Chapter 6 Notes: State Language Policies, School Language Practices, and the English Learner Achievement Gap

English Learner (EL) Achievement Gap

• Significant achievement gap exists between EL and English-Only (EO) students from kindergarten.
• 2005 NAEP Grade 4 Reading:
  • 7% of EL students scored at or above proficient.
  • 32% of EO students scored at or above proficient.
• 2005 NAEP Grade 8 Reading:
  • 4% of EL students scored at or above proficient.
  • 30% of EO students scored at or above proficient.
• Math achievement gap is similarly large. EL students in both grade 4 and 8 have significantly lower scores than EO students.

Government Involvement

• Federal and state governments address EL needs through litigation, legislation, and programs.
• Goals for linguistic minority students:
  • Access to core curriculum.
  • English language proficiency.
  • Native-language proficiency.
  • Closing the achievement gap.
  • Cultural competence.
• Goals and methods have evolved over the past 30-40 years.
• Ballot initiatives in California, Arizona, and Massachusetts curtailed native-language instruction for EL students.

State Policies and School Practices

• Chapter examines the relationship between state policies/school practices and the EL achievement gap.
• Focus areas:
  • Overview of state policies.
  • Descriptive data on the achievement gap.
  • Instructional practices across states.
  • Analysis of the achievement gap using 2005 NAEP data.
  • Relationship between the achievement gap and policies/practices, student composition, school structure, and resources.
• Study draws from a larger research project on state and district variability in the EL achievement gap.

States' Role in Educating Linguistic Minority Students

• State legislation varies due to differing historical LM populations.
• Garcia (2005) tabulation:
  • 12 states mandate special services for LM students.
  • 12 states permit special services.
  • 1 state prohibits special services.
  • 26 states have no direct legislation.
• States with larger LM populations (e.g., Texas, California) have a longer history of legislative activity.
  • Massachusetts passed the first bilingual education law in 1971.
  • California's Chacon-Moscone Bilingual-Bicultural Act (1976).
  • Texas's first bilingual education bill (1981).
• Most state legislation focuses on developing English fluency.
• Restrictive-language-policy states (California, Arizona, Massachusetts) emphasize Structured English Immersion (SEI).
• 17 other states allow or require instruction in a language other than English.
• 15 states mandate a cultural component.

Assessing the English Learner Achievement Gap

• Research study questions:
  • What is the average achievement gap in math and reading at grades 4 and 8 between EL and EO students in the United States?
  • How much of the variability in the achievement gap is attributable to students and their families, schools, and states?
  • How much of the school-level and state-level variability in the achievement gap can be accounted for by resources and other school inputs versus school practices, including inputs and practices specifically targeted for English learners?
• Data from the 2005 NAEP (all 50 states and D.C.).
• Data include background information from student, teacher, and principal questionnaires.
• Teachers identified EL/LEP students and provided data on English proficiency and instructional programs.
• NAEP data augmented with state-level data from other sources.
• Variables measured:
  • Student composition.
  • School structure.
  • School resources.
  • School policies and practices.
• Data analyzed using multilevel modeling techniques.

