi )(yi/ti) ], where t is the total number of students, x represents the number of immigrants, y is the number of students in the comparison group, and i indexes schools.13 Put differently, EXY measures the percentage of the students of type Y in the school attended by the ‘average’ immigrant student.

    As shown in Table 1, New York City public schools include a significant proportion of immigrant students. Almost eight percent of the students in the average school are recent immigrants; almost 16 percent are LEP. Schools span the full range, however, in their representation of immigrant or LEP students - some schools have virtually no immigrant (LEP) students; others are almost entirely composed of recent immigrants (LEP).  Similar patterns emerge for other socioeconomic groups.  The average school is more than a third black, more than a third Hispanic, roughly ten percent Asian and sixteen percent white, and more than two thirds poor. The pupil-weighted means are substantively the same.14
    The exposure indices in Table 1 indicate that the typical immigrant is exposed to a different demographic mix of students than the typical New York City public school student.  The classmates of the typical immigrant are less likely to be black, more likely to be Asian and LEP. Further, almost 15 percent of their classmates are recent immigrants themselves.
The differences are starker for LEP students: the classmates of the typical LEP student include even fewer blacks, more poor students, and more Asian students. Fully half of their classmates are Hispanic and a quarter are LEP themselves.  The pattern differs somewhat across school levels. The difference between the exposure index for immigrants and the index for all students is narrowest for high schools and greatest for elementary schools.  As an example, at the elementary school level, immigrants are exposed to significantly fewer blacks while the differences at the high school level are insubstantial.

How segregated are immigrants and LEP students?
    As shown in Table 2, a vast majority of all students and a majority of the recent immigrants themselves attend schools that are less than 20 percent recent immigrant - only 59 schools are more than 20 percent immigrant. There are, in fact, only four schools in our database serving mostly immigrants; three of them are high schools.  Further, schools serving more immigrants also serve a greater proportion of LEP students, poor students, Hispanics and Asians but a smaller proportion of blacks.
The dissimilarity indices in Table 3 indicate that there is some segregation of immigrants. Roughly 32 percent of immigrant students would have to switch schools to create an even distribution across schools. But, this segregation is significantly milder than the segregation indicated by the higher dissimilarity indices for blacks, Hispanics, Asians, and even the poor. Interestingly, the segregation of immigrants is lowest in elementary schools and highest in middle schools even though in New York City, as elsewhere, the choice of elementary school is dictated largely by residential location.  Since most students attend local elementary schools that serve students residing in a geographically defined zone, segregation in elementary schools likely reflects patterns of residential location. One might expect residential segregation to translate into school segregation. At the middle and high school level, however, more choice is available and more students attend secondary schools outside of their neighborhoods, including specialized programs such as ‘newcomer’ schools aimed specifically at immigrants. Thus, segregation also reflects the choices and preferences of students and schools.  The increasing segregation of immigrants is in sharp contrast to the consistently declining segregation of blacks, Hispanics and Asians, poor students, and, even LEP students.

What are immigrants like demographically?
    Our analyses of the socioeconomic characteristics of immigrant students at the elementary and middle school levels yielded interesting results (Table 5). Perhaps most interesting is that the overlap between recent immigrants and LEP students is only partial  (roughly 64 percent of the recent immigrants in middle school are LEP and only 52 percent in elementary school) which explains the divergence in their exposure indices noted above.  The implication is that a distinction needs to be made between recent immigrants and LEP students in forming policy.  In particular, if additional resources intended to assist immigrants are targeted at LEP students, these programs may overlook as much as 42 percent of the recent immigrants. Further, although the popular perception of immigrants is of Hispanic and Asian students challenged primarily by limited language skills, our analysis indicates that a significant portion of the recent immigrants is black and a good portion is white.  Further, poverty among LEP students is significantly higher than among recent immigrants, which is modestly higher than for students overall.

How does the immigrant experience differ with respect to resources and performance?
    As shown in Table 4, on average, immigrant children attend schools with fewer resources: The average immigrant attends a school that spends roughly $7,582 per pupil, compared to the $7,816 spent on average. The spending disparity is highest in elementary schools (more than $285 per pupil), shrinking almost by half in middle and high schools. While this difference may seem small, if the difference in spending is part of the discretionary spending by the school, it could indeed represent a significant portion of the available discretionary funds. In addition, the regression analysis presented below provides further insight into the significance of these resource disparities.
    While pupil teacher ratio shows a similarly pattern (fewer resources at the elementary level, similar resources elsewhere) the data indicate that the teachers of immigrant students are slightly better educated and have slightly more experience.  Again, the pattern for LEP students is different– LEP students attend schools with typical or higher spending and smaller pupil-teacher ratios, but less experienced, less educated teachers in both elementary and middle schools. Both immigrants and LEP students attend larger schools. Whether this reflects the greater breadth of Bilingual/ESL programs available in larger schools or, alternatively, a preference for larger schools is unknown, but worthy of further study. Finally, while immigrants attend elementary schools with higher performance on reading and math tests and only slightly lower performance at middle school, LEP students attend schools with lower performance at both levels.

