Donate
Independent, objective, nonpartisan research

R 114LHR

Authors

R 114LHR

Tagged with:

Publication PDFs

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(12) "R_114LHR.pdf" ["wpmf_size"]=> string(6) "702148" ["wpmf_filetype"]=> string(3) "pdf" ["wpmf_order"]=> string(1) "0" ["searchwp_content"]=> string(135424) "Reclassification of English Learner Students in California January 2014 Laura E. Hill, Margaret Weston, an d Joseph M. Hayes with research support from Da vid Ez ekiel and B elen Chavez http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 2 Summary Former English Learner students who have improved their facility with English to such a degree that they have been reclassified by their school districts as fluent in the English language are among the best performing students in the state. Because these Reclassified Fluent English Proficient (RFEP) students have much better academic outcomes than English Learner (EL) students, policymakers conjecture that reclassifying ELs more quickly might help close the state’s persistent achievement gap between EL and non- EL or English only (EO) students. To substantiate this conjecture —and noting that the standards for reclassification currently vary greatly among school districts—policymakers are interested in assessing whether district s with more rigorous reclassification standards have systematically lower reclassification rates, but also better student outcomes, than districts with less rigorous standards. Because districts determine their own reclassification criteria, it is difficul t to compare reclassification rates, the progress of ELs, and the outcomes for ELs and RFEPs across school districts throughout the state. We hope this report is informative to policymakers interested in Senate Bill 1108, which has as its goal documenting reclassification policies in California’s school districts and their link to student outcomes. In this report, we are able to overcome the key limitations of previous research through the use of California L ongitudinal Pupil Achievement D ata System (CALPAD S) provided under an arrangement with the California Department of Education (CDE). The data enabled us to track students in each California school district from 2007 –08 through 2012– 13, excluding students in charter schools. This report provides the first longitudinal analysis of the transition from EL to RFEP status for all California school districts. Our analysis indicates that RFEP students not only outperform EL students, but also often do as well as native English speakers when it comes to measures of academic outcomes, such as standardized tests and on-time grade progression. We also conducted a survey of school districts, asking detailed questions about their current and former reclassification policies and practices. We found that more than 90 percent of responding districts report using more demanding criteria than are suggested by the State Board of Education (SBE) guidelines for the four reclassification criteria specified in California Education Code Section 313(f). Our analysis of student -level longitudinal data in conjunction with survey responses reveals that districts using more stringent reclassification criteria have lower reclassification rates. For example, the roughly 30 percent of districts that require an EL student to score Proficient or higher on the English Language Arts portion of the California Standards T est have district reclassification rates that are 3 percentage points lower than in districts that require a score of Basic or higher. In other words , if the average annual reclassification rate among districts using Basic or higher (as suggested by the SBE guidelines) were 10 percent, districts that use Proficient or higher (a more rigorous criterion) would have reclassification rates of 7 percent. Th is translates to a 30 percent reduction in the number of students reclassified in districts using the more rigorous criterion. If districts require higher standards on more than one criterion, their reclassification rates are even lower. However, using stricter criteria than suggested in the state guidelines is also associated with slightly better outcomes for RFEP students. For example, requiring a score of Proficient or higher for students reclassified in 3rd grade is associated with larger percentages of these students scoring Proficient or higher on 6th grade standardized tests (a 4 percentage point increase, which would increase the share scoring Proficient or higher from 78 to 82 percent). Stricter criteria are also associated with a greater likelihood of on-time grade progress among students reclassified in the 8th grade (a 5 percentage point increase, increasing the http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 3 likelihood from 90 to 95 percent) and a 3 percentage point increase in the share scoring Proficient or higher on standardized tests (incr easing the likelihood from 14 to 17 