Size of the EL Achievement Gap Nationally and Among States

• Achievement gap measured by comparing mean test scores on NAEP.
• Exclusion rates of EL students varied among schools and states (nationally, 14% excluded in grade 4 reading).
• Exclusion rate differences could affect reported NAEP scores.
• Effect Size (ES) is used as a common metric for comparisons, representing the difference in test scores as a fraction of a standard deviation (SD).
• Effect Size benchmarks (Cohen, 1988):
  • Small effect: ES >= 0.2 SD
  • Medium effect: ES >= 0.5 SD
  • Large effect: ES >= 0.8 SD
• Comprehensive school reform models have shown varying ES results after evaluation.
  • Average ES of 0.12 in 232 evaluation studies.
  • Average ES of 0.50 in interventions implemented for 8 years or longer.
• Table 6.1 shows achievement gaps between EL and EO students in grades 4 and 8.
  • Grade 4 reading gap nationally: 0.92 SD.
  • Grade 4 math gap nationally: 0.79 SD.
  • Both gaps are considered large, with reading being larger.
• States with restrictive language policies and Nevada have achievement gaps at least the size of the national average.
• States with higher proportions of students in native-language instruction (New Mexico, Texas) have smaller achievement gaps.
• High overall achievement levels do not guarantee a small achievement gap (e.g., Massachusetts).
• Achievement gaps in grade 8 are larger than in grade 4 (1.11 SD in reading, 1.00 SD in math).
• Factors contributing to larger gaps:
  • Reclassification of EL students as Fluent English Proficient.
  • New EL students entering in upper grades.
• Achievement gaps vary widely among states in grade 8.
• Texas's achievement gaps are very large at grade 8, particularly due to weak ESL programs in secondary schools.
• Differences in redesignation rates between states like Texas and California can impact grade 8 EL achievement gaps.

Instructional Programs for EL Students

• Instructional program types:
  • No specially designed instruction.
  • Specially designed instruction in English (ESL, simplified English).
  • Native-language instruction.
• Only about half of all EL students received specialized instruction in 2005.
• Provision of specially designed instruction varies widely among states.
• Native-language instruction is even more variable.
• SDI and native-language instruction are sparse in grade 8, especially in math.

Variability in the EL Achievement Gap Among States and Schools

• Study examines the variability in the EL achievement gap among schools and states.
• Determined that:
  • 75% of variability is related to student differences.
  • 15-20% to school differences.
  • 2-5% to state differences.
• States account for 15-20% of the variability in mean achievement.
• States account for a larger proportion of the variability in the mean school EL achievement gap (37-39%).
• Suggests states have more influence over the EL achievement gap than over average school achievement.

State Predictors of Mean Achievement and the EL Achievement Gap

• Four types of state-level factors:
  • Student composition.
  • Structural variables.
  • School resource variables.
  • Policy and practice variables.
• Only two factors predict differences in the mean EL achievement gap in grade 4 reading across states:
  • Mean reading proficiency of EL students.
  • State policies with respect to standards.
    • States with higher mean levels of EL English proficiency in reading had narrower achievement gaps in reading between EL and EO students.
    • Level of state standards: composite measure that assessed the rigor of standards in all academic subjects, as well as whether the state provides supplementary resources, including materials provided for particular student populations, such as EL students. According to the EPERC, 35 states provide such resources.
• Three factors predict differences in the EL achievement gap in grade 4 math across states:
  • States with EL students at higher levels of English reading proficiency had narrower achievement gaps in math.
  • The number of pages of reading that students did in and out of school each day was negatively related to the EL achievement gap.
  • States with higher proportions of Black students had narrower EL achievement gaps.
• Five factors predict differences in the EL achievement gap in grade 8 reading across states:
  • States with higher proportions of Hispanic students had larger EL achievement gaps.
  • States with higher percentages of students on free and reduced lunch had narrower EL achievement gaps.
  • States with EL students at higher levels of English reading proficiency had smaller EL achievement gaps.
  • States with a higher proportion of students receiving specialized designed instruction had lower achievement gaps.
  • The higher the English proficiency of the EL students in the state, the wider the achievement gap associated with an increased proportion of students receiving specialized instruction. Conversely, the lower the English proficiency of the EL students in the state, the smaller the achievement gap associated with an increased proportion of students receiving specialized instruction. Results suggest a varying effect of specialized instruction.
• Only two factors were significant predictors of the EL achievement gap in grade 8 math:
  • States with larger average school enrollment had larger achievement gaps, and
  • States with more equitable school finance systems had smaller EL achievement gaps in grade 8 math.
• Student composition factors account for two-thirds of the variability in mean school achievement in grade 4 reading, with other factors being less important, although state policy variables account for 13% of the variability.
• In contrast, student composition variables explain a relatively much smaller proportion of the variability in the EL achievement gap in reading (33%), while state policy and school practice variables explain about 12% of the variability.
• Grade 4 math shows an opposite pattern-student composition factors play a less important role in explaining mean math achievement (45% versus 68% in reading), while policy and practice variables play a relatively more important role (20% versus 13% in reading).
• In contrast, student composition factors play a more important role in explaining between-state differences in the EL achievement gap in math (55% versus 33% in reading), with policy and practice factors playing a similar role (12%). Grade 8 reading and math show similar patterns to those of grade 4 reading and math.