How do resources and school performance vary with immigrant (LEP) representation?
    Finally, we perform regression analyses of three resource measures and two performance measures described above. The resource regressions describe equity in the distribution of resources across schools, capturing the relationship between resources and the representation of immigrants, controlling for other characteristics of the school and students. The performance regressions describe equity in the distribution of ‘outputs’ across schools, capturing the relationship between output and the representation of immigrants, ceteris paribus.15 The regression coefficients can be interpreted as capturing the difference in the resource (or output) associated with an increase in the representation of immigrants, controlling for the socioeconomic characteristics of the student body.  Note, however, two important caveats.  First, our work does not provide guidance on what these coefficients should be – that rather difficult job is outside the scope of this paper.  Second, these regressions are not specified to capture causal relationships.  The resource equations cannot be interpreted as cost functions or factor demand equations - there are, after all, no prices among the independent variables – and no argument is made that these resource allocations have emerged from cost minimization efforts by schools or school districts.  The output equations cannot be interpreted as production functions – there are, most importantly, no input variables among the dependent variables and no claim is made that the regression equation captures the production of education.
    The regression analyses in Table 6 describe the relationship between resources and the socioeconomic characteristics of the students and indicate that, as suggested earlier, immigrant students get fewer resources, whether measured by expenditures or by pupil-teacher ratio.   At the same time, their teachers are better educated.  All of these coefficients are significant for elementary schools but only the expenditure result, which is more than twice as large, is significant for middle schools.  Once again, LEP students are treated differently – spending is (significantly) higher, class sizes are (significantly) smaller, but teachers are less educated (only significant for elementary schools). Note that the correlation coefficient between Percent LEP and Percent Immigrant is 0.64; thus, while the two variables are correlated, they are perhaps less correlated than many would expect. While we cannot rule out problems of multicollinearity in the model specification, the correlation is not so high as to cause us great concern.16 Other coefficients are consistent with explicit educational policies – spending increases and pupil-teacher ratio declines with the representation of special education and poor students – however, teacher education declines with poverty and is only increasing in the representation of part-time (and not full-time) special education students.
Interestingly, race per se seems to play little direct role in resource allocation.  Coefficients are generally insignificant determinants of expenditures or pupil-teacher ratio; however, teacher education decreases significantly with percent black at both elementary and middle school levels and with percent Hispanic at the elementary school level, even though limited English proficiency and poverty are included variables.
    Finally, regressions were estimated which included variables describing the characteristics of the immigrant population and which also attempt to measure explore the differences in the proportion of recent immigrants in a school. These are included in Tables 8 and 9. Two important findings stand out from the fuller specification in Table 8. First, there is some evidence that the race of the immigrant population matters. In particular, even fewer resources (measured both by expenditures and pupil-teacher ratio) are allocated to schools in which a greater share of the immigrants are black. This provides some suggestive support for advocates for Caribbean immigrants, who claim that these students have needs, unaddressed by the school system, that derive from their immigrant status.  Second, the regressions indicate that spending declines with the share of the immigrants who are LEP, revealing a divergence in the treatment of recent immigrant LEP students and non-recent immigrant LEP students – which includes second generation, non-recent immigrant and Puerto Rican students. These results are discussed in greater detail in Schwartz and Gershberg (2000).
    The regression analyses in Table 7 describe the relationship between school ‘output’ (measured by performance on math and reading tests for fifth and eighth grade) and the characteristics of the students.  In each case, independent variables include measures of test performance for the same school for the prior year and previous grade.17 As in other studies, the regressions indicate that test performance declines with the representation of poor, LEP, black and Hispanic children and, at the elementary school level, increase with percent immigrants.  (White is the omitted category.) These results coincide with perceptions in the education trade media, such as Zehr (2001a), that immigrant students often find U.S. schools less demanding than those they attended in their native countries. This provides additional evidence of the need for policymakers to disentangle language and other immigrant issues.  Note, however, that, under some circumstances, LEP students are exempt from taking the reading and math tests, so these results need to be interpreted with caution.18
    In a fuller specification presented in Table 9, we investigate the interaction between immigrant status and other student characteristics as well as the impact of different concentrations of immigrants.  These analyses reveal that, at the elementary school level, the positive relationship between performance and immigrants only becomes significant as the share of immigrants reaches 5% and the magnitude of that effect then declines mildly with immigrant share.  Second, at the elementary school level, performance increases with the percentage of the immigrants who are LEP while the share of Hispanics becomes completely insignificant.
    For middle schools, the results are rather different, with math scores being negatively associated with the proportion of immigrants, and the magnitude of that affect appearing to increase as representation increases. In addition, the scores for middle school immigrants who are black are worse, all else equal, in both math and reading.  This is particularly troubling given our previous finding that this group may receive fewer resources. Again, Schwartz and Gershberg (2000) discusses these results in greater detail.

