Donate
Independent, objective, nonpartisan research

R 514LHR

Authors

R 514LHR

Tagged with:

Publication PDFs

Database

This is the content currently stored in the post and postmeta tables.

View live version

object(Timber\Post)#3711 (44) { ["ImageClass"]=> string(12) "Timber\Image" ["PostClass"]=> string(11) "Timber\Post" ["TermClass"]=> string(11) "Timber\Term" ["object_type"]=> string(4) "post" ["custom"]=> array(5) { ["_wp_attached_file"]=> string(12) "R_514LHR.pdf" ["wpmf_size"]=> string(7) "3048272" ["wpmf_filetype"]=> string(3) "pdf" ["wpmf_order"]=> string(1) "0" ["searchwp_content"]=> string(91681) "www.ppic.org Pathways to Fluency Examining the Link between Language Reclassifcation Policies and Student Success M ay 2014 Laura E. Hill • Julian R. Betts • Belen Chavez • Andrew C. Zau • Karen Volz Bachofer with research support from Joseph M. Hayes Supported with funding from the Donald Bren Foundation and the Evelyn and Walter Haas, Jr. Fund SUMMARY N early 25 percent of the students attending California’s K–12 public schools are English Learners (ELs). Their EL designation is intended to last only as long as they need supplemen- tal language support to succeed in school. Some students attain English fuency quickly, but others remain ELs for six years or longer. Because outcomes for students reclassifed as English profcient are much better than for students who remain ELs, policymakers are seeking answers to questions about how quickly EL students should be reclassifed, whether reclassifcation crite - ria should be standardized, and the links between reclassifcation and academic success. These issues are especially urgent now that California is implementing a major overhaul of K–12 standards, testing, and funding—as well as many elements of EL instruction. Because the new Local Control Funding Formula (LCFF) provides additional funding to districts with high numbers of ELs, there is more interest than ever in making sure that districts have the right incentives to help these students succeed. In this report, we examine reclassifcation policies and the academic performance of ELs and former ELs in the two largest school districts in California, Los Angeles Unifed and San Diego Unifed, which together serve approximately 15 percent of the state’s EL students. Using longitudinal student data over ten years, we can follow a cohort of 2nd grade students through their 12th grade year. We fnd that students reclassifed in elementary school (grades 2–5) have very strong aca - demic outcomes throughout middle and high school. These students perform as well as or ISTOCK Pathways to Fluency 2 www.ppic.org better than native English speakers on state standardized tests and are as likely or more likely to make on-time grade progress. There is no evidence that the removal of language supports for ELs who are reclassifed hurts their academic progress relative to that of native English speakers. Reclassifcation criteria in both San Diego and Los Angeles are more stringent than minimum guidelines recommended by the State Board of Education (SBE). These more rigorous criteria are associated with somewhat improved outcomes for students but also lower reclassifcation rates. Despite diferences in reclassifcation criteria between the two districts, the factors that predict successful outcomes for EL students reclassifed in elementary school in Los Angeles and San Diego are remarkably similar. The two standardized tests currently used to reclassify students—the California Standards Test (CST) and the California English Language Develop - ment Test (CELDT)—are individually strong predictors of future academic outcomes such as performance on middle school standardized tests and the high school exit exam. Elementary school marks are less useful as predictors. Our fndings lead us to recommend the following: • Use the CELDT as the sole assessment for reclassification decisions until the CST replace - ment is available. More generally, consider allowing districts to reclassify students on the basis of just one test. • In designing new English language development (ELD) tests and reclassification stan - dards, consider the relative rigor of reading and writing requirements—our data suggest that the current CELDT writing requirement is relatively easy in comparison to the read - ing requirement. • Reconsider the use of reclassification criteria that are more rigorous than those suggested in the State Board of Education guidelines. We fnd evidence in the state’s two largest school districts that English learners would beneft as a group from being reclassifed slightly sooner, through an easing of reclassifcation standards. • Consider a uniform standard for reclassification across school districts. Evaluating districts’ successes with ELs is very difcult when classifcation and reclassifcation policies vary. Over the next few years, many elements of EL instruction, funding, and testing will be changing. Many policymakers have long been frustrated with the pace at which EL students are reclassifed as fully English profcient and are concerned that the additional funds directed toward ELs under the LCFF might increase district incentives to delay reclassifcation for stu - dents on the cusp of English fuency. This is an ideal time to draw lessons from the recent past to inform state and local reclassifcation policies in 2014 and beyond. For the full report and related resources, please visit our publication page: www.ppic.org/main/publication.asp?i=1089 3 Pathways to Fluency www.ppic.org Introduction During the 2012–13 school year, more than 1.3 million English Learners attended public schools in California, accounting for about one-quarter of the state’s K–12 stu- dent population. 1 Nearly two-thirds of these students were enrolled in elementary schools. California school districts are charged with the dual goals of ensuring that English Learners acquire full profciency in English as quickly as possible and that they meet the same rigorous grade-level academic standards that all students must meet. 2 Te SBE has issued guidelines about the minimum criteria used to reclassify EL students, but districts are allowed a great deal of latitude in establishing more rigorous criteria. Te state’s interest in reclassifcation has been height- ened by a number of recent changes to K–12 funding, standards, and assessments. Te new LCFF allocates many more dollars per EL student than districts received under the old funding formula. Tis increases the efect of reclas- sifcation decisions for school districts and the importance of making sure that the funding is used to serve ELs efec- tively. In January 2014, the SBE passed temporary regula- tions for how these dollars are to be spent by districts and established an accountability framework, called the Local Control and Accountability Plan (LCAP). 3 Te implementation of the Common Core State Stan- dards (CCSS), changes to English language development (ELD) standards, and development of new assessments are also under way. Until 2013, California used the CST to determine whether ELs met the academic skills portion of the reclassifcation criteria. However, as part of the state’s transition to a new statewide assessment system aligned with the CCSS, there will be very little testing during the 2013–14 academic year, and the individual student results that help determine readiness for reclassifcation will not be provided to districts. A new statewide testing system will phase in during the 2014–15 academic year, but until then, districts will need to decide how to reclassify EL students when their CST results become outdated. In addi- tion, the state’s ELD standards were revised in 2012, and a new English language profciency exam is expected to be fully implemented by 2016. In light of the state’s interest in reclassifcation deci- sions (as evident in Senate Bill 1108, which mandated a statewide analysis of school district reclassifcation poli- cies), the new CCSS and assessments, and the increased per-pupil funding for EL students, the timing is perfect The state’s interest in reclassifcation has been heightened by a number of recent changes to K–12 funding, standards, and assessments. Abbreviations CAHSEE California High School Exit Examination CCSS Common Core State Standards CDE California Department of Education CELDT California English Language Development Test CMA California Modifed Assessment CST California Standards Test EL English Learner ELA English language arts ELD English language development EO English-only or native English speaker FEP fuent English profcient FRPL free/reduced price lunch G PA grade point average IFEP initially fuent English profcient LAUSD Los Angeles Unifed School District LCAP Local Control and Accountabilty Plan LCFF Local Control Funding Formula LEA Local Educational Agency OPL overall profciency level RFEP reclassifed fuent English profcient SBE State Board of Education SDUSD San Diego Unifed School District SEI Structured English Immersion Pathways to Fluency 4 www.ppic.org for a retrospective examination of California’s current reclassifcation guidelines—and an analysis of the relative importance of each reclassifcation criterion in accurately predicting ELs’ readiness for English-only instruction. PPIC recently published Reclassifcation of English Learner Students in California , a report about the relation- ship between reclassifcation policies and student outcomes four years later. In this report, we examine the academic progress of ELs in the two largest school districts in California—the Los Angeles Unifed School District (LAUSD) and the San Diego Unifed School District (SDUSD). Together, Los Angeles and San Diego serve more than 200,000 ELs, about 15 percent of the ELs enrolled in the state. Because of a growing consensus that language acquisition during the elementary school years infuences longer-term academic outcomes, we focus on students enrolled in these districts during the elementary school years and follow their progress through grade 12. Background on San Diego and Los Angeles San Diego Unifed and Los Angeles Unifed are large and diverse California school districts. Although Spanish is the most common language spoken by ELs in both districts (94% in LAUSD and 76% in SDUSD), the distribution of other languages spoken among ELs varies by district. In LAUSD, the other common languages are Armenian (1.1%), Korean (1.0%), and Filipino (1.0%); in SDUSD, the other common languages are Vietnamese (5.6%), Filipino (4.3%), and Somali (2.6%) (Table 1). Students in Los Angeles are more likely to be low- income than students in San Diego (77% versus 61%). Los Angeles students are more likely to be Hispanic and less likely to be white or Asian than students in San Diego. It is important to take into account the varying demographic characteristics of the student population in each district when comparing academic outcomes for ELs and native English speakers. Goals of the Study Tis report has four parts. Te frst part describes how students come to be reclassifed in LAUSD and SDUSD, the data we use for each district, and how we defne our student cohorts. Te second part examines whether attainment of current reclassifcation criteria in elementary school results in better student performance in middle school and high school—and whether the performance of ELs falters in the years afer reclassifcation. Because we are looking at whether students are making the transition to English-only instructional set - tings at the appropriate time and tracking their long-term prospects, our cohort consists of English Learners enrolled in LAUSD and SDUSD during their elementary school years. 4 We focus on students reclassifed in grades 2–5 in both school districts. 5 Tere are important policy questions about Changes to K–12 funding, standards, and assessments have intensifed California’s interest in reclassifying English Learners. AP P HOTO/RIC H PED RONCELLI San Diego Unifed and Los Angeles Unifed are large and diverse California school districts. Although Spanish is the most common language spoken by ELs in both districts, the distribution of other languages spoken among ELs varies by district. 5 Pathways to Fluency www.ppic.org EL students who arrive in the secondary school years, but these are beyond the scope of this report. Using student-level data from both districts, we are able to follow students over time—while they are ELs, at the time of reclassifcation, and for many years beyond. Most previous research has relied on a cross-sectional approach, which shows large diferences in academic achievement between native English-speaking students and ELs but ofen overlooks the confounding factor that the most successful EL students are reclassifed in early grades and “drop out” of the analyses. 6 Specifcally, we explore the following questions: 1. How do reclassifed fuent English profcient (RFEP) students fare on outcome measures such as the CST, the California High School Exit Examination (CAHSEE), retention in grade, and on-time graduation? 2. Do these outcomes vary by primary language or grade level at reclassifcation? 7 Although this report has a wide scope with regard to reclassifcation of ELs, it cannot tackle many important questions. For example, it does not study EL latecomers arriving in higher grades, nor does it study instructional diferences experienced by EL and RFEP students within either district or between districts. Rather, it provides a portrait of the progress of students who were ever ELs, dependent on their language status, in the contexts of the two districts. Te third part of the report explores relationships among individual reclassifcation criteria and a range of outcomes to determine which criteria are the most challenging for ELs and which are strongly associated with short- and long-term academic outcomes following reclassifcation. Because LAUSD and SDUSD have slightly diferent reclassifcation criteria, we can examine the relationship between more rigorous criteria and both aca - demic outcomes and reclassifcation rates. We also explore Table 1. K–12 enrollment in California, LAUSD, and SDUSD, 2012–13 CaliforniaLAU SD SDUSD Percentage Percentage Percentage Total enrollment 6,226,989 655,49413 0 , 270 Race/ethnicity African American 394,6956.361,78 6 9.413 , 261 10. 2 Hispanic 3 , 2 82 ,10 552.7482,534 73.660,616 46.5 Asian/Pacifc Islander 7 2 5, 81911 . 742, 26 4 6.418,799 14 . 4 White 1, 589, 39325.560,266 9. 230,271 23.2 Other 234,9773.88,644 1. 37, 3 2 3 5.6 Free/reduced price meal eligibility a3,472,481 5 7. 5489,777 76.676, 8 4 6 60.7 Fluent English profcient b1, 339, 56 6 21. 5239,753 36.62 7, 0 3 2 20.8 English Learners 1,346,33321. 617 0 , 7 9 7 26 .133, 851 26.0 Top four languages spoken among ELs Spanish, Vietnamese, Filipino, Cantonese 89.6 Spanish, Armenian, Filipino, Korean 96.7 Spanish, Vietnamese, Filipino, Somali 88.0 SOURCE: California Department of Education DataQuest for the 2012–13 school year. NOTES: SDUSD and LAUSD have similar shares of EL students (26%), but LAUSD has a much higher percentage of students who are fuent English profcient (FEP)—students who were either former ELs who have been reclassifed as fuent English profcient (RFEP) or were designated at school entry as fuent in English even though they were speakers of another language (initially fuent English profcient or IFEP). a Free/reduced price meal numbers are from the 2011–12 school year. b FEP numbers include both RFEP students and IFEP students. IFEP students speak a language other than English at home but were determined by their initial CELDT scores to be fuent in English. Pathways to Fluency 6 www.ppic.org the possibility that difcult criteria could unnecessarily delay reclassifcation for students.Finally, the fourth part of the report looks at com- binations of several criteria (such as English profciency level, performance on basic academic skills assessments, and report card grades) to determine which reclassifca- tion criteria are the best predictors of student success. Tis analysis may help these and other districts decide on the best criteria to use to reclassify EL students. How Do EL Students Get Reclassified? Te California Education Code requires that school districts develop policies and procedures to guide the reclassifcation of English Learners. 8 District-level reclassi - fcation standards must be based on four criteria approved by the SBE: performance in basic skills, an assessment of English profciency, teacher evaluation of academic per- formance, and the opinion of a parent or guardian. To be considered for reclassifcation from EL to fuent English profcient (FEP), students should—at a minimum—meet all four criteria. To meet the minimum basic skills recommendations for reclassifcation, students must score at the Basic level or higher on the CST in English language arts (ELA). Students must also demonstrate English profciency by achieving an overall profciency level (OPL) of Early Advanced or higher on the CELDT, and their scores on each subtest—listening, speaking, reading, and writing—must be rated as Inter - mediate or higher. In addition, teachers must certify that students meet district academic performance indicators and are ready to succeed in an English-only instructional program. Te district must advise parents and guardians of their right to participate in the reclassifcation process and encourage them to attend a face-to-face meeting. Districts have great latitude in setting their own reclas- sifcation policies, as long as they take into consideration the guidelines issued by the SBE. Reclassifcation criteria in LAUSD and SDUSD are somewhat more rigorous than the SBE guidelines. In San Diego, basic skills and English profciency requirements are higher; in Los Angeles, the teacher evaluation component specifes minimum report card grades as a condition of reclassifcation. Table 2 details current statewide, LAUSD, and SDUSD reclassifca- tion criteria. It is important to note that for the cohort we study, LAUSD also required marks in math courses of 3 or higher (on a 4-point scale) through the 2005–06 school year. SDUSD’s reclassifcation criteria have not changed since 2002. SDUSD’s reclassifcation rates are largely unchanged The California English Language Development Test and the California Standards Tests The CELDT is a state-mandated assessment of listening, speaking, reading, and writing in English that is administered in kindergarten through grade 12. It is used to identify stu- dents with limited English profciency to determine their levels of profciency and to assess progress in learning English. The CELDT must be administered within 30 days of enroll - ment to all students whose Home Language Survey indicates that a language other than English is spoken at home and annually to all continuing ELs who have not yet been reclas- sifed as FEP. CELDT results provide performance levels— Beginning, Early Intermediate, Intermediate, Early Advanced, and Advanced—for each of the subtests and an OPL. Until 2006, the listening and speaking subtests were combined into one. Starting in 2006–07, higher scores on the CELDT were required for each level of OPL. All students in grades 2 through 11—including ELs and most students receiving special education services—take the CSTs, state-mandated criterion-referenced tests that assess students’ mastery of the California content standards in Eng - lish language arts, mathematics, science, and history-social science. Results are reported as performance levels—Far Below Basic (1), Below Basic (2), Basic (3), Profcient (4), and Advanced (5)—and are used to identify individual students’ learning needs and assess school quality in state and federal accountability systems. Students with special needs who meet eligibility requirements may take the California Modi - fed Assessment (CMA) or California Alternate Performance Assessment (CAPA) rather than the CSTs. Districts that reclas - sify students before the 2nd grade CST scores are available use other assessments to make their decisions. 7 Pathways to Fluency www.ppic.org from 2005–06 to 2012–13 (10.4 to 10.5), whereas LAUSD’s rates increased from 9.5 to 13.7. 9 LAUSD’s increasing reclassifcation rates are probably associated with elimi - nating the use of mathematics grades as a reclassifcation criterion afer 2005–06. Student Data For this study, we use longitudinal student-level data from Los Angeles and San Diego Unifed School Districts for 2002 through 2012. (In this report, when we refer to In San Diego, basic skills and English profciency requirements are higher; in Los Angeles, the teacher evaluation component specifes minimum report card grades as a condition of reclassifcation. Both districts include reclassifcation criteria that go beyond the state minimum. a single year, such as 2002, we mean the 2001–02 school year.) Starting in 2002, we follow 2nd grade students through what would be their 12th grade year if they made on-time progress. 10 Because students transition from EL to RFEP status at various times, comparisons of EL and RFEP students can be complicated. To make the compari- sons as straightforward as possible, we focus our research on students who remain ELs through the end of 5th grade and students who are reclassifed as FEP by the end of 5th grade—and we focus only on students who are observed Table 2. Elementary grade reclassification criteria (2012–13) in both LAUSD and SDUSD are more rigorous than the SBE guidelines SBE guidelines L AU SD Structured English Immersion or M ainstream English Program aSD USD Performance in basic skills ELA CST score of 300 (Basic) or above 2nd grade: score on Literacy Periodic Assessment #2 or #3 of Basic or above b 3rd to 5th/6th grade: score of Basic or above on ELA CST EL A CST score of 333 (mid-Basic) or above Assessment of English profciency CELDT OPL of Early Advanced or higher All CELDT subtests at Intermediate or higher CELDT OPL of Early Advanced or higher All CELDT subtests at Intermediate or higherCELDT OPL of Early Advanced or higher At least three CELDT subtests at Early Advanced or higher; fourth subtest at Intermediate or higher Teacher evaluation of student academic performance Teacher certifcation that the student meets the district’s academic performance indicators Minimum marks of 3 on ELA courses of listening, speaking, reading, and writing Teacher certifcation that the student can be successful in core subject areas in a regular program designed for native and fuent speakers of English Parent or guardian opinion and consultation District provides notice to parents/ guardians of their right to participate in reclassifcation process and encourages them to attend a face-to-face meeting Parent consulted regarding student’s eligibility to reclassify, and letter must be printed and provided immediately; letter requires a parent signatureDistrict notifes parent/guardian of reclassifcation decision and provides opportunity to consult with staf regarding programs to further increase their student’s academic achievement SOURCES: California Department of Education (2012, p. 18). Los Angeles Unifed School District (2011). San Diego Unifed School District (2009). NOTE: The more rigorous reclassifcation criteria are shown in boldface. aBasic bilingual and dual-language programs have diferent teacher evaluation reclassifcation criteria.bPeriodic formative assessments to measure key ELA standards three times per year in grades 2–5 (Ofce of Curriculum, Instruction, and School Support, LAUSD Ofce of the Deputy Superintendent of Instruction, 2011). Pathways to Fluency 8 www.ppic.org in the data from 2002 to 2005. We compare both groups of students to native English speakers who must also be observed in their district data from 2002 to 2005. As a way of making sure that our cohort is not some- how anomalous, we looked at a more recent cohort of elementary school students who entered kindergarten in 2006–07 and remained in the same district through the 2012 school year. Although we cannot follow these stu- dents through to graduation, we can compare their ele- mentary grade outcomes to the earlier cohort to determine whether the longer-term outcomes of the earlier cohort are likely to be relevant for the present day in each district. Native English speakers are a smaller percentage of the main cohort in LAUSD than in SDUSD (top panel, Table 3). Te group of students who were ever ELs—the combination of current ELs and ELs reclassifed by the end of grade 5—is a much larger proportion of students in LAUSD than in SDUSD. 