percent). Among students reclassified in 9th grade, the Proficient criterion is not related to completing a –g requirements or leaving high school without graduating, but it is negatively associated with getting a high sc hool diploma, reducing the chances by 5 percentage points. Districts and policymakers should decide if these mostly positive outcomes are sufficiently large to justify the more demanding reclassification criteria that hinder so many ELs from transitioning to the RFEP population. In addition to these considerations, two major policy shifts occurring in K –12 education are also relevant to the well -being of the EL population. Under the Common Core State Standards and the new assessments being developed to test them, the assessments and criteria for EL reclassification will necessarily change in the coming years. Indeed, the California Standards Test (CST) will no longer be used, even in the current school year. The new reclassification criteria using th e new assessments will need to be crafted carefully. Current law requires analysis to determine the new reclassification criteria and we agree that such research and analysis are needed. The second relevant policy shift is embodied in the Local Control Funding Formula (LCFF), which increases funding for districts with large populations of EL students but does not provide funding for RFEP students. This new funding formula may reduce districts’ incentives to reclassify students.1 However, a statewide reclassification policy would help deter the likelihood of a district restricting EL students from being reclassified when they are ready to transition to RFEP status. Based on the findings of this report, we recommend that there be one standard for reclas sifying EL students statewide, and that the standard be set, for now, using the assessments and levels recommended in the guidelines provided by the State Board of Education. In most cases, this means districts will need to lower their reclassification sta ndards. The implementation of the Common Core State Standards and the Local Control Funding Formula represent new opportunities for success in K –12 education, and the progress and outcomes of both EL and RFEP student s should be closely monitored to insure that these students participate in these opportunities. 1 The new funding formula goes into effect this school year (2013 -2014), but not all of the funding will be available in the first year. Because many RFEP students are also members of low -income families (and low-income students generate funding for district s as well), districts should not face as large a disincentive to reclassify students as they otherwise might. Contents Summary 2 Figures 5 Tables 6 Acronyms 7 Introduction 8 Recl assification and S tudent Outcom es 10 Student C ohort Sel ect ion an d Defin ition 10 Student O utco mes 13 How Does R ecl assification W ork ? 20 PPIC’s “ English Learners Reclassif ication Survey” 22 Reclassification P ol icies 23 Do District R ecl assification Polici es Infl uen ce Recl assific ation Ra tes? 33 Do Recl assification Polici es A ffect Stud ent Outcom es? 38 Longitudi nal RFEP Outcomes 38 Relationships Be tween More Rigorous Reclassifi cation Policies an d RFEP Outcomes 42 How Do We E valuat e t he T rad e-offs Betw een Low er R ecl assification R at es but Bett er Outcom es? 49 References 51 About the A uthors 52 Acknow ledgm ents 52 Technical appendices to this paper are available on the PPIC website: www.ppic.org/content/pubs/other/114LHR_appendix.pdf http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 5 Figures 1. Pe rcent of stud ents scoring B asic or high er on the CST ELA 15 2. On -time or better grade progression, 7th grade cohort 17 3. Final high school outcomes 18 4. Percent of districts that permit reclassification, by grade level 29 5. Most districts allow for student reclassification at more than one point in time per year 30 6. District reclassification rates under different reclassification requirement scenarios 36 7. Share of cohort, by language group, scoring Basic or higher on CST ELA, controlling for district and student cha racteristics 39 8. Share of cohort, by language group, scoring Proficient or higher on CST ELA, controlling for district and student characteristics 40 9. Percent on -time or better in 10th grade, by language group, 7th grade cohort, controlling for district and student characteristics 41 10. Final high school outcomes for 8th grade cohort, controlling for district and student characteristics 42 http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 6 Tables 1. N umb er and cl assific ation st atus of students included in our a n alysis , by cohort 12 2. Re classific ation r a te s for four cohorts of E Ls 13 3. Survey respondents 22 4. S urv ey r es pond ents are broadly repr esen tative of the state’s student demographics 23 5. English proficiency criteria , by grade level 25 6. Percent of districts imposing basic skills requirements, by grade level 25 7. Teacher evaluation criteria, by grade level 27 8. Subjective evaluation criteria used by teachers, by grade level 28 9. “If you were forced to choose one criteria for each grade level, which of the reclassification criteria would you say is most difficult for EL students to meet in your district?” 