Conclusion
    The key findings in this paper are as follows.  To begin, although recent immigrants represent less than ten percent of New York City’s public school students (and LEP students about 17 percent), our analyses provide encouraging news about their distribution across schools. Public schools span the full range in their representation of immigrant or LEP students, however, our results suggest immigrants are not more segregated than blacks, Hispanics, or poor students. Further, the segregation of immigrants is lowest in elementary schools and highest in middle schools, even though the choice of elementary school is dictated largely by residential location while the choice of middle or high school is more likely to reflect preferences of students and schools. Some of this segregation is undoubtedly programmatic. Newcomer schools, for example, educate only recent immigrants.   Nevertheless, immigrants are exposed to a somewhat different set of classmates than the average New York City public school student - the classmates of the typical immigrant are less likely to be black, more likely to be Asian and LEP and almost 15 percent of their classmates are recent immigrants themselves.
    Although the popular perception of immigrants is of Hispanic and Asian students challenged primarily by limited language skills, our analysis indicates that a significant portion of the recent immigrants are black (23 percent in middle schools and 19 percent in elementary schools) and a good many are white (15 percent in middle schools and almost 18 percent in elementary schools). Further, poverty among LEP students (at roughly 90 percent) is significantly higher than among recent immigrants (roughly 80 percent), which is, in turn, modestly higher than for students overall (in the middle 70s).
    Interestingly, our analyses indicate that the LEP and immigrant experiences diverge significantly. While school resources (measured by pupil teacher ratio and spending) generally decline with the representation of immigrants, resources increase with the representation of LEP students. Average education of the teachers increases with percent immigrants, but decreases with percent LEP.  Accordingly,, our analyses of school outputs (measured by math and reading test scores) indicate that while a greater representation of immigrant students indicates higher better ‘output’, a greater representation of LEP students indicates lower performance. The City and State school systems should begin to recognize more explicitly the characteristics of recent immigrants in the policies and programs it implements. For instance, barriers potentially faced by immigrant students to entry in gifted and talented programs (discussed in Gershberg, 2000), should be addressed. School officials and policymakers should support, or at least explore, more programs and policies aimed at supporting recent immigrants in ways beneficial to helping them exit ESL programs as quickly as possible. Newcomer schools and programs are one example, but as mentioned previously, they are not part of an overall plan at either the City or the State level.19
    Finally, we find evidence that not all immigrant groups are treated equally – in particular, recent black immigrants, who are less likely to be limited English proficient, seem to receive fewer resources and perform relatively poorly. Thus, while New York City and State have virtually no organized policies to support immigrant education aside from those policies aimed at English proficiency, it seems that that the issues and experiences of LEP students and recent immigrants are different enough that they merit more refined policy responses

References
Berne, Robert and Leanna Stiefel (1984) The Measurement of Equity in School Finance Baltimore: Johns Hopkins University Press.

Betts, Julian R. (1998) “Educational Crowding Out: Do Immigrants Affect the Educational Attainment of American Minorities?” in Hamermesh, D. and F. Bean, eds., Help or Hindrance? The Economic Implications of Immigration for African Americans. New York: Russell Sage and Magnus Lofstrom (1998) “The Educational Attainment of Immigrants: Trends and Implications.” Cambridge, MA: NBER Working Paper 6757, 1998.

New York City Board of Education (2000) ELL Subcommittee Research Studies Progress Report.

Downes, Thomas A. and Thomas F. Pogue (1994)  “Adjusting School Aid Formulas for the Higher Cost of Educating Disadvantaged Students.”  National Tax Journal. March: 89-110

Duncombe, William, and John Yinger. (1997)  “Why is it so Hard to Help Central City Schools?”  Journal of Policy Analysis and Management.  Vol. 16, No. 1.  Pp. 85-113.