11 Within the ever-EL group, stu- dents reclassifed in elementary school constitute a greater proportion of students in SDUSD (45%) than in LAUSD (19%). Tis suggests that more rigorous reclassifcation criteria were used in LAUSD at this time. We fnd that the 2nd grade CELDT overall scores of students who were ever classifed as ELs are fairly simi- lar across the two districts (2.70 in LAUSD versus 2.51 in SDUSD). 12 However, when we separate the CELDT scores of EL students who remained ELs through 5th grade from those who were reclassifed by the end of 5th grade, a gap emerges—the 2nd grade scores for both groups are much higher in Los Angeles (3.45 and 2.42) than in San Diego (3.08 and 2.07). Tis same pattern is observed in the 2nd grade CST scores (reported as the share of students scoring Basic or above). Te overall similarity of scores for all students who were ever ELs combined with the diferences across two groups suggest that LAUSD’s reclassifcation criteria were more rigorous during the period we study. Tis may explain the higher average performance of EL students Table 3. Differences in test scores among student cohorts, LAUSD and SDUSD LAU SD SDUSD 2nd grade tests 2nd grade tests CST % Basic or above CST % Basic or above No. CELDT overall ELA M ath No.CELDT overall ELA M ath 2nd grade in 2002 RFEP5 4,3843.45 91.088.4 1,95 43.08 75. 2 71. 6 EL5 18 , 3672.423 7. 9 4 7.1 2,356 2.07 23.636.3 Ever EL (EL5 + RFEP5) 22,7512.7048.7 55.44, 310 2. 5148.0 54.0 Native English speakers 7, 0 74 NA69.1 66.9 4,926 NA8 3 .1 78.8 Kindergarten in 2007 RFEP5 10,9 9 63.7296.8 95.34,573 3.2695.394.9 EL5 11, 2 5 52.5649.659.04 ,121 2.0846.9 61. 8 Ever EL (EL5 + RFEP5) 22, 2513.0874 . 478.08,694 2.7072.4 79. 2 Native English speakers 11, 6 7 8 NA85.3 84.99,569 NA86.3 88.5 SOURCE: Authors’ tabulations from LAUSD and SDUSD student-level data, 2001–2012. NOTES: CELDT overall is the overall performance level on the CELDT and can range from 1 to 5. A score of 4 or higher on the frst CELDT taken results in a designation of IFEP (rather than EL). A score of 4 is required to be reclassifed from EL to RFEP. For CST results, we report the share of students scoring Basic or above. Because the CELDT was not administered to all students until 2002, the SDUSD CELDT overall included only 2nd grade students in 2002, not 2nd grade students in 2001. RFEP5 students are those reclassifed by the end of 5th grade. EL5 students are EL students who have not been reclassifed by the end of 5th grade. Results were signifcantly diferent across the language profciency groups. 9 Pathways to Fluency www.ppic.org who were and were not reclassifed by the end of 5th grade in LAUSD: it is plausible that because only the highest- performing ELs were reclassifed, many high-achieving students remained ELs, resulting in a higher average performance for both groups. Note also that native Eng- lish speakers have much higher 2nd grade CST scores in SDUSD than in LAUSD in the main cohort we study (those in 2nd grade in 2002). For the sake of comparison, we include a more recent cohort of students in Table 3. As shown in the bottom panel, almost half of the LAUSD students in the later cohort who were ever classifed as EL were reclassifed by the end of 5th grade. It is evident that the rate at which LAUSD’s EL students are reclassifed in elementary school has increased. 13 Tis is probably because LAUSD dropped the requirement that EL students earn marks of 3 or better in math courses to be reclassifed. SDUSD’s rate of reclassifcation in elementary school has also increased, but to a lesser extent. Notably, the diference in 2nd grade CST ELA and math scores between students in LAUSD and SDUSD in the older cohort disappears in the younger cohort, which faced lower reclassifcation standards. Tis supports the hypothesis that the higher performance of LAUSD students in the older cohort was the result of some high- achieving LAUSD students not being reclassifed, thus raising average achievement in both groups. 14 Te gap in the 2nd grade CELDT scores of students who were ever ELs in SDUSD and LAUSD increased for the 2007 kindergarten cohort, even though CST results are similar among ELs in the two districts. When we divide the EL group into those who remain ELs in 5th grade and those who are reclassifed by the end of 5th grade for the kindergarten 2007 cohort, the CELDT scores for this cohort are much higher in LAUSD. 15 Since we follow students over time, it is important to account for the possibility of students leaving LAUSD and SDUSD at diferent rates, thereby afecting our observation of student outcomes. We explored this possibility and con- cluded that attrition should not have a major efect on our results. 16 We include some results for students who later exit the district for comparison in technical appendix C. Long-Term Outcomes for Students Reclassified in Elementary School In this section, we examine academic outcomes (annual stan - dardized test scores, on-time progression including on-time high school graduation, and high school exit exam scores) to compare the longer-term performance of reclassifed students with that of native English speakers and English Learners. We focus particularly on former English Learners who were reclassifed in grades 2, 3, 4, or 5 and remained in the district for ten years (until 12th grade, if they made on-time prog - ress). We compare outcomes for these reclassifed students to those of EL students who were not reclassifed by the end of 5th grade and also to those of native English speakers. We are interested not only in how students perform at a given grade level but also in their performance over time. Tis is an important issue for reclassifed students because educators need to know if ELs reclassifed in elementary school con - tinue to be strong academic performers in middle and high school or if they might need supplemental services. In both Los Angeles and San Diego, students reclassi- fed in elementary school are among the best academic performers. Teir elementary school outcomes are well above those of English Learner students and in many cases they are on par with those of native English speakers. Te diferences between students who were and were not reclas - sifed in elementary school persist into middle and high school. 17 We see no evidence that the students who were reclassifed in elementary school falter relative to other students later in their educational trajectories. We examine academic outcomes to compare the longer-term performance of reclassifed students with that of native English speakers and English Learners. Pathways to Fluency 10 www.ppic.org Performance on California Standards Tests Students reclassifed in elementary school in both districts are top performers on the English Language Arts portion of the CST (Figure 1). Te mean ELA CST performance levels for reclassifed students in both districts and for native English speakers in SDUSD are all high mid-Basic (above 3.7) in 11th grade. In LAUSD, native English speak- ers in all grades do worse than students reclassifed in elementary school. Native English speakers in LAUSD also underperform their counterparts in SDUSD. Tis gap could be partly explained by the diference in the share of parents without high school diplomas: 10 percent of parents of native English speakers in LAUSD versus 2 percent of parents of native speakers in SDUSD (see techni- cal appendix Tables A1 and A2). Performance levels for EL students who were not reclassifed in elementary school are below Basic in both districts (2.7 in SDUSD and 2.9 in LAUSD in 11th grade, for instance). Math CST scores reveal similar patterns. In LAUSD, students reclassifed in elementary school have much bet- ter math scores than native English speakers, probably because of the math grades required for reclassifcation in LAUSD at the time. In San Diego these students’ math performance is virtually identical to that of native English speakers. Te scores of SDUSD and LAUSD students who remained ELs throughout elementary school are within 0.3 mean performance levels of each other from 2nd to 7th grade. 18 In San Diego, students reclassifed in elementary school and native English speakers have comparable mean performance levels on the math CST by grade level. 19 We see no evidence in either the ELA or math CST scores that the reclassifed students in our study falter at higher grade Grade LAUSD SDUSD 7Grade 2 3 4 5 6 7 8 9 10 11 Grade 2 3 4 5 6 7 8 9 10 11 23 456Grade23 45 67 Mean performance level CST ELA 1.5 2.0 2.5 3.0 3.5 4.0 4.5 CST math Mean performance level 1.5 2.0 2.5 3.0 3.5 4.0 4.5 1.5 2.0 2.5 3.0 3.5 4.0 4.5 1.5 2.0 2.5 3.0 3.5 4.0 4.5 RFEP5 Native English speaker EL5 Figure 1. Students reclassifed in elementary school perform well on the CST in later grades SOURCE: Authors’ calculations from SDUSD and LAUSD student-level data, 2001–2012. NOTES: Students in 8th grade can take many diferent math CSTs, so we present scores only through 7th grade. Students must be making on-time grade progress for their scores to be included here. We examine on-time grade progression separately in Table 4. CST scores are examined for students remaining in the district through 2012. RFEP5 students are those reclassifed by the end of 5th grade. EL5 students are EL students who have not been reclassifed by the end of 5th grade. Mean diferences for language profciency groups are statistically signifcant as are mean grade-level scores within each district in LAUSD and in SDUSD, with one exception: RFEP math scores in grades 6 and 7 are not signifcantly diferent from native English-speaking students’ scores. 11 Pathways to Fluency www.ppic.org levels relative to native English speaking students. Declines in these students’ mean performance levels in the high school grades are similar to declines among native English speakers. For example, between 9th and 11th grade in LAUSD, ELA CST scores fell by 0.09 CST mean performance levels for native English speakers but only by 0.03 CST mean perfor - mance levels for students reclassifed in elementary school. In SDUSD, the decline between 9th and 11th grade on ELA CST mean performance levels was 0.23 for native English speakers and 0.17 for these reclassifed students. Tis is an important fnding: removal of language supports for ELs who are reclas - sifed does not seem to have hurt their academic progress relative to that of native English speakers. On-Time Progression toward Graduation In both districts, English Learner students not reclassifed by the end of 5th grade are the least likely to make on-time (or better) grade progress relative to other student language pro - fciency groups (Table 4). In Los Angeles, students reclassifed by the end of 5th grade are more likely than native English speakers to progress on time, whereas in San Diego, out - comes are similar for the two groups. In both districts, there is a substantial drop-of in on-time grade progression from 9th to 10th grade. Tis is a greater issue in LAUSD, where only 73 percent of ELs not reclassifed in elementary school are on time by grade 10 (compared to 82% of native English speakers). In SDUSD, 83 percent of ELs not reclassifed in elementary school and approximately 95 percent of students who were reclassifed in elementary school and native Eng- lish speakers are making on-time grade progress by 10th grade. In high school, grade progression is determined by unit accumulation: failing one or more classes can mean that a student is recorded as being in 9th grade for two years in a row. A similar pattern in 9th to 10th grade progression is observed statewide (Hill, Weston, and Hayes 2014). In both districts, English Learner students not reclassifed by the end of 5th grade are the least likely to make on-time (or better) grade progress relative to other student language profciency groups. Table 4. Students reclassified in elementary school maintain high rates of on-time grade progression through high school Percentage on-time advancement to: Percentage on-time graduation a 9th grade10 th grade 11th grade12 t h g r a d e B y 2 012 L AU SD RFEP5 9890 88 85 82 EL5 897368 64 58 Native English speaker 948277 74 66 SD USD RFEP5 98959392 78 EL5 9283 8183 55 Native English speaker 98959392 78 SOURCE: Authors’ calculations from SDUSD and LAUSD student-level data, 2001–2012.aSDUSD’s graduation requirements appear to be more challenging than LAUSD’s. SDUSD students must have a minimum grade point average (GPA) of 2.0 and must successfully complete three years of science and three years of math. (See Table 1 in Betts, Zau, and Bachofer 2013.) LAUSD requires a minimum GPA of 1.0, two years of science, and two years of math (LAUSD 2010). Because we do not have dropout data for LAUSD, we exclude dropouts from our graduation calculations here and throughout the report. When we include dropouts in the SDUSD data, graduation rates decrease only slightly (see technical appendix Table C1). On-time grade-level progression and graduation were examined for students remaining in their district through 2012. RFEP5 students are those reclassifed by the end of 5th grade. EL5 students are EL students who have not been reclassifed by the end of 5th grade. Language profciency group results are signifcantly diferent from each other for on-time grade progression in each grade in LAUSD and SDUSD, with one exception: in SDUSD, RFEP5 students are not signifcantly diferent from native English-speaking students. On-time graduation rates are statistically diferent for all language profciency groups in both districts. Pathways to Fluency 12 www.ppic.org Students reclassifed in elementary school maintain high rates of on-time grade progression throughout high school, and we see no evidence that their strong perfor- mance erodes relative to other language profciency groups. Tis is especially true in Los Angeles, where native English speakers and students not reclassifed in elementary school steadily lose ground from 9th to 12th grade relative to EL students who are reclassifed in elementary school. We also compare graduation rates, but it is important to note that graduation requirements are more rigorous in SDUSD than in LAUSD. 20 Graduating on time requires sufcient credit accumulation, minimum GPA require - ments, and passing specifc courses in addition to passing the CAHSEE. We fnd that in LAUSD, reclassifed students in the cohort we study are more likely to graduate than native English speakers (82% versus 66%). In SDUSD, these two groups of students are equally likely to graduate (78%). Performance on the CAHSEE In both Los Angeles and San Diego, students reclassifed in elementary school are more likely than any other group to pass the CAHSEE in 10th grade and by 12th grade (Figure 2). In fact, nearly 100 percent do so by 12th grade. Students who are not reclassifed in elementary school have low 10th grade CAHSEE passage rates in both dis- tricts (61% in LAUSD and 51% in SDUSD) but do make great strides by the end of grade 12, with approximately 80 percent passing in both districts. 21 Outcomes for Reclassified Students, by Language Spoken and Grade Level at Reclassification Here, we briefy summarize analyses that examine aca- demic outcomes for students reclassifed by the end of 5th grade by their primary language and their elementary school grade level at reclassifcation. We fnd that Spanish- speaking students reclassifed in elementary school have less positive outcomes than those who speak other lan- guages. However, the performance of Spanish speakers is still stronger than that of native English speakers in LAUSD (although not always in SDUSD). (See technical appendix Tables C2, C3, and C4 and technical appendix Figure C3 for full results.) When we examine outcomes according to the grade level in which students are reclas- sifed (grades 2 through 5), we fnd that students who are reclassifed in 2nd or 3rd grade have better outcomes than those reclassifed in 5th grade, on average, in both districts. (See technical appendix Figure C4 for CST results and further discussion of other outcomes.) Students Reclassified Early Have Good Outcomes and Reclassified Students Do Not Lose Ground We have found that in both Los Angeles and San Diego, students reclassifed by the end of 5th grade are among the best performers, and we see no evidence that their performance falters at higher grade levels relative to native English speakers. We also fnd evidence that students reclassifed at the end of elementary school (5th grade) have slightly lower academic outcomes than students reclassi- fed earlier. Research using a wider range of reclassifcation grade levels has found evidence that students reclassifed in high school do not do as well as students reclassifed at earlier grade levels (Hill, Weston, and Hayes 2014). Tis suggests an explanation for the apparent narrowing of the performance gap between reclassifed students and other Percentage passing CAHSEE overall Passing CAHSEE by 2012 Passing CAHSEE in 2010100 90 80 70 60 50 40 30 20 100 Figure 2. Virtually all students reclassifed in elementary school pass the CAHSEE by grade 12 SOURCE: Authors’ calculations from SDUSD and LAUSD individual student data, 2001–2012. NOTES: SDUSD data include students in grade 2 in 2001 or 2002; LAUSD data focus on those in grade 2 in 2002. The 2010 CAHSEE includes students taking the CAHSEE who should be 10th graders but have been retained in 9th grade. Not all retained 9th graders in LAUSD took the 2010 CAHSEE. Students making on-time progress would be in the 12th grade in 2012. CAHSEE scores are examined for students still in their district in 2012. RFEP5 students are those reclassifed by the end of 5th grade. EL5 students are EL students who have not been reclassifed by the end of 5th grade. Language profciency group results are statistically signifcantly diferent from each other for CAHSEE passage in both 2010 and 2012 in both districts. RFEP5 EL5 Native English speaker Native English speaker RFEP5 EL5 LAUSD SDUSD 13 Pathways to Fluency www.ppic.org students in high school in cross-sectional research fnd- ings: the RFEPs in these studies include students reclassi- fed in high school, who do not perform as well as students reclassifed in elementary or middle school (Hill 2012; Gándara and Rumberger 2006). Is the stronger performance of reclassifed students simply a result of skimming of the highest-achieving EL students (who may or may not have benefted from EL instruction), or do reclassifed students make academic gains because of English Learner instruction and subse- quent placement in mainstream instructional programs? Or is it simply that the student characteristics associated with elementary school reclassifcation are also associated with strong academic outcomes in middle and high school? Te CST scores of English Learner students not reclassi- fed by the end of 5th grade do improve by grade level more than native English speaker scores do, which suggests that ELD instruction benefts both ELs and reclassifed students. Although these questions are beyond the scope of this report, it is important for districts and the state to consider them more fully. Without a doubt, the use of reclassifcation criteria means that the best-performing students leave the English Learner group—indeed, many native English speakers do not meet the minimum CST scores that EL students are required to meet to be reclassifed. Tis is most clear in LAUSD, where students reclassifed in elementary school outperform not only native English speakers but also stu- dents who are initially fuent in two languages. In SDUSD, reclassifed students also perform well but do not generally exceed the performance levels of native English speakers. Te performance gap between the two districts is most likely explained by the relative difculty of reclassifcation criteria and, perhaps, also by the lower socioeconomic sta - tus of native English speakers in LAUSD relative to SDUSD. Reclassification Criteria and Long-Term Outcomes Districts have latitude in setting reclassifcation policies as long as they use the four criteria required in state law and take into consideration the SBE’s guidelines. To develop efec - tive policies, districts need to understand the relationship between the thresholds they establish for each criterion (or the use of additional criteria) and student outcomes. In this section, we examine the reclassifcation criteria that are most difcult for students to meet in each district. We also exam - ine middle and high school academic outcomes to under - stand how they are afected by more rigorous reclassifcation thresholds in each district—in particular, we look at the efect of higher CST cut-of scores and the use of course grades. Which Criteria Are Most Challenging? In a recent statewide survey of school districts’ reclassif- cation policies, respondents reported that the basic skills criterion (CST) was most difcult for their EL students to meet (53% for elementary grades, 62% for middle school grades, and 68% for high school grades). In elementary school, the English profciency requirement (CELDT) was a close second (40%) (Hill, Weston, and Hayes 2014). To fnd out whether these district perceptions are borne out by the data on student performance, we have adapted a technique used by Robinson (2011) to determine which of a district’s reclassifcation criteria are actually the most difcult for students to meet. Since EL students who are not reclassifed may not meet any of the reclassifcation criteria, we need to create a measure of which criterion is, on average, the most difcult for students to overcome. 