Donate
Independent, objective, nonpartisan research

Academic Progress for English Learners: The Role Of School Language Environment and Course Placement in Grades 6-12

Database

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

View live version

object(Timber\Post)#3742 (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(131) "academic-progress-for-english-learners-the-role-of-school-language-environment-and-course-placement-in-grades-6-12-january-2019.pdf" ["wpmf_size"]=> string(6) "831879" ["wpmf_filetype"]=> string(3) "pdf" ["wpmf_order"]=> string(1) "0" ["searchwp_content"]=> string(138842) "JANUARY 2019 Laura Hill, Julian Betts, Megan Hopkins, Magaly Lavadenz, Karen Bachofer, Joseph Hayes, Andrew Lee, Marco A. Murillo, Tara Vahdani, and Andrew C. Zau Supported with funding from the William T. Grant Foundation Academic Progress for English Learners The Role of School Language Environment and Course Placement in Grades 6–12 © 2019 Public Policy Institute of California 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 office. 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 reflect the views of the authors and do not necessarily reflect the views of our funders or of the staff, officers, advisory councils, or board of directors of the Public Policy Institute of California. SUMMARY CONTENTS Introduction 5 Who Are Secondary ELs? 7 How Are Course Assignments Implemented? 11 How Do Rates of Progress Differ for Long-Term and Late-Arriving English Learners? 19 Are School Language Environments Related to EL Student Progress? 25 Are Course Placement and ELD Programs Related to EL Student Progress? 32 Conclusion 39 References 41 About the Authors 43 Acknowledgments 43 Technical appendices to this report are available on the PPIC website. PPIC.ORG Over 40 percent of students in California’s K–12 education system speak a language other than English at home. Almost half of these students are considered English Learners (ELs), meaning they require additional language and academic support to succeed in school. In middle and high school, ELs face the dual challenge of attaining fluency in English while mastering the academic courses leading to a high school diploma. This report examines two important types of ELs in Los Angeles and San Diego districts’ middle and high schools: long-term ELs, who have spent several years in US schools without being reclassified fluent in English, and late-arriving ELs, who first enroll in the district in grade 6 or higher and who enter with little English fluency. Our analysis uses student data from 2006–07 to 2015–16 and incorporates interviews conducted in 2017–18 with staff and teachers at the two districts. Overall Trends  Both districts saw a decline in enrollment of long-term and late-arriving ELs in middle and high schools. A principal reason is that the rate of reclassifying ELs as fluent in English before middle school has risen. Our interviews suggest that lower EL enrollment in middle and high schools may have led districts to group ELs of different English proficiency levels into the same class or to place them in general education coursework without integrated language supports.  The share of late-arriving ELs within the EL student population has increased. In recent years, many of these students have been refugees or unaccompanied minors. Generally, long-term ELs have more English fluency and higher standardized test scores than late-arriving ELs, but latearriving ELs make greater academic gains over time.  Both districts faced challenges in assigning long-term and late-arriving ELs to appropriate English Language Development (ELD) coursework, as defined by district guidelines. In both districts, over 50 percent of long-term and late-arriving ELs were placed correctly in recent years. Many other longterm and late-arriving ELs also received ELD courses, but at too high or low a level, while some did not receive ELD courses at all. School Language Environment and Course Placement  Higher proportions of EL students at a school should not be seen as detrimental to the academic performance of fluent English speakers. In both districts, the percentage of ELs at a school was not related to the academic outcomes of native or initially fluent English speakers. We found mixed evidence on the relationship between the percentage of ELs in a school and the academic performance of ELs. Academic Progress for English Learners 3  Long-term ELs who do not take an ELD course have reduced academic performance. In both districts, higher rates of no ELD placement are associated with slower growth for long-term ELs on statewide tests of English language arts. These results suggest that efforts should be made to support ongoing ELD.  In San Diego, we found some evidence that the presence of EL support teachers at schools led to better grades for long-term and late-arriving ELs, and faster English proficiency growth for late-arriving ELs. These results suggest school-level EL support teachers could help ELs progress more quickly. Several recent reforms in California aim to improve educational outcomes for English Learners—and there are more changes to come. Given the diversity and size of this student group, it is critical that policymakers take into account the varying needs of English Learners. PPIC.ORG Academic Progress for English Learners 4 Introduction English Learners are an important part of California’s K–12 student population. More than 40 percent of California school children speak a language other than English at home, according to the California Department of Education (CDE 2017–18).1 And more than one-fifth of California’s students are classified as English Learners (ELs); these students are identified by a standardized English language proficiency test as needing additional language and academic supports in order to succeed in school. A full 38 percent of California’s students are either current or former ELs, including students who were once classified as ELs but have since met reclassification criteria to be considered Reclassified Fluent English Proficient (RFEP). Los Angeles Unified School District (LAUSD) and San Diego Unified School District (SDUSD) are the largest school districts in California, serving 621,000 and 126,000 students, respectively, in 2017–18. These districts also serve substantial numbers of EL students: 143,000 in Los Angeles and 29,000 in San Diego, or approximately 23 percent of all students in each district. Combined, the two districts serve 14 percent of the state’s EL students and 4 percent of the nation’s EL students (CDE 2017–18; National Center for Education Statistics 2018). In this report, we focus on the educational experiences of two very important, but very different, groups of ELs in Los Angeles and San Diego district secondary schools (i.e., middle and high schools):2  The first group, long-term English Learners, or LTELs, make up the majority of ELs in grades 6–12 in California. LTELs are students who have been designated as ELs for several years without having been reclassified as fluent in English.3 Many have been English Learners since kindergarten.  The second group, late-arriving English Learners, or LAELs (also called newcomers), are ELs who arrive in US schools in grade 6 or later and have been in the country for no more than one year upon first enrollment. Some LAELs may have had formal education in their home countries, while others may not be literate in their home language. The heterogeneous backgrounds of long-term and late-arriving ELs raise important policy questions about whether districts and schools can deliver instruction effectively to all of these students. Moreover, we find that each school district recently made changes to its instructional offerings for these student groups. For example, LAUSD developed and mandated new courses for long-term ELs in 2014, where before there had been none; SDUSD piloted similar coursework in 2012 and offered, but did not require, districtwide implementation. For late-arriving ELs, both districts have been moving away from specialized, stand-alone instruction, even as they have experienced increases in their number of newcomers, many of whom were unaccompanied minors and refugees. This report begins by broadly describing long-term and late-arriving ELs, including socioeconomic and linguistic characteristics, district enrollment and school-level patterns. We then examine these students’ assignment to English language development (ELD) coursework and rates of linguistic and academic progress. Next, we explore whether aspects of the school language environment or ELD course placement are associated with rates of student progress on a variety of key academic outcomes. Finally, we highlight actions that schools, districts, and the state can take to support the progress of long-term and late-arriving ELs. By ELD coursework, we refer to both designated English language development classes, which focus explicitly on developing ELs’ English proficiency, as well as integrated ELD, or classes designed to integrate ELD strategies into the language arts content area. These two approaches comprise the comprehensive framework to EL instruction outlined in the California Department of Education’s English Language Arts/English Language 1 In citations to the CDE DataQuest database, the years refer to the school year. 2 EL students who do not fall into either of these categories are excluded from our analyses. 3 As noted below and in Technical Appendix A, the precise definitions of LTEL differ across the two districts and the California Department of Education. PPIC.ORG Academic Progress for English Learners 5 Development Framework for California Public Schools (2015), which states that ELs of all proficiency levels and ages require both designated and integrated ELD. In our analyses, we focus on designated ELD course placement and/or placement into integrated English courses, such as literacy, language arts, and literature. We do not study placement into integrated ELD courses in other content areas, like math and science. Our mixed-methods research approach combines quantitative and qualitative analyses. Our statistical analysis uses individual students’ academic records to examine course placement trends and school language environment. Our qualitative analysis involved nearly 100 semi-structured interviews with district officials and staff at 17 purposefully selected middle and high schools.4 Our qualitative analyses focused on understanding educators’ assessments of long-term and late-arriving ELs’ needs and characteristics, school course placement practices, and the institutional supports (or the lack thereof) that promoted and/or hindered their progress (see Technical Appendix C for a more detailed description). Methodological notes and limitations As with all research, this study has some limitations. We describe three of them here. First, the years included in our quantitative and qualitative datasets do not align. While our quantitative data included student-level data from 2005–06 to 2015–16, our interview data were collected in 2017–18. Although we asked interviewees to provide retrospective accounts, these accounts often did not include more than a few years. Thus, our qualitative data are limited in the extent to which they can help us directly explain the trends we observed in the quantitative data. Second, our focus on district and school course placement implementation did not examine classroom practices or their relative quality. Future research should consider employing additional qualitative methods, such as classroom observations and document reviews of curriculum and lesson plans, to determine how classroom instruction is structured and how teachers engage students in both designated and integrated ELD. For example, the presence of specific designated ELD courses designed for long-term ELs in LAUSD does not imply that the curricular approach of these classes was facilitating students’ language development or academic success. A closer look at classroom practices would help us understand course quality and student engagement. Additionally, including students’ perspectives may also yield important insights about their experiences and engagement in ELD courses. Finally, we did not attempt to tease apart whether a student was enrolled in designated or integrated ELD, or both, or the association that enrollment in each type of ELD had with particular outcomes; thus, we cannot make claims about whether designated or integrated ELD was more or less important for predicting success. 4 School sites were selected based on multiple criteria, including the percentage of ELs served, test score gains on English language proficiency exams, and language heterogeneity. Although we selected schools that represented a range of these characteristics, they are not representative of all schools in LAUSD or SDUSD. PPIC.ORG Academic Progress for English Learners 6 Who Are Secondary ELs? In California, 50 percent of English Learners at the secondary school level are long-term ELs (CDE 2016–17), which the state defines as students who have been English Learners for six or more years.5 Given that ELs often need five to seven years to become proficient enough in English to meet reclassification criteria (NASEM 2017), it is not surprising that there are many long-term ELs in middle and high schools. The research literature indicates that long-term ELs often have limited access to the academic coursework required for high school graduation and admission to postsecondary education (Callahan 2005; Kanno and Kangas 2014; Parrish et al. 2006); that they may experience some negative socioemotional and academic effects of being labeled EL (Flores, Kleyn, and Menken 2015; Thompson 2015); and that they tend to have elevated high school dropout rates (NASEM 2017; Silver, Saunders, and Zarate 2008; Watt and Roessingh 1994). Long-term ELs are not a homogenous group. They may have moved between English-as-a-second-language (ESL), bilingual, and mainstream instructional programs, and they may have moved back and forth between the US and their country of origin (Menken, Kleyn, and Chae 2012). Although many LTELs have strong oral English proficiency, they often have low GPAs and low test scores in English language arts and/or mathematics, which may be related to their inconsistent educational experiences and feelings of stigmatization in LTEL-specific courses that they perceive as being too easy for them (Menken and Kleyn 2010). Whereas long-term ELs constitute a large proportion of the state’s overall secondary EL population, late-arriving ELs, whom we define as EL students arriving within the last year and scoring at the beginning level on their first English proficiency test, constitute 6 percent of ELs in grades 6–12 (CDE 2016–17). Secondary LAELs often face significant challenges, especially if they experienced limited or interrupted formal schooling in their home countries (Umansky et al. 2018), because of the sheer amount of time needed to develop English proficiency and learn academic content in English. Longitudinal research shows that students’ age of arrival to the US is negatively associated with their likelihood of becoming proficient in English (Conger 2009). However, other research cautions against positioning age and second language acquisition as a biologically absolute relationship, instead framing the relationship as one that is influenced by social and educational contexts (Stevens 2015; MacSwan and Pray 2005; Marinova, Marshall, and Snow 2000). For example, Callahan (2005) finds better academic outcomes for secondary LAEL students with uninterrupted education than for those without, a finding confirmed by Hodara (2015) among ELs in the community college system. Overall, heterogeneity among late-arriving ELs in home language proficiency and literacy could influence achievement growth once they enter US schools. Because we wish to learn more about how both student groups progress, and about the factors associated with their progress, we turn to the two largest school districts in the state. In 2017–18, the California Department of Education (CDE) reports that Los Angeles Unified School District had 47,000 ELs in grades 6–12, while San Diego Unified School District had 9,000 such students. Both districts also have high poverty rates (82% in Los Angeles and 60% in San Diego).6 The poverty rates for secondary ELs are higher than for students on average and are more similar across the districts than the overall poverty rates: 94 percent in Los Angeles and 90 percent in San Diego. Despite these similarities, other characteristics of the districts’ EL populations differ. Whereas school staff in both districts indicated that most long-term ELs are Spanish speakers who were either born in the United States or hail 5 Starting in 2013, California students who have been classified as EL for six or more years are termed LTELs, and districts are subject to new requirements to identify these students in order to direct services toward them (Assembly Bill 2193). See Section 313.1 of the Education Code. Later in the paper, we use a definition that more closely maps to each district’s own LTEL definition. 6 The poverty rate is defined as the percentage of students eligible for free or reduced-price meals, which translates into a family income under 185 percent of the federal poverty level ($46,435 for a family of four in 2018). PPIC.ORG Academic Progress for English Learners 7 from Mexico or Central America, they described their late-arriving EL populations differently. Specifically, LAELs in Los Angeles were described as primarily from Mexico or Central American countries, yet many SDUSD school staff noted serving refugees from Africa, the Middle East, and Southeast Asia. Regardless of students’ countries of origin, staff in both districts indicated that many LAELs had previously experienced limited or interrupted formal schooling, as well as considerable trauma, prior to their arrival in the US. In addition, staff in both districts, but especially LAUSD, indicated that some LAELs were expected to work at least part time in order to contribute to family expenses. Using 2017–18 data reported by CDE, we find that 93 percent of secondary school ELs in LAUSD are Spanish speakers, and some ELs also speak indigenous languages. The next most common languages in the district are Armenian and Filipino, with slightly more than 1 percent of secondary ELs speaking each of these languages. In SDUSD, only 76 percent of secondary ELs speak Spanish. Vietnamese (4%), Filipino (4%), Somali (3%), and Arabic (2%) are the next most common languages. No other language is spoken by more than 1 percent of secondary ELs in the district. Further, an important distinction in SDUSD is that the percentage in our secondary school sample whose home language is Spanish is 84 percent for long-term ELs but only 64 percent for latearriving ELs. Time Trends In our analyses, we follow long-term and late-arriving EL students during grades 6–12. When we are studying academic outcomes for ELs, we include students in our sample even after they are reclassified so that we can examine factors associated with their academic successes and struggles. If we excluded students after reclassification, we would lose more successful students from our sample and bias our conclusions about what is positively associated with EL student progress. For that reason, our counts of students differ from those reported by CDE. Further, state education code defines LTELs somewhat differently than they are defined in each school district. To align better with the districts’ definitions, our analyses define LTELs as secondary school students who have completed five or more years of schooling as ELs.7 Our quantitative analysis shows that both districts saw a decline in their numbers of long-term and late-arriving ELs from 2007–08 to 2015–16 (Figure 1). This decline outpaced overall declines in student enrollment. In LAUSD, EL enrollment at the secondary level declined 60 percent between 2007–08 and 2015–16 (compared to 9% for the district as a whole). In SDUSD, secondary EL enrollment decreased 46 percent in the same time period (compared to 2% for all district students). 7 The LTEL definition used in this report conforms to LAUSD’s own LTEL definition and also recognizes that SDUSD considers ELs who have completed five years of schooling as seriously at risk of becoming LTEL in one more year. The statistical analyses described in this report are little changed if we instead use the official definition that an EL must have completed six, rather than five, years of schooling before being considered an LTEL. Our study population includes EL students who become LTELs after grade 6. PPIC.ORG Academic Progress for English Learners 8 FIGURE 1 Los Angeles and San Diego secondary schools both saw declining enrollments of long-term and late-arriving ELs 90 LAUSD English Learners, thousands 80 Long-term EL 70 Late-arriving EL 60 50 40 30 20 10 0 2007-08 2008-09 2009-10 2010-11 2011-12 2012-13 2013-14 2014-15 2015-16 9 SDUSD 8 English Learners, thousands 7 Long-term EL 6 Late-arriving EL 5 4 3 2 1 0 2007-08 2008-09 2009-10 2010-11 2011-12 2012-13 2013-14 2014-15 2015-16 SOURCE: Authors’ tabulations from school district data. NOTE: Here and later in the report we define LTELs as ELs who have completed five or more years as an EL. Reclassified long-term and latearriving ELs are not included in this figure. We found that part of the decline in secondary EL students was linked to increased reclassification rates in elementary school. In SDUSD, the percentage of ELs in elementary school who were still classified as ELs in grade 6 fell from 62 percent in 2006–07 to 41 percent in 2015–16. In LAUSD, this percentage fell from 59 to 39 percent, from 2006–07 to 2015–16.8 8 We tested for changes in the characteristics of 6th-grade LTELs over our study period and found almost no changes in test scores and the share speaking Spanish at home. However, in LAUSD, we did find that while the number of 6th-grade special education LTEL students remained fairly constant, as a proportion of LTEL students, they more than doubled, going from 11 percent of all 6th-grade LTELs in 2006–07 to 25 percent in 2015–16. In SDUSD, both counts and percentages of special education students were stable over time. PPIC.ORG Academic Progress for English Learners 9 We observed year-to-year fluctuations in the numbers of late-arriving ELs in each district, though there was an overall decline across the study period. Notably, LAELs represented an increasing share of the EL population in both districts, especially between 2013–14 and 2015–16. In Los Angeles, the school staff we interviewed indicated that the district’s LAEL population in recent years has included many unaccompanied minors from Central America and Mexico, some of whom spoke the indigenous languages of those countries and for whom obtaining school transcripts was often difficult. In San Diego, school staff noted an increase in the district’s LAEL population from about 2012 through 2015, when there was increased federal support for refugee resettlement, yet said they had received far fewer LAELs, especially refugees from Muslim countries, since 2016. School-Level Patterns The ways in which districts serve long-term and late-arriving EL students may depend on how these students are distributed across schools. For example, if large numbers of these student groups attend just a few secondary schools in the district, then resources and support for English language development coursework may tend to focus on those particular schools, with less attention paid to schools with smaller EL populations, or vice versa. Previous research suggests that the concentration of non-English speakers in a school is a potential determinant of student outcomes for English Learners (Conger, Schwartz, and Stiefel 2011). Using propensity score matching, Callahan et al. (2009) find that greater concentrations of ELs are associated with stronger academic achievement, relative to lower EL concentrations. Bui (2013) similarly finds that larger shares of ELs in a school are associated with increased reclassification, less grade retention, and increased math scores, but not higher reading and language scores. In Table 1, we explore the variation across each district’s schools by type of EL student in grades 6–12.9 In Los Angeles, all schools have at least some ELs, and EL students make up 23 percent of the average school’s population. All Los Angeles schools have at least some long-term and late-arriving ELs as well. The share of LTELs ranges from 12 to 25 percent of students across schools, and the share of LAELs ranges from 3 to 5 percent. In San Diego, the average school’s shares of each student group are similar to those in Los Angeles, with ELs making up 23 percent of the average school population, but the ranges are much larger. The San Diego schools where ELs constitute fewer than 3 percent of students are atypical schools serving special populations. TABLE 1 Los Angeles and San Diego have similar shares of each EL student group across middle and high schools Share of LTELs (%) Share of LAELs (%) Mean Min Max Mean Min LAUSD 18.8 11.5 24.9 3.8 2.6 SDUSD 19.5 1.1 48.7 1.6 0 SOURCE: Authors’ tabulations from school district data, 2006–07 to 2015–16. NOTES: For LTELs, LAELs, and ELs, currently or in the past. Among schools in our sample. Max 5.3 22.4 Share of all ELs (%) Mean 22.9 23.0 Min 15.1 1.8 Max 32.7 71.6 Prior research finds that the particular home language spoken is associated with academic outcomes for ELs and reclassified ELs (Hill et al. 2014), as is fluency in the home language as measured at school entry (Salazar and Hayes 2010). Friesen and Krauth (2011) find that being in a school with many other Chinese-speaking ELs is beneficial for Chinese-speaking ELs but that for Punjabi-speaking ELs, higher concentrations of Punjabi speakers 9 We exclude charter schools not operated by the district. PPIC.ORG Academic Progress for English Learners 10 are associated with lower academic outcomes. An ethnographic study of adolescent Spanish-speaking EL students shows these students acquired English proficiency more quickly if they collaborated on academic tasks with bilingual Spanish speakers (Carhill-Poza 2015). The studies cited above suggest that the language composition of fellow students may have a link to the academic performance of English Learners. We hypothesize that the language composition of ELs could also be connected to the academic performance of students who were never ELs (either English-only speakers or those determined to be fluent in English when they enter the school). Thus we use three measures to examine the language spoken by students at individual schools: the percentage of all students who are English Learners, the percentage of all students who speak the same home language, and the homogeneity of languages spoken among EL students.10 How Are Course Assignments Implemented? Our focus on both long-term and late-arriving ELs—whose backgrounds and needs may be quite different—leads us to ask how these student groups progress and what sorts of “malleable” factors (i.e., factors that are controllable by district officials at least to some degree) are associated with their success. Some school districts may place these two student groups in the same courses with one another, yet their learning needs are different. Both LAUSD and SDUSD have district guidelines in place that outline differentiated English language development coursework for LTELs and LAELs, as is suggested by Olsen (2010) and Menken, Kleyn, and Chae (2012). We seek to understand whether and how LTELs and LAELs were assigned to their intended English language development courses. As noted above, ELD courses refer to both designated ELD and integrated English classes, as mandated by the California Department of Education in 2015. Overall, we find the implementation of course placement policies varied across the two districts, which was in part related to legal mandates and available staffing and resources. In both districts, correct course placement rates (i.e., those that followed district guidelines) were more than 50 percent for long-term and late-arriving ELs in recent years. Many of those not placed correctly were in ELD courses that were either too high or too low for their proficiency level based on district criteria. In interviews, school and district staff remarked on several challenges in implementing policies depending on the proportion of students who are ELs at the school, as well as the size of the school. District Policies Los Angeles Unified and San Diego Unified each have English Learner Master Plans (with the most recent editions adopted in 2012 in LAUSD and 2015 in SDUSD) that outline district policies and practices for English Learner instruction.11 ELs in secondary schools are assigned to ELD courses according to the principles outlined in the master plans, with supplemental memos published nearly every year that rearticulate course assignment criteria. We used all available data to calculate our best estimates of whether EL students were placed correctly, according to each district’s master plan and supplemental memos. Technical Appendix B describes how we employed the course placement criteria in our coding. Of course, there are times when teachers and staff exerted their discretion on placement decisions and those choices are impossible for us to code in the data. 10 We measure language homogeneity by computing a Herfindahl index, which ranges from zero to one. A value of zero means that no ELs speak the same home language and a measure of one means that all ELs speak the same home language. See Technical Appendix D for a more detailed description. 11 LAUSD’s new English Learner Master Plan was completed in spring 2018, after our last year of data collection. PPIC.ORG Academic Progress for English Learners 11 In LAUSD, secondary EL students were assigned to designated ELD courses that focus on developing English proficiency and were thus intended to be taught separately from non-EL students. Additionally, ELs who had been in the US two or more years were required to be enrolled in a grade-level-appropriate English language arts class with integrated ELD support. Prior to 2013–14, LAUSD assigned students to ELD courses based on test scores and whether or not they passed previous ELD courses. Starting in 2013–14, the district began assigning and advancing EL students to ELD courses based on the number of years spent in US schools. The district also adopted two designated ELD courses for long-term ELs in 2013–14 and mandated that ELs who had been in US schools for five or more years be placed into at least one of these courses. These two courses (Literacy and Language, and Advanced English Language Development) carried high school graduation credit (and could be counted toward a–g credits, which are required to be eligible to apply to UC and CSU); placement was based on LTEL status and reading level. SDUSD’s English Learner Master Plan also provides guidance for secondary ELs’ assignment to designated and integrated ELD coursework, and these guidelines have changed over time. From 2005–06 to 2007–08, the plan stated that secondary ELs enrolled in the district for less than one year should be placed in LAEL-specific ELD courses; all other ELs were to be placed in ELD courses based on their level of English language proficiency, as demonstrated by their scores on a standardized English language development test. From 2008–09 to 2015–16, all ELs were placed in ELD coursework based on two criteria: length of enrollment in US schools and level of English language proficiency. Our interviewees confirmed that length of enrollment tended to take precedence, and ELs were often advanced to the next designated ELD course regardless of their English proficiency level, an approach that was similar to LAUSD’s. Central office staff in San Diego indicated that this recent push to move ELs through the designated ELD sequence was more common in high schools than in middle schools, so that ELs would have quicker access to academic coursework that carried graduation credit. In 2015–16, an academic language development (ALD) course was added to SDUSD’s English Learner Master Plan (though it was piloted as early as 2012–13) as a designated ELD option for LTELs enrolled for five or more years and who were performing below grade level on standardized tests. Unlike in Los Angeles, where similar courses carried credit toward graduation, San Diego’s ALD course was an elective designed to provide supplemental support to LTELs in acquiring the academic language skills needed to access content in all of their coursework, not just English; thus concurrent enrollment in a grade-level English language arts course was also required. Course Placement Rates In Figure 2, we report correct course placement rates for current long-term and late-arriving ELs; that is, the extent to which these students were placed in appropriate ELD courses as outlined in district guidelines. In San Diego, LTELs and LAELs were correctly placed in appropriate ELD courses at similar rates, between 50 and 60 percent beginning in 2008–09. In Los Angeles, the correct course placement for LTELs hovered around 40 percent until 2013–14, when course placement rates peaked at 64 percent before declining to 50 percent the following year. LAELs in Los Angeles had higher rates of correct course placement across the period than LTELs or other ELs who were not yet LTELs (not shown), although the correct LAEL placement rate declined slightly in the final year of our data. In both districts, the sharp increases we observed in correct course placement rates (in 2008–09 in San Diego for both groups and in 2013–14 in Los Angeles for LTELs) correspond to the years in which placement criteria changed to include the number of years in US schools as the primary placement criterion, rather than test scores or English language proficiency level, as described above. We identify additional reasons for these observed changes, in LTEL placement rates in particular, in the next section. PPIC.ORG Academic Progress for English Learners 12 FIGURE 2 Long-term and late-arriving ELs are correctly placed in ELD courses at similar rates in San Diego, but there is more of a gap in Los Angeles Correct ELD course placement 90% LAUSD 80% 70% 60% 50% 40% 30% 20% 10% 0% 2007-08 2008-09 2009-10 2010-11 2011-12 2012-13 2013-14 2014-15 2015-16 Long-term EL Late-arriving EL Correct ELD course placement 90% SDUSD 80% 70% 60% 50% 40% 30% 20% 10% 0% 2007-08 2008-09 2009-10 2010-11 2011-12 2012-13 2013-14 2014-15 2015-16 SOURCE: Authors’ tabulations from school district data. NOTE: Estimates combine initial and subsequent course placement. Reclassified ELs are not included. Long-term EL Late-arriving EL Figure 3 shows the percentages of long-term EL students in each district who were placed into ELD classes at too high or too low a level, who were placed correctly, and who received no ELD course at all. We find that for the most part, more than half of LTEL students were enrolled in ELD coursework—an important exception is LTEL students in Los Angeles prior to 2013–14. But many were in course levels that were higher or lower than dictated by district policy. In the final year of our data, 5 percent of LTELs were in ELD courses that were either too high or too low in San Diego, and the same was true for 25 percent of LTELs in Los Angeles. Twenty-one percent of LTELs in Los Angeles and 39 percent of LTELs in San Diego were in no ELD course in 2015–16. Earlier in the study period, over half of long-term ELs in Los Angeles received no ELD courses, and the rate in San Diego was only about a third. PPIC.ORG Academic Progress for English Learners 13 We also observed a few interesting patterns for LTELs’ ELD course placement rates by grade level (not shown), particularly in San Diego. In SDUSD, 94 percent of 6th-grade LTELs were enrolled in ELD coursework in 2015–16, whereas more than 90 percent of 7th and 8th graders were not. In grades 9–12, over 80 percent of LTELs in San Diego received some form of ELD instruction. FIGURE 3 The share of long-term ELs placed into ELD courses that were too high or too low has risen in Los Angeles in recent years, but it has stayed steady in San Diego Course placement rates LAUSD Long-term ELs 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 2007-08 2008-09 2009-10 2010-11 2011-12 2012-13 2013-14 2014-15 2015-16 Correct Too High Too Low No ELD 90% SDUSD Long-term ELs 80% Course placement rates 70% 60% 50% 40% 30% 20% 10% 0% 2007-08 2008-09 2009-10 2010-11 2011-12 2012-13 2013-14 2014-15 SOURCE: Authors’ tabulations from school district data. NOTE: Estimates combine initial and subsequent course placement. Reclassified ELs are not included. 