31 10. “In your opinion, how important are each of the reclassification criteria in the ultimate decision to reclassify a student?” 31 11. District reclassification policy 34 12. Per centag e point change for stud ent outcomes and district recl assification rat es associated with specific recl assification requirements 45 13. End of high school outcomes, cohort 4 47 http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 7 Acronyms ADA API AYP CAHSEE CALPADS CDE CELDT CST DIBELS EL Average Daily Attendance Academic Performance Index Annual Yearly Progress California High School Exit Exam California Longitudinal Pupil Achievement Data System California Department of Education California English Language Development Test California Standards Test Dynamic Indicators of Basic Early Literacy Skills English Learner ELA ELD EO IFEP LCFF LEA LTEL OPL RFEP SBE SEI English Language Arts English Language Development English Only Initially Fluent English Proficient Local Control Funding Formula Local Education Agency Long -Term English Learner Overall Performance Level Reclassified Fluent English Proficient State Board of Education Structured English Immersion http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 8 Introduction California’s K –12 public schools include some 1.4 million English Learners (ELs), representing about 22 percent of the student body. Researchers have found a persistent achievement gap between these ELs and their native -English -speaking peers. However, the EL designation is intended to be temporary. Once ELs become proficient in English, they are reclassif ied and analysts have found that the achievement gap between former EL students —i.e., Reclassified Fluent English Proficient (RFEP) students— and native English speakers is small or even positive, with RFEP students sometimes outperforming native English sp eakers (Saunders and Marcelletti, 2013; Hill, 2012; EdSource, 2008; Gándara and Rumberger, 2006). Because reclassified ELs perform so much better than ELs on a host of academic outcomes, policymakers are avidly interested in designing policies that help EL students transition quickly into unsupported academic instruction. This endeavor assumes that once the students are reclassified, their increased access to academic instruction will increase their academic performance and, correspondingly, that bringing m ore students into the RFEP student group will reduce the achievement gap. However, one must consider the possibility that the gap between RFEP and EL student performance may be the result of reclassifying only the very strongest students among the EL group . Currently, California school districts are allowed to determine their own reclassification policies. The State Board of Education has issued guidelines for the four reclassification criteria specified in California’s Education Code Section 313(f).2 In October 2012, Governor Jerry Brown signed SB 1108 (Chapter 434, Statutes of 2012) , which reflects an interest among policymakers in understanding the variance in reclassification policies across school districts and their association with student outcomes. Until now, there has been a lack of general knowledge about how and when districts reclassify their EL students. In addition, there is no consensus about the ideal reclassification policy for ensuring the success of EL and RFEP students. At the same time, the governor’s 2013 –2014 budget has seriously changed the way school districts are funded, significantly increasing per -pupil funding for EL and low -income students while allowing districts to exert their own control over how those dollars are spent.3 Some are concerned that the extra funding will not be spent on EL students and that districts will have incentives to keep EL students from being reclassified (extra funding is not available for RFEP students).4 However, most EL students (estimates range from 74 to 85 percent) are also members of low -income families (Rose, Sonstelie, and Weston, 2012; California Legislative Analyst’s Office, 2007), which means that even if they are reclassified, these students will still generate extra funding for their school districts. Would lowering reclassification standards reduce the achievement gap by increasing the number of RFEP students, or would it increase the number of RFEP students but lower their performance as a group? 