22 Tis requires detailed knowledge of each district’s reclassi - fcation criteria and many observations of student perfor- mance. For each EL student in each grade, we calculate the distance between that student’s test scores or course marks We fnd that Spanish-speaking students reclassifed in elementary school have less positive outcomes than reclassifed students who speak other languages. Pathways to Fluency 14 www.ppic.org and the scores or grades that would allow that student to be reclassifed under his or her district’s criteria. 23 Te distance between the score or grade and the reclassifcation requirement determines the difculty of the requirement. 24 LAUSD has a lower ELA CST threshold than SDUSD (300 rather than 333) and allows two CELDT subtests to be below Early Advanced (SDUSD allows one). However, LAUSD requires a mark of 3 or better (on a 1- to 4-point scale) in ELA courses as a condition of reclassifcation (until 2006–07, it also required a 3 or better in math). SDUSD does not have any course mark requirement for reclassifca - tion, and such a requirement is not suggested in the SBE guidelines. In Figure 3, we show the biggest reclassifca- tion challenges for LAUSD’s ELs, taking into account the lower cut-point requirement on the CST but using the Early Advanced CELDT cut-of for overall profciency level and all subtests. In elementary grades, the reading component of the CELDT is the bigger barrier for students. Te CST is not a major obstacle until 5th grade, at which point it is the big - gest constraint for about 40 percent of ELs and remains the most common difculty through the end of middle school. However, in analyses where we also included the ELA and math mark requirements for LAUSD elementary school students, we found that marks are the most common stum - bling block for 4th and 5th graders; nearly 70 percent had a math or ELA mark as their most difcult requirement (see technical appendix Table D1). 25 Te ELA writing mark was the biggest challenge for about one-quarter of 4th and 5th grade students. Math marks were the biggest problem for 16 percent of 4th graders and 24 percent of 5th graders. With the mark requirement in place, the relative importance of the CST fell dramatically, and the CELDT reading subtest fell in importance by approximately 40 percentage points in 4th grade to 20 percent and approximately 30 percentage points in 5th grade to 15 percent. In discussions with district staf, we learned that LAUSD has recently convened meet - ings to discuss students who meet all reclassifcation criteria except ELA marks. In some cases, staf may decide to reclas- sify students with marks below 3. When we analyze San Diego’s reclassifcation require - ments, we fnd that the ELA CST requirement of 333 or higher (which is 33 points higher than both the LAUSD and suggested SBE thresholds) is the main criterion preventing students from being reclassifed at grades 4 through 7, with a peak of nearly 60 percent in grade 5 (Figure 4). For grades 2 and 3, as well as grades 8 through 10, the CELDT subtest Percentage with binding reclassifcation constraint CELDT overall CELDT writing CELDT reading CELDT listening speaking CST ELA100 90 80 70 60 50 40 30 20 100 Figure 3. In Los Angeles, the CELDT reading requirement is the most common obstacle to reclassifcation in elementary grades, but the CST becomes as big a barrier in later grades SOURCE: Authors’ calculations using LAUSD individual student data, 2002–2012. See note 23. NOTES: Students must be in 2nd grade in 2002 and must have been in the sample with complete observations of test scores for three consecutive years. Binding reclassifcation criteria for LAUSD do not include grade requirement criteria. We do not show these results because we have only ELA and math marks for grades 4 and 5. Those are shown in technical appendix Table D1. Grade 2 3 4 5 6 7 8 9 10 11Percentage with binding reclassifcation constraint 100 90 80 70 60 50 40 30 20 100 Figure 4. In San Diego, the CELDT reading requirement and the ELA CST requirement are the main obstacles to reclassifcation SOURCE: Authors’ calculations using SDUSD individual student data, 2001–2012. See note 23. NOTES: Binding reclassifcation criteria, SDUSD. Students must have been in 2nd grade in 2001 or 2002 and in the sample with complete test score observations for three consecutive years. Grade 2 3 4 5 6 7 8 9 10 11 CELDT overall CELDT writing CELDT reading CELDT listening speaking CST ELA 15 Pathways to Fluency www.ppic.org score in reading is furthest away from the reclassifcation threshold. 26 Rarely is the requirement of an overall perfor - mance level of Early Advanced or the score on the writing subtest the biggest obstacle for ELs. Listening/speaking is the third-most-challenging criterion. 27 Assessing the Effect of Rigorous Reclassification Criteria As we have seen, some of the reclassifcation criteria in LAUSD and SDUSD are more stringent than those sug - gested by the state’s guidelines. Tese diferences in district criteria can help us answer some important questions. Sup - pose that some of the students who met the state’s suggested reclassifcation criteria but not the more rigorous require - ments set by their districts had indeed been reclassifed. Would they have fared worse academically than those who met all of the district criteria? To explore this question, we divide the students who were reclassifed in one district into two groups: those who would have met the other district’s more stringent requirements and those who would not. 28 What If SDUSD Lowered Its CST Requirement? We begin by looking at what would have happened in SDUSD if the district had allowed students scoring between 300 and 332 on the ELA CST to be reclassifed. To do this, we compare LAUSD students who scored between 300 and 332 on the ELA CST to LAUSD students who scored between 333 and 366. 29 (Tese scores are from the year each LAUSD student was reclassifed.) First, we fnd that 25 percent of the LAUSD students who were reclassi- fed in elementary school scored between 300 and 332 in the fnal year before they were reclassifed and therefore would not have been reclassifed in SDUSD. We fnd that in middle school, the reclassifed students who scored in the higher range just before reclassifcation in elementary school obtained higher scores on the grade 8 ELA CST and grade 7 math CST than those who were reclassifed with lower scores (Figure 5). However, these diferences are relatively small—only about 0.35 mean performance level on the grade 7 math CST and 0.40 on the grade 8 ELA CST. Results for other academic outcomes are similar. Students with higher CST scores (333–366) when they are reclassifed in elementary school pass the CAHSEE at higher rates than students with scores in the lower range (300–332). Tis gap narrows as students continue to take the CAHSEE in grades 11 and 12; by 12th grade, 92 percent of those reclassifed with lower CST scores pass the CAHSEE as compared to 97 percent of those with higher scores. (See technical appendix Figure E1.) Tere is almost no difer- ence in on-time grade progression in high school or on- time graduation between students scoring in the high and low ranges on the CST in their year of reclassifcation. (See technical appendix Figure E2.) What If SDUSD Added Course Mark Requirements? Now we examine what might happen if SDUSD added academic mark requirements to its reclassifcation crite- ria. 30 If we apply LAUSD’s ELA and math mark require- ment to reclassifed EL students in SDUSD (using their estimated grades), we make a striking fnd: most of them CST mean performance level 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 Figure 5. Reclassifed LAUSD students who meet SDUSD’s CST requirement have slightly higher CST scores in middle school SOURCE: Authors’ calculations using LAUSD individual student data, 2002–2012. NOTE: Mean diferences for students predicted to satisfy SDUSD CST requirements are statistically signifcant in LAUSD. 8th grade ELA CST 7th grade math CST Reclassifcation CST 300–332 score Reclassifcation CST 333–336 score Diference Some of the reclassifcation criteria in LAUSD and SDUSD are more stringent than those suggested by the state’s guidelines. Pathways to Fluency 16 www.ppic.org would not have been reclassifed in the year SDUSD reclas- sifed them. Te percentage of students who would not have been reclassifed ranged from 71.2 percent in grade 3 to 92.5 percent in grade 4. 31 Students reclassifed in elementary school in San Diego who met LAUSD’s ELA mark criteria scored higher in both the 8th grade ELA and 7th grade math components of the CST than students who did not (Figure 6). For the ELA CST, students who met the Los Angeles criteria averaged a performance level of 4.3 versus 3.4 for those who did not meet the criteria. For the math CST, students in San Diego who met the Los Angeles criteria averaged a performance level of 4.2 versus 3.4 among those who did not meet the criteria. Tese diferences in middle school CST scores based on meeting the Los Angeles ELA GPA requirement (or not) are larger than those observed using high and low CST cut-ofs (Figure 5). In the longer term, the diference in outcomes can be quite large. Among reclassifed students in SDUSD who met the LAUSD criteria, the passage rate on 10th grade CAHSEE was 97 percent, as opposed to 86.5 percent for those who did not meet the criteria. However, both groups had very high passage rates by the end of 12th grade (100% versus 97.8%). 32 Graduation rates ten years afer grade 2 were also very similar overall (89.2% versus 88.3%), although there is some variation depending on the grade level at which students were reclassifed. SDUSD students who were reclassifed in grade 2 without meeting the LAUSD criteria graduated from high school at a higher rate than students reclassifed in grades 3 through 5 (92.6% versus 83.3%). 33 Are Rigorous Criteria Worth It? EL students in Los Angeles fnd the ELA course mark requirement for reclassifcation the most difcult criterion to meet. In San Diego, where grades are not a reclassifca- tion criterion, the ELA CST and the CELDT reading sub- test are the most difcult criteria. In general, we fnd that outcomes for students who were reclassifed in elementary school by meeting more challenging district-adopted reclassifcation standards are, in many cases, only slightly better than outcomes for students who were not reclassi- fed. Given that large numbers of students are apparently held back from reclassifcation because of the more difcult criteria—with very little long-term gain—these fndings suggest that setting more challenging reclassifcation crite- ria may not be benefcial. When we compared reclassifed students in one district who would and would not have been reclassifed using the other district’s criteria, we found better outcomes for those reclassifed who met the higher reclassifcation criteria. If LAUSD had used SDUSD’s higher CST cut-of, 25 percent of LAUSD’s students reclassifed during elemen - tary school would have faced delayed reclassifcation. Similarly, we estimate that over 70 percent of SDUSD’s students reclassifed during elementary school would have faced delayed reclassifcation if SDUSD had used LAUSD’s ELA and math mark requirements. Our fndings help to explain why, for the cohorts we study, students in the two districts who were ever classifed as EL have similar CST and CELDT scores, but once we divide these students into those who were reclassifed in elementary school and those who were not, LAUSD students have higher CST and CELDT scores than SDUSD students. It appears that LAUSD’s reclassifcation policies keep some of the top- achieving EL students in EL status. CST mean performance level 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 Figure 6. Reclassifed SDUSD students who meet LAUSD’s ELA course mark criteria have higher CST scores in middle school SOURCE: Authors’ calculations using SDUSD individual student data, 2001–2012. NOTE: Mean diferences for students predicted to satisfy LAUSD GPA requirements are statistically signifcant in SDUSD. 8th grade ELA CST 7th grade math CST Reclassifcation with low marks Reclassifcation with high marks Diference 17 Pathways to Fluency www.ppic.org It is important to note that using each of the two districts’ reclassifed students to estimate diferences in outcomes in the other district is an imperfect exercise. Ideally, we would instead examine outcomes for reclas- sifed students in a third district that did not increase its reclassifcation criteria beyond the state-recommended minimum requirements. It is possible that such a compari- son would have revealed larger gains (and bigger reduc- tions in the share of high-performing EL students being reclassifed) as a result of the more rigorous requirements in our two districts. Finally, this exercise cannot account for diferences in English Language Development pro- grams and supports across the two districts. Which Reclassification Criteria Best Predict Student Success? SBE guidelines for reclassifcation advise districts to use CELDT and CST scores but allow the use of student grades in making reclassifcation decisions. 34 However, these guidelines are about to change. New ELD standards were adopted in 2012, and a replacement for the CELDT should be implemented by 2016. California is limiting statewide administration of the CST in 2013–14 in preparation for the introduction of the Smarter Balanced Assessments in 2014–15. Given these imminent changes to the state testing system, now is the perfect time to determine which of these three factors (CST, CELDT, and course marks) best predict success in middle and high school for elementary school students who are English Learners. 35 Will grades alone be sufcient? Tey are somewhat subjective, and grading stan- dards can vary across districts and even across teachers within a district. We also consider whether the practice in California of using both “basic skills” and “English prof- ciency” requirements through two separate tests produces complementary or redundant information. We compared the predictive power of three indicators, measured in grade 5: the CELDT subtest performance levels, the ELA CST performance level, and ELA course marks. 36 We estimated models of fve middle- and high- school outcomes as functions of these three indicators for EL students not reclassifed in elementary school: ELA CST scores in grade 8, math CST scores in grade 7, CAHSEE passage in grade 10, grade retention before grade 10, and on-time graduation. For each EL student, we used variables from 1 to 5 for performance levels on the CST and the CELDT and from 1 to 4 to capture marks (in each of read- ing, writing, listening, and speaking in Los Angeles, and in each of reading, writing, and oral language in San Diego). For both districts, when we attempted to explain test-score-based outcomes in 7th through 10th grades, CST test scores had the most explanatory power, followed by CELDT scores and ELA marks. Taken together, these three indicators captured one-third to one-half of the variation in grade 7 math CST achievement, grade 8 ELA CST achievement, and CAHSEE passage in grade 10. It is These fndings suggest that setting more challenging reclassifcation criteria may not be benefcial. With changes to the state testing system ahead, now is the perfect time to determine which factors best predict academic success for English Learners. MELANIE STET SON F REE MA N/TH E CHR I STIAN SCIENCE MONITO R Pathways to Fluency 18 www.ppic.org perhaps unsurprising that grade 5 CST scores are strong predictors—like the outcome measures we used in our model, they are test-based measures of academic achieve- ment. Figure 7 shows the percentage of the variation in stu- dent outcomes that we could explain using any one of these variables and all three variables together. Using CST scores alone led to models that explained slightly less variation than when we used all three predictors. However, for the two high school outcomes (whether a student had been retained a grade by the end of grade 10 or graduated within ten years of entering grade 2), the explanatory power of these variables, alone or together, was much weaker. We also estimated models that included all three combinations of pairs of these predictor variables: CST, CELDT, and GPA. Technical appendix Figure G1 shows that all three of these pairs perform almost as well as the model that includes all three sets of variables. We found that in LAUSD, using GPA together with either CST or CELDT scores explained almost as much of the variation in outcomes as did the full model with all three sets of vari - ables. But in SDUSD, the models using CST and CELDT scores were the best in explaining middle school CST scores and CAHSEE passage, and all of the models performed about equally well in explaining grade retention and graduation on time. For the cohorts of students we studied, SDUSD (unlike LAUSD) was not yet using a standards- based report card that stipulates the criteria teachers must use to assign grades. It could be that standards-based report cards tend to contain more reliable information on student performance. What insights does this analysis provide for state and district policymakers? Although one of the best predic- tors, the CST is no longer being administered in California schools; its replacement, which will be phased in during 2014–15, could have even better predictive power. In the interim, using only one indicator, such as the CELDT, would not produce markedly worse predictions of suc- cessful reclassifed student outcomes than would a more complex measure. Policy Implications In the context of the coming overhaul of the state testing system associated with the implementation of the CCSS, policymakers’ interest in instituting standardized reclas- sifcation criteria across the state, and the funding incen- tives of the LCFF, we have provided a timely review of the measures used by the state’s two largest school districts to determine when EL students are ready for English-only instruction. CST ELA grade 8 CST math grade 7 CAHSEE passage grade 10 Retained in or before grade 10 Graduated on time Percentage of variation explained 50 45 40 35 30 25 20 15 10 5 0 Figure 7. The percentage of variation in student outcomes explained by grade 5 test scores and the ELA GPA of English Learners varies SOURCE: Authors’ calculations from SDUSD and LAUSD individual student data, 2001–2012. NOTES: SDUSD includes those in grade 2 in 2001 or 2002. LAUSD data focus on those in grade 2 in 2002. CELDT subtests ELA GPA CST CELDT subtests, GPA, and CSTLAUSD CST ELAgrade 8 CST math grade 7 CAHSEE passage grade 10 Retained in or before grade 10 Graduated on time SDUSD 50 45 40 35 30 25 20 15 10 5 0 19 Pathways to Fluency www.ppic.org We found that in both LAUSD and SDUSD, students who fnished 2nd grade in 2002 and were reclassifed as fuent English profcient by the end of grade 5 generally did very well in middle and high school, performing about the same as or better than native English speakers on a variety of academic outcomes. Further, we found no evidence that reclassifed students’ performance faltered relative to that of native English speakers. Te key question is whether English Learners—in Los Angeles, San Diego, and other districts with more rigorous reclassifcation criteria than those suggested by the state—would beneft from being reclassifed slightly sooner, through an easing of reclas- sifcation standards, which would allow districts to con- centrate their resources on the most linguistically needy English Learners. Our fndings in the two largest districts in California lead us to believe that the answer is yes. We also believe that, in the longer term, standardizing reclas- sifcation policies across districts would allow educators and policymakers to compare outcomes across the state— something that will be vitally important as the LCFF is implemented. We end with several recommendations to help ensure a successful transition to new policies. 1. Even though the CST will not be administered in 2013–14, our results suggest that districts can make accurate reclassifcation decisions using only the CELDT. When replacements for the CST and CELDT become available, the state should consider allowing dis - tricts to reclassify students on the basis of just one test. Despite the diferences in student population (languages spoken, share of low-income students, and racial/ethnic distribution) and diferent reclassifcation criteria across the two school districts, the predictors of successful aca- demic outcomes for ever–EL students are similar. We doubt that California will ever abandon the CELDT or its successor, because schools need an objective method to evaluate the language abilities of new arrivals to the district whose frst language is not English, and these students can arrive at any time of the school year. Tests such as the CST and its successor cannot accomplish this goal. Further, the CELDT helps educators measure the progress of English Learners year to year. Because both the CELDT and the CST can predict EL students’ subsequent outcomes quite well, it makes sense for the state to consider whether an EL student who demonstrates sufcient mas- tery of English on either test should be reclassifed without having to face a second hurdle. 2. In the two largest districts in California, EL students fnd the CELDT writing requirement less challenging than the CELDT reading requirement. In designing new tests and reclassifcation standards, the state should consider the relative rigor of its reading and writing requirements. Our data suggest that the current CELDT writing require - ment is relatively easy compared to the CELDT reading requirement. Indeed, in LAUSD, where ELA marks on report cards are part of the reclassifcation criteria, the writing grade requirement is more challenging for EL students to meet than the required mark in reading. Given that the replace - ment exam for the CELDT is currently being developed, this might be an ideal time for the state to reconsider its relative expectations about reading and writing for EL students. 