2015-16 Correct Too High Too Low No ELD PPIC.ORG Academic Progress for English Learners 14 Roughly 10 percent of late-arriving ELs in both districts received no ELD coursework in 2015–16 (Figure 4). In Los Angeles, correct placement rates for LAELs were relatively high over the entire period, but rates have been declining somewhat in recent years. In 2015–16, about 15 percent of LAELs were in courses that were either a higher or lower level than suggested by LAUSD policy. In San Diego, LAEL students were often placed in courses that were too high (33% in 2015–16), based on district guidelines. Looking at course placement for LAELs by grade level (not shown), we found that there was low correct course placement for LAELs in grades 7 and 8 in San Diego (similar to LTELs in the district). In Los Angeles, high school LAELs were less likely to be placed in ELD courses than LAELs at the middle school level, and that was especially true for 12th graders (29% of whom were not in ELD). FIGURE 4 Late-arriving ELs in San Diego were more likely to be placed into ELD courses that were too high, compared to in Los Angeles Course placement rates 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% LAUSD Late-arriving ELs 2007-08 2008-09 2009-10 2010-11 2011-12 2012-13 2013-14 2014-15 2015-16 Correct Too High Too Low No ELD Course placement rates 90% SDUSD Late-arriving ELs 80% 70% 60% 50% 40% 30% 20% 10% 0% 2007-08 2008-09 2009-10 2010-11 2011-12 2012-13 2013-14 2014-15 2015-16 SOURCE: Authors’ tabulations from school district data. NOTE: Estimates combine initial and subsequent course placement. Reclassified ELs are not included. Correct Too High Too Low No ELD PPIC.ORG Academic Progress for English Learners 15 District Differences in Course Placement Implementation Our interviews suggest that the differential course placement rates we observed between the two districts for LTELs and LAELs may be attributed, at least in part, to differences in district policy, staffing, and resources that shaped the implementation of course placement guidelines. It is important to note, however, that our interviews took place in 2017–18, whereas our quantitative data end with the 2015–16 school year. Though some interviewees described prior implementation processes, most discussed changes to course placement implementation over the last few years; thus, our interview data are limited in the extent that we can explain changes in course placement trends over time. Generally, our findings suggest that, whereas LAUSD’s course placement processes have become more centralized and streamlined over the last few years, with additional resources and staffing put in place to support implementation, SDUSD has afforded schools more flexibility with course placement implementation in the context of declining overall resources, and thus there has been more limited EL support staff in the district. One example of this difference between districts relates to the implementation of designated ELD courses for LTELs. Insights from Los Angeles Unified School District In 2013–14, we observed an increase of almost 20 percentage points in correct placement for long-term ELs in Los Angeles (and a corresponding 20 percentage point decrease in the number of LTELs receiving no ELD; see Figure 3). That same year, LAUSD began implementing a voluntary agreement with the US Department of Education’s Office for Civil Rights (OCR) to improve ELs’ access to ELD courses. Prior to the agreement, designated ELD courses for LTELs were not mandatory, as noted by a district administrator: “They were called EL Language Strategy classes, and they were optional. So, if a school chose to include them on the master [schedule] as an elective for ELs to take, that was great. But it was up to the site.” Given that these designated ELD courses were not yet mandatory, LTELs tended to receive only integrated ELD. “Prior to LTEL courses, students . . . finished [designated] ELD 4 and classified as Preparing to Predesignate (PRP), and continued to get support through sheltered instruction in their content classes [i.e., integrated ELD],” said a district administrator. As a result of the OCR agreement, however, the district developed two designated ELD courses for LTELs and mandated their implementation starting in 2013–14. “The OCR did a review and the district came to an agreement to provide additional support with [designated ELD] LTEL classes. Now, LTEL courses are mandatory,” noted a district administrator. In addition to augmenting the district’s designated ELD course offerings, several interviewees in LAUSD noted that, in recent years, ELD course placement was heavily monitored by district-level staff. The district recently implemented a new data management system and EL Dashboard to help with course placement monitoring, identifying whether ELs were placed in the appropriate courses and their progress toward reclassification. Some interviewees described the usefulness of the new data system and dashboard in streamlining the placement process for all ELs. Further, district-level positions, such as data specialists and coordinators, were created to facilitate communication with and support to schools related to course placement. Communication tools included data reports with information related to ELs who were either not enrolled in ELD or were incorrectly placed, as noted by a high school principal: “The district does a pretty good job of forcing you to comply. [We get] emails saying that you’re out of compliance, that there are students who are not taking their [required] courses because sometimes a couple of kids slipped through the cracks.” Although these LTEL courses were described as mandatory, and resources and staff were put in place to support course placement, our quantitative analysis showed that about 21 percent of LTELs in LAUSD were not enrolled in any ELD courses as of 2015–16, the last year of our data (see Figure 3). Indeed, exceptions to the district’s PPIC.ORG Academic Progress for English Learners 16 LTEL policy were described by a central office staff member, who suggested that LTELs in the latter years of high school might be placed in academic content courses instead of designated ELD.12 This staff member noted, “At the end of the day, if I have an 11th grader who has still to finish the a–g [high school graduation] requirements, I’ve got to take them out of that LTEL class. And I’ve got to give him . . . his graduation requirements. I will make the annotation to the district. . . . I’m sorry, but the a–g weighs more than the LTEL class, and that I got from the district as well as from training.” Although our data are limited to a small sample of schools, we did not hear from any interviewees in LAUSD that designated ELD was not offered when it should have been, according to district guidelines. Interviewees from SDUSD did indicate as much, as discussed below. Instead, Los Angeles schools with small EL populations (i.e., where ELs represented between 4 and 11 percent of the student body) tended to rely on double or multiple rostering in designated ELD courses. An EL coordinator described this process: “When we have classes where there are not enough ELs, especially in schools with small EL populations, those courses are usually double rostered. So, ELD 1, 2, 3, and 4 might be together [in the same class] because they may have only 18 ELs at the school.” Even though the courses may have been rostered in this way, student transcripts still reflected the appropriate ELD level, thus indicating correct placement. Insights from San Diego Unified School District In San Diego, correct course placement rates for LTELs remained relatively constant between 2008–09 and 2015–16, with as many as 39 percent of LTELs enrolled in no ELD courses (see Figure 3). Although central office staff indicated that the district developed a designated ELD course (the academic language development course, or ALD) for LTELs as early as 2012–13, it was an elective course that did not carry a–g credit, and its implementation was not mandatory. As such, when LTELs in San Diego completed the district’s designated ELD course sequence, they were meant to be placed in mainstream credit-bearing English classes with integrated ELD supports. A central office staff member described LTEL course placement in the district: “The majority of our [ELs], when they finish our ELD sequence, are not reclassified, so we have all of our LTELs who aren’t reclassified in mainstream [English] classes. They’re not in our ELD sequence, and they may or may not have access to an ALD class.” Of the San Diego schools in our interview sample, only one offered an ALD course in 2017–18. School leaders at our eight other interview sites opted not to offer the course, either due to budgetary constraints and/or because they did not have enough LTELs to offer a course. Three principals stated that they had eliminated ALD at the beginning of the 2017–18 school year as a result of districtwide budget cuts, including this middle school principal: “We got a lot of funds cut from the school, so there’s no way I could offer a class for seven [LTEL students].” Overall, these findings suggest that integrated ELD has been the primary instructional approach for LTELs in San Diego. In contrast with Los Angeles, where the district undertook efforts to centralize data management and streamline support across schools, San Diego school sites have been given more flexibility, at least in recent years, with respect to ELD instruction. A district leader described school flexibility in this way: “We’re giving schools autonomy; there’s no fixed route. We became more flexible in terms of how schools provide support. ELD 1/2 remained double blocks, but 3/4 are single periods. We added an ELD literacy block class for levels 1/2, 3/4, and 5/6, so if schools want their level 1s and 2s to have three periods of ELD, they can, or 3s and 4s can have two periods. But, if schools believe their population can take it, they don’t have to do that. We also added advanced coursework/pathways, like AP Spanish Language and Literature, so some students can go onto Honors if they’re ready. We’re encouraging schools to really look at their populations and work with us to determine their pathways.” This school-level 12 Although we allowed for such placements to be considered “correct” in our quantitative analyses. PPIC.ORG Academic Progress for English Learners 17 autonomy was evident with respect to the offering of ALD courses for LTELs described above, though principals indicated making decisions based on resources, not necessarily on the needs of their LTEL populations. Our qualitative findings also suggest that, in the context of budgetary constraints, SDUSD schools may have been given even more control over course placement in the last few years. Amid declining resources, district-level EL support staff positions were reconfigured or cut altogether, and staff in several of the schools in our interview sample noted no longer receiving direct support from the district related to EL course placement. This trend was especially evident among schools that had small EL populations; schools with larger EL populations either still received some level of district support (two schools) or used site resources to maintain the level of support they received previously (one school). Relatedly, we found that the four schools in our sample that served small EL populations could not afford to continue offering designated ELD classes and opted not to do so, as one middle school EL coordinator noted: “We used to have ELD classes when we had more kids . . . then we didn’t have enough ELs to really justify it.” Similarly, a high school counselor shared, “There are no options [for designated ELD], so they just go into the mainstream classes just like everybody else.” While these mainstream courses presumably offered ELs integrated ELD, these findings suggest that designated ELD was not offered to either LAELs or LTELs in some San Diego schools. Nonetheless, some schools still attempted to offer alternative linguistic supports in the absence of designated ELD courses. For example, in one San Diego high school serving a small EL population, both LAELs and LTELs were placed in a “Spanish for Spanish speakers” course sequence, which included three levels and was followed by AP Spanish. The school principal described this practice as an attempt to address ELs’ linguistic needs in the absence of a designated ELD option at the school. Still, the teachers who taught these classes recognized that they were intended to facilitate fluency in Spanish, not English, as well as serve as the foreign language requirement for graduation. The practice of double or multiple rostering in LAUSD may help to explain why course placement rates for LAELs in Los Angeles were overall higher than in San Diego, especially if San Diego schools were not offering designated ELD. Nonetheless, even though interviewees in San Diego did not discuss double rostering, it may have been used in schools outside of our sample. What predicts a school’s level of correct ELD course placement? Our interviews with district teachers and staff in both LAUSD and SDUSD suggest that correct course placement is made more challenging when there are fewer EL students at a school. Here we report on whether those findings are consistent with our student-level data. In each school district, we examine the relationship between each of our language environment variables (the percentage of ELs in the school, the percentage speaking the same home language among all students, and the degree of language homogeneity among EL students). We also examine the relationship between correct course placement at the school level and the number of EL students per grade level. In LAUSD, we find that as the number of EL students per grade increases, correct placement rates increase as well, but the number of EL students per grade is not a statistically significant predictor in SDUSD (see Table D3 in the technical appendices). Somewhat surprisingly, in both districts, as the percentage of all students speaking the same home language rises, correct course placement falls. At least in the case of LAUSD, the number of ELs per grade is more important given both the size of the coefficient and the range of the variable. For example, an increase of one standard deviation in the number of ELs per grade is associated with a 4 percentage point increase in correct course placement, while an increase of one standard deviation in the percentage speaking the same home language among all students results in a 1 percentage point decrease in correct course placement. PPIC.ORG Academic Progress for English Learners 18 How Do Rates of Progress Differ for Long-Term and LateArriving English Learners? Prior research suggests that EL students’ progress may depend in part on their peers as well as on instruction. Before testing for the association of peer groups and course assignment with linguistic and academic outcomes, we examine performance of LAELs and LTELs across the two districts below. Long-term ELs are, by definition, students who have had a great deal of exposure to English language development instruction and to American schooling. Late-arriving ELs, in contrast, are students with no exposure to American schooling prior to their first year in LAUSD or SDUSD, and they may have had limited or varied amounts of schooling in their country of origin. Note that we further restrict our focus to late-arriving ELs who have very limited English proficiency (scoring at the lowest level on California’s test of English language proficiency). On average, long-term EL students have higher test scores than LAEL students. But, late-arriving EL students appear to make more progress on average in improving their test scores on English language proficiency tests and California Standards Tests on both English language arts and math from one year to the next, and in closing the gap with long-term EL students. English Proficiency Unsurprisingly, we find that LTELs have substantially higher scores on the California English Language Development Test (CELDT) than LAELs (Figure 5). The CELDT is a state-mandated English proficiency assessment of listening, speaking, reading, and writing administered annually to all current ELs statewide as well as to all new enrollees who speak a language other than English at home, in kindergarten through 12th grade.13 But if students who report a home language other than English do well on the test they are deemed Initially Fluent English Proficient (IFEP), do not receive ELD supports, and are not tested again.14 Overall scale scores across grades 6–12 for the CELDT for late-arriving ELs in San Diego (455) are somewhat higher than the overall score for LAELs in Los Angeles (447). These scale scores are translated into overall proficiency levels, which are in turn used to determine when EL students are ready for reclassification.15 These average overall scores mean that most LAELs are scoring at the beginning proficiency level in both districts.16 The average long-term EL score in LAUSD (567) places the average LTEL in the intermediate or early advanced level, depending on the grade level. The average score of 556 for LTELs in SDUSD places the average student squarely in the intermediate level, across all grades 6–12. 13 The CELDT identifies students with limited English proficiency, determines their levels of proficiency, and assesses progress in learning English. It must be administered within 30 days of enrollment 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 been reclassified as Fluent English Proficient. CELDT results provide performance levels—beginning, early intermediate, intermediate, early advanced, and advanced—for each subtest and an overall proficiency level (OPL). Until 2006, the listening and speaking subtests were combined. Starting in 2006–07, higher CELDT scores were required for each OPL. 14 Results from the CELDT given at a student’s entry determine if a student who speaks a language other than English at home will be classified as an English Learner or a student that is initially fluent in English. The CELDT was replaced by the English Language Proficiency Assessments for California (ELPAC) in 2017–18. 15 Scale scores that determine proficiency levels vary somewhat by grade level, with high scores required as students advance through the grades. See CELDT Initial/Annual Score Ranges. 16 Using the 2016–17 and 2017–18 CELDT guide, OPL scores for 9th graders (CDE 2016). PPIC.ORG Academic Progress for English Learners 19 FIGURE 5 In both districts, average CELDT scores are higher for long-term ELs than for late-arriving ELs 600 Mean CELDT scores 500 400 300 200 100 0 LAUSD SOURCE: Authors’ tabulations from school district data, 2006–07 to 2015–16. NOTE: Reclassified long-term and late-arriving ELs are included. SDUSD Long-term EL Late-arriving EL The gap between the average CELDT scores for long-term and late-arriving ELs is over 100 points at both districts. Table 2 illustrates that late-arriving ELs tend to close the gap with long-term EL students over time. LTELs gain about 20 points a year on the CELDT in both school districts, but LAELs gain substantially more (59 points in SDUSD and 38 points in LAUSD). A gain of 20 points is about one-third of the gain necessary to go from one proficiency level to the next, whereas a gain of 59 points is the approximate gain required to advance one proficiency level (CDE 2016–17; CDE 2016). Because LTELs start at relatively higher levels, it is likely more difficult for them to make the same annual gains observed among LAELs. TABLE 2 Late-arriving ELs have larger annual gains in English language proficiency than long-term ELs Average annual point growth in CELDT scores SDUSD Long-term ELs 19 Late-arriving ELs 59 SOURCE: Authors’ tabulations from school district data, 2006–07 to 2015–16. NOTE: Reclassified long-term and late-arriving ELs are included. LAUSD 21 38 In line with our quantitative findings, interviewees in both districts suggested that long-term ELs often struggle to move past intermediate or early advanced proficiency levels, despite having higher average scores. Many attributed these struggles to LTELs’ lack of motivation, given their limited success in passing the exam previously: “We have students who don’t take the CELDT seriously because they have done it for the last 10 years and they’re like, ‘Oh, here we go again, I’m not English proficient, but I don’t want to take this assessment,’” noted a SDUSD high school principal. Several interviewees expressed their view that some LTELs appeared to lack motivation in their classes and did not see the importance of improving their English proficiency PPIC.ORG Academic Progress for English Learners 20 in order to reclassify because they felt they already spoke English: “I noticed some students, they get bored with taking the same class if they haven’t reclassified and they feel as if they don’t need it because they’re already speaking English fluently,” said a LAUSD school counselor. Academic Performance in English Language Arts and Math Nearly all California school children are assessed annually on their mastery of English language arts (ELA) and math. Given that these assessments are administered in English, for EL students they also implicitly measure English proficiency. In 2012–13 and in previous years, the state used the California Standards Test (CST) as its standardized assessment; since 2014–15, the state has administered the California Assessment of Student Performance and Progress (CAASPP, sometimes referred to as the Smarter Balanced assessment).17 We begin by examining performance using the CST’s English language arts test from 2006–07 to 2012–13.18 We then examine scores on the CAASPP in English language arts and math from 2014–15 to 2015–16. Our analysis of the CST for English language arts (Figure 6) shows that LTEL students have higher scores than LAEL students, but both English Learner groups score substantially below students who were never ELs. This figure transforms the five proficiency levels of the CST (far below basic, below basic, basic, proficient, and advanced) into a scale from 1 to 5 and takes the average. Figure 6 also shows lower scores for never ELs in Los Angeles relative to in San Diego. Average proficiency levels for long-term ELs are similar between the two districts, therefore, the performance gap between never ELs and long-term ELs is larger in San Diego (1.33 performance levels) than the gap in Los Angeles (0.85 performance levels). The proficiency levels for late-arriving ELs are lower in Los Angeles than in San Diego; the gap between LAEL and LTEL students is three-quarters of a performance level in Los Angeles, whereas it is only half a performance level at San Diego. FIGURE 6 Both long-term and late-arriving EL students score below never ELs in English language arts 4.5 Mean CST ELA Performance Level 4.0 Never EL 3.5 Long-term EL 3.0 Late-arriving EL 2.5 2.0 1.5 1.0 0.5 0.0 CST ELA CST ELA LAUSD SDUSD SOURCE: Authors’ tabulations from school district data , 2006–07 to 2012–13. NOTES: Reclassified long-term and late-arriving ELs are included. We convert raw CST scores into proficiency levels and average them. 17 During 2013–14, the state did not administer a test of basic skills. The new CAASPP is linked to California’s new Common Core State Standards. 18 We do not examine CST math scores because students took math subject tests (i.e., algebra, geometry) rather than grade-level tests, making comparisons by grade level or age, at least above grade 7, very complicated. PPIC.ORG Academic Progress for English Learners 21 To show changes in achievement across grades, we convert the raw CST test scores into a standardized score (known as a “z-score”) that equals 0 for students at the state mean, 1 for students one standard deviation above the state mean, and so on. Figure 7 shows the mean of these standardized scores by grade level. As with the CELDT, it appears that LAEL students make more progress in test scores during their time in the school district than do LTELs. Although this is a snapshot from 2012–13 and does not follow individual students over time, this figure nonetheless hints at the relative progress of students in the various language groups as they grow older. Never ELs make small (or no gains) relative to the mean performance of all students statewide from one grade level to the next. In Los Angeles, LAEL students narrow the performance gap with LTEL students across grades. In San Diego, we also see LAELs catching up to LTELs, but the gap varies more grade by grade than in Los Angeles. FIGURE 7 Late-arriving ELs narrow the performance gap with long-term ELs in English language arts between grades 6 and 11 0.50 LAUSD Mean CST ELA z-score 0.00 6 7 8 9 10 11 -0.50 -1.00 -1.50 Never EL Long-term EL Late-arriving EL -2.00 Grade SDUSD 1.00 0.50 Mean CST ELA z-score 0.00 6 7 8 9 10 11 -0.50 -1.00 Never EL Long-term EL Late-arriving EL -1.50 -2.00 Grade SOURCE: Authors’ tabulations from school district data. NOTE: These test scores are taken from 2012–13, the last year of the CST. We standardize the test scores across the two districts by using state means and standard deviations for the CST ELA by year. Thus a score of 0 means that students are at the state average, while a score of 0.5 would mean that students are half a standard deviation above the state mean in a given grade. The lines are simulated cohorts of students rather than students being followed over time. EL students remain in their original language status group even after reclassification. Reclassified long-term and late-arriving ELs are included. In SDUSD, in 2012–13, almost no late-arriving ELs took the CST. PPIC.ORG Academic Progress for English Learners 22 Prior research (e.g., Salazar and Hayes 2010) has pointed to the importance of initial proficiency in students’ own home languages in determining what progress they will make in learning English. Upon entering school, students take a language proficiency test to gauge their literacy in their own home language, the results of which we distill into “fluent” or “not fluent.” In Figure 8 we present a comparison of individual students’ year-to-year growth in CST ELA scores, but this time distinguishing by home-language fluency level for LTELs and LAELs. Among students who are fluent in their home language, average growth in CST scores is greater compared to students who are not fluent in their home language upon entering the school. However, the difference is only substantial in SDUSD. Among LTELs in San Diego, students who were initially fluent in their home language have faster gains on the test of English language arts, but in Los Angeles, initially fluent and non-fluent students have the same growth. Among LAELs, initial fluency in their home language is much more important for growth in San Diego than in Los Angeles. We should note, however, that the approaches to assessing home language fluency differ in each district. FIGURE 8 EL status is a more consistent factor in growth in ELA scores than home language fluency, particularly in Los Angeles 0.20 CST-ELA, with home language fluency 0.15 LAUSD SDUSD 0.10 Growth in CST-ELA z-score 0.05 0.00 -0.05 Never EL Fluent Not Fluent Long-term EL SOURCE: Authors’ tabulations from school district data, 2006–07 to 2012–13. NOTE: Reclassified long-term and late-arriving ELs are included. Fluent Not Fluent Late-arriving EL Results on the California Assessment of Student Performance and Progress show a similar pattern for English language arts. The CAASPP has only been administered for four years, and our data include two of those years (2014–15 and 2015–16 school years). As with the CST, we find that never ELs have higher ELA scores than LTELs, who in turn have higher scores than LAELs. In our simulated cohorts of students, where we compare scores by grade level at one point in time, all language groups have higher scores at higher grade levels, and the simulated gains are greatest for LAELs.19 Because the CAASPP administers math tests by grade level and not subject, we can make comparisons across all the grade levels administered. These math scores reveal a slightly different pattern. In Los Angeles, LAELs have 19 The CAASPP is administered in grades 3–8 and grade 11. PPIC.ORG Academic Progress for English Learners 23 somewhat higher CAASPP scores than do LTEL students—but both sets of scores are substantially lower than those for never ELs. In San Diego, we see the same pattern in grades 6 and 7 but in grades 8 and 11—the two other secondary grades with CAASPP math testing—LTELs have slightly higher math performance than LAELs. Once again, both types of EL students had CAASPP math scores considerably below those of never EL students. Grade Point Average There are many other important academic outcomes to consider beyond standardized tests, and here we examine grade point average (GPA) for students in grades 9–12 (Figure 9). In both school districts, late-arriving ELs have higher GPAs than do long-term ELs, and both EL student groups have lower GPAs than never ELs. It is important to note that these two districts do not have standardized policies on grading for ELs. FIGURE 9 Late-arriving ELs have higher high school GPAs than long-term ELs GPA 3.0 2.5 2.0 Never EL Long-term EL Late-arriving EL 1.5 1.0 0.5 0.0 LAUSD SDUSD SOURCE: Authors’ tabulations from school district data, 2006–07 to 2015–16. NOTE: Reclassified long-term and late-arriving ELs are included. Insights from Interviews In both districts, but especially in San Diego, the school staff we interviewed described an increasing focus on graduation rates and ensuring that ELs were enrolled in credit-bearing courses. As such, both LAELs and LTELs were sometimes placed in courses that exceeded their English abilities and often received low grades, such as Ds and Fs. Indeed, some SDUSD teachers expressed a desire to be able to offer “no grade” in lieu of an F, especially for LAELs, because a failing grade did not reflect the gains they were making. Despite LAELs struggling in their coursework, staff indicated that, especially in recent years, ELs were rarely allowed to repeat designated ELD courses and instead were advanced to the next course in the sequence so they could more quickly move into credit-bearing classes. By SDUSD policy, any student failing a high school course is advanced to the next course and makes up the failed course in summer school or through the credit-recovery program. In Los Angeles, staff noted that LTELs were required to repeat LTEL-specific designated ELD courses until they passed them, with the exception of 11th graders who were waived from this policy so they could complete graduation course requirements. This practice of repeating courses may have resulted in student disengagement PPIC.ORG Academic Progress for English Learners 24 and underperformance, as a majority of LAUSD interviewees, and a modest proportion of SDUSD school staff, described LTELs’ boredom, lack of motivation, or poor behavior, especially when repeating designated ELD courses: “The biggest challenge has been behavior management, simply because LTELs have been in these type of classes for a while, and then they think, ‘Well, I'm not going to pass [the CELDT] again’ and then they just don’t try or give their best,” said an LAUSD middle school teacher. In addition, many school staff attributed LTELs’ lack of motivation and underperformance to the ELD curriculum for LTELs, which they described as repetitive and not engaging or interactive. In fact, California recently enacted two new bills that aim to help English Learner students gain more access to the core academic curriculum. The first, AB 2735, requires that school districts allow English Learners full access to all of a school’s standard instructional program, including a–g courses as well as AP classes. It does grant districts an exception for newcomer students. The second, AB 2121, requires districts to permit a fifth year of high school for newcomers and students in the Migrant Education Program in order to provide more opportunities to earn required graduation credits.20 Credit recovery through summer instruction (Johnson 2018) or online learning may provide another option for English Learner students to make up for course periods taken up by ELD. Are School Language Environments Related to EL Student Progress? Below, we examine student progress in more detail, with an eye toward understanding which factors are predictive of student success. First we study whether the language background of peers at the school level is correlated with academic outcomes for English Learners, and if those correlations vary by whether the student is a long-term or a late-arriving EL. Taken as a whole, we did not find consistent relationships between the language environment in the school and various academic outcomes. Certainly there were exceptions, but for the most part we found that the academic progress of English Learners seems quite adaptable to changes in the language environment of the school. In a similar vein, an important finding of this analysis concerns students who were never English Learners. The percentage of the student body that is EL bears no significant negative relation to any of the outcomes we modeled for this group (English and mathematics performance on the former state test), and in the case of LAUSD, it was positively associated with English. Overview of Our Approach We pose three main questions regarding each school’s language environment:  Is academic progress correlated with the percentage of students who are EL?  