2 Section 313(f) reads as follows, “The reclassification procedures developed by the department shall utilize multiple criteria in determining whe ther to reclassify a pupil as proficient in English, including, but not limited to, all of the following: (1) Assessment of l anguage proficiency using an objective assessment instrument, including, but not limited to, the English language development test that is developed or acquired pursuant to Section 60810. (2) Teacher evaluation, including, but not limited to, a review of the pupil’s curriculum mastery. (3) Parental opinion and consultation. (4) Comparison of the performance of the pupil in basic skill s against an empirically established range of performance in basic skills based upon the performance of English-proficient pupils of the same age, that demonstrates whether the pupil is sufficiently proficient in English to participate effectively in a cur riculum designed for pupils of the same age whose native language is English.” 3 Draft regulations were released by the SBE in October, 2013. Emergency regulations were adopted January 16, 2014, and final r egulations are expected later in 2014 . 4 Students classified as both EL and low -income students do not generate additional funding, compared to those who are classified only as low- income students. http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 9 Understanding the relationships between reclassification policies, reclassification rates, and reclassified student outcomes is a necessary prerequisite before considering recommendations about whether more EL students should be reclassified, and more quickly, than is currently the norm. In summer 2013, PPIC conducted a survey of school districts, documenting reclassification policies and practices across a variety of the state’s school districts. The researchers also analyzed six years of longitudinal student-level data captured in CDE’s California Longitudinal Pupil Achievement Data System (CALPADS). This report presents the results of this research. In the following pages, we  Explore differences in academic outcomes for ELs, RFEPs, and other language groups, tracking these students as they progress though school.  Report reclassification policies and practices in school districts throughout the state.  Examine whether school districts who report using more rigorous reclassification policies and practices have lower reclassification rates than districts who report using less rigorous criteria.  Consider whether reclassified students in school districts who report more rigorous reclassification policies and practices have better outcomes than districts who report more relaxed policies.  Conclu de with preliminary recommendations for reclassification policies. This researc h is an important first step in understanding the relationship between reclassification policies and student outcomes in California’s school districts. We hope this report is helpful to legislators, CDE, and the State Board of Education as they contemplate a response to SB 1108. We should note that this report investigates just one policy lever —reclassification policies—and does not consider other important issues. For example, it does not contribute to valuable discussions about the validity and reliability of the standardized tests used to establish English proficiency. 5 Nor does it provide analysis related to EL services and program delivery or teacher preparation. However, given that decisions about how ELs are taught and served are left to the local level, setting uniform reclassification policy is one of the few levers currently available to the state. 5 For a discussion of this issue, see Abedi (2008). http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 10 Recla ssification and S tudent O utcom es Research on outcomes among California’s EL and RFEP students usually examines cross- sectional differences among student groups—RFEPs compared to ELs or English only (EO) students —(e.g., Saunders and Marcelletti, 2013; Hill, 2012; EdSource, 2008; and Gándara and Rumberger, 2006). Other studies track students through longitudinal data, although these studies generally focus on only one or two districts at a time (e.g., Flores, Painter, and Pachon, 2009; Hopkins et al., 2012; Robinson, 2011). The cross- sectional research has found large differences between EO students and ELs, but this work often overlooks the confounding factor that the most successful EL students are reclassified as RFEPs and are not included in either comparison group. When RFEP and EL students are combined into an “ever -EL” student group, the gap between ever -EL and EO students is considerably smaller and has declined somewhat over time (Saunders and Marcelletti, 2013). However, even cross -sectional research that refines comparison groups for ELs cannot account for “time since reclassification” or new entrants to the EL population. Indeed, the inability to control for those potentially confounding factors in cross -section al research may explain why RFEP students outperform EO students in elementary grades, but have worse outcomes by the end of high school (Gándara and Rumberger, 2006; Hill 2012). In 4th grade, for example, RFEP students have much higher