3. Districts should carefully consider whether their reclassifcation standards need to be more rigorous than the state-recommended minimum. A difcult policy question is whether either San Diego or Los Angeles is setting reclassifcation criteria too high or low. In part, the answer depends on how well we expect English Learners to perform academically once reclassi- fed. We examined reclassifed students in one district to see whether higher reclassifcation standards in the other district held back from reclassifcation students who would have fared poorly, had they been reclassifed. We found that more rigorous reclassifcation requirements in both districts—those related to CST scores in San Diego and academic marks in Los Angeles—were associated with slightly better outcomes in later grades, but the diferences were small. Further, we found that there is a downside to Pathways to Fluency 20 www.ppic.org these additional requirements: they prevented or delayed the reclassifcation of large numbers of students.We looked for variations across EL subgroups and found that students reclassifed at earlier grade levels had somewhat better outcomes in secondary school than those reclassifed later in the elementary school years. It is possible that students are benefting from English-only classroom instruction in the early grades or from not being labeled as ELs in later elementary school. Or it could be that students reclassifed early in elementary school had initial advantages in language profciency and academic preparation. Our current research cannot distinguish among these possibilities. 4. Afer careful consultation with districts, the state should consider establishing a uniform set of reclassi- fcation criteria for all school districts. Te issue of difering reclassifcation criteria is especially important, given that California’s new Local Control We found that more rigorous reclassifcation requirements in both districts were associated with slightly better outcomes in later grades, but the diferences were small. The key question is whether English Learners would beneft from being reclassifed slightly sooner, allowing districts to concentrate on the most linguistically needy. SA NDY HU FFAKER /CO R BIS 21 Pathways to Fluency www.ppic.org Technical appendices to this report are available on the PPIC website: www.ppic.org/content/pubs/other/514LHR_appendix.pdf Funding Formula increases funding for English Learners, which could create a disincentive for districts to reclassify students. Te state needs to consider whether it makes sense for some districts to have more rigorous reclassifcation criteria than others. A standard set of criteria could improve fairness for students and make it much easier to monitor the progress of students who have ever been English Learners. In sum, the process through which EL students are reclassifed as fuent English profcient is quite complex. Not only must students reach thresholds on two difer- ent tests, but individual districts can and do set their own requirements, which can be quite diferent from the SBE guidelines and from those of other districts. As the state implements new standards and assessments, and as the new funding formula takes efect, it should also consider making changes to reclassifcation criteria, including considering allowing EL students to meet the requirement on either the Smarter Balanced Assessment or the replace- ment for the CELDT. Te establishment of statewide reclassifcation criteria at a reasonable level of difculty could allow districts to concentrate their LCFF dollars on their lowest-performing students without slowing the academic progress of ELs who are performing well enough to be reclassifed in elementary school. Pathways to Fluency 22 www.ppic.org Notes 1 California Department of Education (CDE) DataQuest. 2 California Department of Education. 3 It is expected that these regulations will be adapted depending on their success in the frst year of LCFF implementation. 4 Further, statewide in 2012–13, two-thirds of English Learners were enrolled in the elementary school grades (K–5) (California Department of Education DataQuest). 5 In this cohort, very few students in SDUSD and almost none in LAUSD were reclassifed in 2nd grade. Statewide, reclassifcation before 2nd grade is rare (Hill, Weston, and Hayes 2014). 6 See Saunders and Marcelletti (2013), Hill (2012), EdSource (2008), and Gándara and Rumberger (2006) for examples. When RFEP and EL students are combined into an “ever-EL” group, the gap between ever-EL and native English-speaking students is considerably smaller and has declined somewhat over time (Saunders and Marcelletti 2013). However, even cross-sectional research that refnes comparison groups for ELs cannot account for the time since reclassifcation or for new entrants to the EL population. 7 Previous research has demonstrated a narrowing of the achievement gap following reclassifcation (Silver, Saunders, and Zarate 2008), but there is reason to believe that the timing of reclassifcation also matters. ELs who are reclassifed quickly have better long-term academic outcomes than those who con- tinue in EL status for many years (Flores, Painter, and Pachon 2009). Not surprisingly, there is a great deal of interest in under- standing the role of reclassifcation standards in the ultimate success of ELs (Parish et al. 2006). 8 California Education Code Section 313. 9 Ofcial CDE reclassifcation rates reported. 10 Because SDUSD student data are available earlier and the size of the student population is smaller, the cohort of students for SDUSD combines 2nd graders from 2001 and 2nd graders from 2002 and follows students to 2011 and 2012, respectively. 11 Te size of the ever-EL group may itself afect educational out- comes for the ever-EL group. However, determining the efect of peers on educational outcomes for ELs and RFEPs is beyond the scope of this report. 12 Te OPL on the CELDT can range from 1 to 5. Te correspon- dence between numbers and performance levels is as follows: 1 = Beginning, 2 = Early Intermediate, 3 = Intermediate, 4 = Early Advanced, and 5 = Advanced. 13 Te overall reclassifcation rates in LAUSD have increased (9.5% to 13.7%) during this period, whereas those in SDUSD have remained steady (10.4% to 10.5%). 14 Note that 2nd grade CST scores increase for all student lan- guage profciency groups between 2002 and 2008 in these two districts (Table 3) and statewide (California Department of Education 2013). 15 It is possible that EL instruction between kindergarten and 2nd grade is more benefcial in LAUSD or that there are unobserved diferences between students who were ever clas- sifed as ELs in the two districts. 16 In technical appendix Tables A1 (LAUSD) and A2 (SDUSD), the demographic characteristics of students who remain in the district (“stayers”) and leave the district (“leavers”) are displayed separately. We fnd that students in LAUSD reclassifed in elementary school who remain in the district are somewhat more likely to be low-income (as measured by free/reduced price meal eligibility) and have somewhat less educated parents, but the diferences are slight. In SDUSD, students reclassifed in elementary school who remain in the district have slightly better educated parents than those who leave the district. 17 Recall that our focus is on ELs reclassifed in elementary school. Some EL students who were not reclassifed by the end of 5th grade may have been reclassifed in middle or high school. 18 We examine math CST scores only through 7th grade because beginning in 8th grade, students take a variety of math CST tests within each grade. For example, in 9th grade, students typi - cally take the CST for Algebra I, Geometry, or (least commonly) Algebra II. Combining the mean performance levels across examinations measuring knowledge of diferent math content is not appropriate. 19 A major diference between the two districts is the relatively stable math CST scores across grades in SDUSD versus the steady decline in math CST scores from the peak in 4th grade in LAUSD (across all language profciency groups) to 7th grade. For example, the mean performance level among LAUSD reclas- sifed students in 4th grade is 4.2, but by 7th grade it has fallen to 3.6. Our requirement that students must remain in their districts to be included in our sample does not afect our results much— we conduct the same analyses for students who eventually leave 23 Pathways to Fluency www.ppic.org their districts (technical appendix Figures C1 and C2) and fnd that the patterns by profciency group are similar, but perfor- mance levels are somewhat lower in comparison to students who remain in their districts. 20 SDUSD’s graduation requirements appear to be more chal- lenging than LAUSD’s. SDUSD students must have a minimum GPA of 2.0 and must successfully complete three years of science and three years of math. (See Table 1 in Betts, Zau, and Bachofer 2013.) LAUSD requires a minimum GPA of 1.0, two years of science, and two years of math (LAUSD 2010). Because we do not have dropout data for LAUSD, we exclude dropouts from our graduation calculations here and throughout the report. When we include dropouts in the SDUSD data, graduation rates decrease only slightly (see technical appendix Table C1). 21 Technical appendix Table C4 shows results for IFEP students as well. In San Diego, IFEP students pass the CAHSEE at higher rates than other students (96% by the end of 12th grade), but in Los Angeles, RFEP students pass CAHSEE by the end of 12th grade at higher rates than IFEP students (98% versus 94%). 22 We refer to this criterion as the “binding constraint” because it is the criterion that the student is furthest from meeting. Within each grade level, we count the number of times a par- ticular reclassifcation criterion is the furthest from being met for all the students in that grade. 23 Our measure of distance has been standardized for each reclassifcation criterion by transforming students’ scores into z-scores. For example, we take a student’s overall score on the CELDT and subtract it from the score required (Early Advanced) and divide it by the standard deviation of the CELDT score for EL students in that grade in that district. We iden- tify the binding constraint as the requirement for which the student’s score is the greatest number of standard deviations below the required level. It is important to note that the students included in our analyses for Figures 3 and 4 are those in grade 2 in 2002 (or 2001 as well in SDUSD) who remain EL students in the given grade a number of years later. Tus, the fgures do not include students who have been reclassifed. Te reclassi- fcation of students out of the group over time may, in part, be responsible for the gradual changes across grades in the binding constraint to reclassifcation for students who are still ELs in the given grade. 24 In both districts, it appears that reclassifcation decisions are made according to the policies in place. In LAUSD, about 2 per- cent of all students should have been reclassifed when they were not, and the same was true for about 4 percent of students in SDUSD. Tere were almost no observations of reclassifcations that did not meet the reclassifcation policy requirements—we saw 13 observations in total (or 0%) in LAUSD, and in SDUSD, about 2 percent of reclassifcations should not have been made. 25 We have no marks for 2nd graders and marks for only some of our 3rd graders, so we exclude them from this analysis. Imple- menting the grade level cut-of requirement for middle and high school is difcult because of the way course names are recorded at higher grade levels. 26 For instance, the CELDT reading subtest requirement was the binding constraint for 43 percent of 10th graders. 27 Recall that SDUSD students may score less than Early Advanced on one CELDT subtest, but for ease of exposition, we do not allow for that possibility here. 28 Because both districts go beyond the state’s recommended minimum reclassifcation requirement, students in one district who would have failed to meet the other district’s requirement are not the perfect comparison group—they are still in some sense above average. If anything, this probably biases our analy- sis toward fnding that higher reclassifcation requirements did not lead to big changes in outcomes. 29 A category of similar size. 30 Since marks are subjective and since LAUSD used a standards- based report card whereas SDUSD did not, we use the relation- ships we observe between test scores and marks in LAUSD to approximate marks for SDUSD students had they been enrolled in LAUSD. We use CELDT and CST scores to predict marks for EL and RFEP students in LAUSD. Coefcients from those models are used in estimating the marks of SDUSD students who were reclassifed in elementary school. Tese regressions are found in technical appendix F. 31 We reanalyzed the SDUSD data using the more recent LAUSD requirement that uses marks in ELA but not math. Te results were virtually the same, with only three more students predicted to have met the LAUSD mark requirements. 32 See technical appendix Figure E3. 33 Tere were diferences between students who met the LAUSD coursemark criteria and those who did not. Tose who met the criteria tended to have higher CELDT OPL scores and higher CST scores. Tey also difered demographically, with those who met the LAUSD course mark criteria less likely to be Hispanic and with parents with higher education levels. Pathways to Fluency 24 www.ppic.org 24 References Assembly Bill 484, Pupil Assessments: Measures of Academic Performance and Progress (MAPP). Available at http:// leginfo.legislature.ca.gov/faces/billNavClient.xhtml?bill_id =20132014 0A B 4 8 4 . Betts, J. R., A. C. Zau, and K. V. Bachofer. 2013. College Readi- ness as a Graduation Requirement: An Assessment of San Diego’s Challenges , San Francisco: Public Policy Institute of California. California Department of Education. “Facts about English Learners in California—CalEdFacts.” Available at www.cde. c a .gov/d s/s d /cb/c efel f ac t s . a sp. California Department of Education. 2012. “California English Language Development Test (CELDT), 2012–2013,” CELDT Information Guide. Available at w w w.cde.ca.gov/ta/tg/el /documents/celdtinfoguide1213.pdf/. California Department of Education. 2013. “State Schools Chief Tom Torlakson Releases 2013 STAR Results,” News release, August 8. 34 Many school districts have decided to use students’ grades as an element in the reclassifcation decision (Hill, Weston, and Hayes 2014). 35 Assembly Bill 484 requires that districts administer the computer-based feld test of the new Common Core Assess- ments in 2013–14, but the results of these tests will not be given to schools. A survey of California school districts estimated that only 33 percent to 40 percent of school districts are able to ofer the computer-based tests in 2013–14. Te new tests will be given in all districts in 2014–15 (Fensterwald 2013). 36 Full results are shown in technical appendix Tables G1–G10. California Department of Education DataQuest. Available at http://data1.cde.ca.gov/dataquest/. California Education Code Section 313. Available at w w w.leginfo.ca.gov/cgi-bin/displaycode?section=edc&group =00001-01000&file=313-313.5. EdSource. 2008. English Learners in California: What the Num- bers Say , Mountain View, CA. Fensterwald, J. 2013. “What’s Next for Standardized Testing in California?” EdSource (blog), October 22. Flores, E., G. Painter, and H. Pachon. 2009. ¿Qué Pasa? Are Eng- lish Language Learning Students Remaining in English Learning Classes Too Long? Los Angeles: Tomás Rivera Policy Institute Full Report. Gándara, P., and R. Rumberger. 2006. Resource Needs for Cali- fornia’s English Learners . Linguistic Minority Research Institute, University of California. Hill, L. E. 2012. California’s English Learner Students , San Fran- cisco: Public Policy Institute of California. Hill, L., M. Weston, and J. Hayes. 2014. Reclassifcation of Eng- lish Learner Students in California . San Francisco: Public Policy Institute of California. Los Angeles Unifed School District. 2010. 2010–2011 Gradu- ation Requirements and Minimum College Admission “A–G” Requirements. Available at http://home.lausd.net/ourpages /auto/2011/12/21/53468052/2010-2011GRADUATION REQUIREMENTSCHART.pdf. Los Angeles Unifed School District. 2011. Reclassifcation of English Learners, Grades 2–12, Bulletin BUL-5619.0, October 17. Los Angeles Unifed School District. 2014. “Comprehensive Assessment Program.” Available at http://notebook.lausd.net /portal/page?_pageid=33,167651&_dad=ptl. Ofce of Curriculum, Instruction, and School Support, LAUSD Ofce of the Deputy Superintendent of Instruction. 2011 . LAUSD Literacy Periodic Assessments Grades K–5 , October. Robinson, J. P. 2011. “Evaluating Criteria for EL Reclassifcation: A Causal-Efects Approach Using a Binding-Score Regression Discontinuity Design with Instrumental Variables,” Education Evaluation and Policy Analysis 33 (3): 267–92. 25 Pathways to Fluency www.ppic.org San Diego Unifed School District. 2009. Master Plan for English Learners . Saunders, W. M., and D. J. Marcelletti. 2013. “Te Gap Tat Can’t Go Away: Te Catch-22 of Reclassifcation in Monitoring the Progress of English Learners,” Educational Evaluation and Policy Analysis 35 (139). Senate Bill 1108, An act to add Section 313.5 to the Education Code, relating to English Learners. Available at www.leginfo .c a . gov/pu b/11-12/ bi l l /s en /s b _1101-115 0/s b _110 8 _ bi l l _ 2 012 0 419 _amended_sen_v97.pdf. Silver, D., M. Saunders, and E. Zarate. 2008. “What Factors Predict High School Graduation in the Los Angeles Unifed School District,” Santa Barbara, CA: California Dropout Research Project Report. Acknowledgments Te authors wish to thank Hans Johnson, Margaret Weston, Karina Jaquet, and Jill Cannon for their help in developing the ideas in this report. We thank Kathy Hayes, Cynthia Lim, Josh Klarins, Valerie Brewington, and Marciela Sanchez Robles at LAUSD for their insight into the data and policies and their feedback about the research. We thank the SDUSD Ofce of Language Acquisi- tion, especially Mary Waldron, Debra Dougherty, and Shelby Madden, for providing extensive background information about English Learner programs and practices in San Diego and for their careful review of early drafs of this report. In addition, we thank Ron Rode, Peter Bell, and Dina Policar from the SDUSD Ofce of Accountability for their assistance and advice. We also wish to thank Caroline Danielson, Rachel Ehlers, Kathy Hayes, Laurel Beck, Lynette Ubois, and Kenji Hakuta for comments on early drafs and Mary Severance, Patricia Bedrosian, Jenny Miyasaki, and Kate Reber for editorial support. Any errors are our own. Pathways to Fluency 26 www.ppic.org 26 About the Authors Laura E. Hill is a research fellow at PPIC, where her research interests include immigrant well- being, English Learners in K–12 schools, race and ethnicity, and youth. She has been a research associate at the SPHERE Institute and a National Institute of Aging postdoctoral fellow. She holds a Ph.D. in demography from the University of California, Berkeley. Julian R. Betts is a Bren fellow and an adjunct fellow at PPIC. He is also professor and former chair of economics at the University of California, San Diego, where he is executive director of the San Diego Education Research Alliance. He is also a research associate at the National Bureau of Eco - nomic Research and UC San Diego campus director of the University of California Educational Eval - uation Center. He holds a Ph.D. in economics from Queen’s University, Kingston, Ontario, Canada. Belen Chavez is an analyst at Harris Economics Group and a former research associate at PPIC. She holds an M.A. in economics from Duke University and a B.A. in business economics from the University of California, Irvine. Andrew C. Zau is a senior statistician for the San Diego Education Research Alliance in the Department of Economics at the University of California, San Diego. His research focuses on the efects of policy and reform on student achievement in K–12 education. Previously, he was a research associate at PPIC. He holds an M.P.H. in epidemiology from San Diego State University and a B.S. in bioengineering from the University of California, San Diego. Karen Volz Bachofer is the director of the San Diego Education Research Alliance in the Depart- ment of Economics at the University of California, San Diego. Previously, she was the executive director of the San Diego Unifed School District’s Research and Evaluation Division, where her responsibilities included oversight of national, state, and district assessment and accountability processes and reporting. She holds a Ph.D. in education from the Claremont Graduate School and San Diego State University. 27 Pathways to Fluency www.ppic.org Board of Directors DONNA LUCAS , CHAIRChief Executive Ofcer Lucas Public Afairs MARK BALDASSAREPresident and CEO Public Policy Institute of California RUBEN BARRALESPresident and CEO GROW Elect MAR Í A BLANCOVice President, Civic Engagement California Community Foundation BRIGITTE BRENAttorney WA LT E R B. HEWLETTMember, Board of Directors The William and Flora Hewlett Foundation PHIL ISENBERGVice Chair Delta Stewardship Council M AS MASUMOTOAuthor and Farmer STEVEN A. MERKSAMERSenior Partner Nielsen, Merksamer, Parrinello, Gross & Leoni, LLP KIM POLESEChairman ClearStreet, Inc. THOMAS C. SUT TONRetired Chairman and CEO Pacifc Life Insurance Company PPIC is a public charity. It does not take or support positions on any ballot measures or on any local, state, or federal legislation, nor does it endorse, support, or oppose any political parties or candidates for public ofce. PPIC was established in 1994 with an endowment from William R. Hewlett. Copyright © 2014 Public Policy Institute of California. All rights reserved. San Francisco, CA Mark Baldassare is President and Chief Executive Ofcer of PPIC. Donna Lucas is Chair of the Board of Directors. Short sections of text, not to exceed three paragraphs, may be quoted without written permission provided that full attribution is given to the source. Research publications refect the views of the authors and do not necessarily refect the views of the staf, ofcers, or Board of Directors of the Public Policy Institute of California. Library of Congress Cataloging-in-Publication Data are available for this publication. I S B N 978 -1-5 8213 -157-3 PUBLIC POLIC Y INSTITUTE OF CALIFORNIA 500 Washington Street, Suite 600 San Francisco, California 94111 Telephone 415.291.4400 Fa x 415 . 2 91. 4 4 01 PPIC S ACRAMENTO CENTER Senator Ofce Building 1121 L Street, Suite 801 Sacramento, California 95814 Telephone 916.440.1120 Fax 916.440.1121 Additional resources related to K–12 education are available at www.ppic.org. The Public Policy Institute of California is dedicated to informing and improving public policy in California through independent, objective, nonpartisan research." } ["___content":protected]=> string(102) "

R 514LHR