Gershberg, Alec Ian (2000). “New Immigrants and the New School Governance in New York: Defining the Issues,” mimeo, New York: The New School University.

Hanushek, Eric A. (1986).  “The economics of schooling:  Production and efficiency in public schools.”  Journal of Economic Literature, 24(3), 1141-1177.

Iatarola, Patrice and Leanna Stiefel (2000)  “Intradistrict Equity in New York City” unpublished manuscript, New York University.

Kao, Grace and Marta Tienda (1995). “Optimism and Achievment: The Educational Performance of Immigrant Youth,” Social Science Quarterly, Vol. 76, No. 1: 1-19

McDonnell, Lorraine M. and Paul Hill, (1993) Newcomers in American Schools: Meeting the Educational Needs of Immigrant Youth, Santa Monica: RAND.

Ogbu, John U. (1991). “Immigrant and Involuntary Minorities in Comparative Perspective,” in Gibson, M. and J. Ogbu, eds, Minority Status and Schooling: A Comparative Study of Immigrant and Involuntary Minorities. New York: Garland Publishing, Inc.: 3-36

_____. (1982). “Cultural discontinuities and Schooling,” Anthropology & Education Quarterly 13 (4): 290-307.

_____ and Herbert Simons (1998). “Voluntary and Involuntary Minorities: A Cultural-Ecological Theory of School Performance with Some Implications for Education,” Anthropology & Education Quarterly 29 (2): 155-188.

Parrish, Thomas B. (1994):  “A Cost Analysis of Alternative Instructional Models for Limited English Proficient Students in California.”  Journal of Education Finance. 19 256-278.

Rivera-Batiz, Francisco. (1996) “The Education of Immigrant Children in New York City.”  ERIC Clearinghouse on Urban Education.  Number 117.

_____  (1995) “The Education of Immigrant Children: The Case of New York City.”  Unpublished Manuscript, International Center for Migration, Ethnicity and Citizenship, The New School for Social Research, New York.

Ruiz-de-Velasco, Jorge and Michael Fix (2000).  Overlooked and Underserved: Immigrant Students in U.S. Secondary Schools. Washington, DC: The Urban Institute.

Schwartz, Amy Ellen (1999) “School Districts and Spending in the Schools” Selected Papers on School Finance 1997-1999 U.S. Department of Education, National Center for Education Statistics. NCES 1999-334.ed. Fowler, William J. Jr. Washington, DC

_____ and Alec Gershberg (2000) “ The Immigrant Experience in New York City Public Schools” Draft. New York University.

Suarez-Orozco, Marcelo (1991). “Immigrant Adaptation to Schooling: A Hispanic Case,” in Gibson, M. and J. Ogbu, eds, Minority Status and Schooling: A Comparative Study of Immigrant and Involuntary Minorities. New York: Garland Publishing, Inc.: 37-62

Tanners, Lisa.  “Immigrant Students in New York City Schools.” Urban Education.  Vol. 32 No. 2, May 1997: 233-255.

Vernez, Georges and Allan Abrahamse. (1996) How Immigrants Fare in U.S. Education. Santa Monica: RAND, 1996.

Zehr, Mary Ann (2001a). “Immigrant Students Find U.S. Schools Less Demanding,” Education Week, January 17.

_____ (2001b). “New York City Modifies Bilingual Education,” Education Week, January 17.

_____ (2001c). “Study: Dearth of Programs for Oder Immigrant Students,” Education Week, January 24.
 

Table 1. Descriptive Statistics and Exposure Indices for Demographic Variables
 
 

Label

N
Mean, Unweighted Enrollment Weighted Mean Immigrant Weighted Mean  LEP Weighted Mean *

Minimum

Maximum
    All Schools
Percent Female 1097 49.2 49.0 48.8 48.8 6.0 100.0
Percent Black 1097 36.6 35.3 28.8 23.9 0.0 97.6
Percent Hispanic 1097 37.7 37.3 39.2 51.1 1.3 99.4
Percent Asian 1097 9.9 11.5 16.6 13.3 0.0 94.3
Percent White 1097 15.8 15.9 15.4 11.7 0.0 93.8
Percent Free Lunch 1097 71.3 66.3 67.6 74.0 5.9 100.0
Percent Recent Immigrants 1097 7.8 9.0 14.6 12.4 0.0 96.3
Percent Limited English Proficiency 1047 15.7 16.6 22.8 26.5 0.1 100.0
  *A smaller number of observations were used to compute this measure due to incomplete LEP data.

