Goldsmith, Pat Antonio
All Segregation Is Not Equal: The Impact of Latino and Black School Composition
University of Wisconsin-Parkside
What is the effect of proportion Latino on the test scores of white, black, and Latino high school students?
Journal Name or Institutional Affiliation:
Vol. 46, No. 1, pp. 83-105
- Results reveal that segregation concentrates disadvantages for Latinos and Blacks, but surprisingly, proportion Latino tends to positively influence test scores over the high school years.
- Proportion Black, in contrast, does not affect test scores except for a negative effective for Blacks in science.
- Predominantly Black schools stand out by having a higher percentage of students in single parent families and predominantly White schools.
- Models indicate proportion Latino and proportion Black do not influence the chance of dropping out as long as control variables are included in the models.
- Measures suggest predominately Latino schools are of lower quality.
- Predominantly Latino schools cluster students of immigrant parents, and students learn more in schools with proportionately more immigrant parents (at least in the subjects of reading and history).
- The effects of proportion Latino are more positive than the effects of proportion Black and that part of the reason this occurs is that segregated Latino schools bring together many students of immigrant parents.
- The magnitude of the coefficients indicates that proportion Latino has a positive effect on Latino's reading and science test scores while it does not affect whites' reading and science test scores.
- Proportion Latino has a positive effect on Latinos' test scores in all subjects, although the positive effect observed on science test scores only exists net of neighborhood and class size differences. Proportion Latino also has a positive effect on blacks' and whites' test scores in math and history but not in reading and science.
- Proportion Black has a negative effect on blacks' science test scores but has no effect on Whites' test scores.
- Proportion Latino usually has a positive effect on test scores and proportion Black usually has no effect. Specifically, there is a positive effect of proportion Latino on Latinos' test scores in all subjects (although weaker in science). This positive effect extends to both less and more acculturated Latinos, and it extends to Blacks' and Whites' test scores in math and history.
- Findings suggest that when immigrant parents are clustered together, as they are in some segregated Latino schools, then students throughout the school benefit. This could explain why Blacks' and Whites' test scores sometimes improve as the schools' proportion f Latino rises.
- The analysis shows no effect of the schools' proportion of single-parent families on test scores.
- Researchers should not treat the effects of proportion Latino and proportion black as identical.
- Integration of Latinos with Whites would reduce some of the inequalities between schools, it would not hurt and in some areas would help the test score of whites, but it would hurt Latinos unless some of the helpful features of predominantly Latino schools could be copied.
Journal Article Empirical Research
Composition, Hispanics, Latinos, Math, Racial Composition, Reading, SES, Science, Segregation, Social Studies
Secondary Survey Data
Method of Analysis:
Unit of Analysis:
- The study uses NELS and zip code level data from the 1990 U.S. Census. Merged the NELS and the 1990 census to account for the characteristics of students' residential zip code areas.
- Sample is limited to students that can be matched with valid secondary school data and to students in schools with at least five sampled students. Additionally, student without valid test scores in 1988 and 1992 are omitted.
- The final sample includes approximately 10,050 students in 944 schools.
- DV: Tests scores (IRT-estimated number right in four different subject areas: reading, math, science, history).
- IV: Nativity/language background (students of immigrants parents, years in the US, English is 2d language), group proportions (proportion Latino, proportion black, and proportion other nonwhite), race/ethnicity/gender (dummy for Mexican, Puerto Rican, Cuban and other Latinos origins, black, male), neighborhoods variables (proportion high school, proportion college, median family income, unemployment rate, fraction prof./mngr., school-level peer effects (school mean SES, proportion single parents), school quality (students per teacher, class size, proportion advanced deg., private school), family variables (SES, Parent HS, Parent BA, Mother/father family, single mother family, single father family, other nontraditional family, no family data), other controls (east, west, Midwest, 8th grade test, missing test control, intercept).
- Test the idea that the effect of proportion Latino may be more positive than proportion black due to: (1) Latinos are less likely than African Americans to be living in single-parent families. (2) Latinos are more likely to be immigrants or the children of immigrants than are African Americans. And immigrants minorities work harder in school because immigrants compare their educational and occupational opportunities to people in their homeland whereas native-born minorities compare their opportunities to those of whites.
- Also ran logistic regression models predicting whether a student has dropped out by twelfth grade.