" ["_permalink":protected]=> string(144) "https://www.ppic.org/publication/pathways-to-fluency-examining-the-link-between-language-reclassification-policies-and-student-success/r_514lhr/" ["_next":protected]=> array(0) { } ["_prev":protected]=> array(0) { } ["_css_class":protected]=> NULL ["id"]=> int(8903) ["ID"]=> int(8903) ["post_author"]=> string(1) "1" ["post_content"]=> string(0) "" ["post_date"]=> string(19) "2017-05-20 02:41:58" ["post_excerpt"]=> string(0) "" ["post_parent"]=> int(4343) ["post_status"]=> string(7) "inherit" ["post_title"]=> string(8) "R 514LHR" ["post_type"]=> string(10) "attachment" ["slug"]=> string(8) "r_514lhr" ["__type":protected]=> NULL ["_wp_attached_file"]=> string(12) "R_514LHR.pdf" ["wpmf_size"]=> string(7) "3048272" ["wpmf_filetype"]=> string(3) "pdf" ["wpmf_order"]=> string(1) "0" ["searchwp_content"]=> string(91681) "www.ppic.org Pathways to Fluency Examining the Link between Language Reclassifcation Policies and Student Success M ay 2014 Laura E. Hill • Julian R. Betts • Belen Chavez • Andrew C. Zau • Karen Volz Bachofer with research support from Joseph M. Hayes Supported with funding from the Donald Bren Foundation and the Evelyn and Walter Haas, Jr. Fund SUMMARY N early 25 percent of the students attending California’s K–12 public schools are English Learners (ELs). Their EL designation is intended to last only as long as they need supplemen- tal language support to succeed in school. Some students attain English fuency quickly, but others remain ELs for six years or longer. Because outcomes for students reclassifed as English profcient are much better than for students who remain ELs, policymakers are seeking answers to questions about how quickly EL students should be reclassifed, whether reclassifcation crite - ria should be standardized, and the links between reclassifcation and academic success. These issues are especially urgent now that California is implementing a major overhaul of K–12 standards, testing, and funding—as well as many elements of EL instruction. Because the new Local Control Funding Formula (LCFF) provides additional funding to districts with high numbers of ELs, there is more interest than ever in making sure that districts have the right incentives to help these students succeed. In this report, we examine reclassifcation policies and the academic performance of ELs and former ELs in the two largest school districts in California, Los Angeles Unifed and San Diego Unifed, which together serve approximately 15 percent of the state’s EL students. Using longitudinal student data over ten years, we can follow a cohort of 2nd grade students through their 12th grade year. We fnd that students reclassifed in elementary school (grades 2–5) have very strong aca - demic outcomes throughout middle and high school. These students perform as well as or ISTOCK Pathways to Fluency 2 www.ppic.org better than native English speakers on state standardized tests and are as likely or more likely to make on-time grade progress. There is no evidence that the removal of language supports for ELs who are reclassifed hurts their academic progress relative to that of native English speakers. Reclassifcation criteria in both San Diego and Los Angeles are more stringent than minimum guidelines recommended by the State Board of Education (SBE). These more rigorous criteria are associated with somewhat improved outcomes for students but also lower reclassifcation rates. Despite diferences in reclassifcation criteria between the two districts, the factors that predict successful outcomes for EL students reclassifed in elementary school in Los Angeles and San Diego are remarkably similar. The two standardized tests currently used to reclassify students—the California Standards Test (CST) and the California English Language Develop - ment Test (CELDT)—are individually strong predictors of future academic outcomes such as performance on middle school standardized tests and the high school exit exam. Elementary school marks are less useful as predictors. Our fndings lead us to recommend the following: • Use the CELDT as the sole assessment for reclassification decisions until the CST replace - ment is available. More generally, consider allowing districts to reclassify students on the basis of just one test. • In designing new English language development (ELD) tests and reclassification stan - dards, consider the relative rigor of reading and writing requirements—our data suggest that the current CELDT writing requirement is relatively easy in comparison to the read - ing requirement. • Reconsider the use of reclassification criteria that are more rigorous than those suggested in the State Board of Education guidelines. We fnd evidence in the state’s two largest school districts that English learners would beneft as a group from being reclassifed slightly sooner, through an easing of reclassifcation standards. • Consider a uniform standard for reclassification across school districts. Evaluating districts’ successes with ELs is very difcult when classifcation and reclassifcation policies vary. Over the next few years, many elements of EL instruction, funding, and testing will be changing. Many policymakers have long been frustrated with the pace at which EL students are reclassifed as fully English profcient and are concerned that the additional funds directed toward ELs under the LCFF might increase district incentives to delay reclassifcation for stu - dents on the cusp of English fuency. This is an ideal time to draw lessons from the recent past to inform state and local reclassifcation policies in 2014 and beyond. For the full report and related resources, please visit our publication page: www.ppic.org/main/publication.asp?i=1089 3 Pathways to Fluency www.ppic.org Introduction During the 2012–13 school year, more than 1.3 million English Learners attended public schools in California, accounting for about one-quarter of the state’s K–12 stu- dent population. 1 Nearly two-thirds of these students were enrolled in elementary schools. California school districts are charged with the dual goals of ensuring that English Learners acquire full profciency in English as quickly as possible and that they meet the same rigorous grade-level academic standards that all students must meet. 2 Te SBE has issued guidelines about the minimum criteria used to reclassify EL students, but districts are allowed a great deal of latitude in establishing more rigorous criteria. Te state’s interest in reclassifcation has been height- ened by a number of recent changes to K–12 funding, standards, and assessments. Te new LCFF allocates many more dollars per EL student than districts received under the old funding formula. Tis increases the efect of reclas- sifcation decisions for school districts and the importance of making sure that the funding is used to serve ELs efec- tively. In January 2014, the SBE passed temporary regula- tions for how these dollars are to be spent by districts and established an accountability framework, called the Local Control and Accountability Plan (LCAP). 3 Te implementation of the Common Core State Stan- dards (CCSS), changes to English language development (ELD) standards, and development of new assessments are also under way. Until 2013, California used the CST to determine whether ELs met the academic skills portion of the reclassifcation criteria. However, as part of the state’s transition to a new statewide assessment system aligned with the CCSS, there will be very little testing during the 2013–14 academic year, and the individual student results that help determine readiness for reclassifcation will not be provided to districts. A new statewide testing system will phase in during the 2014–15 academic year, but until then, districts will need to decide how to reclassify EL students when their CST results become outdated. In addi- tion, the state’s ELD standards were revised in 2012, and a new English language profciency exam is expected to be fully implemented by 2016. In light of the state’s interest in reclassifcation deci- sions (as evident in Senate Bill 1108, which mandated a statewide analysis of school district reclassifcation poli- cies), the new CCSS and assessments, and the increased per-pupil funding for EL students, the timing is perfect The state’s interest in reclassifcation has been heightened by a number of recent changes to K–12 funding, standards, and assessments. Abbreviations CAHSEE California High School Exit Examination CCSS Common Core State Standards CDE California Department of Education CELDT California English Language Development Test CMA California Modifed Assessment CST California Standards Test EL English Learner ELA English language arts ELD English language development EO English-only or native English speaker FEP fuent English profcient FRPL free/reduced price lunch G PA grade point average IFEP initially fuent English profcient LAUSD Los Angeles Unifed School District LCAP Local Control and Accountabilty Plan LCFF Local Control Funding Formula LEA Local Educational Agency OPL overall profciency level RFEP reclassifed fuent English profcient SBE State Board of Education SDUSD San Diego Unifed School District SEI Structured English Immersion Pathways to Fluency 4 www.ppic.org for a retrospective examination of California’s current reclassifcation guidelines—and an analysis of the relative importance of each reclassifcation criterion in accurately predicting ELs’ readiness for English-only instruction. PPIC recently published Reclassifcation of English Learner Students in California , a report about the relation- ship between reclassifcation policies and student outcomes four years later. In this report, we examine the academic progress of ELs in the two largest school districts in California—the Los Angeles Unifed School District (LAUSD) and the San Diego Unifed School District (SDUSD). Together, Los Angeles and San Diego serve more than 200,000 ELs, about 15 percent of the ELs enrolled in the state. Because of a growing consensus that language acquisition during the elementary school years infuences longer-term academic outcomes, we focus on students enrolled in these districts during the elementary school years and follow their progress through grade 12. Background on San Diego and Los Angeles San Diego Unifed and Los Angeles Unifed are large and diverse California school districts. Although Spanish is the most common language spoken by ELs in both districts (94% in LAUSD and 76% in SDUSD), the distribution of other languages spoken among ELs varies by district. In LAUSD, the other common languages are Armenian (1.1%), Korean (1.0%), and Filipino (1.0%); in SDUSD, the other common languages are Vietnamese (5.6%), Filipino (4.3%), and Somali (2.6%) (Table 1). Students in Los Angeles are more likely to be low- income than students in San Diego (77% versus 61%). Los Angeles students are more likely to be Hispanic and less likely to be white or Asian than students in San Diego. It is important to take into account the varying demographic characteristics of the student population in each district when comparing academic outcomes for ELs and native English speakers. Goals of the Study Tis report has four parts. Te frst part describes how students come to be reclassifed in LAUSD and SDUSD, the data we use for each district, and how we defne our student cohorts. Te second part examines whether attainment of current reclassifcation criteria in elementary school results in better student performance in middle school and high school—and whether the performance of ELs falters in the years afer reclassifcation. Because we are looking at whether students are making the transition to English-only instructional set - tings at the appropriate time and tracking their long-term prospects, our cohort consists of English Learners enrolled in LAUSD and SDUSD during their elementary school years. 4 We focus on students reclassifed in grades 2–5 in both school districts. 5 Tere are important policy questions about Changes to K–12 funding, standards, and assessments have intensifed California’s interest in reclassifying English Learners. AP P HOTO/RIC H PED RONCELLI San Diego Unifed and Los Angeles Unifed are large and diverse California school districts. Although Spanish is the most common language spoken by ELs in both districts, the distribution of other languages spoken among ELs varies by district. 5 Pathways to Fluency www.ppic.org EL students who arrive in the secondary school years, but these are beyond the scope of this report. Using student-level data from both districts, we are able to follow students over time—while they are ELs, at the time of reclassifcation, and for many years beyond. Most previous research has relied on a cross-sectional approach, which shows large diferences in academic achievement between native English-speaking students and ELs but ofen overlooks the confounding factor that the most successful EL students are reclassifed in early grades and “drop out” of the analyses. 6 Specifcally, we explore the following questions: 1. How do reclassifed fuent English profcient (RFEP) students fare on outcome measures such as the CST, the California High School Exit Examination (CAHSEE), retention in grade, and on-time graduation? 2. Do these outcomes vary by primary language or grade level at reclassifcation? 7 Although this report has a wide scope with regard to reclassifcation of ELs, it cannot tackle many important questions. For example, it does not study EL latecomers arriving in higher grades, nor does it study instructional diferences experienced by EL and RFEP students within either district or between districts. Rather, it provides a portrait of the progress of students who were ever ELs, dependent on their language status, in the contexts of the two districts. Te third part of the report explores relationships among individual reclassifcation criteria and a range of outcomes to determine which criteria are the most challenging for ELs and which are strongly associated with short- and long-term academic outcomes following reclassifcation. Because LAUSD and SDUSD have slightly diferent reclassifcation criteria, we can examine the relationship between more rigorous criteria and both aca - demic outcomes and reclassifcation rates. We also explore Table 1. K–12 enrollment in California, LAUSD, and SDUSD, 2012–13 CaliforniaLAU SD SDUSD Percentage Percentage Percentage Total enrollment 6,226,989 655,49413 0 , 270 Race/ethnicity African American 394,6956.361,78 6 9.413 , 261 10. 2 Hispanic 3 , 2 82 ,10 552.7482,534 73.660,616 46.5 Asian/Pacifc Islander 7 2 5, 81911 . 742, 26 4 6.418,799 14 . 4 White 1, 589, 39325.560,266 9. 230,271 23.2 Other 234,9773.88,644 1. 37, 3 2 3 5.6 Free/reduced price meal eligibility a3,472,481 5 7. 5489,777 76.676, 8 4 6 60.7 Fluent English profcient b1, 339, 56 6 21. 5239,753 36.62 7, 0 3 2 20.8 English Learners 1,346,33321. 617 0 , 7 9 7 26 .133, 851 26.0 Top four languages spoken among ELs Spanish, Vietnamese, Filipino, Cantonese 89.6 Spanish, Armenian, Filipino, Korean 96.7 Spanish, Vietnamese, Filipino, Somali 88.0 SOURCE: California Department of Education DataQuest for the 2012–13 school year. NOTES: SDUSD and LAUSD have similar shares of EL students (26%), but LAUSD has a much higher percentage of students who are fuent English profcient (FEP)—students who were either former ELs who have been reclassifed as fuent English profcient (RFEP) or were designated at school entry as fuent in English even though they were speakers of another language (initially fuent English profcient or IFEP). a Free/reduced price meal numbers are from the 2011–12 school year. b FEP numbers include both RFEP students and IFEP students. IFEP students speak a language other than English at home but were determined by their initial CELDT scores to be fuent in English. Pathways to Fluency 6 www.ppic.org the possibility that difcult criteria could unnecessarily delay reclassifcation for students.Finally, the fourth part of the report looks at com- binations of several criteria (such as English profciency level, performance on basic academic skills assessments, and report card grades) to determine which reclassifca- tion criteria are the best predictors of student success. Tis analysis may help these and other districts decide on the best criteria to use to reclassify EL students. How Do EL Students Get Reclassified? Te California Education Code requires that school districts develop policies and procedures to guide the reclassifcation of English Learners. 8 District-level reclassi - fcation standards must be based on four criteria approved by the SBE: performance in basic skills, an assessment of English profciency, teacher evaluation of academic per- formance, and the opinion of a parent or guardian. To be considered for reclassifcation from EL to fuent English profcient (FEP), students should—at a minimum—meet all four criteria. To meet the minimum basic skills recommendations for reclassifcation, students must score at the Basic level or higher on the CST in English language arts (ELA). Students must also demonstrate English profciency by achieving an overall profciency level (OPL) of Early Advanced or higher on the CELDT, and their scores on each subtest—listening, speaking, reading, and writing—must be rated as Inter - mediate or higher. In addition, teachers must certify that students meet district academic performance indicators and are ready to succeed in an English-only instructional program. Te district must advise parents and guardians of their right to participate in the reclassifcation process and encourage them to attend a face-to-face meeting. Districts have great latitude in setting their own reclas- sifcation policies, as long as they take into consideration the guidelines issued by the SBE. Reclassifcation criteria in LAUSD and SDUSD are somewhat more rigorous than the SBE guidelines. In San Diego, basic skills and English profciency requirements are higher; in Los Angeles, the teacher evaluation component specifes minimum report card grades as a condition of reclassifcation. Table 2 details current statewide, LAUSD, and SDUSD reclassifca- tion criteria. It is important to note that for the cohort we study, LAUSD also required marks in math courses of 3 or higher (on a 4-point scale) through the 2005–06 school year. SDUSD’s reclassifcation criteria have not changed since 2002. SDUSD’s reclassifcation rates are largely unchanged The California English Language Development Test and the California Standards Tests The CELDT is a state-mandated assessment of listening, speaking, reading, and writing in English that is administered in kindergarten through grade 12. It is used to identify stu- dents with limited English profciency to determine their levels of profciency and to assess progress in learning English. The CELDT must be administered within 30 days of enroll - ment to all students whose Home Language Survey indicates that a language other than English is spoken at home and annually to all continuing ELs who have not yet been reclas- sifed as FEP. CELDT results provide performance levels— Beginning, Early Intermediate, Intermediate, Early Advanced, and Advanced—for each of the subtests and an OPL. Until 2006, the listening and speaking subtests were combined into one. Starting in 2006–07, higher scores on the CELDT were required for each level of OPL. All students in grades 2 through 11—including ELs and most students receiving special education services—take the CSTs, state-mandated criterion-referenced tests that assess students’ mastery of the California content standards in Eng - lish language arts, mathematics, science, and history-social science. Results are reported as performance levels—Far Below Basic (1), Below Basic (2), Basic (3), Profcient (4), and Advanced (5)—and are used to identify individual students’ learning needs and assess school quality in state and federal accountability systems. Students with special needs who meet eligibility requirements may take the California Modi - fed Assessment (CMA) or California Alternate Performance Assessment (CAPA) rather than the CSTs. Districts that reclas - sify students before the 2nd grade CST scores are available use other assessments to make their decisions. 7 Pathways to Fluency www.ppic.org from 2005–06 to 2012–13 (10.4 to 10.5), whereas LAUSD’s rates increased from 9.5 to 13.7. 9 LAUSD’s increasing reclassifcation rates are probably associated with elimi - nating the use of mathematics grades as a reclassifcation criterion afer 2005–06. Student Data For this study, we use longitudinal student-level data from Los Angeles and San Diego Unifed School Districts for 2002 through 2012. (In this report, when we refer to In San Diego, basic skills and English profciency requirements are higher; in Los Angeles, the teacher evaluation component specifes minimum report card grades as a condition of reclassifcation. Both districts include reclassifcation criteria that go beyond the state minimum. a single year, such as 2002, we mean the 2001–02 school year.) Starting in 2002, we follow 2nd grade students through what would be their 12th grade year if they made on-time progress. 10 Because students transition from EL to RFEP status at various times, comparisons of EL and RFEP students can be complicated. To make the compari- sons as straightforward as possible, we focus our research on students who remain ELs through the end of 5th grade and students who are reclassifed as FEP by the end of 5th grade—and we focus only on students who are observed Table 2. Elementary grade reclassification criteria (2012–13) in both LAUSD and SDUSD are more rigorous than the SBE guidelines SBE guidelines L AU SD Structured English Immersion or M ainstream English Program aSD USD Performance in basic skills ELA CST score of 300 (Basic) or above 2nd grade: score on Literacy Periodic Assessment #2 or #3 of Basic or above b 3rd to 5th/6th grade: score of Basic or above on ELA CST EL A CST score of 333 (mid-Basic) or above Assessment of English profciency CELDT OPL of Early Advanced or higher All CELDT subtests at Intermediate or higher CELDT OPL of Early Advanced or higher All CELDT subtests at Intermediate or higherCELDT OPL of Early Advanced or higher At least three CELDT subtests at Early Advanced or higher; fourth subtest at Intermediate or higher Teacher evaluation of student academic performance Teacher certifcation that the student meets the district’s academic performance indicators Minimum marks of 3 on ELA courses of listening, speaking, reading, and writing Teacher certifcation that the student can be successful in core subject areas in a regular program designed for native and fuent speakers of English Parent or guardian opinion and consultation District provides notice to parents/ guardians of their right to participate in reclassifcation process and encourages them to attend a face-to-face meeting Parent consulted regarding student’s eligibility to reclassify, and letter must be printed and provided immediately; letter requires a parent signatureDistrict notifes parent/guardian of reclassifcation decision and provides opportunity to consult with staf regarding programs to further increase their student’s academic achievement SOURCES: California Department of Education (2012, p. 18). Los Angeles Unifed School District (2011). San Diego Unifed School District (2009). NOTE: The more rigorous reclassifcation criteria are shown in boldface. aBasic bilingual and dual-language programs have diferent teacher evaluation reclassifcation criteria.bPeriodic formative assessments to measure key ELA standards three times per year in grades 2–5 (Ofce of Curriculum, Instruction, and School Support, LAUSD Ofce of the Deputy Superintendent of Instruction, 2011). Pathways to Fluency 8 www.ppic.org in the data from 2002 to 2005. We compare both groups of students to native English speakers who must also be observed in their district data from 2002 to 2005. As a way of making sure that our cohort is not some- how anomalous, we looked at a more recent cohort of elementary school students who entered kindergarten in 2006–07 and remained in the same district through the 2012 school year. Although we cannot follow these stu- dents through to graduation, we can compare their ele- mentary grade outcomes to the earlier cohort to determine whether the longer-term outcomes of the earlier cohort are likely to be relevant for the present day in each district. Native English speakers are a smaller percentage of the main cohort in LAUSD than in SDUSD (top panel, Table 3). Te group of students who were ever ELs—the combination of current ELs and ELs reclassifed by the end of grade 5—is a much larger proportion of students in LAUSD than in SDUSD. 