Is academic progress correlated with the percentage of all students who speak the same home language as the student in question?  Is academic progress correlated with the degree of homogeneity of languages spoken by ELs? 20 The Migrant Education Program is a federal program intended to facilitate keeping the children of migrant workers continuously enrolled and progressing in US schools, despite frequent moves. PPIC.ORG Academic Progress for English Learners 25 In this section and the subsequent section on course placement and ELD programs, we measure a variety of academic outcomes in order to understand what is correlated with LTEL and LAEL student success. Our analyses are not causal, but we do use linear regression models to take many factors into account at the same time. We study whether there is a relationship between the above language and ELD course measures and six outcomes. Three outcomes compare changes in annual test scores in math and reading (the California Standards Test) and GPA between EL students and never ELs.21 For ELs alone, we also study their progress on the state’s annual test of English fluency, the California English Language Development Test. The final two outcomes measure events that can only happen once in a student’s career, if they happen at all: reclassification and on-time graduation. The California Standards Test scores are available in math and English language arts for students in grades 2 through 11 from spring 2002 through spring 2013. By converting all CST scores into z-scores using statewide means and standard deviations, these variables compare student performance with a statewide norm. When modeling changes in z-scores for individual students, we are measuring their movement in the statewide distribution of achievement. We use ELA scores for grades 6–11 and math scores for grades 6 and 7.22 Reclassification of EL students in our sample can happen anytime between 6th and 12th grade. Graduating on time is estimated for 12th graders.23 Those who left the district before grade 12 (and in the case of San Diego were not known dropouts) were excluded as we lacked information on these former students’ graduation or reclassification statuses after they left the host districts. For each outcome we model, we take into account student demographic characteristics (e.g., gender, parental education, special education status, and home language) and the student’s performance on the same measure in the prior year. The results of the regression analyses included here and in the next section on ELD course placement and programs are from models that also allow for district-wide nonlinear trends in student performance and that interact grade and school indicators. This means that any correlation we find between a language environment or EL course measure and student progress has already taken into account the student’s past performance and demographic measures, district-wide trends, and differences between each combination of school and grade level that are assumed not to change over time. Technical Appendix D explains our regression analyses in greater detail. The language environment could affect the learning environment for different EL types in varied ways. Further, prior research has suggested that an increased percentage of ELs who speak languages other than English in the school could have negative impacts on students for whom English is their first language or for reclassified students. For both of these reasons, we allow the correlation of these “language environment” variables with a given measure of academic progress to vary between never ELs, LAELs and LTELs. While we do not present results here, we tested each language variable independently before including them as explanatory variables simultaneously.24 Annual Outcomes Our results for the school language environment suggest a few points of agreement across the two school districts:  The percentage of ELs at a school has no negative statistical link to academic outcomes for never EL students. 21 In SDUSD we measure GPA from grades 6 through 12, while in LAUSD we model GPA for grades 9 through 12. 22 Starting in 8th grade, students take CST math tests specific to the courses in which they are enrolled, making their CST math scores impossible to compare. 23 In SDUSD, when we predict graduation, we are able to include students who have dropped out. These dropout data are not available in LAUSD. 24 Language variables that were not significant independently but were significant when modeled simultaneously are highlighted in gray in Tables D4A and D4B the technical appendices. PPIC.ORG Academic Progress for English Learners 26  As the percentage of students in the school who speak the same home language as the student in question rises, never ELs have more positive growth in test scores and GPA.  Greater language homogeneity among ELs has little statistical relationship with academic outcomes, and there are no consistencies between the two school districts on this measure. The figures below illustrate these points of agreement and other findings that are more idiosyncratic to the individual districts. Each chart shows the predicted effect of changing from the 25th percentile to the 75th percentile for each language environment variable. In Figures 10–15, the charts only display statistically significant results, meaning no vertical bars are shown for factors that were not significant predictors for a given group of students. If the effect was statistically significant, the height of the bar shows the size of the predicted effect. The vertical axes are measured in standard deviations. The exception is GPA; in both districts the standard deviation for GPA was almost 1, so we did not convert the familiar GPA variable into standard-deviation units. Thus, a bar for GPA that is 1 unit high implies that a change from the 25th percentile to the 75th percentile in the given variable is predicted to increase GPA by a full grade point. The Percentage of English Learners Figure 10 shows the predicted effects of the percentage of students at the school who are EL on tests of ELA, math, and English proficiency, as well as GPA. It is worth studying the link between academic progress and the percentage of students who are EL because for ELs there are countervailing factors. For example, it might help ELs to have the percentage of ELs rise somewhat if it means the school reaches a critical mass of ELs and it can provide ELD classes specific to each EL’s needs. But a higher percentage of ELs could potentially make it easier for ELs to converse in a language other than English outside of class. The leftmost third of Figure 10 shows the percentage of ELs at the school is not significantly related to three of the academic outcomes for never EL students, and is positively related to CST ELA scores in Los Angeles. This is an important finding because it means that both districts appear to manage instruction such that having more English Learners in a school does not negatively influence learning outcomes for native English speakers and others who spoke English fluently when they first entered the districts.25 Turning to the middle part of the figure (LTELs), the correlation of the percentage of ELs with LTELs’ performance varies by outcome and district. In Los Angeles both math and ELA CST test scores for LTELs are positively related to the percentage of ELs in the school and in San Diego CELDT is negatively related. Finally turning to the rightmost part of the figure, LAELs in Los Angeles appear to improve their performance on the ELA test as the percentage of ELs rises. This is not the case in San Diego, where there is no such positive association and where performance on the CELDT falls. For both ELD course placement factors and language environment factors, an increase in the percentage of students who are EL could in theory either hinder or boost rates of learning in a school. We tested for non-linear effects of each of these variables, but we did not find any convincing evidence in favor of this idea. Technical Appendix D provides additional details. 25 A non-statistically-significant finding could mean that the variable really does not matter, or it could mean that the effect is estimated imprecisely. In the case of the percentage of ELs and the insignificant relation to outcomes for never ELs, the former seems like the more likely explanation. The estimates are very precise. Technical Appendix D explains this in more detail. PPIC.ORG Academic Progress for English Learners 27 Predicted effect of a change from 25th to 75th percentile in share EL at school FIGURE 10 Higher shares of ELs in a school are not negatively associated with academic outcomes for never ELs 0.10 0.05 0.00 -0.05 CST ELA Z-Score CELDT Score CST Math Z-Score Annual GPA -0.10 -0.15 -0.20 LAUSD SDUSD LAUSD SDUSD LAUSD SDUSD Never EL SOURCE: Author calculations. Long-term EL Late-arriving EL NOTES: Chart shows the predicted effect of a change from the 25th percentile to the 75th percentile of ELs at the school level. Only statistically significant regression results (p<.05 or greater) are displayed. Reclassified long-term and late-arriving ELs are included. Full results found in Table D4A in the technical appendices. Interviews suggested that in both districts, schools serving larger EL populations faced particular challenges related to ensuring progress for late-arriving ELs. These challenges may be heightened in SDUSD, given that San Diego welcomes more refugees than any other county in the state, and the number of refugees entering the district has increased in recent years. However, in our quantitative analysis we found a negative association with higher percentages of ELs and outcomes for LAELs in only one case out of eight: CELDT scores for LAELs in San Diego. This negative association may be related to ELD teachers in San Diego feeling unequipped to provide the supports necessary to address LAELs’ extensive and varied academic and socioemotional needs, which they attributed to students experiencing interrupted formal education, trauma in their home countries or during their immigration journeys, or “culture shock,” as one ELD teacher described it. Several teachers characterized such needs as “desperate” and not “normal” compared to the LAELs they served in the past, and felt that these issues presented significant barriers to students’ adjustment, as well as challenges to their teaching. Although the topic of course placement was not explicitly referenced in this context, interviewees at schools with large EL populations in both districts advocated for newcomers, especially those with limited formal education, to have a separate course dedicated to their academic and social transition. Without these kinds of additional supports (which were sometimes offered, but have recently been eliminated in many schools in SDUSD due to lack of funding), it was challenging for students and teachers to achieve desired learning outcomes. The Percentage of All Students Speaking the Same Home Language The second variable of interest, the percentage of all students at the school who speak the same language at home as the student in question, is worth studying because a greater percentage of the student body speaking one’s own language may be an indicator of ELs’ ability to speak their home language even outside the ELD classroom, such as in other courses with never ELs or reclassified students. At the same time, it could allow the school to design ELD classes targeted to that student’s language group or English proficiency level. PPIC.ORG Academic Progress for English Learners 28 Figure 11 shows the relation between this variable and our four annual outcomes for the three groups of students. In cases where the percentage of all students speaking the same home language as the student in question matters, the association is almost always positive. For our EL student groups, having more students speaking their same home language does not appear to slow English acquisition. For never ELs, the three relevant outcomes (CST English language arts and math, and GPA) show a small positive relationship with the percentage of all students at the school who speak the same language at home as the student in question. Future analyses may test for whether this finding is positive just for English speakers among never ELs or if outcomes are positive even for never ELs who speak languages other than English at home. In San Diego LAELs had better performance on the three tests, but not GPA, when the percentage of all students speaking the same home language was higher, which argues against the existence of perverse effects of language enclaves for these students. Simulated change from 25th to 75th percentile in % speaking same home language FIGURE 11 Higher shares of students speaking students’ home language do not have a consistent negative association with test scores and GPA 0.37 0.32 0.27 0.22 0.17 0.12 0.07 0.02 -0.03 -0.08 CST ELA Z-Score CELDT Score CST Math Z-Score Annual GPA LAUSD SDUSD LAUSD SDUSD LAUSD SDUSD Never EL Long-term EL Late-arriving EL SOURCE: Author calculations. NOTES: Chart shows the predicted effect of a change from the 25th percentile to the 75th percentile in percent speaking the same home language. Only statistically significant regression results (p<.05 or greater) are displayed. Reclassified long-term and late-arriving ELs are included. Full results found in Table D4A in the technical appendices. We also asked school site staff about language enclave possibilities. Across the two districts, interviewees had varying beliefs. In particular, several teachers in SDUSD commented that their non-Spanish-speaking students, especially those from Asian and African backgrounds, were progressing much faster than their Spanish-speaking students because employing their home language to communicate within the classroom or larger community was not often possible, as it was for students whose home language was Spanish. At the same time, there was some acknowledgement among teachers in both districts that students who spoke the same home language could serve as language models and supports for each other, so language enclaves could also be beneficial to students’ language development. PPIC.ORG Academic Progress for English Learners 29 Language Homogeneity Among ELs Finally, we study the relation between language homogeneity among ELs and annual student outcomes to test the possibility that schools may be able to improve English fluency for ELs more quickly if the ELs at the school are linguistically similar. But higher homogeneity may also create an “enclave” effect within which English Learners feel quite comfortable conducting most of their school day in a language other than English. The measure of language homogeneity that we use, known as a Herfindahl index, can be interpreted as the probability that any two ELs picked randomly at a given school would speak the same home language. Figure 12 shows the predicted effect of an increase of the Herfindahl index from the 25th percentile to the 75th percentile. For example, for Los Angeles LTELs, this corresponds to a 16 percentage point increase in the probability that any two EL students picked at random in a school would share the same home language (ranging from a 84% chance to a 100% chance). For San Diego LTELs, the predicted effect corresponds to a 32 percentage point increase in the probability that any two EL students picked at random in a school would share the same home language (ranging from a 53% chance to an 85% chance). The predicted effect is statistically significant in only 7 of 22 cases, and the direction of the effects often varies across outcomes or between the two districts. We conclude, at least for these two districts, that there is not a systematic correlation between EL language homogeneity and annual academic outcomes for any of the three student language groups.26 A limitation of these results is that they are estimated imprecisely, meaning that the range of possible true effects can be quite large. This appears to be due to the very high mean Herfindahl index value (0.8 to 0.89 out of a possible maximum of 1 in LAUSD and 0.59 to 0.67 in SDUSD).27 We discuss this further in Technical Appendix D. 26 Specifically, in LAUSD, more EL student language homogeneity is associated with more positive growth on the CST ELA for both LTELs and LAELs. It is also associated with slower growth on the middle school math CST for never ELs. In SDUSD, never ELs and LTELs both have slower growth on the CST ELA the more language homogeneity there is among ELs. 27 Standard deviations range from 0.14 to 0.22. PPIC.ORG Academic Progress for English Learners 30 Predicted effect of a change from 25th to 75th percentile in language homogeneity FIGURE 12 There does not appear to be a consistent correlation between EL language homogeneity and annual academic outcomes 0.05 CST ELA Z-Score 0.00 CELDT Score CST Math Z-Score -0.05 Annual GPA -0.10 -0.15 LAUSD SDUSD LAUSD SDUSD LAUSD SDUSD Never EL Long-term EL Late-arriving EL SOURCE: Author calculations. NOTES: Chart shows the predicted effect of a change from the 25th percentile to the 75th percentile in language homogeneity. Only statistically significant regression results (p<.05 or greater) are displayed. Reclassified long-term and late-arriving ELs are included. Full results found in Table D4A in the technical appendices. Though we found no quantitative relationship between language homogeneity and outcomes, our interviewees, particularly in SDUSD where the population is more linguistically diverse, suggested that having a large range of home languages at a school presented instructional challenges. Several teachers in the district described employing translation tools and applications, such as Rosetta Stone, for students who spoke less common languages. Longer-Term Outcomes We also examined the relationship between the three measures of a school’s language environment and whether the given EL was ever reclassified as fluent in English as well as whether the given student graduated on time, defined as graduating within four years of first entering grade 9.28 It is more difficult to study whether school language environment matters for these longer-term outcomes than for annual outcomes. In the above results, we searched for a link between the language environment in a given school year and students’ annual performance in that same year. Here, we look for an association between language variables, such as the percentage of students who are EL in a given year, and outcomes that are typically observed several years later. The difficulty created is that we are simultaneously searching for a relation between school characteristics in several different years and a single outcome several years later. With a weaker temporal linkage, it might be more difficult to find statistically significant relationships even if they existed. Our results suggest a few points of agreement across the two school districts for the role of school language environment on these two longer-term outcomes (see full results in Table D4B in the technical appendices):  The percentage of students speaking the same language at the school as the student in question is associated with greater chances of graduating on time for never EL students. 28 As mentioned above, in LAUSD, we do not know which students have dropped out versus which students have moved to other districts. Students who have dropped out are included in the San Diego student data. PPIC.ORG Academic Progress for English Learners 31  Greater language homogeneity among ELs has little statistical relationship with reclassification or graduating on time. In LAUSD, we find that the percentage of students speaking the same language at the school as the student in question is also positively associated with ever being reclassified for LTEL students, but negatively associated with reclassification for LAELs. The percentage of ELs at the school was also positively related to reclassification for LAELs in Los Angeles. In SDUSD, higher shares of ELs are negatively associated with reclassification for both EL groups and negatively associated with graduation for LAELs. School language environment did come up as a potential predictor of reclassification in our school district interviews in SDUSD. Several interviewees described feeling that community context played a major role in reclassification. Interviewees suggested that because LTELs could readily communicate in Spanish at home, with peers in school, and throughout the community, they had few opportunities to develop their academic English skills beyond the intermediate level. Are Course Placement and ELD Programs Related to EL Student Progress? In this section, we examine the role of English Language Development course assignment, including specialized courses for LTELs and LAELs, and whether that correlates with linguistic and academic success for EL students. We also look at the relationship between EL support teachers and EL student outcomes in SDUSD. Overall, by far the most consistent and powerful finding was that in both districts when long-term ELs were not put into any ELD class, their gains on the CST English language arts test declined.29 Both districts have sizeable minorities of long-term ELs who receive no ELD class, and therefore our finding on the relation between not receiving an ELD class and test scores on the CST ELA test is quite relevant. This may have further implications, given pressures in both districts to ensure that students take a–g courses as soon as possible to meet the new graduation standards. While we found a few cases in which ELs being placed into an ELD class that was too low or too high seemed to matter, the results varied. Overview of Our Approach Our main instructional measure is whether ELs are correctly placed into ELD courses, but we also have some information on ELD course offerings and programs, and we study the link between these and student progress as well. As described in the previous section on course assignments, each school district has its own policies that determine how long-term and late-arriving ELs are assigned to the ELD coursework intended to improve EL students’ fluency and literacy in English. That section documents how each district has struggled to assign all EL students to their intended courses. Here, we examine the relationship to academic outcomes for students who are placed in ELD courses that are too low or too high, or that receive no ELD, according to district policy. It is important to note that these measures do not capture ELD pedagogy or how effectively districts and schools are implementing their ELD course designs. As above, we examine growth in CST ELA, CELDT, CST math, and GPA for students in each district, as well as whether students were ever reclassified and whether they graduated on time. 29 We found a similar result for LAELs in San Diego but not LAUSD. PPIC.ORG Academic Progress for English Learners 32 We model our academic outcomes based on the rates of low, high, and no ELD placement at the school level and compare our results to correct ELD course placement. While individual student outcomes may be more closely linked to their own placement, we noted earlier that each district allows teachers and administrators some discretion in EL course assignment. Therefore, if an LTEL student is not enrolled in an ELD course, it may be because it was determined that the student was ready to tackle the demands of an English-only course without additional ELD support. This student may have better academic outcomes than if the student had been placed in the “correct” ELD course. Using individual ELD placement may lead us to conclude that “no ELD” placement is better for long-term EL students than correct course placement were we to use students’ individual placement rates. Further, because our interviews suggested that the ability of districts to correctly place their EL students depends in some ways on the language composition of students at their schools, we model course placement while also controlling for our three school-level language variables (percentage of ELs, share of students speaking the same home language, and language homogeneity among ELs). Finally, we expect that course placement at the school level could differentially affect not only LAELs and LTELs, but also never ELs. If more EL students are correctly placed (as opposed to being without ELD), this may also affect academic outcomes for never ELs, who would share fewer courses with EL students. Similarly, if a greater share of ELs are properly placed into the ELD courses designed for them, this could be beneficial to both LTELs and LAELs, who would each be more likely to be enrolled in the courses specifically designed for them. Annual Outcomes As with our analysis of the school language environment, there is an important point of commonality across the school districts when it comes to the association with school-level ELD course placement and annual academic outcomes:  Higher rates of no ELD placement are associated with slower growth on CST ELA for long-term EL students. Figures 13–15 show the predicted effect of a change from the 25th percentile to the 75th percentile in the share of ELs at the school placed too low or too high, or not placed into any ELD course whatsoever. For example, among LTELs in LAUSD, the 25th percentile for “too low” placement is 1 percent, and the 75th percentile is 9 percent, so we simulate a change of 8 percentage points. Figures 13 and 14 show that often there is no significant link between placements at too high or low a level, and the exceptions vary in direction. However, a much more consistent picture emerges when we examine Figure 15, which shows the predicted effect of the change from the 25th percentile to the 75th percentile in the share of ELs not taking any ELD course. CST ELA scores are predicted to fall for LTELs in both districts, and in San Diego, never EL and LAEL students as well. Interestingly, CELDT scores largely do not appear to be affected, with the exception of a positive association for LAELs in San Diego. In both districts, LAELs’ GPA growth increases when rates of school-wide no ELD placement are higher. PPIC.ORG Academic Progress for English Learners 33 FIGURE 13 There seems to be no consistent association between the share of EL students placed too low in ELD courses and annual outcomes Predicted effect of a change from 25th to 75th percentile in share placed too low 0.35 0.30 0.25 0.20 0.15 0.10 0.05 0.00 -0.05 -0.10 CST ELA Z-Score CELDT Score CST Math Z-Score Annual GPA LAUSD SDUSD Never EL LAUSD SDUSD Long-term EL LAUSD SDUSD Late-arriving EL SOURCE: Author calculations. NOTES: Chart shows the predicted effect of a change from the 25th percentile to the 75th percentile in being placed in ELD coursework that is too low. Only statistically significant regression results (p<.05 or greater) are displayed. Reclassified long-term and late-arriving ELs are included. Full results found in Table D4A in the technical appendices. FIGURE 14 There seems to be no consistent association between the share of EL students placed too high in ELD courses and annual outcomes Predicted effect of a change from 25th to 75th percentile in share placed too high 0.1 0.08 0.06 0.04 0.02 0 -0.02 -0.04 -0.06 -0.08 -0.1 CST ELA Z-Score CELDT Score CST Math Z-Score Annual GPA LAUSD SDUSD Never EL LAUSD SDUSD Long-term EL LAUSD SDUSD Late-arriving EL SOURCE: Author calculations. NOTES: Chart shows the predicted effect of a change from the 25th percentile to the 75th percentile in being placed in ELD coursework that is too high. Only statistically significant regression results (p<.05 or greater) are displayed. Reclassified long-term and late-arriving ELs are included. Full results found in Table D4A in the technical appendices. PPIC.ORG Academic Progress for English Learners 34 FIGURE 15 Higher shares of EL students placed in no ELD courses are associated with lower English language arts scores Predicted effect of a change from 25th to 75th percentile in share not placed in ELD 0.1 0.08 0.06 0.04 0.02 0 -0.02 -0.04 -0.06 -0.08 -0.1 CST ELA Z-Score CELDT Score CST Math Z-Score Annual GPA LAUSD SDUSD Never EL LAUSD SDUSD Long-term EL LAUSD SDUSD Late-arriving EL SOURCE: Author calculations. NOTES: Chart shows the predicted effect of a change from the 25th percentile to the 75th percentile in being placed in no ELD coursework. Only statistically significant regression results (p<.05 or greater) are displayed. Reclassified long-term and late-arriving ELs are included. Full results found in Table D4A in the technical appendices. Our tentative conclusion is that ELs do not suffer systematically from placement into too high or too low an ELD course, but not placing an EL in designated ELD coursework (or “no ELD” as in our models) at all is associated with significantly lower gains for LTELs on the state’s test of English language arts. Given that some teachers in both districts referred to pressures, associated with the a–g graduation requirements, to place all students into ELA classes that earn a–g credit, our result raises some concerns for ELs when they are not enrolled in ELD courses. And despite the fact that it appears that the share of LAEL students within each district is fairly constant in recent years, about half of teachers commented that fluctuations in LAEL enrollment made it difficult to plan for courses and consistently secure teachers. Further, about half of all interviewed teachers described not feeling adequately prepared to address students’ non-academic needs, as one teacher in SDUSD stated, “Academically, emotionally, even health care . . . if you sit down [with] one of these kids . . . there [are] just so many issues that it’s hard to address every single one as a teacher.” Similarly, teachers in LAUSD indicated that LAEL courses required differentiated instruction to account for the varying amounts of prior schooling newcomers had in their countries of origin. As one teacher in LAUSD indicated, “So, [even when] they can read or write really well in Spanish, I [also] have students that have never been to school, so I have to deal with a student who was an excellent student back in El Salvador, who is an A student, super interested in learning and competent, and a student . . . who has no idea of what's going on, he barely knows how to read and write in Spanish. How in the world can you bridge that gap? What do you do with all these variety of students and all these needs? [It’s] really very hard.” Thus, even though LAELs might have been placed into the correct ELD courses, teachers in both districts may have struggled to address their needs. An additional issue raised in interviews was the suggestion that, even when students were assigned to designated ELD classes with the proper course title, they might also have students assigned to different ELD courses enrolled in the same class period. Interviewees in both districts described struggling to serve ELs at different English proficiency levels in the same classroom. At schools with low proportions of ELs in Los Angeles, low EL PPIC.ORG Academic Progress for English Learners 35 enrollment at specific language proficiency levels resulted in the consolidation of ELD courses (i.e., the double or multiple rostering described above). At schools with small EL populations in San Diego, ELs were often integrated into general education classrooms, and teachers indicated that their class sizes were becoming too diverse and too large to manage. In all cases, teachers struggled to accommodate ELs’ needs. In LAUSD, we were able to examine in our student-level data the extent to which LTELs and LAELs were placed in the same ELD and grade-level English language arts classrooms with students enrolled in different courses. We found that in more recent years, increasing proportions of LAELs in the district were in the same classroom with students enrolled in different courses (up from 10% in 2002–03 to 35% in 2015–16; see Figure B2 in the technical appendices). This did not appear to be an issue for LTELs in LAUSD during our study period. Unfortunately, we were not able to investigate this question in SDUSD.30 Longer-Term Outcomes In models of whether students graduated on time, we found two points of commonality between the districts (see full regression results in Table D4B in the technical appendices).  If a greater proportion of EL students are placed in courses that are too high for them, graduation rates are higher for never EL students.  