Table 2. Distribution of Students by Representation of Immigrants

Percent   Number of Percent  Percent
Immigrants Number Number Immigrant of Total Percent Free Lunch Percent Percent
(Range) of Schools of Students Students Immigrants LEP Eligible Black Hispanic

Below 5% 481 344,982 9,439 10.3 8.6 70.3 46.1 33.8
5 to 10% 299 291,231 21,275 23.1 15.8 71.3 36.1 40.7
10 to 15% 185 217,725 26,388 28.7 21.9 72.6 27.9 43.8
15 to 20% 78 104,540 17,915 19.5 25.7 71.3 17.2 38.3
20 to 30% 46 53,864 12,282 13.3 33.0 75.3 13.6 35.8
30 to 40% 8 8,984 3,141 3.4 35.6 86.4 26.0 23.7
40 to 50% 1 644 287 0.3 45.7 83.0 21.1 15.7
50 to 60% 1 302 175 0.2 85.0 71.3 13.6 36.8
60 to 70% 1 305 206 0.2 94.8 93.7 12.8 36.1
70 to 80% 0
80 to 90% 0
90 to 100% 2 975 930 1.0 89.8 84.8 4.7 53.1
TOTAL 1,102 1,023,552 92,038 100.0 45.6 78.0 21.9 35.8

Table 3. Dissimilarity Indices

Label All Schools Elementary
Schools Middle
Schools High
Schools
Immigrants 0.3185 0.3090 0.3632 0.3404
Female 0.0644 0.0345 0.0411 0.1363
Black 0.5347 0.5925 0.5269 0.4335
Hispanic 0.4788 0.5053 0.4854 0.4273
Asian 0.5212 0.5583 0.5165 0.4532
White 0.6533 0.7025 0.6233 0.5817
LEP 0.3711 0.3721 0.3783 0.3596
Free Lunch Eligible 0.5234 0.4960 0.4213 0.4283
Table 4.  Resources and Performance
 
 

Label

N

Mean Pupil
Weighted
Mean Immigrant
Weighted Mean LEP
Weighted
Mean*

Minimum

Maximum
    Elementary School
Total School Register 691 778.6 939.8 1,017.9 1,031.6 42.0 2,672.0
Read 5th Grd: mean N.C.E. 641 50.1 49.9 51.1 48.2 4.0 86.0
Math 5th Grd: mean N.C.E. 644 56.3 56.2 58.3 54.6 1.0 87.5
Read 4th Grd:% 50+ pctile 97 660 53.8 53.2 55.4 49.7 11.3 100.0
Math 4th Grd:% 50+ pctile 97 661 63.6 63.2 65.8 59.3 14.7 100.0
Total Spending Per Pupil 689 8,216.5 7,921.8 7,636.3 7,901.8 5,537.2 19,441.0
Pupil Teacher Ratio 689 16.0 16.4 16.7 16.1 6.6 31.5
Teacher Experience % 5 year + 669 61.4 61.5 62.2 60.4 6.7 92.9
Teacher Education % Masters + 669 78.8 78.8 80.7 77.8 41.7 100.0
    Middle School
Total School Register 233 847.9 1,168.9 1,285.5 1,236.9 59.0 2,250.0
Read 8th Grade: mean N.C.E. 224 50.5 50.9 50.4 48.7 24.0 81.8
Math 8th Grade: mean N.C.E. 224 52.7 53.8 53.2 51.4 31.2 84.6
Read 7th Grade:% 50+ pctile 97 217 43.3 45.3 44.0 40.8 5.3 95.1
Math 7th Grade:% 50+ pctile 97 218 49.4 51.5 49.7 45.9 6.9 100.0
Total Spending Per Pupil 231 8,701.1 8,095.2 7,931.5 8,242.5 4,761.5 22,414.4
Pupil Teacher Ratio 231 14.7 15.1 15.1 14.6 7.7 22.0
Teacher Experience % 5 year + 194 62.9 65.2 65.4 63.8 0.0 100.0
Teacher Education % Masters + 194 77.5 78.7 79.3 77.7 50.0 100.0
    High
School
Total School Register 173 1,658.8 2,761.2 2,932.9 2,966.4 27.0 5,021.0
Total Spending Per Pupil 173 8,105.8 7,427.2 7,284.9 7,408.9 5,360.3 17,170.7
Pupil Teacher Ratio 173 17.0 18.4 18.3 18.0 7.1 21.9

         * A smaller number of observations were used to compute this measure due to incomplete LEP data.
 