11 Within the ever-EL group, stu- dents reclassifed in elementary school constitute a greater proportion of students in SDUSD (45%) than in LAUSD (19%). Tis suggests that more rigorous reclassifcation criteria were used in LAUSD at this time. We fnd that the 2nd grade CELDT overall scores of students who were ever classifed as ELs are fairly simi- lar across the two districts (2.70 in LAUSD versus 2.51 in SDUSD). 12 However, when we separate the CELDT scores of EL students who remained ELs through 5th grade from those who were reclassifed by the end of 5th grade, a gap emerges—the 2nd grade scores for both groups are much higher in Los Angeles (3.45 and 2.42) than in San Diego (3.08 and 2.07). Tis same pattern is observed in the 2nd grade CST scores (reported as the share of students scoring Basic or above). Te overall similarity of scores for all students who were ever ELs combined with the diferences across two groups suggest that LAUSD’s reclassifcation criteria were more rigorous during the period we study. Tis may explain the higher average performance of EL students Table 3. Differences in test scores among student cohorts, LAUSD and SDUSD LAU SD SDUSD 2nd grade tests 2nd grade tests CST % Basic or above CST % Basic or above No. CELDT overall ELA M ath No.CELDT overall ELA M ath 2nd grade in 2002 RFEP5 4,3843.45 91.088.4 1,95 43.08 75. 2 71. 6 EL5 18 , 3672.423 7. 9 4 7.1 2,356 2.07 23.636.3 Ever EL (EL5 + RFEP5) 22,7512.7048.7 55.44, 310 2. 5148.0 54.0 Native English speakers 7, 0 74 NA69.1 66.9 4,926 NA8 3 .1 78.8 Kindergarten in 2007 RFEP5 10,9 9 63.7296.8 95.34,573 3.2695.394.9 EL5 11, 2 5 52.5649.659.04 ,121 2.0846.9 61. 8 Ever EL (EL5 + RFEP5) 22, 2513.0874 . 478.08,694 2.7072.4 79. 2 Native English speakers 11, 6 7 8 NA85.3 84.99,569 NA86.3 88.5 SOURCE: Authors’ tabulations from LAUSD and SDUSD student-level data, 2001–2012. NOTES: CELDT overall is the overall performance level on the CELDT and can range from 1 to 5. A score of 4 or higher on the frst CELDT taken results in a designation of IFEP (rather than EL). A score of 4 is required to be reclassifed from EL to RFEP. For CST results, we report the share of students scoring Basic or above. Because the CELDT was not administered to all students until 2002, the SDUSD CELDT overall included only 2nd grade students in 2002, not 2nd grade students in 2001. RFEP5 students are those reclassifed by the end of 5th grade. EL5 students are EL students who have not been reclassifed by the end of 5th grade. Results were signifcantly diferent across the language profciency groups. 9 Pathways to Fluency www.ppic.org who were and were not reclassifed by the end of 5th grade in LAUSD: it is plausible that because only the highest- performing ELs were reclassifed, many high-achieving students remained ELs, resulting in a higher average performance for both groups. Note also that native Eng- lish speakers have much higher 2nd grade CST scores in SDUSD than in LAUSD in the main cohort we study (those in 2nd grade in 2002). For the sake of comparison, we include a more recent cohort of students in Table 3. As shown in the bottom panel, almost half of the LAUSD students in the later cohort who were ever classifed as EL were reclassifed by the end of 5th grade. It is evident that the rate at which LAUSD’s EL students are reclassifed in elementary school has increased. 13 Tis is probably because LAUSD dropped the requirement that EL students earn marks of 3 or better in math courses to be reclassifed. SDUSD’s rate of reclassifcation in elementary school has also increased, but to a lesser extent. Notably, the diference in 2nd grade CST ELA and math scores between students in LAUSD and SDUSD in the older cohort disappears in the younger cohort, which faced lower reclassifcation standards. Tis supports the hypothesis that the higher performance of LAUSD students in the older cohort was the result of some high- achieving LAUSD students not being reclassifed, thus raising average achievement in both groups. 14 Te gap in the 2nd grade CELDT scores of students who were ever ELs in SDUSD and LAUSD increased for the 2007 kindergarten cohort, even though CST results are similar among ELs in the two districts. When we divide the EL group into those who remain ELs in 5th grade and those who are reclassifed by the end of 5th grade for the kindergarten 2007 cohort, the CELDT scores for this cohort are much higher in LAUSD. 15 Since we follow students over time, it is important to account for the possibility of students leaving LAUSD and SDUSD at diferent rates, thereby afecting our observation of student outcomes. We explored this possibility and con- cluded that attrition should not have a major efect on our results. 16 We include some results for students who later exit the district for comparison in technical appendix C. Long-Term Outcomes for Students Reclassified in Elementary School In this section, we examine academic outcomes (annual stan - dardized test scores, on-time progression including on-time high school graduation, and high school exit exam scores) to compare the longer-term performance of reclassifed students with that of native English speakers and English Learners. We focus particularly on former English Learners who were reclassifed in grades 2, 3, 4, or 5 and remained in the district for ten years (until 12th grade, if they made on-time prog - ress). We compare outcomes for these reclassifed students to those of EL students who were not reclassifed by the end of 5th grade and also to those of native English speakers. We are interested not only in how students perform at a given grade level but also in their performance over time. Tis is an important issue for reclassifed students because educators need to know if ELs reclassifed in elementary school con - tinue to be strong academic performers in middle and high school or if they might need supplemental services. In both Los Angeles and San Diego, students reclassi- fed in elementary school are among the best academic performers. Teir elementary school outcomes are well above those of English Learner students and in many cases they are on par with those of native English speakers. Te diferences between students who were and were not reclas - sifed in elementary school persist into middle and high school. 17 We see no evidence that the students who were reclassifed in elementary school falter relative to other students later in their educational trajectories. We examine academic outcomes to compare the longer-term performance of reclassifed students with that of native English speakers and English Learners. Pathways to Fluency 10 www.ppic.org Performance on California Standards Tests Students reclassifed in elementary school in both districts are top performers on the English Language Arts portion of the CST (Figure 1). Te mean ELA CST performance levels for reclassifed students in both districts and for native English speakers in SDUSD are all high mid-Basic (above 3.7) in 11th grade. In LAUSD, native English speak- ers in all grades do worse than students reclassifed in elementary school. Native English speakers in LAUSD also underperform their counterparts in SDUSD. Tis gap could be partly explained by the diference in the share of parents without high school diplomas: 10 percent of parents of native English speakers in LAUSD versus 2 percent of parents of native speakers in SDUSD (see techni- cal appendix Tables A1 and A2). Performance levels for EL students who were not reclassifed in elementary school are below Basic in both districts (2.7 in SDUSD and 2.9 in LAUSD in 11th grade, for instance). Math CST scores reveal similar patterns. In LAUSD, students reclassifed in elementary school have much bet- ter math scores than native English speakers, probably because of the math grades required for reclassifcation in LAUSD at the time. In San Diego these students’ math performance is virtually identical to that of native English speakers. Te scores of SDUSD and LAUSD students who remained ELs throughout elementary school are within 0.3 mean performance levels of each other from 2nd to 7th grade. 18 In San Diego, students reclassifed in elementary school and native English speakers have comparable mean performance levels on the math CST by grade level. 19 We see no evidence in either the ELA or math CST scores that the reclassifed students in our study falter at higher grade Grade LAUSD SDUSD 7Grade 2 3 4 5 6 7 8 9 10 11 Grade 2 3 4 5 6 7 8 9 10 11 23 456Grade23 45 67 Mean performance level CST ELA 1.5 2.0 2.5 3.0 3.5 4.0 4.5 CST math Mean performance level 1.5 2.0 2.5 3.0 3.5 4.0 4.5 1.5 2.0 2.5 3.0 3.5 4.0 4.5 1.5 2.0 2.5 3.0 3.5 4.0 4.5 RFEP5 Native English speaker EL5 Figure 1. Students reclassifed in elementary school perform well on the CST in later grades SOURCE: Authors’ calculations from SDUSD and LAUSD student-level data, 2001–2012. NOTES: Students in 8th grade can take many diferent math CSTs, so we present scores only through 7th grade. Students must be making on-time grade progress for their scores to be included here. We examine on-time grade progression separately in Table 4. CST scores are examined for students remaining in the district through 2012. RFEP5 students are those reclassifed by the end of 5th grade. EL5 students are EL students who have not been reclassifed by the end of 5th grade. Mean diferences for language profciency groups are statistically signifcant as are mean grade-level scores within each district in LAUSD and in SDUSD, with one exception: RFEP math scores in grades 6 and 7 are not signifcantly diferent from native English-speaking students’ scores. 11 Pathways to Fluency www.ppic.org levels relative to native English speaking students. Declines in these students’ mean performance levels in the high school grades are similar to declines among native English speakers. For example, between 9th and 11th grade in LAUSD, ELA CST scores fell by 0.09 CST mean performance levels for native English speakers but only by 0.03 CST mean perfor - mance levels for students reclassifed in elementary school. In SDUSD, the decline between 9th and 11th grade on ELA CST mean performance levels was 0.23 for native English speakers and 0.17 for these reclassifed students. Tis is an important fnding: removal of language supports for ELs who are reclas - sifed does not seem to have hurt their academic progress relative to that of native English speakers. On-Time Progression toward Graduation In both districts, English Learner students not reclassifed by the end of 5th grade are the least likely to make on-time (or better) grade progress relative to other student language pro - fciency groups (Table 4). In Los Angeles, students reclassifed by the end of 5th grade are more likely than native English speakers to progress on time, whereas in San Diego, out - comes are similar for the two groups. In both districts, there is a substantial drop-of in on-time grade progression from 9th to 10th grade. Tis is a greater issue in LAUSD, where only 73 percent of ELs not reclassifed in elementary school are on time by grade 10 (compared to 82% of native English speakers). In SDUSD, 83 percent of ELs not reclassifed in elementary school and approximately 95 percent of students who were reclassifed in elementary school and native Eng- lish speakers are making on-time grade progress by 10th grade. In high school, grade progression is determined by unit accumulation: failing one or more classes can mean that a student is recorded as being in 9th grade for two years in a row. A similar pattern in 9th to 10th grade progression is observed statewide (Hill, Weston, and Hayes 2014). In both districts, English Learner students not reclassifed by the end of 5th grade are the least likely to make on-time (or better) grade progress relative to other student language profciency groups. Table 4. Students reclassified in elementary school maintain high rates of on-time grade progression through high school Percentage on-time advancement to: Percentage on-time graduation a 9th grade10 th grade 11th grade12 t h g r a d e B y 2 012 L AU SD RFEP5 9890 88 85 82 EL5 897368 64 58 Native English speaker 948277 74 66 SD USD RFEP5 98959392 78 EL5 9283 8183 55 Native English speaker 98959392 78 SOURCE: Authors’ calculations from SDUSD and LAUSD student-level data, 2001–2012.aSDUSD’s graduation requirements appear to be more challenging than LAUSD’s. SDUSD students must have a minimum grade point average (GPA) of 2.0 and must successfully complete three years of science and three years of math. (See Table 1 in Betts, Zau, and Bachofer 2013.) LAUSD requires a minimum GPA of 1.0, two years of science, and two years of math (LAUSD 2010). Because we do not have dropout data for LAUSD, we exclude dropouts from our graduation calculations here and throughout the report. When we include dropouts in the SDUSD data, graduation rates decrease only slightly (see technical appendix Table C1). On-time grade-level progression and graduation were examined for students remaining in their district through 2012. RFEP5 students are those reclassifed by the end of 5th grade. EL5 students are EL students who have not been reclassifed by the end of 5th grade. Language profciency group results are signifcantly diferent from each other for on-time grade progression in each grade in LAUSD and SDUSD, with one exception: in SDUSD, RFEP5 students are not signifcantly diferent from native English-speaking students. On-time graduation rates are statistically diferent for all language profciency groups in both districts. Pathways to Fluency 12 www.ppic.org Students reclassifed in elementary school maintain high rates of on-time grade progression throughout high school, and we see no evidence that their strong perfor- mance erodes relative to other language profciency groups. Tis is especially true in Los Angeles, where native English speakers and students not reclassifed in elementary school steadily lose ground from 9th to 12th grade relative to EL students who are reclassifed in elementary school. We also compare graduation rates, but it is important to note that graduation requirements are more rigorous in SDUSD than in LAUSD. 20 Graduating on time requires sufcient credit accumulation, minimum GPA require - ments, and passing specifc courses in addition to passing the CAHSEE. We fnd that in LAUSD, reclassifed students in the cohort we study are more likely to graduate than native English speakers (82% versus 66%). In SDUSD, these two groups of students are equally likely to graduate (78%). Performance on the CAHSEE In both Los Angeles and San Diego, students reclassifed in elementary school are more likely than any other group to pass the CAHSEE in 10th grade and by 12th grade (Figure 2). In fact, nearly 100 percent do so by 12th grade. Students who are not reclassifed in elementary school have low 10th grade CAHSEE passage rates in both dis- tricts (61% in LAUSD and 51% in SDUSD) but do make great strides by the end of grade 12, with approximately 80 percent passing in both districts. 21 Outcomes for Reclassified Students, by Language Spoken and Grade Level at Reclassification Here, we briefy summarize analyses that examine aca- demic outcomes for students reclassifed by the end of 5th grade by their primary language and their elementary school grade level at reclassifcation. We fnd that Spanish- speaking students reclassifed in elementary school have less positive outcomes than those who speak other lan- guages. However, the performance of Spanish speakers is still stronger than that of native English speakers in LAUSD (although not always in SDUSD). (See technical appendix Tables C2, C3, and C4 and technical appendix Figure C3 for full results.) When we examine outcomes according to the grade level in which students are reclas- sifed (grades 2 through 5), we fnd that students who are reclassifed in 2nd or 3rd grade have better outcomes than those reclassifed in 5th grade, on average, in both districts. (See technical appendix Figure C4 for CST results and further discussion of other outcomes.) Students Reclassified Early Have Good Outcomes and Reclassified Students Do Not Lose Ground We have found that in both Los Angeles and San Diego, students reclassifed by the end of 5th grade are among the best performers, and we see no evidence that their performance falters at higher grade levels relative to native English speakers. We also fnd evidence that students reclassifed at the end of elementary school (5th grade) have slightly lower academic outcomes than students reclassi- fed earlier. Research using a wider range of reclassifcation grade levels has found evidence that students reclassifed in high school do not do as well as students reclassifed at earlier grade levels (Hill, Weston, and Hayes 2014). Tis suggests an explanation for the apparent narrowing of the performance gap between reclassifed students and other Percentage passing CAHSEE overall Passing CAHSEE by 2012 Passing CAHSEE in 2010100 90 80 70 60 50 40 30 20 100 Figure 2. Virtually all students reclassifed in elementary school pass the CAHSEE by grade 12 SOURCE: Authors’ calculations from SDUSD and LAUSD individual student data, 2001–2012. NOTES: SDUSD data include students in grade 2 in 2001 or 2002; LAUSD data focus on those in grade 2 in 2002. The 2010 CAHSEE includes students taking the CAHSEE who should be 10th graders but have been retained in 9th grade. Not all retained 9th graders in LAUSD took the 2010 CAHSEE. Students making on-time progress would be in the 12th grade in 2012. CAHSEE scores are examined for students still in their district in 2012. RFEP5 students are those reclassifed by the end of 5th grade. EL5 students are EL students who have not been reclassifed by the end of 5th grade. Language profciency group results are statistically signifcantly diferent from each other for CAHSEE passage in both 2010 and 2012 in both districts. RFEP5 EL5 Native English speaker Native English speaker RFEP5 EL5 LAUSD SDUSD 13 Pathways to Fluency www.ppic.org students in high school in cross-sectional research fnd- ings: the RFEPs in these studies include students reclassi- fed in high school, who do not perform as well as students reclassifed in elementary or middle school (Hill 2012; Gándara and Rumberger 2006). Is the stronger performance of reclassifed students simply a result of skimming of the highest-achieving EL students (who may or may not have benefted from EL instruction), or do reclassifed students make academic gains because of English Learner instruction and subse- quent placement in mainstream instructional programs? Or is it simply that the student characteristics associated with elementary school reclassifcation are also associated with strong academic outcomes in middle and high school? Te CST scores of English Learner students not reclassi- fed by the end of 5th grade do improve by grade level more than native English speaker scores do, which suggests that ELD instruction benefts both ELs and reclassifed students. Although these questions are beyond the scope of this report, it is important for districts and the state to consider them more fully. Without a doubt, the use of reclassifcation criteria means that the best-performing students leave the English Learner group—indeed, many native English speakers do not meet the minimum CST scores that EL students are required to meet to be reclassifed. Tis is most clear in LAUSD, where students reclassifed in elementary school outperform not only native English speakers but also stu- dents who are initially fuent in two languages. In SDUSD, reclassifed students also perform well but do not generally exceed the performance levels of native English speakers. Te performance gap between the two districts is most likely explained by the relative difculty of reclassifcation criteria and, perhaps, also by the lower socioeconomic sta - tus of native English speakers in LAUSD relative to SDUSD. Reclassification Criteria and Long-Term Outcomes Districts have latitude in setting reclassifcation policies as long as they use the four criteria required in state law and take into consideration the SBE’s guidelines. To develop efec - tive policies, districts need to understand the relationship between the thresholds they establish for each criterion (or the use of additional criteria) and student outcomes. In this section, we examine the reclassifcation criteria that are most difcult for students to meet in each district. We also exam - ine middle and high school academic outcomes to under - stand how they are afected by more rigorous reclassifcation thresholds in each district—in particular, we look at the efect of higher CST cut-of scores and the use of course grades. Which Criteria Are Most Challenging? In a recent statewide survey of school districts’ reclassif- cation policies, respondents reported that the basic skills criterion (CST) was most difcult for their EL students to meet (53% for elementary grades, 62% for middle school grades, and 68% for high school grades). In elementary school, the English profciency requirement (CELDT) was a close second (40%) (Hill, Weston, and Hayes 2014). To fnd out whether these district perceptions are borne out by the data on student performance, we have adapted a technique used by Robinson (2011) to determine which of a district’s reclassifcation criteria are actually the most difcult for students to meet. Since EL students who are not reclassifed may not meet any of the reclassifcation criteria, we need to create a measure of which criterion is, on average, the most difcult for students to overcome. 