If a greater share of EL students are not in ELD at the school level, graduation rates are higher for LAEL students in both districts. Otherwise, findings on longer-term outcomes are highly mixed. In Los Angeles, more EL students not enrolled in ELD is associated with lower graduation rates for LTELs; the opposite is the case in San Diego. In SDUSD, a higher share of ELs placed in ELD courses that are too low is associated with lower graduation rates for never EL students. Higher proportions of EL students placed too high is associated with lower graduation rates for LTELs. And no ELD placement in SDUSD is positively associated with graduation rates for never ELs and LTELs as well as LAELs (as previously described). In modeling the probability that ELs were ever reclassified, the results differed between the two districts to such an extent that we do not believe that there are clear implications for these two districts, let alone other districts.31 Specialized LTEL and Newcomer Courses Recall that EL course design, especially for LTEL students, changed in LAUSD with two new courses for LTELs, which were mandatory beginning in 2013–14. SDUSD piloted LTEL coursework in 2012–13 (called Academic Language Development, or ALD) and then formally introduced it as optional coursework in 2015–16. There are also special courses designed for LAELs (“newcomer courses”). In San Diego, these were historically courses that allowed LAELs in their first year of schooling to take specialized ELD and academic courses together with other first-year LAELs. More recently SDUSD has moved to enroll first-year LAELs in regular ELD and academic courses. Our interview data suggest this move was influenced by declining enrollment in the district and budget cuts, as well as an overall district philosophy that emphasized EL integration into general education courses. 30 The data required to determine which students were in the same classroom are not available for SDUSD. 31 In SDUSD, when an EL student was in a school where a higher share of ELs were placed in too low an ELD class, the probability of reclassification was lower. Similarly we found negative predicted effects when the share of ELs not placed into any ELD class rose. These negative outcomes related to ELD misplacement applied to both LTELs and LAELs. On the other hand, for LAELs but not LTELs, a higher share of ELs placed in too high an ELD class was associated with greater probability of reclassification. For LAUSD, in only three cases was the course placement variable statistically significant, and the signs were opposite those found in SDUSD. A higher share of ELs placed into too low an ELD class was positively linked to reclassification of ELs in LAUSD, and a higher share of ELs placed into too high an ELD class was negatively associated with reclassification for LTELs. PPIC.ORG Academic Progress for English Learners 36 To test for an association between specialized courses for LTELs and LAELs and the academic progress of each language group, we created indicators for whether any students had enrolled in LTEL or newcomer courses in a given school and a given year. We then interacted these indicator variables with indicators for whether the individual student was never EL, a LAEL or LTEL, and estimated a version of the earlier models that added these six interacted variables as additional explanatory variables. Because these specialized courses were available in both districts for only a subset of years, we focused on testing for a relationship between the provision of these courses and annual outcomes, rather than longer-term outcomes. Table D5 in the technical appendices shows the coefficients of our regression model in which we add indicators for any newcomer classes to the earlier model, which includes the language environment and ELD course placement variables. For SDUSD we also include the LTEL course indicators. For newcomer courses, we found no significant link to any of the four outcomes in Los Angeles, with two exceptions. LAELs in the district had significantly higher GPA if their school offered newcomer courses, but LTELs had slightly lower CST math scores. In San Diego, on the other hand, we found three significant associations, and all three were positive and sizeable. In San Diego district schools with newcomer courses, LTELs had higher CELDT score gains and higher CST math score gains, and never EL students had higher math score gains as well. We cannot know for sure if these results are causal, but one potential explanation is that both LTELs and never ELs gained academically when LAELs in their first year were taught most academic courses separately from other students, where teachers could more effectively differentiate their instruction. This is potentially important given SDUSD’s decision to roll back specialized programs for newcomers. In the case of courses for long-term ELs, we caution that our evidence base is very weak, so much so that we do not present quantitative results for Los Angeles.32 For San Diego, our analysis of ALD coursework for LTELs was only slightly more successful. ALD courses began to appear in transcripts only in 2015–16 and only in some schools, so we lacked the variation to obtain precise estimates. We found no significant effect of ALD courses in San Diego for CELDT scores for either EL group, or for GPA for any of our three language groups. That said, the best interpretation of these results is that we will need more years of data to answer with any certainty whether ALD courses “matter.” An additional factor to consider in Los Angeles is that the very high and increasing percentage of special education students among LTELs may have created a mismatch between course design and student need. Our interview data suggest respondents in both districts had concerns about the utility of these courses for LTELs. In Los Angeles, where schools offered the newly created designated ELD courses for LTELs, most interviewees suggested that students perceived these classes to be “boring” and in turn dismissed them. One middle school principal commented: “I think our kids need more hands-on . . . more [real-world] application. I find that they have a lot of . . . scripted things to do and it's not real to them and they get bored. They get bored. The prompts are not made for them. They need something they can connect to.” Similarly, an ALD teacher in San Diego said, “The way that that curriculum, and/or text is laid out, is that it builds on . . . it’s a lot of … very repetitive, and it’s constant repeat, or constant. . . a big part of what I’m trying to do is find that balance of how can we be repetitive without being boring.” Another teacher noted, “The ALD just wasn’t engaging enough, I couldn’t even make it 32 In LAUSD, our efforts to measure associations between the new LTEL courses and academic outcomes for LTELs, LAELs, and never ELs are frustrated by our measure of correct course placement. In LAUSD, all LTELs are supposed to be placed in one of two LTEL courses, either Advanced ELD or Language and Literacy. Our measure of correct course placement for EL students at the school level is largely a measure of correct course placement for LTELs, since LTELs are the majority of secondary ELs. Our models that measure both whether a school offers the LTEL courses and the share of EL students that are correctly placed are unable to produce results for the association between those two variables due to their colinearity. Instead, since the LTEL courses were implemented beginning in 2013–14, we estimated our standard model for just the 2013–14 to 2015–16 period. We find that the only change for LTELs from our main model is that higher percentages of EL students being placed too high is correlated with positive CELDT gains. We cannot estimate CST results because the assessment was last given in 2012–13. PPIC.ORG Academic Progress for English Learners 37 engaging enough.” Although the new designated ELD courses for LTELs were often described as either boring or poorly designed, we do not have classroom observational data to expand on this claim. English Learner Support Teachers in San Diego In San Diego, schools with large numbers of ELs were assigned EL support teachers (ELST), who worked with school administrators to facilitate ELs’ placement in appropriate coursework, monitored ELs’ progress in acquiring English fluency, identified students ready to reclassify, ensured that the reclassification process was correctly implemented, and supported classroom teachers who served EL students. This program began in 2004– 05 and slowly ramped up until the early part of this decade; due to financial constraints the number of EL support teachers has declined in the last five years. When asked why the number of these teachers in the district had dropped from 120 to 46 in recent years, a district leader we interviewed cited “resources and politics.” Most recently, the EL support teacher position has been eliminated and in March 2017 replaced with two new job titles: English language instructional resource teacher (ELI-RT) and English language development instructional specialist (ELD specialist). Rather than having one EL support teacher assigned to one to two schools, the district’s approach now involves having a resource teacher (ELI-RT) serve a cluster of about four to five schools. Under this new configuration, EL coordinators at school sites (who can be teachers or administrators) became the primary points of contact for the district. One district staff member described this approach as “haphazard,” going on to say, “We’re lucky if we have somebody to communicate with [at the schools]. . . . We have EL coordinators but you know once again so many are so brand new and they get bumped out of the position, and it’s like every year we feel like we’re training people all over again.” Overall, we find several strands of evidence that San Diego’s EL support teachers have historically been associated with better outcomes for ELs and never ELs as well. To assess whether the presence of EL support teachers in a school leads to better outcomes for either EL group or for never ELs, we added a measure of EL support teacher intensity to our earlier models, which already controlled for the school’s language environment and the nature of ELD course placements. Our measure is the ratio of EL support teachers (full-time equivalents) in a school to the number of ELs in the school in the given year. This ratio peaked at 0.01 in 2012–13 and declined to 0.008 in 2013–14.33 In the analysis below we discuss the predicted effects of increasing the EL support teacher to EL ratio from 0 to 0.01. To make the wording less awkward, we refer to this below as “implementation of ELST support.” We found that for both EL groups, implementation of ELST support is associated with significantly higher growth in GPA. The predicted effect was to increase GPA of late-arriving and long-term ELs by 0.15 and 0.05, respectively. CELDT scores were predicted to rise by about one-tenth of a standard deviation for LAELs, but the result was not statistically significant when we additionally controlled for the accuracy of ELD course placements. One interpretation of this is that having an EL support teacher in a school facilitated more accurate ELD placements, which in turn boosted LAELs’ performance. CST ELA scores were not affected. One other significant finding concerns never EL students, for whom math achievement was predicted to rise by 0.02 of a standard deviation after the implementation of ELST support.34 33 We were not able to locate information on EL support teacher school assignments after 2013–14, a period during which our understanding is that the number of EL support teachers declined, so we studied the school years 2006–07 to 2013–14. 34 Table D6 in the technical appendices provides details. PPIC.ORG Academic Progress for English Learners 38 We also examined the two longer-term outcomes, whether EL students were ever reclassified and whether they graduated on time. Two results were significant. Implementation of ELST support was associated with a 12 percentage point reduction in the probability that late-arriving ELs would ever be reclassified, but no impact on the probability that they would graduate on time. Turning to long-term ELs, implementation of ELST support was associated with no change in the probability of reclassification, but a 4.9 percentage point increase in the probability of graduating on time. Our interview data also positioned EL support teachers as central players in ensuring EL course placement aligned with district policy, as they facilitated the flow of information from the district to schools, as well as between schools when ELs matriculated. Although some communication still happened to facilitate articulation between middle and high school among staff at “the big feeder schools,” it was “not as systematized as it should be,” according to a resource teacher we interviewed. One district staff member also shared that support teachers were “hugely beneficial. . . . They helped us in making sure that kids were where they were supposed to be. . . . They just . . . added so much to the site in every respect.” The elimination of site-based EL support teachers thus may impact how well schools are able to follow course placement guidelines and the fidelity to the district’s EL Master Plan, which is designed to move students toward reclassification and graduation. Conclusion English Learners are a diverse group of students—including long-term ELs who have been educated in US schools, sometimes since kindergarten, and late-arriving ELs, some of whom arrive with a high degree of literacy in their home language and others with much less fluency. In the two districts studied in this report—Los Angeles Unified and San Diego Unified—school staff and administrators described difficulties in serving ELs in middle and high schools, a perspective that was echoed in our quantitative analysis. Teachers mentioned particular challenges in serving late-arriving ELs, with many having experienced trauma prior to arrival in the United States, interrupted schooling, and/or financial strain. In San Diego, district and school staff reported that many latearriving ELs were refugees; in Los Angeles, many were unaccompanied minors. Academic outcomes across these two student groups also vary. Our findings revealed that long-term EL students had relatively stronger academic outcomes on average. But unsurprisingly, given that progress in English proficiency is often more rapid at lower proficiency levels, late-arriving ELs made greater gains in the academic outcomes we measured. Policymakers have recognized the importance of improving outcomes for English Learners and have undertaken recent funding, instructional, and accountability reforms. Moving forward, it is critical that implementation and new policies consider the diverse linguistic and academic needs of EL students and how these factors contribute to their long-term success. Below we outline how some of our key findings can help guide policymaking. In our examination of the school language environment, we generally found that the academic progress of English Learners seems quite adaptable to factors such as the share of ELs in a school, the share of all students speaking the same home language, and the language homogeneity of EL students. Importantly, we found that students who were never identified as EL do not have slower rates of growth in schools with higher proportions of ELs. This finding suggests fluent English speakers are not negatively affected by having large numbers of ELs as school peers, at least in these two districts. PPIC.ORG Academic Progress for English Learners 39 Results were somewhat less consistent for the relationship between the percentage of ELs at a school and the academic outcomes of EL students. In Los Angeles, we found some evidence that these two student groups make faster academic progress in schools where a higher percentage of students are learning English. Notably, in our school interviews, EL instructional staff often noted that placing ELs in appropriate coursework was more challenging when the percentage of ELs at the school was lower. However, in San Diego, there is some evidence that longer-term outcomes (i.e., reclassification and on-time graduation) are negatively associated with higher shares of ELs at the school level. Another goal of this report was to examine how EL students are placed into English language development courses and whether course placement is associated with academic outcomes. In San Diego, both student groups were correctly placed (i.e., placed in accordance with district guidelines) between 50 and 60 percent of the time. In Los Angeles, long-term ELs were placed correctly 40 to 60 percent of the time; correct course placement for late-arriving ELs was higher, in the 70 percent range. Many of those not placed correctly were in ELD courses that were either too high or too low for their proficiency level. Trends in long-term EL students receiving no ELD instruction—in recent years this rate has declined in Los Angeles and risen somewhat in San Diego—may be influenced by district-specific factors. For example, in Los Angeles, the district has increased district-level oversight and support for EL students’ course assignments, at least partially in response to pressure from a settlement with the US Department of Education’s Office for Civil Rights. Meanwhile, in San Diego, the district leaves more discretion to school-level staff. Further, interviewees in San Diego schools serving small EL populations noted that in the most recent school year, most if not all ELD courses designed specifically for long-term ELs were eliminated at their schools due to districtwide budget cuts. Instead, ELs were placed in grade-level English courses, when possible with teachers whom administrators deemed well-prepared to differentiate instruction for ELs. These patterns are important because we found that having no specific ELD instruction is associated with slower growth on state tests of academic English for long-term ELs in both school districts, and for late-arriving ELs in San Diego. In addition, many of the school staff we interviewed believed that appropriate ELD supports were crucial for EL progress. At the same time, some staff noted that placing EL students in courses required for graduation may sometimes be given precedence over placing them in ELD coursework. Thus, the practice of placing EL students in general English coursework without offering ELD instruction should be examined with great care—especially since eliminating ELD courses that focus explicitly on developing ELs’ English proficiency runs counter to California’s ELA/ELD Framework policy. Finally, in San Diego, we found some evidence that the presence of EL support teachers at school sites led to better grades for both long-term and late-arriving ELs, and faster English proficiency growth for late-arriving ELs. Budgetary issues have led the district to replace these teachers with two new types of support teachers, but at lower levels of support. In interviews, school staff praised the EL support teachers and many expressed a desire to restore funding for this position. If the district’s financial shortfall eases, bringing back school-level EL support teachers to previous levels could help ELs’ academic progress. English Learners in middle and high school face unique challenges as they strive to gain English language proficiency and meet the other academic requirements necessary to gain a high school diploma. For California, helping ELs advance both linguistically and academically remains a high priority—and understanding the factors that contribute to their progress and struggles will be critical to promoting their long-term success. PPIC.ORG Academic Progress for English Learners 40 REFERENCES Bui, Sa A. 2013. “How Do Limited English Proficient Students Affect Each Other’s Educational Outcomes? Evidence from Student Panel Data“ working paper. CDE (California Department of Education). 2016. 2016–17 and 2017–18 CELDT Information Guide. CDE (California Department of Education). 2016–17. DataQuest. CDE (California Department of Education). 2017–18. DataQuest. California State Legislature. 2012. Assembly Bill 2193: An Act to Add Sections 313.1 and 313.2 to the Education Code, relating to English Learners. Callahan, Rebecca. 2005. “Tracking and High School English Learners: Limiting Opportunity to Learn.” American Educational Research Journal 42 (2): 305–328. Callahan, Rebecca, Lindsey Wilkinson, Chandra Muller, and Michelle Frisco. 2008. “ESL Placement and Schools: Effects on Immigrant Achievement.” Educational Policy 23 (2): 355–384. Carhill-Poza, Avary. 2015. “Opportunities and Outcomes: The Role of Peers in Developing the Oral Academic English Proficiency of Adolescent English Learners.” Modern Language Journal 99 (4): 678–695. Conger, Dylan. 2009. “Testing, Time Limits, and English Learners: Does Age of Entry Affect How Quickly Students Can Learn English?” Social Science Research 38 (2): 383-396. Conger, Dylan, Amy E. Schwartz, and Leanna Stiefel. 2011. “The Effect of Immigrant Communities on Foreign-born Student Achievement.” International Migration Review 45 (3): 675–701. Flores, Nelson, Tatyana Kleyn, and Kate Menken. 2015. “Looking Holistically in a Climate of Partiality: Identities of Students Labeled Long-Term English Language Learners.” Journal of Language, Identity & Education 14 (2): 113–132. Friesen, Jane, and Brian Krauth. 2011. “Ethnic Enclaves in the Classroom.” Labour Economics 18 (5): 656–663. Hill, Laura, Julian R. Betts, Belen Chavez, Andrew C. Zau, and Karen V. Bachofer. 2014. Pathways to Fluency: Examining the Link between Language Reclassification Policies and Student Success. Public Policy Institute of California. Hodara, Michelle. 2015. “The Effects of English as a Second Language Courses on Language Minority Community College Students.” Educational Evaluation and Policy Analysis 37 (2): 243–270. Johnson, Angela S. 2018 “Can Summer School Help ELs Succeed?” EdWeek, January 15. Kanno, Yasuko, and Sarah E. N. Kangas. 2014. “‘I’m Not Going to Be, Like, for the AP’: English Language Learners’ Limited Access to Advanced College-Preparatory Courses in High School.” American Educational Research Journal 51 (5): 848–878. Los Angeles Unified School District. 2011. Reclassification of English Learners, Grades 2–12. Bulletin BUL-5619.0, October 17. MacSwan, Jeff, and Lisa Pray. 2005. “Learning English Bilingually: Age of Onset of Exposure and Rate of Acquisition among English Language Learners in a Bilingual Education Program.” Bilingual Research Journal 29 (3): 653–678. Marinova-Todd, Stefka H., D. Bradford Marshall, and Catherine E. Snow. 2000. “Three Misconceptions about Age and L2 Learning.” TESOL Quarterly 34 (1): 9–34. Menken, Kate and Tatyana Kleyn. 2010. “The Long-Term Impact of Subtractive Schooling in the Educational Experiences of Secondary English Language Learners.” International Journal of Bilingual Education and Bilingualism 13 (4) 399–417. Menken, Kate, Tatyana Kleyn, and Nabin Chae. 2012. “Spotlight on Long-Term English Language Learners: Characteristics and Prior Schooling Experience of an Invisible Population.” International Multilingual Research Journal 6 (2): 121–142. Miles, Matthew B., and A. Michael Huberman. 1994. Qualitative Data Analysis: An Expanded Sourcebook. Thousand Oaks, CA: Sage Publications Inc. NASEM (National Academies of Sciences, Engineering, and Medicine). 2017. Promoting the Educational Success of Children and Youth Learning English: Promising Futures. Edited by Ruby Takanishi and Suzanne Le Menestrel. Washington, DC: The National Academies Press. National Center for Education Statistics. 2018. English Learners in Public Schools. National Center for Education Statistics, updated April 2018. Olsen, Laurie. 2010. Reparable Harm: Fulfilling the Unkept Promise of Educational Opportunity for California’s Long Term English Learners. Long Beach, CA: Californians Together. PPIC.ORG Academic Progress for English Learners 41 Parish, Thomas B., Amy Merickel, Maria Pérez, Robert Linquanti, Miguel Socias, Angeline Spain, Cecilia Speroni, Phil Esra, Leslie Brock, and Danielle Delancey. 2006. Effects of the Implementation of Proposition 227 on the Education of English Learners, K–12: Findings from a Five-Year Evaluation. Prepared by American Institutes for Research and WestEd for the California Department of Education. Patton, Michael Q. 1990. Qualitative Evaluation and Research Methods. 2nd edition. Newbury Park, CA: Sage Publications Inc. Salazar, Jesus Jose, and Katherine G. Hayes. 2010. “The Role of Home Language Proficiency and Its Impact on English Learners' Academic Achievement.” Paper presented at the annual meeting of the American Educational Research Association. San Diego Unified School District. 2009. Master Plan for English Learners. Office of Language Acquisition. San Diego Unified School District. 2010. Master Plan for English Learners. Office of Language Acquisition, updated April 2010. Silver, David, Marisa Saunders, and Estela Zarate. 2008. “What Factors Predict High School Graduation in the Los Angeles Unified School District” California Dropout Research Project. Stevens, Gillian. 2015. “Trajectories of English Acquisition among Foreign-born Spanish-Language Children in the United States.” International Migration Review 49 (4) 981–1000. Strauss, Anselm L., and Juliet M. Corbin. 1990. Basics of Qualitative Research: Grounded Theory Procedures and Techniques. Newbury Park, CA: Sage Publications Inc. Thompson, Karen D. 2015. “Questioning the Long-Term English Learner Label: How Categorization Can Blind Us to Students' Abilities.” Teachers College Record 117 (12): 1–50. Umansky, Ilana, Megan Hopkins, Dafney B. Dabach, Lorna Porter, Karen Thompson, and Delia Pompa. 2018. Understanding and Supporting the Educational Needs of Recently Arrived Immigrant English Learner Students: Lessons for State and Local Education Agencies. Council of Chief State School Officers. Watt, David, and Hetty Roessingh. 1994. “Some You Win, Most You Lose: Tracking ESL Student Drop Out in High School (1988– 1993).” English Quarterly 26 (3): 5–7. PPIC.ORG Academic Progress for English Learners 42 ABOUT THE AUTHORS Laura Hill is a policy director and senior fellow at the Public Policy Institute of California. Her areas of expertise are K–12 education and immigration. She is currently researching English Learners in California schools, special education, and ESL students in community colleges. Her recent publications examine student achievement on the state’s K–12 assessments and the link between language reclassification policies and student success. Prior to joining PPIC, she was a research associate at the SPHERE Institute and a National Institute of Aging postdoctoral fellow. She holds a PhD in demography and an MA in economics from the University of California, Berkeley. Julian Betts is a professor and former chair of the Department of Economics at the University of California, San Diego, and is executive director of the department’s San Diego Education Research Alliance (SanDERA). He is also an adjunct fellow at PPIC and research associate at the National Bureau of Economic Research. He has performed student-level research using San Diego Unified School District data over the past 18 years. His research on determinants of student achievement and evaluation of interventions has examined differences in outcomes between EL and non-EL students. He has extensive experience with experimental and quasi-experimental methods. He holds a PhD in economics from Queen’s University, Kingston, Ontario, Canada. Megan Hopkins is assistant professor in the Department of Education Studies at the University of California, San Diego. She is an expert in mixed methods as well as EL policy and program implementation. She has a research agenda focused on supporting school districts undergoing demographic and/or policy change. Magaly Lavadenz is professor in the Department of Educational Leadership and Founding Director of Loyola Marymount University’s Center for Equity for English Learners. Her research interests include preparation and professional support of Latino and bilingual teachers, the experiences of the Central American immigrant community, public policy affecting language use, biliteracy, and education of Latinos and immigrant communities. Karen Bachofer is director of SanDERA. Previously she was executive director of San Diego Unified School District’s Research and Evaluation Division, where her responsibilities included direct oversight of national, state, and district assessment, and accountability processes and reporting. She also oversaw internal, external, and commissioned evaluation activities, including Title III reporting and production of the district’s annual English Learner Progress Report. Her SanDERA research has included studies of the impact of EL reclassification policies and student learning trajectories. She holds a PhD in education from the Claremont Graduate School and San Diego State University. Joseph Hayes is a research associate at the Public Policy Institute of California, where he studies population change, educational policy, and corrections issues. Recent projects have focused on estimates of the undocumented immigration population, English Learner reclassification policies in California public schools, and the changing composition of the state’s prison and parole populations. He holds an MS in agricultural economics from the University of Wisconsin, Madison. Andrew Lee is a research associate at the Public Policy Institute of California, focusing on K–12 education, specifically as it relates to English Learners in California schools. Prior to joining PPIC, he contributed to research on health policy and agricultural development policy at Acumen LLC and the UN Food and Agriculture Organization, respectively. He holds an MA in development practice from the University of California, Berkeley, and a BA in economics and English from the University of Michigan, Ann Arbor. Marco A. Murillo is a senior research associate at Loyola Marymount University’s Center for Equity for English Learners (CEEL). His research interests focus on the secondary school experience of immigrant youth as well as issues relating to college readiness and access for low-income, students of color. He holds a BA in political science and history from the University of California, San Diego, and a PhD in education (urban schooling) from the University of California, Los Angeles. Tara Vahdani is a research associate at CEEL, where she works on a broad range of projects related to equity and excellence in English Learner education. Prior to joining CEEL, she engaged in qualitative and quantitative methods as a research assistant at the Civil Rights Project/Proyecto Derechos Civiles, and later UC San Diego, where she earned a BA, studying education and political science with an emphasis in public policy. Andrew C. Zau is senior statistician at SanDERA. His research focuses on the effects of policy on student achievement in K–12 education. He has 16 years of experience analyzing SDUSD student-level longitudinal data and has played key roles in evaluations of the impact of reading reforms, diagnostic math testing, and school choice in San Diego, using experimental and quasi-experimental approaches. He holds an MPH in epidemiology from San Diego State University and a BS in bioengineering from the University of California, San Diego. ACKNOWLEDGMENTS We gratefully acknowledge additional research support from Mariko Cavey, Leslie Gautsch, Hayley Weddle, Kathryn Wiley, Iwunze Ugo, and Nidia Hernandez. Caroline Danielson, Kenji Hakuta, Rebecca Callahan, Kathy Hayes, Julien Lafortune, and Lynette Ubois provided very helpful reviews, and Vicki Hsieh provided much-appreciated editorial support, and Kit Reber provided production support. We thank the many staff in central office and school sites in both districts for sharing their insights. In particular we thank Sandra Cephas at SDUSD and Hilda Maldonado at LAUSD for many helpful conversations and for sharing materials on the history of EL support in their respective districts. PPIC.ORG Academic Progress for English Learners 43 PUBLIC POLICY INSTITUTE OF CALIFORNIA Board of Directors Steven A. Merksamer, Chair Senior Partner Nielsen Merksamer Parrinello Gross & Leoni LLP Mark Baldassare President and CEO Public Policy Institute of California Ruben Barrales Senior Vice President, External Relations Wells Fargo María Blanco Executive Director University of California Immigrant Legal Services Center Louise Henry Bryson Chair Emerita, Board of Trustees J. Paul Getty Trust A. Marisa Chun Partner McDermott Will & Emery LLP Chet Hewitt President and CEO Sierra Health Foundation Phil Isenberg Former Chair Delta Stewardship Council Donna Lucas Chief Executive Officer Lucas Public Affairs Mas Masumoto Author and Farmer Leon E. Panetta Chairman The Panetta Institute for Public Policy Gerald L. Parsky Chairman Aurora Capital Group Kim Polese Chairman ClearStreet, Inc. Gaddi H. Vasquez Senior Vice President, Government Affairs Edison International Southern California Edison PPIC.ORG Academic Progress for English Learners 45 The Public Policy Institute of California is dedicated to informing and improving public policy in California through independent, objective, nonpartisan research. Public Policy Institute of California 500 Washington Street, Suite 600 San Francisco, CA 94111 T: 415.291.4400 F: 415.291.4401 PPIC.ORG PPIC Sacramento Center Senator Office Building 1121 L Street, Suite 801 Sacramento, CA 95814 T: 916.440.1120 F: 916.440.1121" } ["___content":protected]=> string(328) "

Academic Progress for English Learners: The Role Of School Language Environment and Course Placement in Grades 6-12