scores on standard ized tests than native English speakers, but among 10th graders, native English speakers have slightly higher scores than RFEP students. It is important that we understand whether the apparent decline of 10th grade RFEP students’ scores is real or whether 10th grade scores are lower because more recently reclassified students in the cohort have lower scores than students reclassified at earlier grades. In our research, we follow the outcomes of individual students over time; thus we can consider the role of “time since reclassification” in the academic outcomes of RFEP students. We begin this section by describing our longitudinal student- level data and the cohorts of students we construct for analysis. We then examine standardized test scores, on- time grade progression, and high school outcomes across our student cohorts and language groups. Student C ohort Selection and Defin ition In this report, we are able to overcome the key limitations of previous research through the use of the state’s longitudinal student-level data system (CALPADS). The data enabled us to track students in each of the state’s school districts from 2007 –08 through 2012– 13, excluding students in charter schools. This report provides the first longitudinal analysis of the transition from EL to RFEP status for all California school districts. Because we can use data for all of the state’s school districts and can follow students across many years of data, our research is uniquely situated in its ability to compare the outcomes of RFEPs with the outcomes of ELs, EOs, and students who have a primary language other than English but are designated as Initially Fluent English Proficient (IFEP). We can also simultaneously consider a student’s ba ckground characteristics, the grade in which a student was reclassified, the role of district characteristics (such as demographic composition and performance on key measures), and finally, the association between district reclassification policies and rec lassified students’ outcomes. http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 11 Using CALPADS, we constructed four cohorts of students based on the grade they were in during the 2007 –08 school year (grades 2, 4, 7, or 8) and then divided each cohort into groups based on language status: EL, IFE P, RFEP, o r EO. For our purposes, ELs must have been ELs in kindergarten and must not have been reclassified while we observe them (school years 2007– 08 through 2012– 13). We have divided the RFEP students in each cohort into three groups: those reclassified in 2008 –09 (our “target” reclassification year), those reclassified beforehand (“pre-target”), and those reclassified afterward (“post-target”). Because reclassification policies can change over time, examining student progress relative to policies in place at one point in time (2008– 09) allows us to draw connections between reclassification criteria and later student outcomes. EL or RFEP students who started school in California after kindergarten are not included in our cohorts. Most ELs in California started school in the United States in kindergarten, rather than arriving from abroad as older students in later grades.6 Table 1 summarizes the number of students and their classification status in each of our four cohorts. The 2nd grade cohort allows us to observe a large cohort of students who were reclassified at the first opportunity in some districts —at the beginning of 3rd grade. Many districts do not reclassify before this point. The 4th grade cohort (students reclassi fied in 5th grade) represents another important reclassification year —this is commonly understood to be the modal reclassification grade . 7 The 7th grade cohort represents students who are L ong-T erm English Learners (LTELs), and we can observe their outcomes to the beginning of their 12th grade year. 8 The final cohort, the 8th grade cohort, also includes LTELs, enabling us to observe their end -of -high- school outcomes. 9 6 Published CDE data suggest that the share of EL students who arrived from abroad within the last year is relatively small by grade. The share peaks in number in grade 2 —representing approximately 4,000 (or 2%) of 2nd grade ELs —and peaks in percentage in grade 9, representing about 6% of 9th grade ELs (Hill, 2012). Their cumulative percentage by grade depends on their reclassification rates. Future extensions of this research could include EL students who arrived in the United States after the kindergarten year, but we do not include them h ere. 7 Ther e are at least two reasons why this is considered the modal reclassification grade: 1) elementary schools wish to transition students to middle school as RFEPs in order to avoid having students become Long -Term ELs, and 2) one component of the redesignation criteria is easiest to achieve in the prior year (the CST ELA, administered in the 4th grade). 