22 Tis requires detailed knowledge of each district’s reclassi - fcation criteria and many observations of student perfor- mance. For each EL student in each grade, we calculate the distance between that student’s test scores or course marks We fnd that Spanish-speaking students reclassifed in elementary school have less positive outcomes than reclassifed students who speak other languages. Pathways to Fluency 14 www.ppic.org and the scores or grades that would allow that student to be reclassifed under his or her district’s criteria. 23 Te distance between the score or grade and the reclassifcation requirement determines the difculty of the requirement. 24 LAUSD has a lower ELA CST threshold than SDUSD (300 rather than 333) and allows two CELDT subtests to be below Early Advanced (SDUSD allows one). However, LAUSD requires a mark of 3 or better (on a 1- to 4-point scale) in ELA courses as a condition of reclassifcation (until 2006–07, it also required a 3 or better in math). SDUSD does not have any course mark requirement for reclassifca - tion, and such a requirement is not suggested in the SBE guidelines. In Figure 3, we show the biggest reclassifca- tion challenges for LAUSD’s ELs, taking into account the lower cut-point requirement on the CST but using the Early Advanced CELDT cut-of for overall profciency level and all subtests. In elementary grades, the reading component of the CELDT is the bigger barrier for students. Te CST is not a major obstacle until 5th grade, at which point it is the big - gest constraint for about 40 percent of ELs and remains the most common difculty through the end of middle school. However, in analyses where we also included the ELA and math mark requirements for LAUSD elementary school students, we found that marks are the most common stum - bling block for 4th and 5th graders; nearly 70 percent had a math or ELA mark as their most difcult requirement (see technical appendix Table D1). 25 Te ELA writing mark was the biggest challenge for about one-quarter of 4th and 5th grade students. Math marks were the biggest problem for 16 percent of 4th graders and 24 percent of 5th graders. With the mark requirement in place, the relative importance of the CST fell dramatically, and the CELDT reading subtest fell in importance by approximately 40 percentage points in 4th grade to 20 percent and approximately 30 percentage points in 5th grade to 15 percent. In discussions with district staf, we learned that LAUSD has recently convened meet - ings to discuss students who meet all reclassifcation criteria except ELA marks. In some cases, staf may decide to reclas- sify students with marks below 3. When we analyze San Diego’s reclassifcation require - ments, we fnd that the ELA CST requirement of 333 or higher (which is 33 points higher than both the LAUSD and suggested SBE thresholds) is the main criterion preventing students from being reclassifed at grades 4 through 7, with a peak of nearly 60 percent in grade 5 (Figure 4). For grades 2 and 3, as well as grades 8 through 10, the CELDT subtest Percentage with binding reclassifcation constraint CELDT overall CELDT writing CELDT reading CELDT listening speaking CST ELA100 90 80 70 60 50 40 30 20 100 Figure 3. In Los Angeles, the CELDT reading requirement is the most common obstacle to reclassifcation in elementary grades, but the CST becomes as big a barrier in later grades SOURCE: Authors’ calculations using LAUSD individual student data, 2002–2012. See note 23. NOTES: Students must be in 2nd grade in 2002 and must have been in the sample with complete observations of test scores for three consecutive years. Binding reclassifcation criteria for LAUSD do not include grade requirement criteria. We do not show these results because we have only ELA and math marks for grades 4 and 5. Those are shown in technical appendix Table D1. Grade 2 3 4 5 6 7 8 9 10 11Percentage with binding reclassifcation constraint 100 90 80 70 60 50 40 30 20 100 Figure 4. In San Diego, the CELDT reading requirement and the ELA CST requirement are the main obstacles to reclassifcation SOURCE: Authors’ calculations using SDUSD individual student data, 2001–2012. See note 23. NOTES: Binding reclassifcation criteria, SDUSD. Students must have been in 2nd grade in 2001 or 2002 and in the sample with complete test score observations for three consecutive years. Grade 2 3 4 5 6 7 8 9 10 11 CELDT overall CELDT writing CELDT reading CELDT listening speaking CST ELA 15 Pathways to Fluency www.ppic.org score in reading is furthest away from the reclassifcation threshold. 26 Rarely is the requirement of an overall perfor - mance level of Early Advanced or the score on the writing subtest the biggest obstacle for ELs. Listening/speaking is the third-most-challenging criterion. 27 Assessing the Effect of Rigorous Reclassification Criteria As we have seen, some of the reclassifcation criteria in LAUSD and SDUSD are more stringent than those sug - gested by the state’s guidelines. Tese diferences in district criteria can help us answer some important questions. Sup - pose that some of the students who met the state’s suggested reclassifcation criteria but not the more rigorous require - ments set by their districts had indeed been reclassifed. Would they have fared worse academically than those who met all of the district criteria? To explore this question, we divide the students who were reclassifed in one district into two groups: those who would have met the other district’s more stringent requirements and those who would not. 28 What If SDUSD Lowered Its CST Requirement? We begin by looking at what would have happened in SDUSD if the district had allowed students scoring between 300 and 332 on the ELA CST to be reclassifed. To do this, we compare LAUSD students who scored between 300 and 332 on the ELA CST to LAUSD students who scored between 333 and 366. 29 (Tese scores are from the year each LAUSD student was reclassifed.) First, we fnd that 25 percent of the LAUSD students who were reclassi- fed in elementary school scored between 300 and 332 in the fnal year before they were reclassifed and therefore would not have been reclassifed in SDUSD. We fnd that in middle school, the reclassifed students who scored in the higher range just before reclassifcation in elementary school obtained higher scores on the grade 8 ELA CST and grade 7 math CST than those who were reclassifed with lower scores (Figure 5). However, these diferences are relatively small—only about 0.35 mean performance level on the grade 7 math CST and 0.40 on the grade 8 ELA CST. Results for other academic outcomes are similar. Students with higher CST scores (333–366) when they are reclassifed in elementary school pass the CAHSEE at higher rates than students with scores in the lower range (300–332). Tis gap narrows as students continue to take the CAHSEE in grades 11 and 12; by 12th grade, 92 percent of those reclassifed with lower CST scores pass the CAHSEE as compared to 97 percent of those with higher scores. (See technical appendix Figure E1.) Tere is almost no difer- ence in on-time grade progression in high school or on- time graduation between students scoring in the high and low ranges on the CST in their year of reclassifcation. (See technical appendix Figure E2.) What If SDUSD Added Course Mark Requirements? Now we examine what might happen if SDUSD added academic mark requirements to its reclassifcation crite- ria. 30 If we apply LAUSD’s ELA and math mark require- ment to reclassifed EL students in SDUSD (using their estimated grades), we make a striking fnd: most of them CST mean performance level 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 Figure 5. Reclassifed LAUSD students who meet SDUSD’s CST requirement have slightly higher CST scores in middle school SOURCE: Authors’ calculations using LAUSD individual student data, 2002–2012. NOTE: Mean diferences for students predicted to satisfy SDUSD CST requirements are statistically signifcant in LAUSD. 8th grade ELA CST 7th grade math CST Reclassifcation CST 300–332 score Reclassifcation CST 333–336 score Diference Some of the reclassifcation criteria in LAUSD and SDUSD are more stringent than those suggested by the state’s guidelines. Pathways to Fluency 16 www.ppic.org would not have been reclassifed in the year SDUSD reclas- sifed them. Te percentage of students who would not have been reclassifed ranged from 71.2 percent in grade 3 to 92.5 percent in grade 4. 31 Students reclassifed in elementary school in San Diego who met LAUSD’s ELA mark criteria scored higher in both the 8th grade ELA and 7th grade math components of the CST than students who did not (Figure 6). For the ELA CST, students who met the Los Angeles criteria averaged a performance level of 4.3 versus 3.4 for those who did not meet the criteria. For the math CST, students in San Diego who met the Los Angeles criteria averaged a performance level of 4.2 versus 3.4 among those who did not meet the criteria. Tese diferences in middle school CST scores based on meeting the Los Angeles ELA GPA requirement (or not) are larger than those observed using high and low CST cut-ofs (Figure 5). In the longer term, the diference in outcomes can be quite large. Among reclassifed students in SDUSD who met the LAUSD criteria, the passage rate on 10th grade CAHSEE was 97 percent, as opposed to 86.5 percent for those who did not meet the criteria. However, both groups had very high passage rates by the end of 12th grade (100% versus 97.8%). 32 Graduation rates ten years afer grade 2 were also very similar overall (89.2% versus 88.3%), although there is some variation depending on the grade level at which students were reclassifed. SDUSD students who were reclassifed in grade 2 without meeting the LAUSD criteria graduated from high school at a higher rate than students reclassifed in grades 3 through 5 (92.6% versus 83.3%). 33 Are Rigorous Criteria Worth It? EL students in Los Angeles fnd the ELA course mark requirement for reclassifcation the most difcult criterion to meet. In San Diego, where grades are not a reclassifca- tion criterion, the ELA CST and the CELDT reading sub- test are the most difcult criteria. In general, we fnd that outcomes for students who were reclassifed in elementary school by meeting more challenging district-adopted reclassifcation standards are, in many cases, only slightly better than outcomes for students who were not reclassi- fed. Given that large numbers of students are apparently held back from reclassifcation because of the more difcult criteria—with very little long-term gain—these fndings suggest that setting more challenging reclassifcation crite- ria may not be benefcial. When we compared reclassifed students in one district who would and would not have been reclassifed using the other district’s criteria, we found better outcomes for those reclassifed who met the higher reclassifcation criteria. If LAUSD had used SDUSD’s higher CST cut-of, 25 percent of LAUSD’s students reclassifed during elemen - tary school would have faced delayed reclassifcation. Similarly, we estimate that over 70 percent of SDUSD’s students reclassifed during elementary school would have faced delayed reclassifcation if SDUSD had used LAUSD’s ELA and math mark requirements. Our fndings help to explain why, for the cohorts we study, students in the two districts who were ever classifed as EL have similar CST and CELDT scores, but once we divide these students into those who were reclassifed in elementary school and those who were not, LAUSD students have higher CST and CELDT scores than SDUSD students. It appears that LAUSD’s reclassifcation policies keep some of the top- achieving EL students in EL status. CST mean performance level 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 Figure 6. Reclassifed SDUSD students who meet LAUSD’s ELA course mark criteria have higher CST scores in middle school SOURCE: Authors’ calculations using SDUSD individual student data, 2001–2012. NOTE: Mean diferences for students predicted to satisfy LAUSD GPA requirements are statistically signifcant in SDUSD. 8th grade ELA CST 7th grade math CST Reclassifcation with low marks Reclassifcation with high marks Diference 17 Pathways to Fluency www.ppic.org It is important to note that using each of the two districts’ reclassifed students to estimate diferences in outcomes in the other district is an imperfect exercise. Ideally, we would instead examine outcomes for reclas- sifed students in a third district that did not increase its reclassifcation criteria beyond the state-recommended minimum requirements. It is possible that such a compari- son would have revealed larger gains (and bigger reduc- tions in the share of high-performing EL students being reclassifed) as a result of the more rigorous requirements in our two districts. Finally, this exercise cannot account for diferences in English Language Development pro- grams and supports across the two districts. Which Reclassification Criteria Best Predict Student Success? SBE guidelines for reclassifcation advise districts to use CELDT and CST scores but allow the use of student grades in making reclassifcation decisions. 34 However, these guidelines are about to change. New ELD standards were adopted in 2012, and a replacement for the CELDT should be implemented by 2016. California is limiting statewide administration of the CST in 2013–14 in preparation for the introduction of the Smarter Balanced Assessments in 2014–15. Given these imminent changes to the state testing system, now is the perfect time to determine which of these three factors (CST, CELDT, and course marks) best predict success in middle and high school for elementary school students who are English Learners. 35 Will grades alone be sufcient? Tey are somewhat subjective, and grading stan- dards can vary across districts and even across teachers within a district. We also consider whether the practice in California of using both “basic skills” and “English prof- ciency” requirements through two separate tests produces complementary or redundant information. We compared the predictive power of three indicators, measured in grade 5: the CELDT subtest performance levels, the ELA CST performance level, and ELA course marks. 36 We estimated models of fve middle- and high- school outcomes as functions of these three indicators for EL students not reclassifed in elementary school: ELA CST scores in grade 8, math CST scores in grade 7, CAHSEE passage in grade 10, grade retention before grade 10, and on-time graduation. For each EL student, we used variables from 1 to 5 for performance levels on the CST and the CELDT and from 1 to 4 to capture marks (in each of read- ing, writing, listening, and speaking in Los Angeles, and in each of reading, writing, and oral language in San Diego). For both districts, when we attempted to explain test-score-based outcomes in 7th through 10th grades, CST test scores had the most explanatory power, followed by CELDT scores and ELA marks. Taken together, these three indicators captured one-third to one-half of the variation in grade 7 math CST achievement, grade 8 ELA CST achievement, and CAHSEE passage in grade 10. It is These fndings suggest that setting more challenging reclassifcation criteria may not be benefcial. With changes to the state testing system ahead, now is the perfect time to determine which factors best predict academic success for English Learners. MELANIE STET SON F REE MA N/TH E CHR I STIAN SCIENCE MONITO R Pathways to Fluency 18 www.ppic.org perhaps unsurprising that grade 5 CST scores are strong predictors—like the outcome measures we used in our model, they are test-based measures of academic achieve- ment. Figure 7 shows the percentage of the variation in stu- dent outcomes that we could explain using any one of these variables and all three variables together. Using CST scores alone led to models that explained slightly less variation than when we used all three predictors. However, for the two high school outcomes (whether a student had been retained a grade by the end of grade 10 or graduated within ten years of entering grade 2), the explanatory power of these variables, alone or together, was much weaker. We also estimated models that included all three combinations of pairs of these predictor variables: CST, CELDT, and GPA. Technical appendix Figure G1 shows that all three of these pairs perform almost as well as the model that includes all three sets of variables. We found that in LAUSD, using GPA together with either CST or CELDT scores explained almost as much of the variation in outcomes as did the full model with all three sets of vari - ables. But in SDUSD, the models using CST and CELDT scores were the best in explaining middle school CST scores and CAHSEE passage, and all of the models performed about equally well in explaining grade retention and graduation on time. For the cohorts of students we studied, SDUSD (unlike LAUSD) was not yet using a standards- based report card that stipulates the criteria teachers must use to assign grades. It could be that standards-based report cards tend to contain more reliable information on student performance. What insights does this analysis provide for state and district policymakers? Although one of the best predic- tors, the CST is no longer being administered in California schools; its replacement, which will be phased in during 2014–15, could have even better predictive power. In the interim, using only one indicator, such as the CELDT, would not produce markedly worse predictions of suc- cessful reclassifed student outcomes than would a more complex measure. Policy Implications In the context of the coming overhaul of the state testing system associated with the implementation of the CCSS, policymakers’ interest in instituting standardized reclas- sifcation criteria across the state, and the funding incen- tives of the LCFF, we have provided a timely review of the measures used by the state’s two largest school districts to determine when EL students are ready for English-only instruction. CST ELA grade 8 CST math grade 7 CAHSEE passage grade 10 Retained in or before grade 10 Graduated on time Percentage of variation explained 50 45 40 35 30 25 20 15 10 5 0 Figure 7. The percentage of variation in student outcomes explained by grade 5 test scores and the ELA GPA of English Learners varies SOURCE: Authors’ calculations from SDUSD and LAUSD individual student data, 2001–2012. NOTES: SDUSD includes those in grade 2 in 2001 or 2002. LAUSD data focus on those in grade 2 in 2002. CELDT subtests ELA GPA CST CELDT subtests, GPA, and CSTLAUSD CST ELAgrade 8 CST math grade 7 CAHSEE passage grade 10 Retained in or before grade 10 Graduated on time SDUSD 50 45 40 35 30 25 20 15 10 5 0 19 Pathways to Fluency www.ppic.org We found that in both LAUSD and SDUSD, students who fnished 2nd grade in 2002 and were reclassifed as fuent English profcient by the end of grade 5 generally did very well in middle and high school, performing about the same as or better than native English speakers on a variety of academic outcomes. Further, we found no evidence that reclassifed students’ performance faltered relative to that of native English speakers. Te key question is whether English Learners—in Los Angeles, San Diego, and other districts with more rigorous reclassifcation criteria than those suggested by the state—would beneft from being reclassifed slightly sooner, through an easing of reclas- sifcation standards, which would allow districts to con- centrate their resources on the most linguistically needy English Learners. Our fndings in the two largest districts in California lead us to believe that the answer is yes. We also believe that, in the longer term, standardizing reclas- sifcation policies across districts would allow educators and policymakers to compare outcomes across the state— something that will be vitally important as the LCFF is implemented. We end with several recommendations to help ensure a successful transition to new policies. 1. Even though the CST will not be administered in 2013–14, our results suggest that districts can make accurate reclassifcation decisions using only the CELDT. When replacements for the CST and CELDT become available, the state should consider allowing dis - tricts to reclassify students on the basis of just one test. Despite the diferences in student population (languages spoken, share of low-income students, and racial/ethnic distribution) and diferent reclassifcation criteria across the two school districts, the predictors of successful aca- demic outcomes for ever–EL students are similar. We doubt that California will ever abandon the CELDT or its successor, because schools need an objective method to evaluate the language abilities of new arrivals to the district whose frst language is not English, and these students can arrive at any time of the school year. Tests such as the CST and its successor cannot accomplish this goal. Further, the CELDT helps educators measure the progress of English Learners year to year. Because both the CELDT and the CST can predict EL students’ subsequent outcomes quite well, it makes sense for the state to consider whether an EL student who demonstrates sufcient mas- tery of English on either test should be reclassifed without having to face a second hurdle. 2. In the two largest districts in California, EL students fnd the CELDT writing requirement less challenging than the CELDT reading requirement. In designing new tests and reclassifcation standards, the state should consider the relative rigor of its reading and writing requirements. Our data suggest that the current CELDT writing require - ment is relatively easy compared to the CELDT reading requirement. Indeed, in LAUSD, where ELA marks on report cards are part of the reclassifcation criteria, the writing grade requirement is more challenging for EL students to meet than the required mark in reading. Given that the replace - ment exam for the CELDT is currently being developed, this might be an ideal time for the state to reconsider its relative expectations about reading and writing for EL students. 