Table 5.  Recent Immigrant and LEP Student Demographics, Elementary and Middle School 1997-98
 

Label N Mean Immigrant Weighted Mean LEP Weighted Mean* N Mean  Immigrant Weighted Mean  LEP Weighted Mean*
Elementary Schools     Middle Schools
Pct. Recent Immigrants who are:
     Black 686 26.7 18.8   218 30.6 23.1
     Hispanic 686 37 35.5   218 31.1 31.7
     Asian 686 20.6 27.1   218 22.9 29.4
     White 686 15.3 18.2   218 15 15.4
     LEP 686 46.5 52.3   218 53.4 63.9
     Free Lunch 686 78.3 81.3   218 74.9 80.4
Pct. Limited English Prof. who are:
     Black 681 10.5   4.6 227 11.7   5.8
     Hispanic 681 62.6   70.8 227 58.8   64.6
     Asian 681 15.5   15.5 227 18   19.8
     White 681 11.1   8.9 227 11.3   9.7
     Free Lunch Eligible 691 86.1  90.4 228 86.2  89.5
 

* A smaller  number of observations was used to compute this measure due to incomplete LEP data.

Table 6.  School-level Educational Resource Equity Regressions, New York City School District 1997-98

  Elementary Schools  Middle  Schools
  Pupil   Pupil
 Expenditure Teacher Teacher Expenditure Teacher Teacher
 Per Pupil Ratio Education Per Pupil Ratio Education
 (1) (2) (3) (4) (5) (6)
Intercept 7,385.23* 16.67* 82.60* 4,940.16* 20.11* 67.09*
 (868.66) (1.48) (8.30) (1,355.73) (1.77) (10.90)
Pct Female -29.52* 0.08* 0.05 8.71 0.00 0.39*
 (17.27) (0.03) (0.17) (26.51) (0.03) (0.22)
Pct FT Special Ed.  171.91* -0.20* -0.06 160.97* -0.19* 0.08
 (7.14) (0.01) (0.07) (19.13) (0.02) (0.13)
Pct PT Special Ed. 133.74* -0.16* 0.67* 131.33* -0.07* 0.70*
 (16.69) (0.03) (0.16) (29.93) (0.04) (0.21)
Pct Free Lunch 1.17 -0.02* -0.09* 9.79 -0.04* -0.10*
 (2.92) (0.00) (0.03) (6.97) (0.01) (0.05)
Pct LEP 22.65* -0.06* -0.16* 50.08* -0.05* -0.13
  (5.12) (0.01) (0.05) (18.19) (0.02) (0.13)
Pct Black 3.78 -0.01 -0.07* 4.04 0.00 -0.12*
 (2.74) (0.00) (0.03) (5.95) (0.01) (0.04)
Pct Hispanic -1.22 0.01 -0.05* -10.66 0.01 -0.06
 (3.14) (0.01) (0.03) (7.77) (0.01) (0.05)
Pct Asian 1.89 0.01* 0.10* 12.18 0.01 0.04
 (3.53) (0.01) (0.03) (9.67) (0.01) (0.07)
Pct Immigrant -28.63* 0.05* 0.38* -80.33* 0.03 0.29
 (8.85) (0.02) (0.08) (26.65) (0.03) (0.18)
R-square 0.64 0.59 0.40 0.50 0.60 0.41
N 670 670 664 208 208 189

All regressions are weighted by number of students
*  indicates significant at the 10% level or higher
Standard errors in parentheses.
Table 7. School-level Education Outcome Equity, New York City School District, 1997-98
Dependent Variable - Average Normal Curve Equivalents (NCE )
 Reading
Fifth
 Grade Math
Fifth
Grade Reading
Eighth Grade Math
Eighth Grade
 (1) (2) (3) (4)
Intercept 34.20* 35.65* 31.72* 38.69*
 (4.07) (5.00) (3.61) (4.08)
Lagged Test Score in Reading 0.33*  0.43*
 (0.01)  (0.02)
Lagged Test Score in Math 0.38*  0.44*
  (0.02)  (0.02)
Percent Female 0.12 0.10 0.01 -0.02
 (0.08) (0.10) (0.07) (0.08)
Percent Full Time Special Ed. -0.03 -0.01 -0.06 -0.04
 (0.03) (0.04) (0.05) (0.05)
Percent Part Time Special Ed. -0.08 -0.08 -0.01 -0.23*
 (0.08) (0.09) (0.08) (0.09)
Percent Free Lunch -0.05* -0.05* 0.03 0.00
 (0.01) (0.02) (0.02) (0.02)
Percent LEP -0.05* -0.05* -0.01 0.06
 (0.02) (0.03) (0.04) (0.05)
Percent Black -0.04* -0.08* -0.01 -0.05*
 (0.01) (0.02) (0.01) (0.02)
Percent Hispanic -0.03* -0.05* -0.03* -0.07*
 (0.01) (0.02) (0.02) (0.02)
Percent Asian -0.01 0.02 -0.02 -0.04
 (0.02) (0.02) (0.03) (0.03)
Percent Immigrant 0.12* 0.12* -0.03 -0.06
 (0.04) (0.05) (0.07) (0.07)
R-square 0.83 0.85 0.90 0.91
F 306 334 194 214
N 617 621 233 235