" ["_permalink":protected]=> string(276) "https://www.ppic.org/publication/academic-progress-for-english-learners-the-role-of-school-language-environment-and-course-placement-in-grades-6-12/academic-progress-for-english-learners-the-role-of-school-language-environment-and-course-placement-in-grades-6-12-january-2019/" ["_next":protected]=> array(0) { } ["_prev":protected]=> array(0) { } ["_css_class":protected]=> NULL ["id"]=> int(18106) ["ID"]=> int(18106) ["post_author"]=> string(1) "4" ["post_content"]=> string(0) "" ["post_date"]=> string(19) "2019-01-23 10:06:13" ["post_excerpt"]=> string(0) "" ["post_parent"]=> int(18052) ["post_status"]=> string(7) "inherit" ["post_title"]=> string(115) "Academic Progress for English Learners: The Role Of School Language Environment and Course Placement in Grades 6-12" ["post_type"]=> string(10) "attachment" ["slug"]=> string(127) "academic-progress-for-english-learners-the-role-of-school-language-environment-and-course-placement-in-grades-6-12-january-2019" ["__type":protected]=> NULL ["_wp_attached_file"]=> string(131) "academic-progress-for-english-learners-the-role-of-school-language-environment-and-course-placement-in-grades-6-12-january-2019.pdf" ["wpmf_size"]=> string(6) "831879" ["wpmf_filetype"]=> string(3) "pdf" ["wpmf_order"]=> string(1) "0" ["searchwp_content"]=> string(138842) "JANUARY 2019 Laura Hill, Julian Betts, Megan Hopkins, Magaly Lavadenz, Karen Bachofer, Joseph Hayes, Andrew Lee, Marco A. Murillo, Tara Vahdani, and Andrew C. Zau Supported with funding from the William T. Grant Foundation Academic Progress for English Learners The Role of School Language Environment and Course Placement in Grades 6–12 © 2019 Public Policy Institute of California 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 office. 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 reflect the views of the authors and do not necessarily reflect the views of our funders or of the staff, officers, advisory councils, or board of directors of the Public Policy Institute of California. SUMMARY CONTENTS Introduction 5 Who Are Secondary ELs? 7 How Are Course Assignments Implemented? 11 How Do Rates of Progress Differ for Long-Term and Late-Arriving English Learners? 19 Are School Language Environments Related to EL Student Progress? 25 Are Course Placement and ELD Programs Related to EL Student Progress? 32 Conclusion 39 References 41 About the Authors 43 Acknowledgments 43 Technical appendices to this report are available on the PPIC website. PPIC.ORG Over 40 percent of students in California’s K–12 education system speak a language other than English at home. Almost half of these students are considered English Learners (ELs), meaning they require additional language and academic support to succeed in school. In middle and high school, ELs face the dual challenge of attaining fluency in English while mastering the academic courses leading to a high school diploma. This report examines two important types of ELs in Los Angeles and San Diego districts’ middle and high schools: long-term ELs, who have spent several years in US schools without being reclassified fluent in English, and late-arriving ELs, who first enroll in the district in grade 6 or higher and who enter with little English fluency. Our analysis uses student data from 2006–07 to 2015–16 and incorporates interviews conducted in 2017–18 with staff and teachers at the two districts. Overall Trends  Both districts saw a decline in enrollment of long-term and late-arriving ELs in middle and high schools. A principal reason is that the rate of reclassifying ELs as fluent in English before middle school has risen. Our interviews suggest that lower EL enrollment in middle and high schools may have led districts to group ELs of different English proficiency levels into the same class or to place them in general education coursework without integrated language supports.  The share of late-arriving ELs within the EL student population has increased. In recent years, many of these students have been refugees or unaccompanied minors. Generally, long-term ELs have more English fluency and higher standardized test scores than late-arriving ELs, but latearriving ELs make greater academic gains over time.  Both districts faced challenges in assigning long-term and late-arriving ELs to appropriate English Language Development (ELD) coursework, as defined by district guidelines. In both districts, over 50 percent of long-term and late-arriving ELs were placed correctly in recent years. Many other longterm and late-arriving ELs also received ELD courses, but at too high or low a level, while some did not receive ELD courses at all. School Language Environment and Course Placement  Higher proportions of EL students at a school should not be seen as detrimental to the academic performance of fluent English speakers. In both districts, the percentage of ELs at a school was not related to the academic outcomes of native or initially fluent English speakers. We found mixed evidence on the relationship between the percentage of ELs in a school and the academic performance of ELs. Academic Progress for English Learners 3  Long-term ELs who do not take an ELD course have reduced academic performance. In both districts, higher rates of no ELD placement are associated with slower growth for long-term ELs on statewide tests of English language arts. These results suggest that efforts should be made to support ongoing ELD.  In San Diego, we found some evidence that the presence of EL support teachers at schools led to better grades for long-term and late-arriving ELs, and faster English proficiency growth for late-arriving ELs. These results suggest school-level EL support teachers could help ELs progress more quickly. Several recent reforms in California aim to improve educational outcomes for English Learners—and there are more changes to come. Given the diversity and size of this student group, it is critical that policymakers take into account the varying needs of English Learners. PPIC.ORG Academic Progress for English Learners 4 Introduction English Learners are an important part of California’s K–12 student population. More than 40 percent of California school children speak a language other than English at home, according to the California Department of Education (CDE 2017–18).1 And more than one-fifth of California’s students are classified as English Learners (ELs); these students are identified by a standardized English language proficiency test as needing additional language and academic supports in order to succeed in school. A full 38 percent of California’s students are either current or former ELs, including students who were once classified as ELs but have since met reclassification criteria to be considered Reclassified Fluent English Proficient (RFEP). Los Angeles Unified School District (LAUSD) and San Diego Unified School District (SDUSD) are the largest school districts in California, serving 621,000 and 126,000 students, respectively, in 2017–18. These districts also serve substantial numbers of EL students: 143,000 in Los Angeles and 29,000 in San Diego, or approximately 23 percent of all students in each district. Combined, the two districts serve 14 percent of the state’s EL students and 4 percent of the nation’s EL students (CDE 2017–18; National Center for Education Statistics 2018). In this report, we focus on the educational experiences of two very important, but very different, groups of ELs in Los Angeles and San Diego district secondary schools (i.e., middle and high schools):2  The first group, long-term English Learners, or LTELs, make up the majority of ELs in grades 6–12 in California. LTELs are students who have been designated as ELs for several years without having been reclassified as fluent in English.3 Many have been English Learners since kindergarten.  The second group, late-arriving English Learners, or LAELs (also called newcomers), are ELs who arrive in US schools in grade 6 or later and have been in the country for no more than one year upon first enrollment. Some LAELs may have had formal education in their home countries, while others may not be literate in their home language. The heterogeneous backgrounds of long-term and late-arriving ELs raise important policy questions about whether districts and schools can deliver instruction effectively to all of these students. Moreover, we find that each school district recently made changes to its instructional offerings for these student groups. For example, LAUSD developed and mandated new courses for long-term ELs in 2014, where before there had been none; SDUSD piloted similar coursework in 2012 and offered, but did not require, districtwide implementation. For late-arriving ELs, both districts have been moving away from specialized, stand-alone instruction, even as they have experienced increases in their number of newcomers, many of whom were unaccompanied minors and refugees. This report begins by broadly describing long-term and late-arriving ELs, including socioeconomic and linguistic characteristics, district enrollment and school-level patterns. We then examine these students’ assignment to English language development (ELD) coursework and rates of linguistic and academic progress. Next, we explore whether aspects of the school language environment or ELD course placement are associated with rates of student progress on a variety of key academic outcomes. Finally, we highlight actions that schools, districts, and the state can take to support the progress of long-term and late-arriving ELs. By ELD coursework, we refer to both designated English language development classes, which focus explicitly on developing ELs’ English proficiency, as well as integrated ELD, or classes designed to integrate ELD strategies into the language arts content area. These two approaches comprise the comprehensive framework to EL instruction outlined in the California Department of Education’s English Language Arts/English Language 1 In citations to the CDE DataQuest database, the years refer to the school year. 2 EL students who do not fall into either of these categories are excluded from our analyses. 3 As noted below and in Technical Appendix A, the precise definitions of LTEL differ across the two districts and the California Department of Education. PPIC.ORG Academic Progress for English Learners 5 Development Framework for California Public Schools (2015), which states that ELs of all proficiency levels and ages require both designated and integrated ELD. In our analyses, we focus on designated ELD course placement and/or placement into integrated English courses, such as literacy, language arts, and literature. We do not study placement into integrated ELD courses in other content areas, like math and science. Our mixed-methods research approach combines quantitative and qualitative analyses. Our statistical analysis uses individual students’ academic records to examine course placement trends and school language environment. Our qualitative analysis involved nearly 100 semi-structured interviews with district officials and staff at 17 purposefully selected middle and high schools.4 Our qualitative analyses focused on understanding educators’ assessments of long-term and late-arriving ELs’ needs and characteristics, school course placement practices, and the institutional supports (or the lack thereof) that promoted and/or hindered their progress (see Technical Appendix C for a more detailed description). Methodological notes and limitations As with all research, this study has some limitations. We describe three of them here. First, the years included in our quantitative and qualitative datasets do not align. While our quantitative data included student-level data from 2005–06 to 2015–16, our interview data were collected in 2017–18. Although we asked interviewees to provide retrospective accounts, these accounts often did not include more than a few years. Thus, our qualitative data are limited in the extent to which they can help us directly explain the trends we observed in the quantitative data. Second, our focus on district and school course placement implementation did not examine classroom practices or their relative quality. Future research should consider employing additional qualitative methods, such as classroom observations and document reviews of curriculum and lesson plans, to determine how classroom instruction is structured and how teachers engage students in both designated and integrated ELD. For example, the presence of specific designated ELD courses designed for long-term ELs in LAUSD does not imply that the curricular approach of these classes was facilitating students’ language development or academic success. A closer look at classroom practices would help us understand course quality and student engagement. Additionally, including students’ perspectives may also yield important insights about their experiences and engagement in ELD courses. Finally, we did not attempt to tease apart whether a student was enrolled in designated or integrated ELD, or both, or the association that enrollment in each type of ELD had with particular outcomes; thus, we cannot make claims about whether designated or integrated ELD was more or less important for predicting success. 4 School sites were selected based on multiple criteria, including the percentage of ELs served, test score gains on English language proficiency exams, and language heterogeneity. Although we selected schools that represented a range of these characteristics, they are not representative of all schools in LAUSD or SDUSD. PPIC.ORG Academic Progress for English Learners 6 Who Are Secondary ELs? In California, 50 percent of English Learners at the secondary school level are long-term ELs (CDE 2016–17), which the state defines as students who have been English Learners for six or more years.5 Given that ELs often need five to seven years to become proficient enough in English to meet reclassification criteria (NASEM 2017), it is not surprising that there are many long-term ELs in middle and high schools. The research literature indicates that long-term ELs often have limited access to the academic coursework required for high school graduation and admission to postsecondary education (Callahan 2005; Kanno and Kangas 2014; Parrish et al. 2006); that they may experience some negative socioemotional and academic effects of being labeled EL (Flores, Kleyn, and Menken 2015; Thompson 2015); and that they tend to have elevated high school dropout rates (NASEM 2017; Silver, Saunders, and Zarate 2008; Watt and Roessingh 1994). Long-term ELs are not a homogenous group. They may have moved between English-as-a-second-language (ESL), bilingual, and mainstream instructional programs, and they may have moved back and forth between the US and their country of origin (Menken, Kleyn, and Chae 2012). Although many LTELs have strong oral English proficiency, they often have low GPAs and low test scores in English language arts and/or mathematics, which may be related to their inconsistent educational experiences and feelings of stigmatization in LTEL-specific courses that they perceive as being too easy for them (Menken and Kleyn 2010). Whereas long-term ELs constitute a large proportion of the state’s overall secondary EL population, late-arriving ELs, whom we define as EL students arriving within the last year and scoring at the beginning level on their first English proficiency test, constitute 6 percent of ELs in grades 6–12 (CDE 2016–17). Secondary LAELs often face significant challenges, especially if they experienced limited or interrupted formal schooling in their home countries (Umansky et al. 2018), because of the sheer amount of time needed to develop English proficiency and learn academic content in English. Longitudinal research shows that students’ age of arrival to the US is negatively associated with their likelihood of becoming proficient in English (Conger 2009). However, other research cautions against positioning age and second language acquisition as a biologically absolute relationship, instead framing the relationship as one that is influenced by social and educational contexts (Stevens 2015; MacSwan and Pray 2005; Marinova, Marshall, and Snow 2000). For example, Callahan (2005) finds better academic outcomes for secondary LAEL students with uninterrupted education than for those without, a finding confirmed by Hodara (2015) among ELs in the community college system. Overall, heterogeneity among late-arriving ELs in home language proficiency and literacy could influence achievement growth once they enter US schools. Because we wish to learn more about how both student groups progress, and about the factors associated with their progress, we turn to the two largest school districts in the state. In 2017–18, the California Department of Education (CDE) reports that Los Angeles Unified School District had 47,000 ELs in grades 6–12, while San Diego Unified School District had 9,000 such students. Both districts also have high poverty rates (82% in Los Angeles and 60% in San Diego).6 The poverty rates for secondary ELs are higher than for students on average and are more similar across the districts than the overall poverty rates: 94 percent in Los Angeles and 90 percent in San Diego. Despite these similarities, other characteristics of the districts’ EL populations differ. Whereas school staff in both districts indicated that most long-term ELs are Spanish speakers who were either born in the United States or hail 5 Starting in 2013, California students who have been classified as EL for six or more years are termed LTELs, and districts are subject to new requirements to identify these students in order to direct services toward them (Assembly Bill 2193). See Section 313.1 of the Education Code. Later in the paper, we use a definition that more closely maps to each district’s own LTEL definition. 6 The poverty rate is defined as the percentage of students eligible for free or reduced-price meals, which translates into a family income under 185 percent of the federal poverty level ($46,435 for a family of four in 2018). PPIC.ORG Academic Progress for English Learners 7 from Mexico or Central America, they described their late-arriving EL populations differently. Specifically, LAELs in Los Angeles were described as primarily from Mexico or Central American countries, yet many SDUSD school staff noted serving refugees from Africa, the Middle East, and Southeast Asia. Regardless of students’ countries of origin, staff in both districts indicated that many LAELs had previously experienced limited or interrupted formal schooling, as well as considerable trauma, prior to their arrival in the US. In addition, staff in both districts, but especially LAUSD, indicated that some LAELs were expected to work at least part time in order to contribute to family expenses. Using 2017–18 data reported by CDE, we find that 93 percent of secondary school ELs in LAUSD are Spanish speakers, and some ELs also speak indigenous languages. The next most common languages in the district are Armenian and Filipino, with slightly more than 1 percent of secondary ELs speaking each of these languages. In SDUSD, only 76 percent of secondary ELs speak Spanish. Vietnamese (4%), Filipino (4%), Somali (3%), and Arabic (2%) are the next most common languages. No other language is spoken by more than 1 percent of secondary ELs in the district. Further, an important distinction in SDUSD is that the percentage in our secondary school sample whose home language is Spanish is 84 percent for long-term ELs but only 64 percent for latearriving ELs. Time Trends In our analyses, we follow long-term and late-arriving EL students during grades 6–12. When we are studying academic outcomes for ELs, we include students in our sample even after they are reclassified so that we can examine factors associated with their academic successes and struggles. If we excluded students after reclassification, we would lose more successful students from our sample and bias our conclusions about what is positively associated with EL student progress. For that reason, our counts of students differ from those reported by CDE. Further, state education code defines LTELs somewhat differently than they are defined in each school district. To align better with the districts’ definitions, our analyses define LTELs as secondary school students who have completed five or more years of schooling as ELs.7 Our quantitative analysis shows that both districts saw a decline in their numbers of long-term and late-arriving ELs from 2007–08 to 2015–16 (Figure 1). This decline outpaced overall declines in student enrollment. In LAUSD, EL enrollment at the secondary level declined 60 percent between 2007–08 and 2015–16 (compared to 9% for the district as a whole). In SDUSD, secondary EL enrollment decreased 46 percent in the same time period (compared to 2% for all district students). 7 The LTEL definition used in this report conforms to LAUSD’s own LTEL definition and also recognizes that SDUSD considers ELs who have completed five years of schooling as seriously at risk of becoming LTEL in one more year. The statistical analyses described in this report are little changed if we instead use the official definition that an EL must have completed six, rather than five, years of schooling before being considered an LTEL. Our study population includes EL students who become LTELs after grade 6. PPIC.ORG Academic Progress for English Learners 8 FIGURE 1 Los Angeles and San Diego secondary schools both saw declining enrollments of long-term and late-arriving ELs 90 LAUSD English Learners, thousands 80 Long-term EL 70 Late-arriving EL 60 50 40 30 20 10 0 2007-08 2008-09 2009-10 2010-11 2011-12 2012-13 2013-14 2014-15 2015-16 9 SDUSD 8 English Learners, thousands 7 Long-term EL 6 Late-arriving EL 5 4 3 2 1 0 2007-08 2008-09 2009-10 2010-11 2011-12 2012-13 2013-14 2014-15 2015-16 SOURCE: Authors’ tabulations from school district data. NOTE: Here and later in the report we define LTELs as ELs who have completed five or more years as an EL. Reclassified long-term and latearriving ELs are not included in this figure. We found that part of the decline in secondary EL students was linked to increased reclassification rates in elementary school. In SDUSD, the percentage of ELs in elementary school who were still classified as ELs in grade 6 fell from 62 percent in 2006–07 to 41 percent in 2015–16. In LAUSD, this percentage fell from 59 to 39 percent, from 2006–07 to 2015–16.8 8 We tested for changes in the characteristics of 6th-grade LTELs over our study period and found almost no changes in test scores and the share speaking Spanish at home. However, in LAUSD, we did find that while the number of 6th-grade special education LTEL students remained fairly constant, as a proportion of LTEL students, they more than doubled, going from 11 percent of all 6th-grade LTELs in 2006–07 to 25 percent in 2015–16. In SDUSD, both counts and percentages of special education students were stable over time. PPIC.ORG Academic Progress for English Learners 9 We observed year-to-year fluctuations in the numbers of late-arriving ELs in each district, though there was an overall decline across the study period. Notably, LAELs represented an increasing share of the EL population in both districts, especially between 2013–14 and 2015–16. In Los Angeles, the school staff we interviewed indicated that the district’s LAEL population in recent years has included many unaccompanied minors from Central America and Mexico, some of whom spoke the indigenous languages of those countries and for whom obtaining school transcripts was often difficult. In San Diego, school staff noted an increase in the district’s LAEL population from about 2012 through 2015, when there was increased federal support for refugee resettlement, yet said they had received far fewer LAELs, especially refugees from Muslim countries, since 2016. School-Level Patterns The ways in which districts serve long-term and late-arriving EL students may depend on how these students are distributed across schools. For example, if large numbers of these student groups attend just a few secondary schools in the district, then resources and support for English language development coursework may tend to focus on those particular schools, with less attention paid to schools with smaller EL populations, or vice versa. Previous research suggests that the concentration of non-English speakers in a school is a potential determinant of student outcomes for English Learners (Conger, Schwartz, and Stiefel 2011). Using propensity score matching, Callahan et al. (2009) find that greater concentrations of ELs are associated with stronger academic achievement, relative to lower EL concentrations. Bui (2013) similarly finds that larger shares of ELs in a school are associated with increased reclassification, less grade retention, and increased math scores, but not higher reading and language scores. In Table 1, we explore the variation across each district’s schools by type of EL student in grades 6–12.9 In Los Angeles, all schools have at least some ELs, and EL students make up 23 percent of the average school’s population. All Los Angeles schools have at least some long-term and late-arriving ELs as well. The share of LTELs ranges from 12 to 25 percent of students across schools, and the share of LAELs ranges from 3 to 5 percent. In San Diego, the average school’s shares of each student group are similar to those in Los Angeles, with ELs making up 23 percent of the average school population, but the ranges are much larger. The San Diego schools where ELs constitute fewer than 3 percent of students are atypical schools serving special populations. TABLE 1 Los Angeles and San Diego have similar shares of each EL student group across middle and high schools Share of LTELs (%) Share of LAELs (%) Mean Min Max Mean Min LAUSD 18.8 11.5 24.9 3.8 2.6 SDUSD 19.5 1.1 48.7 1.6 0 SOURCE: Authors’ tabulations from school district data, 2006–07 to 2015–16. NOTES: For LTELs, LAELs, and ELs, currently or in the past. Among schools in our sample. Max 5.3 22.4 Share of all ELs (%) Mean 22.9 23.0 Min 15.1 1.8 Max 32.7 71.6 Prior research finds that the particular home language spoken is associated with academic outcomes for ELs and reclassified ELs (Hill et al. 2014), as is fluency in the home language as measured at school entry (Salazar and Hayes 2010). Friesen and Krauth (2011) find that being in a school with many other Chinese-speaking ELs is beneficial for Chinese-speaking ELs but that for Punjabi-speaking ELs, higher concentrations of Punjabi speakers 9 We exclude charter schools not operated by the district. PPIC.ORG Academic Progress for English Learners 10 are associated with lower academic outcomes. An ethnographic study of adolescent Spanish-speaking EL students shows these students acquired English proficiency more quickly if they collaborated on academic tasks with bilingual Spanish speakers (Carhill-Poza 2015). The studies cited above suggest that the language composition of fellow students may have a link to the academic performance of English Learners. We hypothesize that the language composition of ELs could also be connected to the academic performance of students who were never ELs (either English-only speakers or those determined to be fluent in English when they enter the school). Thus we use three measures to examine the language spoken by students at individual schools: the percentage of all students who are English Learners, the percentage of all students who speak the same home language, and the homogeneity of languages spoken among EL students.10 How Are Course Assignments Implemented? Our focus on both long-term and late-arriving ELs—whose backgrounds and needs may be quite different—leads us to ask how these student groups progress and what sorts of “malleable” factors (i.e., factors that are controllable by district officials at least to some degree) are associated with their success. Some school districts may place these two student groups in the same courses with one another, yet their learning needs are different. Both LAUSD and SDUSD have district guidelines in place that outline differentiated English language development coursework for LTELs and LAELs, as is suggested by Olsen (2010) and Menken, Kleyn, and Chae (2012). We seek to understand whether and how LTELs and LAELs were assigned to their intended English language development courses. As noted above, ELD courses refer to both designated ELD and integrated English classes, as mandated by the California Department of Education in 2015. Overall, we find the implementation of course placement policies varied across the two districts, which was in part related to legal mandates and available staffing and resources. In both districts, correct course placement rates (i.e., those that followed district guidelines) were more than 50 percent for long-term and late-arriving ELs in recent years. Many of those not placed correctly were in ELD courses that were either too high or too low for their proficiency level based on district criteria. In interviews, school and district staff remarked on several challenges in implementing policies depending on the proportion of students who are ELs at the school, as well as the size of the school. District Policies Los Angeles Unified and San Diego Unified each have English Learner Master Plans (with the most recent editions adopted in 2012 in LAUSD and 2015 in SDUSD) that outline district policies and practices for English Learner instruction.11 ELs in secondary schools are assigned to ELD courses according to the principles outlined in the master plans, with supplemental memos published nearly every year that rearticulate course assignment criteria. We used all available data to calculate our best estimates of whether EL students were placed correctly, according to each district’s master plan and supplemental memos. Technical Appendix B describes how we employed the course placement criteria in our coding. Of course, there are times when teachers and staff exerted their discretion on placement decisions and those choices are impossible for us to code in the data. 10 We measure language homogeneity by computing a Herfindahl index, which ranges from zero to one. A value of zero means that no ELs speak the same home language and a measure of one means that all ELs speak the same home language. See Technical Appendix D for a more detailed description. 11 LAUSD’s new English Learner Master Plan was completed in spring 2018, after our last year of data collection. PPIC.ORG Academic Progress for English Learners 11 In LAUSD, secondary EL students were assigned to designated ELD courses that focus on developing English proficiency and were thus intended to be taught separately from non-EL students. Additionally, ELs who had been in the US two or more years were required to be enrolled in a grade-level-appropriate English language arts class with integrated ELD support. Prior to 2013–14, LAUSD assigned students to ELD courses based on test scores and whether or not they passed previous ELD courses. Starting in 2013–14, the district began assigning and advancing EL students to ELD courses based on the number of years spent in US schools. The district also adopted two designated ELD courses for long-term ELs in 2013–14 and mandated that ELs who had been in US schools for five or more years be placed into at least one of these courses. These two courses (Literacy and Language, and Advanced English Language Development) carried high school graduation credit (and could be counted toward a–g credits, which are required to be eligible to apply to UC and CSU); placement was based on LTEL status and reading level. SDUSD’s English Learner Master Plan also provides guidance for secondary ELs’ assignment to designated and integrated ELD coursework, and these guidelines have changed over time. From 2005–06 to 2007–08, the plan stated that secondary ELs enrolled in the district for less than one year should be placed in LAEL-specific ELD courses; all other ELs were to be placed in ELD courses based on their level of English language proficiency, as demonstrated by their scores on a standardized English language development test. From 2008–09 to 2015–16, all ELs were placed in ELD coursework based on two criteria: length of enrollment in US schools and level of English language proficiency. Our interviewees confirmed that length of enrollment tended to take precedence, and ELs were often advanced to the next designated ELD course regardless of their English proficiency level, an approach that was similar to LAUSD’s. Central office staff in San Diego indicated that this recent push to move ELs through the designated ELD sequence was more common in high schools than in middle schools, so that ELs would have quicker access to academic coursework that carried graduation credit. In 2015–16, an academic language development (ALD) course was added to SDUSD’s English Learner Master Plan (though it was piloted as early as 2012–13) as a designated ELD option for LTELs enrolled for five or more years and who were performing below grade level on standardized tests. Unlike in Los Angeles, where similar courses carried credit toward graduation, San Diego’s ALD course was an elective designed to provide supplemental support to LTELs in acquiring the academic language skills needed to access content in all of their coursework, not just English; thus concurrent enrollment in a grade-level English language arts course was also required. Course Placement Rates In Figure 2, we report correct course placement rates for current long-term and late-arriving ELs; that is, the extent to which these students were placed in appropriate ELD courses as outlined in district guidelines. In San Diego, LTELs and LAELs were correctly placed in appropriate ELD courses at similar rates, between 50 and 60 percent beginning in 2008–09. In Los Angeles, the correct course placement for LTELs hovered around 40 percent until 2013–14, when course placement rates peaked at 64 percent before declining to 50 percent the following year. LAELs in Los Angeles had higher rates of correct course placement across the period than LTELs or other ELs who were not yet LTELs (not shown), although the correct LAEL placement rate declined slightly in the final year of our data. In both districts, the sharp increases we observed in correct course placement rates (in 2008–09 in San Diego for both groups and in 2013–14 in Los Angeles for LTELs) correspond to the years in which placement criteria changed to include the number of years in US schools as the primary placement criterion, rather than test scores or English language proficiency level, as described above. We identify additional reasons for these observed changes, in LTEL placement rates in particular, in the next section. PPIC.ORG Academic Progress for English Learners 12 FIGURE 2 Long-term and late-arriving ELs are correctly placed in ELD courses at similar rates in San Diego, but there is more of a gap in Los Angeles Correct ELD course placement 90% LAUSD 80% 70% 60% 50% 40% 30% 20% 10% 0% 2007-08 2008-09 2009-10 2010-11 2011-12 2012-13 2013-14 2014-15 2015-16 Long-term EL Late-arriving EL Correct ELD course placement 90% SDUSD 80% 70% 60% 50% 40% 30% 20% 10% 0% 2007-08 2008-09 2009-10 2010-11 2011-12 2012-13 2013-14 2014-15 2015-16 SOURCE: Authors’ tabulations from school district data. NOTE: Estimates combine initial and subsequent course placement. Reclassified ELs are not included. Long-term EL Late-arriving EL Figure 3 shows the percentages of long-term EL students in each district who were placed into ELD classes at too high or too low a level, who were placed correctly, and who received no ELD course at all. We find that for the most part, more than half of LTEL students were enrolled in ELD coursework—an important exception is LTEL students in Los Angeles prior to 2013–14. But many were in course levels that were higher or lower than dictated by district policy. In the final year of our data, 5 percent of LTELs were in ELD courses that were either too high or too low in San Diego, and the same was true for 25 percent of LTELs in Los Angeles. Twenty-one percent of LTELs in Los Angeles and 39 percent of LTELs in San Diego were in no ELD course in 2015–16. Earlier in the study period, over half of long-term ELs in Los Angeles received no ELD courses, and the rate in San Diego was only about a third. PPIC.ORG Academic Progress for English Learners 13 We also observed a few interesting patterns for LTELs’ ELD course placement rates by grade level (not shown), particularly in San Diego. In SDUSD, 94 percent of 6th-grade LTELs were enrolled in ELD coursework in 2015–16, whereas more than 90 percent of 7th and 8th graders were not. In grades 9–12, over 80 percent of LTELs in San Diego received some form of ELD instruction. FIGURE 3 The share of long-term ELs placed into ELD courses that were too high or too low has risen in Los Angeles in recent years, but it has stayed steady in San Diego Course placement rates LAUSD Long-term ELs 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 2007-08 2008-09 2009-10 2010-11 2011-12 2012-13 2013-14 2014-15 2015-16 Correct Too High Too Low No ELD 90% SDUSD Long-term ELs 80% Course placement rates 70% 60% 50% 40% 30% 20% 10% 0% 2007-08 2008-09 2009-10 2010-11 2011-12 2012-13 2013-14 2014-15 SOURCE: Authors’ tabulations from school district data. NOTE: Estimates combine initial and subsequent course placement. Reclassified ELs are not included. 