3. Districts should carefully consider whether their reclassifcation standards need to be more rigorous than the state-recommended minimum. A difcult policy question is whether either San Diego or Los Angeles is setting reclassifcation criteria too high or low. In part, the answer depends on how well we expect English Learners to perform academically once reclassi- fed. We examined reclassifed students in one district to see whether higher reclassifcation standards in the other district held back from reclassifcation students who would have fared poorly, had they been reclassifed. We found that more rigorous reclassifcation requirements in both districts—those related to CST scores in San Diego and academic marks in Los Angeles—were associated with slightly better outcomes in later grades, but the diferences were small. Further, we found that there is a downside to Pathways to Fluency 20 www.ppic.org these additional requirements: they prevented or delayed the reclassifcation of large numbers of students.We looked for variations across EL subgroups and found that students reclassifed at earlier grade levels had somewhat better outcomes in secondary school than those reclassifed later in the elementary school years. It is possible that students are benefting from English-only classroom instruction in the early grades or from not being labeled as ELs in later elementary school. Or it could be that students reclassifed early in elementary school had initial advantages in language profciency and academic preparation. Our current research cannot distinguish among these possibilities. 4. Afer careful consultation with districts, the state should consider establishing a uniform set of reclassi- fcation criteria for all school districts. Te issue of difering reclassifcation criteria is especially important, given that California’s new Local Control We found that more rigorous reclassifcation requirements in both districts were associated with slightly better outcomes in later grades, but the diferences were small. The key question is whether English Learners would beneft from being reclassifed slightly sooner, allowing districts to concentrate on the most linguistically needy. SA NDY HU FFAKER /CO R BIS 21 Pathways to Fluency www.ppic.org Technical appendices to this report are available on the PPIC website: www.ppic.org/content/pubs/other/514LHR_appendix.pdf Funding Formula increases funding for English Learners, which could create a disincentive for districts to reclassify students. Te state needs to consider whether it makes sense for some districts to have more rigorous reclassifcation criteria than others. A standard set of criteria could improve fairness for students and make it much easier to monitor the progress of students who have ever been English Learners. In sum, the process through which EL students are reclassifed as fuent English profcient is quite complex. Not only must students reach thresholds on two difer- ent tests, but individual districts can and do set their own requirements, which can be quite diferent from the SBE guidelines and from those of other districts. As the state implements new standards and assessments, and as the new funding formula takes efect, it should also consider making changes to reclassifcation criteria, including considering allowing EL students to meet the requirement on either the Smarter Balanced Assessment or the replace- ment for the CELDT. Te establishment of statewide reclassifcation criteria at a reasonable level of difculty could allow districts to concentrate their LCFF dollars on their lowest-performing students without slowing the academic progress of ELs who are performing well enough to be reclassifed in elementary school. Pathways to Fluency 22 www.ppic.org Notes 1 California Department of Education (CDE) DataQuest. 2 California Department of Education. 3 It is expected that these regulations will be adapted depending on their success in the frst year of LCFF implementation. 4 Further, statewide in 2012–13, two-thirds of English Learners were enrolled in the elementary school grades (K–5) (California Department of Education DataQuest). 5 In this cohort, very few students in SDUSD and almost none in LAUSD were reclassifed in 2nd grade. Statewide, reclassifcation before 2nd grade is rare (Hill, Weston, and Hayes 2014). 6 See Saunders and Marcelletti (2013), Hill (2012), EdSource (2008), and Gándara and Rumberger (2006) for examples. When RFEP and EL students are combined into an “ever-EL” group, the gap between ever-EL and native English-speaking students is considerably smaller and has declined somewhat over time (Saunders and Marcelletti 2013). However, even cross-sectional research that refnes comparison groups for ELs cannot account for the time since reclassifcation or for new entrants to the EL population. 7 Previous research has demonstrated a narrowing of the achievement gap following reclassifcation (Silver, Saunders, and Zarate 2008), but there is reason to believe that the timing of reclassifcation also matters. ELs who are reclassifed quickly have better long-term academic outcomes than those who con- tinue in EL status for many years (Flores, Painter, and Pachon 2009). Not surprisingly, there is a great deal of interest in under- standing the role of reclassifcation standards in the ultimate success of ELs (Parish et al. 2006). 8 California Education Code Section 313. 9 Ofcial CDE reclassifcation rates reported. 10 Because SDUSD student data are available earlier and the size of the student population is smaller, the cohort of students for SDUSD combines 2nd graders from 2001 and 2nd graders from 2002 and follows students to 2011 and 2012, respectively. 11 Te size of the ever-EL group may itself afect educational out- comes for the ever-EL group. However, determining the efect of peers on educational outcomes for ELs and RFEPs is beyond the scope of this report. 12 Te OPL on the CELDT can range from 1 to 5. Te correspon- dence between numbers and performance levels is as follows: 1 = Beginning, 2 = Early Intermediate, 3 = Intermediate, 4 = Early Advanced, and 5 = Advanced. 13 Te overall reclassifcation rates in LAUSD have increased (9.5% to 13.7%) during this period, whereas those in SDUSD have remained steady (10.4% to 10.5%). 14 Note that 2nd grade CST scores increase for all student lan- guage profciency groups between 2002 and 2008 in these two districts (Table 3) and statewide (California Department of Education 2013). 15 It is possible that EL instruction between kindergarten and 2nd grade is more benefcial in LAUSD or that there are unobserved diferences between students who were ever clas- sifed as ELs in the two districts. 16 In technical appendix Tables A1 (LAUSD) and A2 (SDUSD), the demographic characteristics of students who remain in the district (“stayers”) and leave the district (“leavers”) are displayed separately. We fnd that students in LAUSD reclassifed in elementary school who remain in the district are somewhat more likely to be low-income (as measured by free/reduced price meal eligibility) and have somewhat less educated parents, but the diferences are slight. In SDUSD, students reclassifed in elementary school who remain in the district have slightly better educated parents than those who leave the district. 17 Recall that our focus is on ELs reclassifed in elementary school. Some EL students who were not reclassifed by the end of 5th grade may have been reclassifed in middle or high school. 18 We examine math CST scores only through 7th grade because beginning in 8th grade, students take a variety of math CST tests within each grade. For example, in 9th grade, students typi - cally take the CST for Algebra I, Geometry, or (least commonly) Algebra II. Combining the mean performance levels across examinations measuring knowledge of diferent math content is not appropriate. 19 A major diference between the two districts is the relatively stable math CST scores across grades in SDUSD versus the steady decline in math CST scores from the peak in 4th grade in LAUSD (across all language profciency groups) to 7th grade. For example, the mean performance level among LAUSD reclas- sifed students in 4th grade is 4.2, but by 7th grade it has fallen to 3.6. Our requirement that students must remain in their districts to be included in our sample does not afect our results much— we conduct the same analyses for students who eventually leave 23 Pathways to Fluency www.ppic.org their districts (technical appendix Figures C1 and C2) and fnd that the patterns by profciency group are similar, but perfor- mance levels are somewhat lower in comparison to students who remain in their districts. 20 SDUSD’s graduation requirements appear to be more chal- lenging than LAUSD’s. SDUSD students must have a minimum GPA of 2.0 and must successfully complete three years of science and three years of math. (See Table 1 in Betts, Zau, and Bachofer 2013.) LAUSD requires a minimum GPA of 1.0, two years of science, and two years of math (LAUSD 2010). Because we do not have dropout data for LAUSD, we exclude dropouts from our graduation calculations here and throughout the report. When we include dropouts in the SDUSD data, graduation rates decrease only slightly (see technical appendix Table C1). 21 Technical appendix Table C4 shows results for IFEP students as well. In San Diego, IFEP students pass the CAHSEE at higher rates than other students (96% by the end of 12th grade), but in Los Angeles, RFEP students pass CAHSEE by the end of 12th grade at higher rates than IFEP students (98% versus 94%). 22 We refer to this criterion as the “binding constraint” because it is the criterion that the student is furthest from meeting. Within each grade level, we count the number of times a par- ticular reclassifcation criterion is the furthest from being met for all the students in that grade. 23 Our measure of distance has been standardized for each reclassifcation criterion by transforming students’ scores into z-scores. For example, we take a student’s overall score on the CELDT and subtract it from the score required (Early Advanced) and divide it by the standard deviation of the CELDT score for EL students in that grade in that district. We iden- tify the binding constraint as the requirement for which the student’s score is the greatest number of standard deviations below the required level. It is important to note that the students included in our analyses for Figures 3 and 4 are those in grade 2 in 2002 (or 2001 as well in SDUSD) who remain EL students in the given grade a number of years later. Tus, the fgures do not include students who have been reclassifed. Te reclassi- fcation of students out of the group over time may, in part, be responsible for the gradual changes across grades in the binding constraint to reclassifcation for students who are still ELs in the given grade. 24 In both districts, it appears that reclassifcation decisions are made according to the policies in place. In LAUSD, about 2 per- cent of all students should have been reclassifed when they were not, and the same was true for about 4 percent of students in SDUSD. Tere were almost no observations of reclassifcations that did not meet the reclassifcation policy requirements—we saw 13 observations in total (or 0%) in LAUSD, and in SDUSD, about 2 percent of reclassifcations should not have been made. 25 We have no marks for 2nd graders and marks for only some of our 3rd graders, so we exclude them from this analysis. Imple- menting the grade level cut-of requirement for middle and high school is difcult because of the way course names are recorded at higher grade levels. 26 For instance, the CELDT reading subtest requirement was the binding constraint for 43 percent of 10th graders. 27 Recall that SDUSD students may score less than Early Advanced on one CELDT subtest, but for ease of exposition, we do not allow for that possibility here. 28 Because both districts go beyond the state’s recommended minimum reclassifcation requirement, students in one district who would have failed to meet the other district’s requirement are not the perfect comparison group—they are still in some sense above average. If anything, this probably biases our analy- sis toward fnding that higher reclassifcation requirements did not lead to big changes in outcomes. 29 A category of similar size. 30 Since marks are subjective and since LAUSD used a standards- based report card whereas SDUSD did not, we use the relation- ships we observe between test scores and marks in LAUSD to approximate marks for SDUSD students had they been enrolled in LAUSD. We use CELDT and CST scores to predict marks for EL and RFEP students in LAUSD. Coefcients from those models are used in estimating the marks of SDUSD students who were reclassifed in elementary school. Tese regressions are found in technical appendix F. 31 We reanalyzed the SDUSD data using the more recent LAUSD requirement that uses marks in ELA but not math. Te results were virtually the same, with only three more students predicted to have met the LAUSD mark requirements. 32 See technical appendix Figure E3. 33 Tere were diferences between students who met the LAUSD coursemark criteria and those who did not. Tose who met the criteria tended to have higher CELDT OPL scores and higher CST scores. Tey also difered demographically, with those who met the LAUSD course mark criteria less likely to be Hispanic and with parents with higher education levels. Pathways to Fluency 24 www.ppic.org 24 References Assembly Bill 484, Pupil Assessments: Measures of Academic Performance and Progress (MAPP). Available at http:// leginfo.legislature.ca.gov/faces/billNavClient.xhtml?bill_id =20132014 0A B 4 8 4 . Betts, J. R., A. C. Zau, and K. V. Bachofer. 2013. College Readi- ness as a Graduation Requirement: An Assessment of San Diego’s Challenges , San Francisco: Public Policy Institute of California. California Department of Education. “Facts about English Learners in California—CalEdFacts.” Available at www.cde. c a .gov/d s/s d /cb/c efel f ac t s . a sp. California Department of Education. 2012. “California English Language Development Test (CELDT), 2012–2013,” CELDT Information Guide. Available at w w w.cde.ca.gov/ta/tg/el /documents/celdtinfoguide1213.pdf/. California Department of Education. 2013. “State Schools Chief Tom Torlakson Releases 2013 STAR Results,” News release, August 8. 34 Many school districts have decided to use students’ grades as an element in the reclassifcation decision (Hill, Weston, and Hayes 2014). 35 Assembly Bill 484 requires that districts administer the computer-based feld test of the new Common Core Assess- ments in 2013–14, but the results of these tests will not be given to schools. A survey of California school districts estimated that only 33 percent to 40 percent of school districts are able to ofer the computer-based tests in 2013–14. Te new tests will be given in all districts in 2014–15 (Fensterwald 2013). 36 Full results are shown in technical appendix Tables G1–G10. California Department of Education DataQuest. Available at http://data1.cde.ca.gov/dataquest/. California Education Code Section 313. Available at w w w.leginfo.ca.gov/cgi-bin/displaycode?section=edc&group =00001-01000&file=313-313.5. EdSource. 2008. English Learners in California: What the Num- bers Say , Mountain View, CA. Fensterwald, J. 2013. “What’s Next for Standardized Testing in California?” EdSource (blog), October 22. Flores, E., G. Painter, and H. Pachon. 2009. ¿Qué Pasa? Are Eng- lish Language Learning Students Remaining in English Learning Classes Too Long? Los Angeles: Tomás Rivera Policy Institute Full Report. Gándara, P., and R. Rumberger. 2006. Resource Needs for Cali- fornia’s English Learners . Linguistic Minority Research Institute, University of California. Hill, L. E. 2012. California’s English Learner Students , San Fran- cisco: Public Policy Institute of California. Hill, L., M. Weston, and J. Hayes. 2014. Reclassifcation of Eng- lish Learner Students in California . San Francisco: Public Policy Institute of California. Los Angeles Unifed School District. 2010. 2010–2011 Gradu- ation Requirements and Minimum College Admission “A–G” Requirements. Available at http://home.lausd.net/ourpages /auto/2011/12/21/53468052/2010-2011GRADUATION REQUIREMENTSCHART.pdf. Los Angeles Unifed School District. 2011. Reclassifcation of English Learners, Grades 2–12, Bulletin BUL-5619.0, October 17. Los Angeles Unifed School District. 2014. “Comprehensive Assessment Program.” Available at http://notebook.lausd.net /portal/page?_pageid=33,167651&_dad=ptl. Ofce of Curriculum, Instruction, and School Support, LAUSD Ofce of the Deputy Superintendent of Instruction. 2011 . LAUSD Literacy Periodic Assessments Grades K–5 , October. Robinson, J. P. 2011. “Evaluating Criteria for EL Reclassifcation: A Causal-Efects Approach Using a Binding-Score Regression Discontinuity Design with Instrumental Variables,” Education Evaluation and Policy Analysis 33 (3): 267–92. 25 Pathways to Fluency www.ppic.org San Diego Unifed School District. 2009. Master Plan for English Learners . Saunders, W. M., and D. J. Marcelletti. 2013. “Te Gap Tat Can’t Go Away: Te Catch-22 of Reclassifcation in Monitoring the Progress of English Learners,” Educational Evaluation and Policy Analysis 35 (139). Senate Bill 1108, An act to add Section 313.5 to the Education Code, relating to English Learners. Available at www.leginfo .c a . gov/pu b/11-12/ bi l l /s en /s b _1101-115 0/s b _110 8 _ bi l l _ 2 012 0 419 _amended_sen_v97.pdf. Silver, D., M. Saunders, and E. Zarate. 2008. “What Factors Predict High School Graduation in the Los Angeles Unifed School District,” Santa Barbara, CA: California Dropout Research Project Report. Acknowledgments Te authors wish to thank Hans Johnson, Margaret Weston, Karina Jaquet, and Jill Cannon for their help in developing the ideas in this report. We thank Kathy Hayes, Cynthia Lim, Josh Klarins, Valerie Brewington, and Marciela Sanchez Robles at LAUSD for their insight into the data and policies and their feedback about the research. We thank the SDUSD Ofce of Language Acquisi- tion, especially Mary Waldron, Debra Dougherty, and Shelby Madden, for providing extensive background information about English Learner programs and practices in San Diego and for their careful review of early drafs of this report. In addition, we thank Ron Rode, Peter Bell, and Dina Policar from the SDUSD Ofce of Accountability for their assistance and advice. We also wish to thank Caroline Danielson, Rachel Ehlers, Kathy Hayes, Laurel Beck, Lynette Ubois, and Kenji Hakuta for comments on early drafs and Mary Severance, Patricia Bedrosian, Jenny Miyasaki, and Kate Reber for editorial support. Any errors are our own. Pathways to Fluency 26 www.ppic.org 26 About the Authors Laura E. Hill is a research fellow at PPIC, where her research interests include immigrant well- being, English Learners in K–12 schools, race and ethnicity, and youth. She has been a research associate at the SPHERE Institute and a National Institute of Aging postdoctoral fellow. She holds a Ph.D. in demography from the University of California, Berkeley. Julian R. Betts is a Bren fellow and an adjunct fellow at PPIC. He is also professor and former chair of economics at the University of California, San Diego, where he is executive director of the San Diego Education Research Alliance. He is also a research associate at the National Bureau of Eco - nomic Research and UC San Diego campus director of the University of California Educational Eval - uation Center. He holds a Ph.D. in economics from Queen’s University, Kingston, Ontario, Canada. Belen Chavez is an analyst at Harris Economics Group and a former research associate at PPIC. She holds an M.A. in economics from Duke University and a B.A. in business economics from the University of California, Irvine. Andrew C. Zau is a senior statistician for the San Diego Education Research Alliance in the Department of Economics at the University of California, San Diego. His research focuses on the efects of policy and reform on student achievement in K–12 education. Previously, he was a research associate at PPIC. He holds an M.P.H. in epidemiology from San Diego State University and a B.S. in bioengineering from the University of California, San Diego. Karen Volz Bachofer is the director of the San Diego Education Research Alliance in the Depart- ment of Economics at the University of California, San Diego. Previously, she was the executive director of the San Diego Unifed School District’s Research and Evaluation Division, where her responsibilities included oversight of national, state, and district assessment and accountability processes and reporting. She holds a Ph.D. in education from the Claremont Graduate School and San Diego State University. 27 Pathways to Fluency www.ppic.org Board of Directors DONNA LUCAS , CHAIRChief Executive Ofcer Lucas Public Afairs MARK BALDASSAREPresident and CEO Public Policy Institute of California RUBEN BARRALESPresident and CEO GROW Elect MAR Í A BLANCOVice President, Civic Engagement California Community Foundation BRIGITTE BRENAttorney WA LT E R B. HEWLETTMember, Board of Directors The William and Flora Hewlett Foundation PHIL ISENBERGVice Chair Delta Stewardship Council M AS MASUMOTOAuthor and Farmer STEVEN A. MERKSAMERSenior Partner Nielsen, Merksamer, Parrinello, Gross & Leoni, LLP KIM POLESEChairman ClearStreet, Inc. THOMAS C. SUT TONRetired Chairman and CEO Pacifc Life Insurance Company PPIC is a public charity. It does not take or support positions on any ballot measures or on any local, state, or federal legislation, nor does it endorse, support, or oppose any political parties or candidates for public ofce. PPIC was established in 1994 with an endowment from William R. Hewlett. Copyright © 2014 Public Policy Institute of California. All rights reserved. San Francisco, CA Mark Baldassare is President and Chief Executive Ofcer of PPIC. Donna Lucas is Chair of the Board of Directors. Short sections of text, not to exceed three paragraphs, may be quoted without written permission provided that full attribution is given to the source. Research publications refect the views of the authors and do not necessarily refect the views of the staf, ofcers, or Board of Directors of the Public Policy Institute of California. Library of Congress Cataloging-in-Publication Data are available for this publication. I S B N 978 -1-5 8213 -157-3 PUBLIC POLIC Y INSTITUTE OF CALIFORNIA 500 Washington Street, Suite 600 San Francisco, California 94111 Telephone 415.291.4400 Fa x 415 . 2 91. 4 4 01 PPIC S ACRAMENTO CENTER Senator Ofce Building 1121 L Street, Suite 801 Sacramento, California 95814 Telephone 916.440.1120 Fax 916.440.1121 Additional resources related to K–12 education are available at www.ppic.org. The Public Policy Institute of California is dedicated to informing and improving public policy in California through independent, objective, nonpartisan research." ["post_date_gmt"]=> string(19) "2017-05-20 09:41:58" ["comment_status"]=> string(4) "open" ["ping_status"]=> string(6) "closed" ["post_password"]=> string(0) "" ["post_name"]=> string(8) "r_514lhr" ["to_ping"]=> string(0) "" ["pinged"]=> string(0) "" ["post_modified"]=> string(19) "2017-05-20 02:41:58" ["post_modified_gmt"]=> string(19) "2017-05-20 09:41:58" ["post_content_filtered"]=> string(0) "" ["guid"]=> string(50) "http://148.62.4.17/wp-content/uploads/R_514LHR.pdf" ["menu_order"]=> int(0) ["post_mime_type"]=> string(15) "application/pdf" ["comment_count"]=> string(1) "0" ["filter"]=> string(3) "raw" ["status"]=> string(7) "inherit" ["attachment_authors"]=> bool(false) }