All regressions are weighted by number of students
* indicates significant at the 10% level or higher
Standard errors in parentheses.
 
 
 

 Table 8: Resource Equity Regressions, Elementary & Middle Schools
 

 Elementary Schools Middle Schools
 

 Expenditure Per Pupil Pupil Teacher Ratio Teacher Education Expenditure Per Pupil Pupil Teacher Ratio Teacher Education
 (1) (2) (3) (4) (5) (6)
Intercept 7960.48* 15.86* 81.71* 5603.65* 19.1* 58.57*
 (887.69) (1.51) (8.48) (1,685.35) (2.2) (13.58)
Percent Female -27.39 0.08* 0.03 11.98 0 0.48*
 (17.26) (0.03) (0.16) (29.1) (0.04) (0.24)
Percent Full Time Special Ed.  169.41* -0.2* -0.07 157.78* -0.19* 0.11
 (7.24) (0.01) (0.07) (19.74) (0.03) (0.14)
Percent Part Time Special Ed. 131.61* -0.15* 0.7* 109.78* -0.05 0.56*
 (16.99) (0.03) (0.16) (33.83) (0.04) (0.25)
Percent Free Lunch 4.34 -0.03* -0.06 8.9 -0.04* -0.18*
 (4.06) (0.01) (0.04) (9.92) (0.01) (0.07)
Percent LEP 30.56* -0.07* -0.13* 58.81* -0.07* -0.08
  (6.06) (0.01) (0.06) (22.01) (0.03) (0.16)
Percent Black 9.38* -0.02* -0.12* 12.22 -0.02 -0.09
 (4.77) (0.01) (0.05) (10.83) (0.01) (0.08)
Percent Hispanic 0.55 0 -0.03 0.13 -0.01 -0.02
 (4.55) (0.01) (0.04) (11.25) (0.01) (0.08)
Percent Asian -0.64 0.02 0.07 4.79 0.03 0.11
 (6.03) (0.01) (0.06) (16.02) (0.02) (0.12)
Percent Immigrants -8.25* 0.02* 0.04 -6.29 0.02* 0.02
   Who are Black (4.03) (0.01) (0.04) (8.42) (0.01) (0.06)
Percent Immigrants -2.35 0 -0.04 -13.06 0.02* -0.06
   Who are Hispanic (3.22) (0.01) (0.03) (8.12) (0.01) (0.06)
Percent Immigrants 0.79 0 0.02 1.27 0 -0.03
   Who are Asian (3.6) (0.01) (0.03) (7.46) (0.01) (0.05)
 
 
 
 
 
 
 
 

 Table 8: Resource Equity Regressions, Elementary & Middle Schools (Continued)
 

 Elementary Schools Middle Schools
 Expenditure Per Pupil Pupil Teacher Ratio Teacher Education Expenditure Per Pupil Pupil Teacher Ratio Teacher Education
 (1) (2) (3) (4) (5) (6)
Percent LEP  -6.48 0.01 0.02 -18.53 0.02 0.02
    Who are Free Lunch Eligible (4.19) (0.01) (0.04) (11.87) (0.02) (0.09)
Percent Immigrants  0.98 0 -0.02 12.87* -0.02 0.07
    Who are Free Lunch Eligible (3.68) (0.01) (0.04) (7.47) (0.01) (0.06)
Percent Immigrants   -6.95* 0.01 0.01 1.01 0.01 0.05
    Who are LEP (2.89) (0.00) (0.03) (6.9) (0.01) (0.05)
Percent Immigrant * -67.96* 0.11 0.72* -74.94 -0.13 0.33
     0-5% immig dummy (38.78) (0.07) (0.37) (103.53) (0.14) (0.05)
Percent Immigrant* -51.93* 0.07* 0.47* -128.4 -0.01 0.25
    5-10% immig dummy (19.35) (0.03) (0.18) (52.32) (0.07) (0.38)
Percent Immigrant* -45.63* 0.07* 0.41* -103.8 0.03 0.28
   10-20% immig dummy (12.71) (0.02) (0.12) (36.25) (0.05) (0.26)
Percent Immigrant* -36.63* 0.05* 0.25* -107.74 0.07 0.23
   20-30% immig dummy (11.16) (0.02) (0.11) (37.23) (0.05) (0.27)
Percent Immigrant* -14.49 0.04 0.54* -65.62 -0.02 -0.06
  30-50% immig dummy (-16.25) (0.03) (0.15) (50.39) (0.07) (0.36)
Percent Immigrant *    -57.31 0 0.01
90-100% immig dummy    (33.52) (0.04) (0.24)
R-square 0.65 0.6 0.42 0.55 0.65 0.44
F 63 51 24 11 16 7
N 668 668 662 198 198 184
 