2015-16 Correct Too High Too Low No ELD PPIC.ORG Academic Progress for English Learners 14 Roughly 10 percent of late-arriving ELs in both districts received no ELD coursework in 2015–16 (Figure 4). In Los Angeles, correct placement rates for LAELs were relatively high over the entire period, but rates have been declining somewhat in recent years. In 2015–16, about 15 percent of LAELs were in courses that were either a higher or lower level than suggested by LAUSD policy. In San Diego, LAEL students were often placed in courses that were too high (33% in 2015–16), based on district guidelines. Looking at course placement for LAELs by grade level (not shown), we found that there was low correct course placement for LAELs in grades 7 and 8 in San Diego (similar to LTELs in the district). In Los Angeles, high school LAELs were less likely to be placed in ELD courses than LAELs at the middle school level, and that was especially true for 12th graders (29% of whom were not in ELD). FIGURE 4 Late-arriving ELs in San Diego were more likely to be placed into ELD courses that were too high, compared to in Los Angeles Course placement rates 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% LAUSD Late-arriving ELs 2007-08 2008-09 2009-10 2010-11 2011-12 2012-13 2013-14 2014-15 2015-16 Correct Too High Too Low No ELD Course placement rates 90% SDUSD Late-arriving ELs 80% 70% 60% 50% 40% 30% 20% 10% 0% 2007-08 2008-09 2009-10 2010-11 2011-12 2012-13 2013-14 2014-15 2015-16 SOURCE: Authors’ tabulations from school district data. NOTE: Estimates combine initial and subsequent course placement. Reclassified ELs are not included. Correct Too High Too Low No ELD PPIC.ORG Academic Progress for English Learners 15 District Differences in Course Placement Implementation Our interviews suggest that the differential course placement rates we observed between the two districts for LTELs and LAELs may be attributed, at least in part, to differences in district policy, staffing, and resources that shaped the implementation of course placement guidelines. It is important to note, however, that our interviews took place in 2017–18, whereas our quantitative data end with the 2015–16 school year. Though some interviewees described prior implementation processes, most discussed changes to course placement implementation over the last few years; thus, our interview data are limited in the extent that we can explain changes in course placement trends over time. Generally, our findings suggest that, whereas LAUSD’s course placement processes have become more centralized and streamlined over the last few years, with additional resources and staffing put in place to support implementation, SDUSD has afforded schools more flexibility with course placement implementation in the context of declining overall resources, and thus there has been more limited EL support staff in the district. One example of this difference between districts relates to the implementation of designated ELD courses for LTELs. Insights from Los Angeles Unified School District In 2013–14, we observed an increase of almost 20 percentage points in correct placement for long-term ELs in Los Angeles (and a corresponding 20 percentage point decrease in the number of LTELs receiving no ELD; see Figure 3). That same year, LAUSD began implementing a voluntary agreement with the US Department of Education’s Office for Civil Rights (OCR) to improve ELs’ access to ELD courses. Prior to the agreement, designated ELD courses for LTELs were not mandatory, as noted by a district administrator: “They were called EL Language Strategy classes, and they were optional. So, if a school chose to include them on the master [schedule] as an elective for ELs to take, that was great. But it was up to the site.” Given that these designated ELD courses were not yet mandatory, LTELs tended to receive only integrated ELD. “Prior to LTEL courses, students . . . finished [designated] ELD 4 and classified as Preparing to Predesignate (PRP), and continued to get support through sheltered instruction in their content classes [i.e., integrated ELD],” said a district administrator. As a result of the OCR agreement, however, the district developed two designated ELD courses for LTELs and mandated their implementation starting in 2013–14. “The OCR did a review and the district came to an agreement to provide additional support with [designated ELD] LTEL classes. Now, LTEL courses are mandatory,” noted a district administrator. In addition to augmenting the district’s designated ELD course offerings, several interviewees in LAUSD noted that, in recent years, ELD course placement was heavily monitored by district-level staff. The district recently implemented a new data management system and EL Dashboard to help with course placement monitoring, identifying whether ELs were placed in the appropriate courses and their progress toward reclassification. Some interviewees described the usefulness of the new data system and dashboard in streamlining the placement process for all ELs. Further, district-level positions, such as data specialists and coordinators, were created to facilitate communication with and support to schools related to course placement. Communication tools included data reports with information related to ELs who were either not enrolled in ELD or were incorrectly placed, as noted by a high school principal: “The district does a pretty good job of forcing you to comply. [We get] emails saying that you’re out of compliance, that there are students who are not taking their [required] courses because sometimes a couple of kids slipped through the cracks.” Although these LTEL courses were described as mandatory, and resources and staff were put in place to support course placement, our quantitative analysis showed that about 21 percent of LTELs in LAUSD were not enrolled in any ELD courses as of 2015–16, the last year of our data (see Figure 3). Indeed, exceptions to the district’s PPIC.ORG Academic Progress for English Learners 16 LTEL policy were described by a central office staff member, who suggested that LTELs in the latter years of high school might be placed in academic content courses instead of designated ELD.12 This staff member noted, “At the end of the day, if I have an 11th grader who has still to finish the a–g [high school graduation] requirements, I’ve got to take them out of that LTEL class. And I’ve got to give him . . . his graduation requirements. I will make the annotation to the district. . . . I’m sorry, but the a–g weighs more than the LTEL class, and that I got from the district as well as from training.” Although our data are limited to a small sample of schools, we did not hear from any interviewees in LAUSD that designated ELD was not offered when it should have been, according to district guidelines. Interviewees from SDUSD did indicate as much, as discussed below. Instead, Los Angeles schools with small EL populations (i.e., where ELs represented between 4 and 11 percent of the student body) tended to rely on double or multiple rostering in designated ELD courses. An EL coordinator described this process: “When we have classes where there are not enough ELs, especially in schools with small EL populations, those courses are usually double rostered. So, ELD 1, 2, 3, and 4 might be together [in the same class] because they may have only 18 ELs at the school.” Even though the courses may have been rostered in this way, student transcripts still reflected the appropriate ELD level, thus indicating correct placement. Insights from San Diego Unified School District In San Diego, correct course placement rates for LTELs remained relatively constant between 2008–09 and 2015–16, with as many as 39 percent of LTELs enrolled in no ELD courses (see Figure 3). Although central office staff indicated that the district developed a designated ELD course (the academic language development course, or ALD) for LTELs as early as 2012–13, it was an elective course that did not carry a–g credit, and its implementation was not mandatory. As such, when LTELs in San Diego completed the district’s designated ELD course sequence, they were meant to be placed in mainstream credit-bearing English classes with integrated ELD supports. A central office staff member described LTEL course placement in the district: “The majority of our [ELs], when they finish our ELD sequence, are not reclassified, so we have all of our LTELs who aren’t reclassified in mainstream [English] classes. They’re not in our ELD sequence, and they may or may not have access to an ALD class.” Of the San Diego schools in our interview sample, only one offered an ALD course in 2017–18. School leaders at our eight other interview sites opted not to offer the course, either due to budgetary constraints and/or because they did not have enough LTELs to offer a course. Three principals stated that they had eliminated ALD at the beginning of the 2017–18 school year as a result of districtwide budget cuts, including this middle school principal: “We got a lot of funds cut from the school, so there’s no way I could offer a class for seven [LTEL students].” Overall, these findings suggest that integrated ELD has been the primary instructional approach for LTELs in San Diego. In contrast with Los Angeles, where the district undertook efforts to centralize data management and streamline support across schools, San Diego school sites have been given more flexibility, at least in recent years, with respect to ELD instruction. A district leader described school flexibility in this way: “We’re giving schools autonomy; there’s no fixed route. We became more flexible in terms of how schools provide support. ELD 1/2 remained double blocks, but 3/4 are single periods. We added an ELD literacy block class for levels 1/2, 3/4, and 5/6, so if schools want their level 1s and 2s to have three periods of ELD, they can, or 3s and 4s can have two periods. But, if schools believe their population can take it, they don’t have to do that. We also added advanced coursework/pathways, like AP Spanish Language and Literature, so some students can go onto Honors if they’re ready. We’re encouraging schools to really look at their populations and work with us to determine their pathways.” This school-level 12 Although we allowed for such placements to be considered “correct” in our quantitative analyses. PPIC.ORG Academic Progress for English Learners 17 autonomy was evident with respect to the offering of ALD courses for LTELs described above, though principals indicated making decisions based on resources, not necessarily on the needs of their LTEL populations. Our qualitative findings also suggest that, in the context of budgetary constraints, SDUSD schools may have been given even more control over course placement in the last few years. Amid declining resources, district-level EL support staff positions were reconfigured or cut altogether, and staff in several of the schools in our interview sample noted no longer receiving direct support from the district related to EL course placement. This trend was especially evident among schools that had small EL populations; schools with larger EL populations either still received some level of district support (two schools) or used site resources to maintain the level of support they received previously (one school). Relatedly, we found that the four schools in our sample that served small EL populations could not afford to continue offering designated ELD classes and opted not to do so, as one middle school EL coordinator noted: “We used to have ELD classes when we had more kids . . . then we didn’t have enough ELs to really justify it.” Similarly, a high school counselor shared, “There are no options [for designated ELD], so they just go into the mainstream classes just like everybody else.” While these mainstream courses presumably offered ELs integrated ELD, these findings suggest that designated ELD was not offered to either LAELs or LTELs in some San Diego schools. Nonetheless, some schools still attempted to offer alternative linguistic supports in the absence of designated ELD courses. For example, in one San Diego high school serving a small EL population, both LAELs and LTELs were placed in a “Spanish for Spanish speakers” course sequence, which included three levels and was followed by AP Spanish. The school principal described this practice as an attempt to address ELs’ linguistic needs in the absence of a designated ELD option at the school. Still, the teachers who taught these classes recognized that they were intended to facilitate fluency in Spanish, not English, as well as serve as the foreign language requirement for graduation. The practice of double or multiple rostering in LAUSD may help to explain why course placement rates for LAELs in Los Angeles were overall higher than in San Diego, especially if San Diego schools were not offering designated ELD. Nonetheless, even though interviewees in San Diego did not discuss double rostering, it may have been used in schools outside of our sample. What predicts a school’s level of correct ELD course placement? Our interviews with district teachers and staff in both LAUSD and SDUSD suggest that correct course placement is made more challenging when there are fewer EL students at a school. Here we report on whether those findings are consistent with our student-level data. In each school district, we examine the relationship between each of our language environment variables (the percentage of ELs in the school, the percentage speaking the same home language among all students, and the degree of language homogeneity among EL students). We also examine the relationship between correct course placement at the school level and the number of EL students per grade level. In LAUSD, we find that as the number of EL students per grade increases, correct placement rates increase as well, but the number of EL students per grade is not a statistically significant predictor in SDUSD (see Table D3 in the technical appendices). Somewhat surprisingly, in both districts, as the percentage of all students speaking the same home language rises, correct course placement falls. At least in the case of LAUSD, the number of ELs per grade is more important given both the size of the coefficient and the range of the variable. For example, an increase of one standard deviation in the number of ELs per grade is associated with a 4 percentage point increase in correct course placement, while an increase of one standard deviation in the percentage speaking the same home language among all students results in a 1 percentage point decrease in correct course placement. PPIC.ORG Academic Progress for English Learners 18 How Do Rates of Progress Differ for Long-Term and LateArriving English Learners? Prior research suggests that EL students’ progress may depend in part on their peers as well as on instruction. Before testing for the association of peer groups and course assignment with linguistic and academic outcomes, we examine performance of LAELs and LTELs across the two districts below. Long-term ELs are, by definition, students who have had a great deal of exposure to English language development instruction and to American schooling. Late-arriving ELs, in contrast, are students with no exposure to American schooling prior to their first year in LAUSD or SDUSD, and they may have had limited or varied amounts of schooling in their country of origin. Note that we further restrict our focus to late-arriving ELs who have very limited English proficiency (scoring at the lowest level on California’s test of English language proficiency). On average, long-term EL students have higher test scores than LAEL students. But, late-arriving EL students appear to make more progress on average in improving their test scores on English language proficiency tests and California Standards Tests on both English language arts and math from one year to the next, and in closing the gap with long-term EL students. English Proficiency Unsurprisingly, we find that LTELs have substantially higher scores on the California English Language Development Test (CELDT) than LAELs (Figure 5). The CELDT is a state-mandated English proficiency assessment of listening, speaking, reading, and writing administered annually to all current ELs statewide as well as to all new enrollees who speak a language other than English at home, in kindergarten through 12th grade.13 But if students who report a home language other than English do well on the test they are deemed Initially Fluent English Proficient (IFEP), do not receive ELD supports, and are not tested again.14 Overall scale scores across grades 6–12 for the CELDT for late-arriving ELs in San Diego (455) are somewhat higher than the overall score for LAELs in Los Angeles (447). These scale scores are translated into overall proficiency levels, which are in turn used to determine when EL students are ready for reclassification.15 These average overall scores mean that most LAELs are scoring at the beginning proficiency level in both districts.16 The average long-term EL score in LAUSD (567) places the average LTEL in the intermediate or early advanced level, depending on the grade level. The average score of 556 for LTELs in SDUSD places the average student squarely in the intermediate level, across all grades 6–12. 13 The CELDT identifies students with limited English proficiency, determines their levels of proficiency, and assesses progress in learning English. It must be administered within 30 days of enrollment 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 been reclassified as Fluent English Proficient. CELDT results provide performance levels—beginning, early intermediate, intermediate, early advanced, and advanced—for each subtest and an overall proficiency level (OPL). Until 2006, the listening and speaking subtests were combined. Starting in 2006–07, higher CELDT scores were required for each OPL. 14 Results from the CELDT given at a student’s entry determine if a student who speaks a language other than English at home will be classified as an English Learner or a student that is initially fluent in English. The CELDT was replaced by the English Language Proficiency Assessments for California (ELPAC) in 2017–18. 15 Scale scores that determine proficiency levels vary somewhat by grade level, with high scores required as students advance through the grades. See CELDT Initial/Annual Score Ranges. 16 Using the 2016–17 and 2017–18 CELDT guide, OPL scores for 9th graders (CDE 2016). PPIC.ORG Academic Progress for English Learners 19 FIGURE 5 In both districts, average CELDT scores are higher for long-term ELs than for late-arriving ELs 600 Mean CELDT scores 500 400 300 200 100 0 LAUSD SOURCE: Authors’ tabulations from school district data, 2006–07 to 2015–16. NOTE: Reclassified long-term and late-arriving ELs are included. SDUSD Long-term EL Late-arriving EL The gap between the average CELDT scores for long-term and late-arriving ELs is over 100 points at both districts. Table 2 illustrates that late-arriving ELs tend to close the gap with long-term EL students over time. LTELs gain about 20 points a year on the CELDT in both school districts, but LAELs gain substantially more (59 points in SDUSD and 38 points in LAUSD). A gain of 20 points is about one-third of the gain necessary to go from one proficiency level to the next, whereas a gain of 59 points is the approximate gain required to advance one proficiency level (CDE 2016–17; CDE 2016). Because LTELs start at relatively higher levels, it is likely more difficult for them to make the same annual gains observed among LAELs. TABLE 2 Late-arriving ELs have larger annual gains in English language proficiency than long-term ELs Average annual point growth in CELDT scores SDUSD Long-term ELs 19 Late-arriving ELs 59 SOURCE: Authors’ tabulations from school district data, 2006–07 to 2015–16. NOTE: Reclassified long-term and late-arriving ELs are included. LAUSD 21 38 In line with our quantitative findings, interviewees in both districts suggested that long-term ELs often struggle to move past intermediate or early advanced proficiency levels, despite having higher average scores. Many attributed these struggles to LTELs’ lack of motivation, given their limited success in passing the exam previously: “We have students who don’t take the CELDT seriously because they have done it for the last 10 years and they’re like, ‘Oh, here we go again, I’m not English proficient, but I don’t want to take this assessment,’” noted a SDUSD high school principal. Several interviewees expressed their view that some LTELs appeared to lack motivation in their classes and did not see the importance of improving their English proficiency PPIC.ORG Academic Progress for English Learners 20 in order to reclassify because they felt they already spoke English: “I noticed some students, they get bored with taking the same class if they haven’t reclassified and they feel as if they don’t need it because they’re already speaking English fluently,” said a LAUSD school counselor. Academic Performance in English Language Arts and Math Nearly all California school children are assessed annually on their mastery of English language arts (ELA) and math. Given that these assessments are administered in English, for EL students they also implicitly measure English proficiency. In 2012–13 and in previous years, the state used the California Standards Test (CST) as its standardized assessment; since 2014–15, the state has administered the California Assessment of Student Performance and Progress (CAASPP, sometimes referred to as the Smarter Balanced assessment).17 We begin by examining performance using the CST’s English language arts test from 2006–07 to 2012–13.18 We then examine scores on the CAASPP in English language arts and math from 2014–15 to 2015–16. Our analysis of the CST for English language arts (Figure 6) shows that LTEL students have higher scores than LAEL students, but both English Learner groups score substantially below students who were never ELs. This figure transforms the five proficiency levels of the CST (far below basic, below basic, basic, proficient, and advanced) into a scale from 1 to 5 and takes the average. Figure 6 also shows lower scores for never ELs in Los Angeles relative to in San Diego. Average proficiency levels for long-term ELs are similar between the two districts, therefore, the performance gap between never ELs and long-term ELs is larger in San Diego (1.33 performance levels) than the gap in Los Angeles (0.85 performance levels). The proficiency levels for late-arriving ELs are lower in Los Angeles than in San Diego; the gap between LAEL and LTEL students is three-quarters of a performance level in Los Angeles, whereas it is only half a performance level at San Diego. FIGURE 6 Both long-term and late-arriving EL students score below never ELs in English language arts 4.5 Mean CST ELA Performance Level 4.0 Never EL 3.5 Long-term EL 3.0 Late-arriving EL 2.5 2.0 1.5 1.0 0.5 0.0 CST ELA CST ELA LAUSD SDUSD SOURCE: Authors’ tabulations from school district data , 2006–07 to 2012–13. NOTES: Reclassified long-term and late-arriving ELs are included. We convert raw CST scores into proficiency levels and average them. 17 During 2013–14, the state did not administer a test of basic skills. The new CAASPP is linked to California’s new Common Core State Standards. 18 We do not examine CST math scores because students took math subject tests (i.e., algebra, geometry) rather than grade-level tests, making comparisons by grade level or age, at least above grade 7, very complicated. PPIC.ORG Academic Progress for English Learners 21 To show changes in achievement across grades, we convert the raw CST test scores into a standardized score (known as a “z-score”) that equals 0 for students at the state mean, 1 for students one standard deviation above the state mean, and so on. Figure 7 shows the mean of these standardized scores by grade level. As with the CELDT, it appears that LAEL students make more progress in test scores during their time in the school district than do LTELs. Although this is a snapshot from 2012–13 and does not follow individual students over time, this figure nonetheless hints at the relative progress of students in the various language groups as they grow older. Never ELs make small (or no gains) relative to the mean performance of all students statewide from one grade level to the next. In Los Angeles, LAEL students narrow the performance gap with LTEL students across grades. In San Diego, we also see LAELs catching up to LTELs, but the gap varies more grade by grade than in Los Angeles. FIGURE 7 Late-arriving ELs narrow the performance gap with long-term ELs in English language arts between grades 6 and 11 0.50 LAUSD Mean CST ELA z-score 0.00 6 7 8 9 10 11 -0.50 -1.00 -1.50 Never EL Long-term EL Late-arriving EL -2.00 Grade SDUSD 1.00 0.50 Mean CST ELA z-score 0.00 6 7 8 9 10 11 -0.50 -1.00 Never EL Long-term EL Late-arriving EL -1.50 -2.00 Grade SOURCE: Authors’ tabulations from school district data. NOTE: These test scores are taken from 2012–13, the last year of the CST. We standardize the test scores across the two districts by using state means and standard deviations for the CST ELA by year. Thus a score of 0 means that students are at the state average, while a score of 0.5 would mean that students are half a standard deviation above the state mean in a given grade. The lines are simulated cohorts of students rather than students being followed over time. EL students remain in their original language status group even after reclassification. Reclassified long-term and late-arriving ELs are included. In SDUSD, in 2012–13, almost no late-arriving ELs took the CST. PPIC.ORG Academic Progress for English Learners 22 Prior research (e.g., Salazar and Hayes 2010) has pointed to the importance of initial proficiency in students’ own home languages in determining what progress they will make in learning English. Upon entering school, students take a language proficiency test to gauge their literacy in their own home language, the results of which we distill into “fluent” or “not fluent.” In Figure 8 we present a comparison of individual students’ year-to-year growth in CST ELA scores, but this time distinguishing by home-language fluency level for LTELs and LAELs. Among students who are fluent in their home language, average growth in CST scores is greater compared to students who are not fluent in their home language upon entering the school. However, the difference is only substantial in SDUSD. Among LTELs in San Diego, students who were initially fluent in their home language have faster gains on the test of English language arts, but in Los Angeles, initially fluent and non-fluent students have the same growth. Among LAELs, initial fluency in their home language is much more important for growth in San Diego than in Los Angeles. We should note, however, that the approaches to assessing home language fluency differ in each district. FIGURE 8 EL status is a more consistent factor in growth in ELA scores than home language fluency, particularly in Los Angeles 0.20 CST-ELA, with home language fluency 0.15 LAUSD SDUSD 0.10 Growth in CST-ELA z-score 0.05 0.00 -0.05 Never EL Fluent Not Fluent Long-term EL SOURCE: Authors’ tabulations from school district data, 2006–07 to 2012–13. NOTE: Reclassified long-term and late-arriving ELs are included. Fluent Not Fluent Late-arriving EL Results on the California Assessment of Student Performance and Progress show a similar pattern for English language arts. The CAASPP has only been administered for four years, and our data include two of those years (2014–15 and 2015–16 school years). As with the CST, we find that never ELs have higher ELA scores than LTELs, who in turn have higher scores than LAELs. In our simulated cohorts of students, where we compare scores by grade level at one point in time, all language groups have higher scores at higher grade levels, and the simulated gains are greatest for LAELs.19 Because the CAASPP administers math tests by grade level and not subject, we can make comparisons across all the grade levels administered. These math scores reveal a slightly different pattern. In Los Angeles, LAELs have 19 The CAASPP is administered in grades 3–8 and grade 11. PPIC.ORG Academic Progress for English Learners 23 somewhat higher CAASPP scores than do LTEL students—but both sets of scores are substantially lower than those for never ELs. In San Diego, we see the same pattern in grades 6 and 7 but in grades 8 and 11—the two other secondary grades with CAASPP math testing—LTELs have slightly higher math performance than LAELs. Once again, both types of EL students had CAASPP math scores considerably below those of never EL students. Grade Point Average There are many other important academic outcomes to consider beyond standardized tests, and here we examine grade point average (GPA) for students in grades 9–12 (Figure 9). In both school districts, late-arriving ELs have higher GPAs than do long-term ELs, and both EL student groups have lower GPAs than never ELs. It is important to note that these two districts do not have standardized policies on grading for ELs. FIGURE 9 Late-arriving ELs have higher high school GPAs than long-term ELs GPA 3.0 2.5 2.0 Never EL Long-term EL Late-arriving EL 1.5 1.0 0.5 0.0 LAUSD SDUSD SOURCE: Authors’ tabulations from school district data, 2006–07 to 2015–16. NOTE: Reclassified long-term and late-arriving ELs are included. Insights from Interviews In both districts, but especially in San Diego, the school staff we interviewed described an increasing focus on graduation rates and ensuring that ELs were enrolled in credit-bearing courses. As such, both LAELs and LTELs were sometimes placed in courses that exceeded their English abilities and often received low grades, such as Ds and Fs. Indeed, some SDUSD teachers expressed a desire to be able to offer “no grade” in lieu of an F, especially for LAELs, because a failing grade did not reflect the gains they were making. Despite LAELs struggling in their coursework, staff indicated that, especially in recent years, ELs were rarely allowed to repeat designated ELD courses and instead were advanced to the next course in the sequence so they could more quickly move into credit-bearing classes. By SDUSD policy, any student failing a high school course is advanced to the next course and makes up the failed course in summer school or through the credit-recovery program. In Los Angeles, staff noted that LTELs were required to repeat LTEL-specific designated ELD courses until they passed them, with the exception of 11th graders who were waived from this policy so they could complete graduation course requirements. This practice of repeating courses may have resulted in student disengagement PPIC.ORG Academic Progress for English Learners 24 and underperformance, as a majority of LAUSD interviewees, and a modest proportion of SDUSD school staff, described LTELs’ boredom, lack of motivation, or poor behavior, especially when repeating designated ELD courses: “The biggest challenge has been behavior management, simply because LTELs have been in these type of classes for a while, and then they think, ‘Well, I'm not going to pass [the CELDT] again’ and then they just don’t try or give their best,” said an LAUSD middle school teacher. In addition, many school staff attributed LTELs’ lack of motivation and underperformance to the ELD curriculum for LTELs, which they described as repetitive and not engaging or interactive. In fact, California recently enacted two new bills that aim to help English Learner students gain more access to the core academic curriculum. The first, AB 2735, requires that school districts allow English Learners full access to all of a school’s standard instructional program, including a–g courses as well as AP classes. It does grant districts an exception for newcomer students. The second, AB 2121, requires districts to permit a fifth year of high school for newcomers and students in the Migrant Education Program in order to provide more opportunities to earn required graduation credits.20 Credit recovery through summer instruction (Johnson 2018) or online learning may provide another option for English Learner students to make up for course periods taken up by ELD. Are School Language Environments Related to EL Student Progress? Below, we examine student progress in more detail, with an eye toward understanding which factors are predictive of student success. First we study whether the language background of peers at the school level is correlated with academic outcomes for English Learners, and if those correlations vary by whether the student is a long-term or a late-arriving EL. Taken as a whole, we did not find consistent relationships between the language environment in the school and various academic outcomes. Certainly there were exceptions, but for the most part we found that the academic progress of English Learners seems quite adaptable to changes in the language environment of the school. In a similar vein, an important finding of this analysis concerns students who were never English Learners. The percentage of the student body that is EL bears no significant negative relation to any of the outcomes we modeled for this group (English and mathematics performance on the former state test), and in the case of LAUSD, it was positively associated with English. Overview of Our Approach We pose three main questions regarding each school’s language environment:  Is academic progress correlated with the percentage of students who are EL?  