School Predictors of Mean Achievement and the EL Achievement Gap

• Four types of school-level factors:
  • Student composition.
  • Structural factors.
  • School resource factors.
  • Practice factors.
• Six school-level factors predict differences in the mean EL achievement gap in grade 4 reading across schools:
  • Schools with higher proportions of students on free and reduced lunch and, second, schools with higher proportions of Black students had smaller EL achievement gaps.
  • Third, schools with higher mean levels of EL English proficiency in reading had narrower achievement gaps in reading between EL and EO students.
  • Fourth, urban schools had slightly smaller EL achievement gaps.
  • Fifth, schools with a higher proportion of EL students excluded from testing had a lower achievement gap.
  • Finally, the lower the English proficiency of the EL students in the school, the smaller the achievement gap associated with an increased proportion of students receiving specialized instruction.
• Four school-level factors predict the EL achievement gap in grade 4 math:
  • First, schools with students of higher socioeconomic status (SES) had larger achievement gaps.
  • Second, schools with higher proportions of Black students had narrower EL achievement gaps.
  • Third, schools with EL students at higher levels of English reading proficiency had narrower achievement gaps in math, suggesting that reading proficiency plays an important role in EL students' math achievement.
  • Fourth, the more years of English instruction that students receive, the smaller the achievement gap.
• Three school-level factors predict the EL achievement gap in grade 8 reading:
  • First, schools with higher proportions of EL students and, second, schools with higher proportions of Black students had smaller EL achievement gaps.
  • Third, schools with EL students at higher levels of English reading proficiency had lower EL achievement gaps in math.
• Finally, five school-level factors predict the EL achievement gap in grade 8 math.
  • First, schools with higher proportions of EL students had lower achievement gaps.
  • Second, schools with higher-SES students had larger achievement gaps.
  • Third, schools with higher proportions of Hispanic students had larger EL achievement gaps, while, fourth, schools with higher proportions of Black students had smaller EL achievement gaps.
  • Fifth, schools with EL students at higher levels of English reading proficiency had lower EL achievement gaps in math (ES= .27).
• Student composition factors account for most of the variability in mean school achievement across all four test areas, ranging from 75% for grade 4 math to 79% for grade 8 math.

Summary and Conclusions

• EL achievement gaps are sizable in all four areas (reading and math in grades 4 and 8) based on 2005 NAEP.
• Gaps are larger in reading than in math and larger in grade 8 than in grade 4.
• Exclusion rates vary, which means that differences in the EL achievement gap for tested students may not represent differences in the achievement gap of all EL students.
• Achievement gaps do not appear to be related to absolute achievement levels of either EO or EL students.
• Instructional practices, particularly offering specially designed instruction for EL students, helped reduce the achievement gaps among states and schools, at least in some cases.
• The study supports that state policies and school practices restricting the use of native-language instruction could limit the ability of states and schools to reduce the EL achievement gap.
• The impact of variables most under control of schools, districts, and state policy to reduce the EL achievement gap is very small by comparison.
• The school-related factors that have the most impact on both student achievement and the EL achievement gap are related to student composition of schools.
• The most effective policy mechanism to improve student achievement and reduce the achievement gap is a more equitable distribution of students among schools.