All regressions are weighted by number of students
* indicates significant at the 10 % level or higher
Standard errors in parentheses

Table 9: Outcome Equity Regressions, Elementary and Middle Schools, 1998

 Elementary Schools Middle Schools
 5th grade avg. NCE 8th grade avg. NCE
 Reading Math Reading Math
 (1) (2) (3) (4)
Intercept 33.94* 35.28* 30.37* 39.73*
 (4.18) (5.13) (4.31) (4.86)
Read 4th Grade:% 50+ percentile 97 0.32*
 (0.01)
Read 7th Grade:% 50+ percentile 97   0.43*
   (0.02)
Math 4th Grade:% 50+ percentile 97  0.38*
  (0.02)
Math 7th Grade:% 50+ percentile 97    0.43*
    (0.02)
Percent Female 0.12 0.1 0.03 0.01
 (0.08) (0.1) (0.08) (0.09)
Percent Full Time Special Ed. -0.02 0 -0.07 -0.05
 (0.03) (0.04) (0.05) (0.05)
Percent Part Time Special Ed. -0.08 -0.08 -0.03 -0.21*
 (0.08) (0.1) (0.09) -0.1)
Percent Free Lunch -0.05* -0.06* 0 0.01
 (0.02) (0.02) (0.02) (0.03)
Percent LEP -0.08* -0.09* -0.02 0.08
 (0.03) (0.03) (0.05) (0.05)
Percent Black -0.04* -0.08* 0.04 0
 (0.02) (0.03) (0.03) (0.03)
Percent Hispanic -0.03 -0.04 0.01 -0.01
 (0.02) (0.03) (0.03) (0.03)
Percent Asian -0.01 0.05 0.04 0.06
 (0.03) (0.03) (0.04) (0.05)
Percent Immigrants 0.01 0.01 -0.04* -0.04*
    Who are Black (0.02) (0.02) (0.02) (0.02)
Percent Immigrants -0.01 -0.01 -0.04* -0.07*
    Who are Hispanic (0.01) (0.02) (0.02) (0.02)
Percent Immigrants -0.01 -0.03 -0.05* -0.06*
    Who are Asian (0.02) (0.02) (0.02) (0.02)

 Table 9: Outcome Equity Regressions, Elementary & Middle Schools (Continued)

 Elementary Schools Middle Schools
 Reading Math Reading Math
 (1) (2) (3) (4)
Percent LEP  -0.01 0 0.02 -0.04
     Who are Free Lunch Eligible (0.02) (0.02) (0.03) (0.03)
Percent  Immigrants  0 0 0 0.01
     Who are Free Lunch Eligible (0.02) (0.02) (0.02) (0.02)
Percent Immigrants  0.03* 0.03* 0.02 0.01
     Who are LEP (0.01) (0.02) (0.02) (0.02)
Percent Immigrant * 0.24 0.15 -0.01 -0.27
    0-5% immig dummy (0.18) (0.22) (0.24) (0.27)
Percent Immigrant* 0.2* 0.19* 0.03 -0.19
   5-10% immig dummy (0.09) (0.11) (0.12) (0.13)
Percent Immigrant* 0.18* 0.16* -0.05 -0.19*
  10-20% immig dummy (0.06) (0.07) (0.08) (0.08)
Percent Immigrant* 0.11* 0.11* -0.05 -0.13
  20-30% immig dummy (0.05) (0.06) (0.09) (0.09)
Percent Immigrant* 0.09 0.11 -0.07 -0.3*
  30-50% immig dummy (0.07) (0.09) (0.12) (0.13)
R-square 0.84 0.85 0.9 0.91
F 153 166 96 105
N 616 620 224 226
 

 All regressions are weighted by number of students
* indicates significant at the 10% level or higher
Standard errors in parantheses
 

Endnotes