Is academic progress correlated with the percentage of all students who speak the same home language as the student in question?  Is academic progress correlated with the degree of homogeneity of languages spoken by ELs? 20 The Migrant Education Program is a federal program intended to facilitate keeping the children of migrant workers continuously enrolled and progressing in US schools, despite frequent moves. PPIC.ORG Academic Progress for English Learners 25 In this section and the subsequent section on course placement and ELD programs, we measure a variety of academic outcomes in order to understand what is correlated with LTEL and LAEL student success. Our analyses are not causal, but we do use linear regression models to take many factors into account at the same time. We study whether there is a relationship between the above language and ELD course measures and six outcomes. Three outcomes compare changes in annual test scores in math and reading (the California Standards Test) and GPA between EL students and never ELs.21 For ELs alone, we also study their progress on the state’s annual test of English fluency, the California English Language Development Test. The final two outcomes measure events that can only happen once in a student’s career, if they happen at all: reclassification and on-time graduation. The California Standards Test scores are available in math and English language arts for students in grades 2 through 11 from spring 2002 through spring 2013. By converting all CST scores into z-scores using statewide means and standard deviations, these variables compare student performance with a statewide norm. When modeling changes in z-scores for individual students, we are measuring their movement in the statewide distribution of achievement. We use ELA scores for grades 6–11 and math scores for grades 6 and 7.22 Reclassification of EL students in our sample can happen anytime between 6th and 12th grade. Graduating on time is estimated for 12th graders.23 Those who left the district before grade 12 (and in the case of San Diego were not known dropouts) were excluded as we lacked information on these former students’ graduation or reclassification statuses after they left the host districts. For each outcome we model, we take into account student demographic characteristics (e.g., gender, parental education, special education status, and home language) and the student’s performance on the same measure in the prior year. The results of the regression analyses included here and in the next section on ELD course placement and programs are from models that also allow for district-wide nonlinear trends in student performance and that interact grade and school indicators. This means that any correlation we find between a language environment or EL course measure and student progress has already taken into account the student’s past performance and demographic measures, district-wide trends, and differences between each combination of school and grade level that are assumed not to change over time. Technical Appendix D explains our regression analyses in greater detail. The language environment could affect the learning environment for different EL types in varied ways. Further, prior research has suggested that an increased percentage of ELs who speak languages other than English in the school could have negative impacts on students for whom English is their first language or for reclassified students. For both of these reasons, we allow the correlation of these “language environment” variables with a given measure of academic progress to vary between never ELs, LAELs and LTELs. While we do not present results here, we tested each language variable independently before including them as explanatory variables simultaneously.24 Annual Outcomes Our results for the school language environment suggest a few points of agreement across the two school districts:  The percentage of ELs at a school has no negative statistical link to academic outcomes for never EL students. 21 In SDUSD we measure GPA from grades 6 through 12, while in LAUSD we model GPA for grades 9 through 12. 22 Starting in 8th grade, students take CST math tests specific to the courses in which they are enrolled, making their CST math scores impossible to compare. 23 In SDUSD, when we predict graduation, we are able to include students who have dropped out. These dropout data are not available in LAUSD. 24 Language variables that were not significant independently but were significant when modeled simultaneously are highlighted in gray in Tables D4A and D4B the technical appendices. PPIC.ORG Academic Progress for English Learners 26  As the percentage of students in the school who speak the same home language as the student in question rises, never ELs have more positive growth in test scores and GPA.  Greater language homogeneity among ELs has little statistical relationship with academic outcomes, and there are no consistencies between the two school districts on this measure. The figures below illustrate these points of agreement and other findings that are more idiosyncratic to the individual districts. Each chart shows the predicted effect of changing from the 25th percentile to the 75th percentile for each language environment variable. In Figures 10–15, the charts only display statistically significant results, meaning no vertical bars are shown for factors that were not significant predictors for a given group of students. If the effect was statistically significant, the height of the bar shows the size of the predicted effect. The vertical axes are measured in standard deviations. The exception is GPA; in both districts the standard deviation for GPA was almost 1, so we did not convert the familiar GPA variable into standard-deviation units. Thus, a bar for GPA that is 1 unit high implies that a change from the 25th percentile to the 75th percentile in the given variable is predicted to increase GPA by a full grade point. The Percentage of English Learners Figure 10 shows the predicted effects of the percentage of students at the school who are EL on tests of ELA, math, and English proficiency, as well as GPA. It is worth studying the link between academic progress and the percentage of students who are EL because for ELs there are countervailing factors. For example, it might help ELs to have the percentage of ELs rise somewhat if it means the school reaches a critical mass of ELs and it can provide ELD classes specific to each EL’s needs. But a higher percentage of ELs could potentially make it easier for ELs to converse in a language other than English outside of class. The leftmost third of Figure 10 shows the percentage of ELs at the school is not significantly related to three of the academic outcomes for never EL students, and is positively related to CST ELA scores in Los Angeles. This is an important finding because it means that both districts appear to manage instruction such that having more English Learners in a school does not negatively influence learning outcomes for native English speakers and others who spoke English fluently when they first entered the districts.25 Turning to the middle part of the figure (LTELs), the correlation of the percentage of ELs with LTELs’ performance varies by outcome and district. In Los Angeles both math and ELA CST test scores for LTELs are positively related to the percentage of ELs in the school and in San Diego CELDT is negatively related. Finally turning to the rightmost part of the figure, LAELs in Los Angeles appear to improve their performance on the ELA test as the percentage of ELs rises. This is not the case in San Diego, where there is no such positive association and where performance on the CELDT falls. For both ELD course placement factors and language environment factors, an increase in the percentage of students who are EL could in theory either hinder or boost rates of learning in a school. We tested for non-linear effects of each of these variables, but we did not find any convincing evidence in favor of this idea. Technical Appendix D provides additional details. 25 A non-statistically-significant finding could mean that the variable really does not matter, or it could mean that the effect is estimated imprecisely. In the case of the percentage of ELs and the insignificant relation to outcomes for never ELs, the former seems like the more likely explanation. The estimates are very precise. Technical Appendix D explains this in more detail. PPIC.ORG Academic Progress for English Learners 27 Predicted effect of a change from 25th to 75th percentile in share EL at school FIGURE 10 Higher shares of ELs in a school are not negatively associated with academic outcomes for never ELs 0.10 0.05 0.00 -0.05 CST ELA Z-Score CELDT Score CST Math Z-Score Annual GPA -0.10 -0.15 -0.20 LAUSD SDUSD LAUSD SDUSD LAUSD SDUSD Never EL SOURCE: Author calculations. Long-term EL Late-arriving EL NOTES: Chart shows the predicted effect of a change from the 25th percentile to the 75th percentile of ELs at the school level. Only statistically significant regression results (p<.05 or greater) are displayed. Reclassified long-term and late-arriving ELs are included. Full results found in Table D4A in the technical appendices. Interviews suggested that in both districts, schools serving larger EL populations faced particular challenges related to ensuring progress for late-arriving ELs. These challenges may be heightened in SDUSD, given that San Diego welcomes more refugees than any other county in the state, and the number of refugees entering the district has increased in recent years. However, in our quantitative analysis we found a negative association with higher percentages of ELs and outcomes for LAELs in only one case out of eight: CELDT scores for LAELs in San Diego. This negative association may be related to ELD teachers in San Diego feeling unequipped to provide the supports necessary to address LAELs’ extensive and varied academic and socioemotional needs, which they attributed to students experiencing interrupted formal education, trauma in their home countries or during their immigration journeys, or “culture shock,” as one ELD teacher described it. Several teachers characterized such needs as “desperate” and not “normal” compared to the LAELs they served in the past, and felt that these issues presented significant barriers to students’ adjustment, as well as challenges to their teaching. Although the topic of course placement was not explicitly referenced in this context, interviewees at schools with large EL populations in both districts advocated for newcomers, especially those with limited formal education, to have a separate course dedicated to their academic and social transition. Without these kinds of additional supports (which were sometimes offered, but have recently been eliminated in many schools in SDUSD due to lack of funding), it was challenging for students and teachers to achieve desired learning outcomes. The Percentage of All Students Speaking the Same Home Language The second variable of interest, the percentage of all students at the school who speak the same language at home as the student in question, is worth studying because a greater percentage of the student body speaking one’s own language may be an indicator of ELs’ ability to speak their home language even outside the ELD classroom, such as in other courses with never ELs or reclassified students. At the same time, it could allow the school to design ELD classes targeted to that student’s language group or English proficiency level. PPIC.ORG Academic Progress for English Learners 28 Figure 11 shows the relation between this variable and our four annual outcomes for the three groups of students. In cases where the percentage of all students speaking the same home language as the student in question matters, the association is almost always positive. For our EL student groups, having more students speaking their same home language does not appear to slow English acquisition. For never ELs, the three relevant outcomes (CST English language arts and math, and GPA) show a small positive relationship with the percentage of all students at the school who speak the same language at home as the student in question. Future analyses may test for whether this finding is positive just for English speakers among never ELs or if outcomes are positive even for never ELs who speak languages other than English at home. In San Diego LAELs had better performance on the three tests, but not GPA, when the percentage of all students speaking the same home language was higher, which argues against the existence of perverse effects of language enclaves for these students. Simulated change from 25th to 75th percentile in % speaking same home language FIGURE 11 Higher shares of students speaking students’ home language do not have a consistent negative association with test scores and GPA 0.37 0.32 0.27 0.22 0.17 0.12 0.07 0.02 -0.03 -0.08 CST ELA Z-Score CELDT Score CST Math Z-Score Annual GPA LAUSD SDUSD LAUSD SDUSD LAUSD SDUSD Never EL Long-term EL Late-arriving EL SOURCE: Author calculations. NOTES: Chart shows the predicted effect of a change from the 25th percentile to the 75th percentile in percent speaking the same home language. Only statistically significant regression results (p<.05 or greater) are displayed. Reclassified long-term and late-arriving ELs are included. Full results found in Table D4A in the technical appendices. We also asked school site staff about language enclave possibilities. Across the two districts, interviewees had varying beliefs. In particular, several teachers in SDUSD commented that their non-Spanish-speaking students, especially those from Asian and African backgrounds, were progressing much faster than their Spanish-speaking students because employing their home language to communicate within the classroom or larger community was not often possible, as it was for students whose home language was Spanish. At the same time, there was some acknowledgement among teachers in both districts that students who spoke the same home language could serve as language models and supports for each other, so language enclaves could also be beneficial to students’ language development. PPIC.ORG Academic Progress for English Learners 29 Language Homogeneity Among ELs Finally, we study the relation between language homogeneity among ELs and annual student outcomes to test the possibility that schools may be able to improve English fluency for ELs more quickly if the ELs at the school are linguistically similar. But higher homogeneity may also create an “enclave” effect within which English Learners feel quite comfortable conducting most of their school day in a language other than English. The measure of language homogeneity that we use, known as a Herfindahl index, can be interpreted as the probability that any two ELs picked randomly at a given school would speak the same home language. Figure 12 shows the predicted effect of an increase of the Herfindahl index from the 25th percentile to the 75th percentile. For example, for Los Angeles LTELs, this corresponds to a 16 percentage point increase in the probability that any two EL students picked at random in a school would share the same home language (ranging from a 84% chance to a 100% chance). For San Diego LTELs, the predicted effect corresponds to a 32 percentage point increase in the probability that any two EL students picked at random in a school would share the same home language (ranging from a 53% chance to an 85% chance). The predicted effect is statistically significant in only 7 of 22 cases, and the direction of the effects often varies across outcomes or between the two districts. We conclude, at least for these two districts, that there is not a systematic correlation between EL language homogeneity and annual academic outcomes for any of the three student language groups.26 A limitation of these results is that they are estimated imprecisely, meaning that the range of possible true effects can be quite large. This appears to be due to the very high mean Herfindahl index value (0.8 to 0.89 out of a possible maximum of 1 in LAUSD and 0.59 to 0.67 in SDUSD).27 We discuss this further in Technical Appendix D. 26 Specifically, in LAUSD, more EL student language homogeneity is associated with more positive growth on the CST ELA for both LTELs and LAELs. It is also associated with slower growth on the middle school math CST for never ELs. In SDUSD, never ELs and LTELs both have slower growth on the CST ELA the more language homogeneity there is among ELs. 27 Standard deviations range from 0.14 to 0.22. PPIC.ORG Academic Progress for English Learners 30 Predicted effect of a change from 25th to 75th percentile in language homogeneity FIGURE 12 There does not appear to be a consistent correlation between EL language homogeneity and annual academic outcomes 0.05 CST ELA Z-Score 0.00 CELDT Score CST Math Z-Score -0.05 Annual GPA -0.10 -0.15 LAUSD SDUSD LAUSD SDUSD LAUSD SDUSD Never EL Long-term EL Late-arriving EL SOURCE: Author calculations. NOTES: Chart shows the predicted effect of a change from the 25th percentile to the 75th percentile in language homogeneity. Only statistically significant regression results (p<.05 or greater) are displayed. Reclassified long-term and late-arriving ELs are included. Full results found in Table D4A in the technical appendices. Though we found no quantitative relationship between language homogeneity and outcomes, our interviewees, particularly in SDUSD where the population is more linguistically diverse, suggested that having a large range of home languages at a school presented instructional challenges. Several teachers in the district described employing translation tools and applications, such as Rosetta Stone, for students who spoke less common languages. Longer-Term Outcomes We also examined the relationship between the three measures of a school’s language environment and whether the given EL was ever reclassified as fluent in English as well as whether the given student graduated on time, defined as graduating within four years of first entering grade 9.28 It is more difficult to study whether school language environment matters for these longer-term outcomes than for annual outcomes. In the above results, we searched for a link between the language environment in a given school year and students’ annual performance in that same year. Here, we look for an association between language variables, such as the percentage of students who are EL in a given year, and outcomes that are typically observed several years later. The difficulty created is that we are simultaneously searching for a relation between school characteristics in several different years and a single outcome several years later. With a weaker temporal linkage, it might be more difficult to find statistically significant relationships even if they existed. Our results suggest a few points of agreement across the two school districts for the role of school language environment on these two longer-term outcomes (see full results in Table D4B in the technical appendices):  The percentage of students speaking the same language at the school as the student in question is associated with greater chances of graduating on time for never EL students. 28 As mentioned above, in LAUSD, we do not know which students have dropped out versus which students have moved to other districts. Students who have dropped out are included in the San Diego student data. PPIC.ORG Academic Progress for English Learners 31  Greater language homogeneity among ELs has little statistical relationship with reclassification or graduating on time. In LAUSD, we find that the percentage of students speaking the same language at the school as the student in question is also positively associated with ever being reclassified for LTEL students, but negatively associated with reclassification for LAELs. The percentage of ELs at the school was also positively related to reclassification for LAELs in Los Angeles. In SDUSD, higher shares of ELs are negatively associated with reclassification for both EL groups and negatively associated with graduation for LAELs. School language environment did come up as a potential predictor of reclassification in our school district interviews in SDUSD. Several interviewees described feeling that community context played a major role in reclassification. Interviewees suggested that because LTELs could readily communicate in Spanish at home, with peers in school, and throughout the community, they had few opportunities to develop their academic English skills beyond the intermediate level. Are Course Placement and ELD Programs Related to EL Student Progress? In this section, we examine the role of English Language Development course assignment, including specialized courses for LTELs and LAELs, and whether that correlates with linguistic and academic success for EL students. We also look at the relationship between EL support teachers and EL student outcomes in SDUSD. Overall, by far the most consistent and powerful finding was that in both districts when long-term ELs were not put into any ELD class, their gains on the CST English language arts test declined.29 Both districts have sizeable minorities of long-term ELs who receive no ELD class, and therefore our finding on the relation between not receiving an ELD class and test scores on the CST ELA test is quite relevant. This may have further implications, given pressures in both districts to ensure that students take a–g courses as soon as possible to meet the new graduation standards. While we found a few cases in which ELs being placed into an ELD class that was too low or too high seemed to matter, the results varied. Overview of Our Approach Our main instructional measure is whether ELs are correctly placed into ELD courses, but we also have some information on ELD course offerings and programs, and we study the link between these and student progress as well. As described in the previous section on course assignments, each school district has its own policies that determine how long-term and late-arriving ELs are assigned to the ELD coursework intended to improve EL students’ fluency and literacy in English. That section documents how each district has struggled to assign all EL students to their intended courses. Here, we examine the relationship to academic outcomes for students who are placed in ELD courses that are too low or too high, or that receive no ELD, according to district policy. It is important to note that these measures do not capture ELD pedagogy or how effectively districts and schools are implementing their ELD course designs. As above, we examine growth in CST ELA, CELDT, CST math, and GPA for students in each district, as well as whether students were ever reclassified and whether they graduated on time. 29 We found a similar result for LAELs in San Diego but not LAUSD. PPIC.ORG Academic Progress for English Learners 32 We model our academic outcomes based on the rates of low, high, and no ELD placement at the school level and compare our results to correct ELD course placement. While individual student outcomes may be more closely linked to their own placement, we noted earlier that each district allows teachers and administrators some discretion in EL course assignment. Therefore, if an LTEL student is not enrolled in an ELD course, it may be because it was determined that the student was ready to tackle the demands of an English-only course without additional ELD support. This student may have better academic outcomes than if the student had been placed in the “correct” ELD course. Using individual ELD placement may lead us to conclude that “no ELD” placement is better for long-term EL students than correct course placement were we to use students’ individual placement rates. Further, because our interviews suggested that the ability of districts to correctly place their EL students depends in some ways on the language composition of students at their schools, we model course placement while also controlling for our three school-level language variables (percentage of ELs, share of students speaking the same home language, and language homogeneity among ELs). Finally, we expect that course placement at the school level could differentially affect not only LAELs and LTELs, but also never ELs. If more EL students are correctly placed (as opposed to being without ELD), this may also affect academic outcomes for never ELs, who would share fewer courses with EL students. Similarly, if a greater share of ELs are properly placed into the ELD courses designed for them, this could be beneficial to both LTELs and LAELs, who would each be more likely to be enrolled in the courses specifically designed for them. Annual Outcomes As with our analysis of the school language environment, there is an important point of commonality across the school districts when it comes to the association with school-level ELD course placement and annual academic outcomes:  Higher rates of no ELD placement are associated with slower growth on CST ELA for long-term EL students. Figures 13–15 show the predicted effect of a change from the 25th percentile to the 75th percentile in the share of ELs at the school placed too low or too high, or not placed into any ELD course whatsoever. For example, among LTELs in LAUSD, the 25th percentile for “too low” placement is 1 percent, and the 75th percentile is 9 percent, so we simulate a change of 8 percentage points. Figures 13 and 14 show that often there is no significant link between placements at too high or low a level, and the exceptions vary in direction. However, a much more consistent picture emerges when we examine Figure 15, which shows the predicted effect of the change from the 25th percentile to the 75th percentile in the share of ELs not taking any ELD course. CST ELA scores are predicted to fall for LTELs in both districts, and in San Diego, never EL and LAEL students as well. Interestingly, CELDT scores largely do not appear to be affected, with the exception of a positive association for LAELs in San Diego. In both districts, LAELs’ GPA growth increases when rates of school-wide no ELD placement are higher. PPIC.ORG Academic Progress for English Learners 33 FIGURE 13 There seems to be no consistent association between the share of EL students placed too low in ELD courses and annual outcomes Predicted effect of a change from 25th to 75th percentile in share placed too low 0.35 0.30 0.25 0.20 0.15 0.10 0.05 0.00 -0.05 -0.10 CST ELA Z-Score CELDT Score CST Math Z-Score Annual GPA LAUSD SDUSD Never EL LAUSD SDUSD Long-term EL LAUSD SDUSD Late-arriving EL SOURCE: Author calculations. NOTES: Chart shows the predicted effect of a change from the 25th percentile to the 75th percentile in being placed in ELD coursework that is too low. Only statistically significant regression results (p<.05 or greater) are displayed. Reclassified long-term and late-arriving ELs are included. Full results found in Table D4A in the technical appendices. FIGURE 14 There seems to be no consistent association between the share of EL students placed too high in ELD courses and annual outcomes Predicted effect of a change from 25th to 75th percentile in share placed too high 0.1 0.08 0.06 0.04 0.02 0 -0.02 -0.04 -0.06 -0.08 -0.1 CST ELA Z-Score CELDT Score CST Math Z-Score Annual GPA LAUSD SDUSD Never EL LAUSD SDUSD Long-term EL LAUSD SDUSD Late-arriving EL SOURCE: Author calculations. NOTES: Chart shows the predicted effect of a change from the 25th percentile to the 75th percentile in being placed in ELD coursework that is too high. Only statistically significant regression results (p<.05 or greater) are displayed. Reclassified long-term and late-arriving ELs are included. Full results found in Table D4A in the technical appendices. PPIC.ORG Academic Progress for English Learners 34 FIGURE 15 Higher shares of EL students placed in no ELD courses are associated with lower English language arts scores Predicted effect of a change from 25th to 75th percentile in share not placed in ELD 0.1 0.08 0.06 0.04 0.02 0 -0.02 -0.04 -0.06 -0.08 -0.1 CST ELA Z-Score CELDT Score CST Math Z-Score Annual GPA LAUSD SDUSD Never EL LAUSD SDUSD Long-term EL LAUSD SDUSD Late-arriving EL SOURCE: Author calculations. NOTES: Chart shows the predicted effect of a change from the 25th percentile to the 75th percentile in being placed in no ELD coursework. Only statistically significant regression results (p<.05 or greater) are displayed. Reclassified long-term and late-arriving ELs are included. Full results found in Table D4A in the technical appendices. Our tentative conclusion is that ELs do not suffer systematically from placement into too high or too low an ELD course, but not placing an EL in designated ELD coursework (or “no ELD” as in our models) at all is associated with significantly lower gains for LTELs on the state’s test of English language arts. Given that some teachers in both districts referred to pressures, associated with the a–g graduation requirements, to place all students into ELA classes that earn a–g credit, our result raises some concerns for ELs when they are not enrolled in ELD courses. And despite the fact that it appears that the share of LAEL students within each district is fairly constant in recent years, about half of teachers commented that fluctuations in LAEL enrollment made it difficult to plan for courses and consistently secure teachers. Further, about half of all interviewed teachers described not feeling adequately prepared to address students’ non-academic needs, as one teacher in SDUSD stated, “Academically, emotionally, even health care . . . if you sit down [with] one of these kids . . . there [are] just so many issues that it’s hard to address every single one as a teacher.” Similarly, teachers in LAUSD indicated that LAEL courses required differentiated instruction to account for the varying amounts of prior schooling newcomers had in their countries of origin. As one teacher in LAUSD indicated, “So, [even when] they can read or write really well in Spanish, I [also] have students that have never been to school, so I have to deal with a student who was an excellent student back in El Salvador, who is an A student, super interested in learning and competent, and a student . . . who has no idea of what's going on, he barely knows how to read and write in Spanish. How in the world can you bridge that gap? What do you do with all these variety of students and all these needs? [It’s] really very hard.” Thus, even though LAELs might have been placed into the correct ELD courses, teachers in both districts may have struggled to address their needs. An additional issue raised in interviews was the suggestion that, even when students were assigned to designated ELD classes with the proper course title, they might also have students assigned to different ELD courses enrolled in the same class period. Interviewees in both districts described struggling to serve ELs at different English proficiency levels in the same classroom. At schools with low proportions of ELs in Los Angeles, low EL PPIC.ORG Academic Progress for English Learners 35 enrollment at specific language proficiency levels resulted in the consolidation of ELD courses (i.e., the double or multiple rostering described above). At schools with small EL populations in San Diego, ELs were often integrated into general education classrooms, and teachers indicated that their class sizes were becoming too diverse and too large to manage. In all cases, teachers struggled to accommodate ELs’ needs. In LAUSD, we were able to examine in our student-level data the extent to which LTELs and LAELs were placed in the same ELD and grade-level English language arts classrooms with students enrolled in different courses. We found that in more recent years, increasing proportions of LAELs in the district were in the same classroom with students enrolled in different courses (up from 10% in 2002–03 to 35% in 2015–16; see Figure B2 in the technical appendices). This did not appear to be an issue for LTELs in LAUSD during our study period. Unfortunately, we were not able to investigate this question in SDUSD.30 Longer-Term Outcomes In models of whether students graduated on time, we found two points of commonality between the districts (see full regression results in Table D4B in the technical appendices).  If a greater proportion of EL students are placed in courses that are too high for them, graduation rates are higher for never EL students.  