8 AB 2193 (Chapter 47, Statutes of 2012) defines an LTEL as “an English Learner who is enrolled in any of grades 6 to 12, inclu sive, has been enrol led in schools in the United States for more than six years, has remained at the same English language proficiency level for two or more consecutive years as determined by the English language development test identified or developed pursuant to Section 60 810, or any successor test, and scores far below Basic or below Basic on the English language arts standards-based achievement test administered pursuant to Section 60640, or any successor test.” (Accessed from http://leginfo.legislature.ca.gov/faces/billNavClient.xhtml?bill_id=201120120AB2193 .) Because we do not use the English language development test or the English language arts standards-based achieve ment test to create the cohort, our LTELs are best thought of as an approximation of official LTELs. 9 Most end -of-high school outcomes such as graduation, finishing a –g course requirements, and reasons for leaving without a diploma are not recorded for al l students until five years after beginning 9th grade. http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 12 TABLE 1 N umb er and cl assific ation st atus of students included in our a n alysis , by cohort Grade 2 Grade 4 Grade 7 Grade 8 EO 85,167 77,337 112,293 67,131 IFEP 12,568 13,642 18,967 11,215 RFEP, pre-target year 7,692 18,038 48,267 33,871 Long-Term English Learners RFEP, target year (2008– 09) 3,246 2,623 1,404 997 RFEP, post -target year 22,583 9,898 4,544 2,363 EL never reclassified (2008– 09 –2012– 13) 19,538 10,400 7,516 4,628 Total 150,794 131,938 192,991 120,205 Because we are interested in understanding the relationship between district reclassification policies and outcomes for students, we generally restrict our sample to students who do not transfer across districts — with one important exception.10 We do include students who transfer from their elementary school district to the appropriate high school or unified school district .11, 12 Similarly, we require that students be present in the data for all six years of our study, with exceptions made for our two oldest cohorts. In our 7th grade cohort, we require that students remain in the data until the third year of the study, when they should be in 9th grade if they are progressing on time. After 9th grade, we do allow the students to be missing, because dropping out (or remaining in school) is an outcome of interest for high school students. We allow an analogous exception for our 8th grade cohort, requiring them to remain in the data only until the 3rd year of the study or 10th grade. Of all of our data cleaning and cohort construction restrictions, the one with the most substantial impact on the representativeness of our sample is our decision to exclude ELs who arrive in the California school system after kindergarten. Excluding these later -arriving ELs significantly affects the share of ELs in the 7th and 8th grade cohorts. In 2007– 08 ELs represented 21 percent of 7th graders an d 18 percent of 8th graders (the first year of our data; see Appendix Table A1 ). When we drop students who are missing and ELs who did not start in kindergarten, the share of ELs falls to 8 percent in both cohorts. Requiring that ELs be in California schools since kindergarten also substantially affects the share of Latinos in our sample, since most ELs speak Spanish, but not the share of economically disadvantaged students. We also exclude special education students.13 Districts may vary in the extent to which ELs are classified as 10 See Appendix Table A1 for details on the number of students not included in the study. 11 In practice, we drop students from the study who transfer out of their district before the appropriate year to complete a transfer from elementary districts to high school districts. For example, a student enrolled in a K –8 district who transfers districts is dropped from the analysis unles s that transfer occurs in the 9th grade. For students who transfer in the appropriate year, we require that their transfer be within the county in order to stay in the data. Transfers are not permitted for students in unified school districts. 12 Future res earch could assess the effect of excluding more mobile students from our analysis. 13 Special education students who are ELs can be reclassified, but often through different assessment instruments. Our survey of district reclassification policies did ask ab out reclassification policies for special education students (see the following section and Appendix B ), but our http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 13 special education students, and perhaps even more so in the case of LTELs who are also special education students. We do find that excluding special education students has a somewhat larger effect on the characteristics of the 7th and 8th grade cohorts.14 Due to these data cleaning and cohort construction restrictions, we believe that this report best describes outcomes and predictors of outcomes for ELs who start in California schools in kindergarten. Additional research should address late -arriving ELs and the role of special education status in EL classification and reclassification. Student O utco mes Our analyses focus on the following student outcomes: scoring Basic or higher on the California Standards English Language Arts (CST ELA) test, scoring Proficient or higher on the CST ELA, and on -time grade progression. For our oldest cohort of students, we also consider earning a high school diploma, meeting a – g requirements, and leaving prior to completing high school.15 Before discussing these outcomes for EL and reclassified students, we consider differences in reclassification rates by grade among our cohorts. Among our four cohorts, across all districts, reclassification rates in our target reclassification years are highest for students in the 8th grade cohort, who are reclassified in the 9th grade (12.5%), although this rate is similar to the reclassification rates in the 5th grade.16 TABLE 2 Re classific ation r at es for four cohorts of E Ls Reclassification rates observed in: 2nd grade cohort 4th grade cohort 7th grade cohort 8th grade cohort 3rd grade 7.1% 5th grade 11.4% 8th grade 10.4% 9th grade 12.5% Rates of reclassification, 2008– 09 – 2012– 13 56.9% 54.6% 44.2% 42.1% Note: Students in all cohorts h ave been EL students since k ind erg art en. Looking at longer term outcomes for ELs, the 2nd and 4th grade cohorts have roughly equal chances of being reclassified during the next five years: 57 percent of 2nd grade ELs are reclassified as RFEP students by 7th grade, and 55 percent of 4th grade ELs are reclassified as RFEP students by the 9th grade. In the 7th and 8th grade cohorts, students who were reclassified in the target year or later are considered Long -T erm English analyses in this report exclude special education students. Of course, districts can also differ is their classification of EL students into special education in the first place. We leave these considerations for later studies. 14 When special education students are excluded, the share of ELs drops by 1 to 2 percentage points across cohorts, but since the share of ELs remaining is already substantially lower in the 7th and 8th grade cohorts, this exclusion has a larger impact on the older cohorts. 15 We hope in future extensions of this research to incorporate other test results, such as CST math and CAHSEE scores and pass rates. 16 Reclassification rates are calculated by dividing the number of new RFEP students in a calendar year by the sum of ELs and new RFEPs. http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 14 Learners. Among students who are still ELs in 7th grade, 44 percent become RFEPs by the end of high school. Rates for the 8th grade cohort are similar (42%). 17 Although these reclassification rates are lower than the reclassification rates in the younger grades, it is notable that sizeable percentages of LTELs ar e ultimately reclassified by the end of high school. In the following sections, we discuss outcomes for reclassified and non- reclassified EL students, as measured by CST ELA scores, on -time grade progression, and final high school outcomes. Student O utcom es: C ST E LA Scor es We examine ELA scores for the youngest three cohorts.18 Descriptively, across all three cohorts, RFEP students, no matter w hen they are reclassified, are more likely than EL students to achieve a score of Basic or higher on the CST ELA, and their performance is consistently better across the years following their reclassification. Recall that we are following individual students over time, and EL students are those who have not been reclassified by the last year in which we observe them (2012 –2013). That RFEP students outperform EL students is not surprising, given that reclassified students exit the EL student population because they meet reclassification criteria that include the CST ELA . 19 However, previous research based on cross- sectional data has questioned whether these differences pers ist over time. More notable is the finding that RFEP students often outperform native English speakers, although the differences are generally small (Figure 1). RFEP students in our 2nd grade cohort are a little more likely than EO students to score Basic or higher on the 6th grade ELA, no matter when they were reclassified. The scores of IFEP students are very similar to those of RFEP students .20 In the 4th grade cohort, RFEP students reclassified before our “target” year perform better than EO and IFEP students when they take the CST ELA in 8th grade, but those reclassified during or after the “target” year do not. In the 7th grade cohort, RFEP students reclassified by the target year (i.e., pre-target students) or during the target year have higher ELA scores in the 11th grade —78 percent and 73 percent, respectively —than EO students (71%). The