If a greater share of EL students are not in ELD at the school level, graduation rates are higher for LAEL students in both districts. Otherwise, findings on longer-term outcomes are highly mixed. In Los Angeles, more EL students not enrolled in ELD is associated with lower graduation rates for LTELs; the opposite is the case in San Diego. In SDUSD, a higher share of ELs placed in ELD courses that are too low is associated with lower graduation rates for never EL students. Higher proportions of EL students placed too high is associated with lower graduation rates for LTELs. And no ELD placement in SDUSD is positively associated with graduation rates for never ELs and LTELs as well as LAELs (as previously described). In modeling the probability that ELs were ever reclassified, the results differed between the two districts to such an extent that we do not believe that there are clear implications for these two districts, let alone other districts.31 Specialized LTEL and Newcomer Courses Recall that EL course design, especially for LTEL students, changed in LAUSD with two new courses for LTELs, which were mandatory beginning in 2013–14. SDUSD piloted LTEL coursework in 2012–13 (called Academic Language Development, or ALD) and then formally introduced it as optional coursework in 2015–16. There are also special courses designed for LAELs (“newcomer courses”). In San Diego, these were historically courses that allowed LAELs in their first year of schooling to take specialized ELD and academic courses together with other first-year LAELs. More recently SDUSD has moved to enroll first-year LAELs in regular ELD and academic courses. Our interview data suggest this move was influenced by declining enrollment in the district and budget cuts, as well as an overall district philosophy that emphasized EL integration into general education courses. 30 The data required to determine which students were in the same classroom are not available for SDUSD. 31 In SDUSD, when an EL student was in a school where a higher share of ELs were placed in too low an ELD class, the probability of reclassification was lower. Similarly we found negative predicted effects when the share of ELs not placed into any ELD class rose. These negative outcomes related to ELD misplacement applied to both LTELs and LAELs. On the other hand, for LAELs but not LTELs, a higher share of ELs placed in too high an ELD class was associated with greater probability of reclassification. For LAUSD, in only three cases was the course placement variable statistically significant, and the signs were opposite those found in SDUSD. A higher share of ELs placed into too low an ELD class was positively linked to reclassification of ELs in LAUSD, and a higher share of ELs placed into too high an ELD class was negatively associated with reclassification for LTELs. PPIC.ORG Academic Progress for English Learners 36 To test for an association between specialized courses for LTELs and LAELs and the academic progress of each language group, we created indicators for whether any students had enrolled in LTEL or newcomer courses in a given school and a given year. We then interacted these indicator variables with indicators for whether the individual student was never EL, a LAEL or LTEL, and estimated a version of the earlier models that added these six interacted variables as additional explanatory variables. Because these specialized courses were available in both districts for only a subset of years, we focused on testing for a relationship between the provision of these courses and annual outcomes, rather than longer-term outcomes. Table D5 in the technical appendices shows the coefficients of our regression model in which we add indicators for any newcomer classes to the earlier model, which includes the language environment and ELD course placement variables. For SDUSD we also include the LTEL course indicators. For newcomer courses, we found no significant link to any of the four outcomes in Los Angeles, with two exceptions. LAELs in the district had significantly higher GPA if their school offered newcomer courses, but LTELs had slightly lower CST math scores. In San Diego, on the other hand, we found three significant associations, and all three were positive and sizeable. In San Diego district schools with newcomer courses, LTELs had higher CELDT score gains and higher CST math score gains, and never EL students had higher math score gains as well. We cannot know for sure if these results are causal, but one potential explanation is that both LTELs and never ELs gained academically when LAELs in their first year were taught most academic courses separately from other students, where teachers could more effectively differentiate their instruction. This is potentially important given SDUSD’s decision to roll back specialized programs for newcomers. In the case of courses for long-term ELs, we caution that our evidence base is very weak, so much so that we do not present quantitative results for Los Angeles.32 For San Diego, our analysis of ALD coursework for LTELs was only slightly more successful. ALD courses began to appear in transcripts only in 2015–16 and only in some schools, so we lacked the variation to obtain precise estimates. We found no significant effect of ALD courses in San Diego for CELDT scores for either EL group, or for GPA for any of our three language groups. That said, the best interpretation of these results is that we will need more years of data to answer with any certainty whether ALD courses “matter.” An additional factor to consider in Los Angeles is that the very high and increasing percentage of special education students among LTELs may have created a mismatch between course design and student need. Our interview data suggest respondents in both districts had concerns about the utility of these courses for LTELs. In Los Angeles, where schools offered the newly created designated ELD courses for LTELs, most interviewees suggested that students perceived these classes to be “boring” and in turn dismissed them. One middle school principal commented: “I think our kids need more hands-on . . . more [real-world] application. I find that they have a lot of . . . scripted things to do and it's not real to them and they get bored. They get bored. The prompts are not made for them. They need something they can connect to.” Similarly, an ALD teacher in San Diego said, “The way that that curriculum, and/or text is laid out, is that it builds on . . . it’s a lot of … very repetitive, and it’s constant repeat, or constant. . . a big part of what I’m trying to do is find that balance of how can we be repetitive without being boring.” Another teacher noted, “The ALD just wasn’t engaging enough, I couldn’t even make it 32 In LAUSD, our efforts to measure associations between the new LTEL courses and academic outcomes for LTELs, LAELs, and never ELs are frustrated by our measure of correct course placement. In LAUSD, all LTELs are supposed to be placed in one of two LTEL courses, either Advanced ELD or Language and Literacy. Our measure of correct course placement for EL students at the school level is largely a measure of correct course placement for LTELs, since LTELs are the majority of secondary ELs. Our models that measure both whether a school offers the LTEL courses and the share of EL students that are correctly placed are unable to produce results for the association between those two variables due to their colinearity. Instead, since the LTEL courses were implemented beginning in 2013–14, we estimated our standard model for just the 2013–14 to 2015–16 period. We find that the only change for LTELs from our main model is that higher percentages of EL students being placed too high is correlated with positive CELDT gains. We cannot estimate CST results because the assessment was last given in 2012–13. PPIC.ORG Academic Progress for English Learners 37 engaging enough.” Although the new designated ELD courses for LTELs were often described as either boring or poorly designed, we do not have classroom observational data to expand on this claim. English Learner Support Teachers in San Diego In San Diego, schools with large numbers of ELs were assigned EL support teachers (ELST), who worked with school administrators to facilitate ELs’ placement in appropriate coursework, monitored ELs’ progress in acquiring English fluency, identified students ready to reclassify, ensured that the reclassification process was correctly implemented, and supported classroom teachers who served EL students. This program began in 2004– 05 and slowly ramped up until the early part of this decade; due to financial constraints the number of EL support teachers has declined in the last five years. When asked why the number of these teachers in the district had dropped from 120 to 46 in recent years, a district leader we interviewed cited “resources and politics.” Most recently, the EL support teacher position has been eliminated and in March 2017 replaced with two new job titles: English language instructional resource teacher (ELI-RT) and English language development instructional specialist (ELD specialist). Rather than having one EL support teacher assigned to one to two schools, the district’s approach now involves having a resource teacher (ELI-RT) serve a cluster of about four to five schools. Under this new configuration, EL coordinators at school sites (who can be teachers or administrators) became the primary points of contact for the district. One district staff member described this approach as “haphazard,” going on to say, “We’re lucky if we have somebody to communicate with [at the schools]. . . . We have EL coordinators but you know once again so many are so brand new and they get bumped out of the position, and it’s like every year we feel like we’re training people all over again.” Overall, we find several strands of evidence that San Diego’s EL support teachers have historically been associated with better outcomes for ELs and never ELs as well. To assess whether the presence of EL support teachers in a school leads to better outcomes for either EL group or for never ELs, we added a measure of EL support teacher intensity to our earlier models, which already controlled for the school’s language environment and the nature of ELD course placements. Our measure is the ratio of EL support teachers (full-time equivalents) in a school to the number of ELs in the school in the given year. This ratio peaked at 0.01 in 2012–13 and declined to 0.008 in 2013–14.33 In the analysis below we discuss the predicted effects of increasing the EL support teacher to EL ratio from 0 to 0.01. To make the wording less awkward, we refer to this below as “implementation of ELST support.” We found that for both EL groups, implementation of ELST support is associated with significantly higher growth in GPA. The predicted effect was to increase GPA of late-arriving and long-term ELs by 0.15 and 0.05, respectively. CELDT scores were predicted to rise by about one-tenth of a standard deviation for LAELs, but the result was not statistically significant when we additionally controlled for the accuracy of ELD course placements. One interpretation of this is that having an EL support teacher in a school facilitated more accurate ELD placements, which in turn boosted LAELs’ performance. CST ELA scores were not affected. One other significant finding concerns never EL students, for whom math achievement was predicted to rise by 0.02 of a standard deviation after the implementation of ELST support.34 33 We were not able to locate information on EL support teacher school assignments after 2013–14, a period during which our understanding is that the number of EL support teachers declined, so we studied the school years 2006–07 to 2013–14. 34 Table D6 in the technical appendices provides details. PPIC.ORG Academic Progress for English Learners 38 We also examined the two longer-term outcomes, whether EL students were ever reclassified and whether they graduated on time. Two results were significant. Implementation of ELST support was associated with a 12 percentage point reduction in the probability that late-arriving ELs would ever be reclassified, but no impact on the probability that they would graduate on time. Turning to long-term ELs, implementation of ELST support was associated with no change in the probability of reclassification, but a 4.9 percentage point increase in the probability of graduating on time. Our interview data also positioned EL support teachers as central players in ensuring EL course placement aligned with district policy, as they facilitated the flow of information from the district to schools, as well as between schools when ELs matriculated. Although some communication still happened to facilitate articulation between middle and high school among staff at “the big feeder schools,” it was “not as systematized as it should be,” according to a resource teacher we interviewed. One district staff member also shared that support teachers were “hugely beneficial. . . . They helped us in making sure that kids were where they were supposed to be. . . . They just . . . added so much to the site in every respect.” The elimination of site-based EL support teachers thus may impact how well schools are able to follow course placement guidelines and the fidelity to the district’s EL Master Plan, which is designed to move students toward reclassification and graduation. Conclusion English Learners are a diverse group of students—including long-term ELs who have been educated in US schools, sometimes since kindergarten, and late-arriving ELs, some of whom arrive with a high degree of literacy in their home language and others with much less fluency. In the two districts studied in this report—Los Angeles Unified and San Diego Unified—school staff and administrators described difficulties in serving ELs in middle and high schools, a perspective that was echoed in our quantitative analysis. Teachers mentioned particular challenges in serving late-arriving ELs, with many having experienced trauma prior to arrival in the United States, interrupted schooling, and/or financial strain. In San Diego, district and school staff reported that many latearriving ELs were refugees; in Los Angeles, many were unaccompanied minors. Academic outcomes across these two student groups also vary. Our findings revealed that long-term EL students had relatively stronger academic outcomes on average. But unsurprisingly, given that progress in English proficiency is often more rapid at lower proficiency levels, late-arriving ELs made greater gains in the academic outcomes we measured. Policymakers have recognized the importance of improving outcomes for English Learners and have undertaken recent funding, instructional, and accountability reforms. Moving forward, it is critical that implementation and new policies consider the diverse linguistic and academic needs of EL students and how these factors contribute to their long-term success. Below we outline how some of our key findings can help guide policymaking. In our examination of the school language environment, we generally found that the academic progress of English Learners seems quite adaptable to factors such as the share of ELs in a school, the share of all students speaking the same home language, and the language homogeneity of EL students. Importantly, we found that students who were never identified as EL do not have slower rates of growth in schools with higher proportions of ELs. This finding suggests fluent English speakers are not negatively affected by having large numbers of ELs as school peers, at least in these two districts. PPIC.ORG Academic Progress for English Learners 39 Results were somewhat less consistent for the relationship between the percentage of ELs at a school and the academic outcomes of EL students. In Los Angeles, we found some evidence that these two student groups make faster academic progress in schools where a higher percentage of students are learning English. Notably, in our school interviews, EL instructional staff often noted that placing ELs in appropriate coursework was more challenging when the percentage of ELs at the school was lower. However, in San Diego, there is some evidence that longer-term outcomes (i.e., reclassification and on-time graduation) are negatively associated with higher shares of ELs at the school level. Another goal of this report was to examine how EL students are placed into English language development courses and whether course placement is associated with academic outcomes. In San Diego, both student groups were correctly placed (i.e., placed in accordance with district guidelines) between 50 and 60 percent of the time. In Los Angeles, long-term ELs were placed correctly 40 to 60 percent of the time; correct course placement for late-arriving ELs was higher, in the 70 percent range. Many of those not placed correctly were in ELD courses that were either too high or too low for their proficiency level. Trends in long-term EL students receiving no ELD instruction—in recent years this rate has declined in Los Angeles and risen somewhat in San Diego—may be influenced by district-specific factors. For example, in Los Angeles, the district has increased district-level oversight and support for EL students’ course assignments, at least partially in response to pressure from a settlement with the US Department of Education’s Office for Civil Rights. Meanwhile, in San Diego, the district leaves more discretion to school-level staff. Further, interviewees in San Diego schools serving small EL populations noted that in the most recent school year, most if not all ELD courses designed specifically for long-term ELs were eliminated at their schools due to districtwide budget cuts. Instead, ELs were placed in grade-level English courses, when possible with teachers whom administrators deemed well-prepared to differentiate instruction for ELs. These patterns are important because we found that having no specific ELD instruction is associated with slower growth on state tests of academic English for long-term ELs in both school districts, and for late-arriving ELs in San Diego. In addition, many of the school staff we interviewed believed that appropriate ELD supports were crucial for EL progress. At the same time, some staff noted that placing EL students in courses required for graduation may sometimes be given precedence over placing them in ELD coursework. Thus, the practice of placing EL students in general English coursework without offering ELD instruction should be examined with great care—especially since eliminating ELD courses that focus explicitly on developing ELs’ English proficiency runs counter to California’s ELA/ELD Framework policy. Finally, in San Diego, we found some evidence that the presence of EL support teachers at school sites led to better grades for both long-term and late-arriving ELs, and faster English proficiency growth for late-arriving ELs. Budgetary issues have led the district to replace these teachers with two new types of support teachers, but at lower levels of support. In interviews, school staff praised the EL support teachers and many expressed a desire to restore funding for this position. If the district’s financial shortfall eases, bringing back school-level EL support teachers to previous levels could help ELs’ academic progress. English Learners in middle and high school face unique challenges as they strive to gain English language proficiency and meet the other academic requirements necessary to gain a high school diploma. For California, helping ELs advance both linguistically and academically remains a high priority—and understanding the factors that contribute to their progress and struggles will be critical to promoting their long-term success. PPIC.ORG Academic Progress for English Learners 40 REFERENCES Bui, Sa A. 2013. “How Do Limited English Proficient Students Affect Each Other’s Educational Outcomes? Evidence from Student Panel Data“ working paper. CDE (California Department of Education). 2016. 2016–17 and 2017–18 CELDT Information Guide. CDE (California Department of Education). 2016–17. DataQuest. CDE (California Department of Education). 2017–18. DataQuest. California State Legislature. 2012. Assembly Bill 2193: An Act to Add Sections 313.1 and 313.2 to the Education Code, relating to English Learners. Callahan, Rebecca. 2005. “Tracking and High School English Learners: Limiting Opportunity to Learn.” American Educational Research Journal 42 (2): 305–328. Callahan, Rebecca, Lindsey Wilkinson, Chandra Muller, and Michelle Frisco. 2008. “ESL Placement and Schools: Effects on Immigrant Achievement.” Educational Policy 23 (2): 355–384. Carhill-Poza, Avary. 2015. “Opportunities and Outcomes: The Role of Peers in Developing the Oral Academic English Proficiency of Adolescent English Learners.” Modern Language Journal 99 (4): 678–695. Conger, Dylan. 2009. “Testing, Time Limits, and English Learners: Does Age of Entry Affect How Quickly Students Can Learn English?” Social Science Research 38 (2): 383-396. Conger, Dylan, Amy E. Schwartz, and Leanna Stiefel. 2011. “The Effect of Immigrant Communities on Foreign-born Student Achievement.” International Migration Review 45 (3): 675–701. Flores, Nelson, Tatyana Kleyn, and Kate Menken. 2015. “Looking Holistically in a Climate of Partiality: Identities of Students Labeled Long-Term English Language Learners.” Journal of Language, Identity & Education 14 (2): 113–132. Friesen, Jane, and Brian Krauth. 2011. “Ethnic Enclaves in the Classroom.” Labour Economics 18 (5): 656–663. Hill, Laura, Julian R. Betts, Belen Chavez, Andrew C. Zau, and Karen V. Bachofer. 2014. Pathways to Fluency: Examining the Link between Language Reclassification Policies and Student Success. Public Policy Institute of California. Hodara, Michelle. 2015. “The Effects of English as a Second Language Courses on Language Minority Community College Students.” Educational Evaluation and Policy Analysis 37 (2): 243–270. Johnson, Angela S. 2018 “Can Summer School Help ELs Succeed?” EdWeek, January 15. Kanno, Yasuko, and Sarah E. N. Kangas. 2014. “‘I’m Not Going to Be, Like, for the AP’: English Language Learners’ Limited Access to Advanced College-Preparatory Courses in High School.” American Educational Research Journal 51 (5): 848–878. Los Angeles Unified School District. 2011. Reclassification of English Learners, Grades 2–12. Bulletin BUL-5619.0, October 17. MacSwan, Jeff, and Lisa Pray. 2005. “Learning English Bilingually: Age of Onset of Exposure and Rate of Acquisition among English Language Learners in a Bilingual Education Program.” Bilingual Research Journal 29 (3): 653–678. Marinova-Todd, Stefka H., D. Bradford Marshall, and Catherine E. Snow. 2000. “Three Misconceptions about Age and L2 Learning.” TESOL Quarterly 34 (1): 9–34. Menken, Kate and Tatyana Kleyn. 2010. “The Long-Term Impact of Subtractive Schooling in the Educational Experiences of Secondary English Language Learners.” International Journal of Bilingual Education and Bilingualism 13 (4) 399–417. Menken, Kate, Tatyana Kleyn, and Nabin Chae. 2012. “Spotlight on Long-Term English Language Learners: Characteristics and Prior Schooling Experience of an Invisible Population.” International Multilingual Research Journal 6 (2): 121–142. Miles, Matthew B., and A. Michael Huberman. 1994. Qualitative Data Analysis: An Expanded Sourcebook. Thousand Oaks, CA: Sage Publications Inc. NASEM (National Academies of Sciences, Engineering, and Medicine). 2017. Promoting the Educational Success of Children and Youth Learning English: Promising Futures. Edited by Ruby Takanishi and Suzanne Le Menestrel. Washington, DC: The National Academies Press. National Center for Education Statistics. 2018. English Learners in Public Schools. National Center for Education Statistics, updated April 2018. Olsen, Laurie. 2010. Reparable Harm: Fulfilling the Unkept Promise of Educational Opportunity for California’s Long Term English Learners. Long Beach, CA: Californians Together. PPIC.ORG Academic Progress for English Learners 41 Parish, Thomas B., Amy Merickel, Maria Pérez, Robert Linquanti, Miguel Socias, Angeline Spain, Cecilia Speroni, Phil Esra, Leslie Brock, and Danielle Delancey. 2006. Effects of the Implementation of Proposition 227 on the Education of English Learners, K–12: Findings from a Five-Year Evaluation. Prepared by American Institutes for Research and WestEd for the California Department of Education. Patton, Michael Q. 1990. Qualitative Evaluation and Research Methods. 2nd edition. Newbury Park, CA: Sage Publications Inc. Salazar, Jesus Jose, and Katherine G. Hayes. 2010. “The Role of Home Language Proficiency and Its Impact on English Learners' Academic Achievement.” Paper presented at the annual meeting of the American Educational Research Association. San Diego Unified School District. 2009. Master Plan for English Learners. Office of Language Acquisition. San Diego Unified School District. 2010. Master Plan for English Learners. Office of Language Acquisition, updated April 2010. Silver, David, Marisa Saunders, and Estela Zarate. 2008. “What Factors Predict High School Graduation in the Los Angeles Unified School District” California Dropout Research Project. Stevens, Gillian. 2015. “Trajectories of English Acquisition among Foreign-born Spanish-Language Children in the United States.” International Migration Review 49 (4) 981–1000. Strauss, Anselm L., and Juliet M. Corbin. 1990. Basics of Qualitative Research: Grounded Theory Procedures and Techniques. Newbury Park, CA: Sage Publications Inc. Thompson, Karen D. 2015. “Questioning the Long-Term English Learner Label: How Categorization Can Blind Us to Students' Abilities.” Teachers College Record 117 (12): 1–50. Umansky, Ilana, Megan Hopkins, Dafney B. Dabach, Lorna Porter, Karen Thompson, and Delia Pompa. 2018. Understanding and Supporting the Educational Needs of Recently Arrived Immigrant English Learner Students: Lessons for State and Local Education Agencies. Council of Chief State School Officers. Watt, David, and Hetty Roessingh. 1994. “Some You Win, Most You Lose: Tracking ESL Student Drop Out in High School (1988– 1993).” English Quarterly 26 (3): 5–7. PPIC.ORG Academic Progress for English Learners 42 ABOUT THE AUTHORS Laura Hill is a policy director and senior fellow at the Public Policy Institute of California. Her areas of expertise are K–12 education and immigration. She is currently researching English Learners in California schools, special education, and ESL students in community colleges. Her recent publications examine student achievement on the state’s K–12 assessments and the link between language reclassification policies and student success. Prior to joining PPIC, she was a research associate at the SPHERE Institute and a National Institute of Aging postdoctoral fellow. She holds a PhD in demography and an MA in economics from the University of California, Berkeley. Julian Betts is a professor and former chair of the Department of Economics at the University of California, San Diego, and is executive director of the department’s San Diego Education Research Alliance (SanDERA). He is also an adjunct fellow at PPIC and research associate at the National Bureau of Economic Research. He has performed student-level research using San Diego Unified School District data over the past 18 years. His research on determinants of student achievement and evaluation of interventions has examined differences in outcomes between EL and non-EL students. He has extensive experience with experimental and quasi-experimental methods. He holds a PhD in economics from Queen’s University, Kingston, Ontario, Canada. Megan Hopkins is assistant professor in the Department of Education Studies at the University of California, San Diego. She is an expert in mixed methods as well as EL policy and program implementation. She has a research agenda focused on supporting school districts undergoing demographic and/or policy change. Magaly Lavadenz is professor in the Department of Educational Leadership and Founding Director of Loyola Marymount University’s Center for Equity for English Learners. Her research interests include preparation and professional support of Latino and bilingual teachers, the experiences of the Central American immigrant community, public policy affecting language use, biliteracy, and education of Latinos and immigrant communities. Karen Bachofer is director of SanDERA. Previously she was executive director of San Diego Unified School District’s Research and Evaluation Division, where her responsibilities included direct oversight of national, state, and district assessment, and accountability processes and reporting. She also oversaw internal, external, and commissioned evaluation activities, including Title III reporting and production of the district’s annual English Learner Progress Report. Her SanDERA research has included studies of the impact of EL reclassification policies and student learning trajectories. She holds a PhD in education from the Claremont Graduate School and San Diego State University. Joseph Hayes is a research associate at the Public Policy Institute of California, where he studies population change, educational policy, and corrections issues. Recent projects have focused on estimates of the undocumented immigration population, English Learner reclassification policies in California public schools, and the changing composition of the state’s prison and parole populations. He holds an MS in agricultural economics from the University of Wisconsin, Madison. Andrew Lee is a research associate at the Public Policy Institute of California, focusing on K–12 education, specifically as it relates to English Learners in California schools. Prior to joining PPIC, he contributed to research on health policy and agricultural development policy at Acumen LLC and the UN Food and Agriculture Organization, respectively. He holds an MA in development practice from the University of California, Berkeley, and a BA in economics and English from the University of Michigan, Ann Arbor. Marco A. Murillo is a senior research associate at Loyola Marymount University’s Center for Equity for English Learners (CEEL). His research interests focus on the secondary school experience of immigrant youth as well as issues relating to college readiness and access for low-income, students of color. He holds a BA in political science and history from the University of California, San Diego, and a PhD in education (urban schooling) from the University of California, Los Angeles. Tara Vahdani is a research associate at CEEL, where she works on a broad range of projects related to equity and excellence in English Learner education. Prior to joining CEEL, she engaged in qualitative and quantitative methods as a research assistant at the Civil Rights Project/Proyecto Derechos Civiles, and later UC San Diego, where she earned a BA, studying education and political science with an emphasis in public policy. Andrew C. Zau is senior statistician at SanDERA. His research focuses on the effects of policy on student achievement in K–12 education. He has 16 years of experience analyzing SDUSD student-level longitudinal data and has played key roles in evaluations of the impact of reading reforms, diagnostic math testing, and school choice in San Diego, using experimental and quasi-experimental approaches. He holds an MPH in epidemiology from San Diego State University and a BS in bioengineering from the University of California, San Diego. ACKNOWLEDGMENTS We gratefully acknowledge additional research support from Mariko Cavey, Leslie Gautsch, Hayley Weddle, Kathryn Wiley, Iwunze Ugo, and Nidia Hernandez. Caroline Danielson, Kenji Hakuta, Rebecca Callahan, Kathy Hayes, Julien Lafortune, and Lynette Ubois provided very helpful reviews, and Vicki Hsieh provided much-appreciated editorial support, and Kit Reber provided production support. We thank the many staff in central office and school sites in both districts for sharing their insights. In particular we thank Sandra Cephas at SDUSD and Hilda Maldonado at LAUSD for many helpful conversations and for sharing materials on the history of EL support in their respective districts. PPIC.ORG Academic Progress for English Learners 43 PUBLIC POLICY INSTITUTE OF CALIFORNIA Board of Directors Steven A. Merksamer, Chair Senior Partner Nielsen Merksamer Parrinello Gross & Leoni LLP Mark Baldassare President and CEO Public Policy Institute of California Ruben Barrales Senior Vice President, External Relations Wells Fargo María Blanco Executive Director University of California Immigrant Legal Services Center Louise Henry Bryson Chair Emerita, Board of Trustees J. Paul Getty Trust A. Marisa Chun Partner McDermott Will & Emery LLP Chet Hewitt President and CEO Sierra Health Foundation Phil Isenberg Former Chair Delta Stewardship Council Donna Lucas Chief Executive Officer Lucas Public Affairs Mas Masumoto Author and Farmer Leon E. Panetta Chairman The Panetta Institute for Public Policy Gerald L. Parsky Chairman Aurora Capital Group Kim Polese Chairman ClearStreet, Inc. Gaddi H. Vasquez Senior Vice President, Government Affairs Edison International Southern California Edison PPIC.ORG Academic Progress for English Learners 45 The Public Policy Institute of California is dedicated to informing and improving public policy in California through independent, objective, nonpartisan research. Public Policy Institute of California 500 Washington Street, Suite 600 San Francisco, CA 94111 T: 415.291.4400 F: 415.291.4401 PPIC.ORG PPIC Sacramento Center Senator Office Building 1121 L Street, Suite 801 Sacramento, CA 95814 T: 916.440.1120 F: 916.440.1121" ["post_date_gmt"]=> string(19) "2019-01-23 18:06:13" ["comment_status"]=> string(6) "closed" ["ping_status"]=> string(6) "closed" ["post_password"]=> string(0) "" ["post_name"]=> string(127) "academic-progress-for-english-learners-the-role-of-school-language-environment-and-course-placement-in-grades-6-12-january-2019" ["to_ping"]=> string(0) "" ["pinged"]=> string(0) "" ["post_modified"]=> string(19) "2019-01-23 10:06:50" ["post_modified_gmt"]=> string(19) "2019-01-23 18:06:50" ["post_content_filtered"]=> string(0) "" ["guid"]=> string(171) "https://www.ppic.org/wp-content/uploads/academic-progress-for-english-learners-the-role-of-school-language-environment-and-course-placement-in-grades-6-12-january-2019.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) }