results are somewhat different if we examine the share of students who score Proficient or higher on the CST ELA. 21 Very few EL students are able to reach the Profi cient level (results are shown in Appendix Figure A1 ). In our 2nd and 4th grade cohorts, RFEP students reclassified prior to our target year tested Proficient in a roughly equal pro portion to IFEP students and higher than EO students. In the 2nd, 4th and 7th grade cohorts, students reclassified in the target year or later do not perform as well as EO or IFEP students. There is little evidence that RFEP students lose ground over time. They do not exhibit steeper declines in test scores than other student language groups (e.g., IFEPs or EOs). Turning to outcomes for ELs, younger EL 17 Of course, many ELs might have dropped out of school or moved out of state between the 10th and 12th grades, so the reclassification rate may actually be lower. As noted in the text and further discussed in the Technical Appendix, we do not allow students in the younger cohorts to exit from the data. 18 In a later section, we analyze factors (including district reclassification criteria) that may predict differences across groups. Hence, we do not in this section test the d ifferences for statistical significance. 19 Future research could compare outcomes in later grades for EO, IFEP, and RFEP students with the same 3rd grade ELA scores. 20 IFEP students are likely to speak a language other than English at home but are designated fluent in English if they pass an English proficiency test in kindergarten. Being bilingual, especially at a young age and with a high degree of fluency, is associated with increased density of grey matter in the brain (Mechelli et al., 2004). T his, along with higher average socioeconomic status (52 percent of IFEP students are low income, in comparison to 79 percent of EL and 63 percent of RFEP students reclassified in 2008 -09), may explain the very high performance of IFEP students. 21 Federal guidelines require that all students score Proficient or higher by 2014. http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 15 cohorts appear to make gains in grades 3 through 6 (2nd grade cohort) and grades 7 and 8 (4th grade cohort), while other student language groups do not. FIGU RE 1 Pe rcent of stude nts scoring B asic or high er on the CST ELA 2nd Grade Cohort 4th G rade Cohort 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 2nd 3rd 4th5th6th Grade IFEP RFEP pre RFEP 2008-09 RFEP post EO EL 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 4th 5th6th7th8th Grade IFEP RFEP pre RFEP 2008-09 RFEP post EO EL http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 16 7th G rade Cohort SOURCE: Authors’ calculations using CALPADS NOTE: Results are not teste d for statistical significance. Student O utcomes: O n-T ime Gra de P rog ression Descriptively, EL students are much less likely than other types of students to advance one grade per school year (Figure 2). RFEP students, regardless of when they were reclassified, are the most successful students in terms of on -time (or better) grade progression to 12th grade: Over 82 percent progressed on time to their final year in high school. (Recall that we do allow students to leave the sample starting with the 10th grade, so these on -time rates are just of those students who remain in the appropriate district each year we observe them.) About 81 percent of IFEP students make on -time progress to the 12th grade. The two groups with the worst outcomes on this measu re are EO and continuing EL students. EO students (80%) are somewhat more likely than continuing EL students (74%) to reach the 12th grade on time. 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 7th 8th9th10th 11th Grade IFEP RFEP pre RFEP 2008-09 RFEP post EO EL http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 17 FIGURE 2 On -time or better grade progression, 7th grade cohort SOURCE: Authors’ calculations using CALPADS NOTE : All students in Figure 2 are in the 7th grade cohort. Results are not tested for statistical significance. The share making on -time grade progression is slightly higher in the 11th than 10th grade for EL, RFEP target -year, and RFEP post targe t- year groups, which is likely due to students dropping out of school between the 10th and 11th grade. CALPADS lacks information on grade retention that may have occurred prior to 2007– 08. This data limitation likely results in an underestimate of the retention patterns for ever -EL students, who are m ore likely to be retained in grades K –3 than non- EL students (Cannon and Lipscomb, 2011) . 22 If we cou ld observe early grade retention, the lines in Figure 2 for the EL and RFEP groups would likely be lower, but they would still show the same rate of on -time progression in the high school years. High S chool O utcom es We constructed our 8th grade cohort in such a way as to observe differences in final high school outcomes .23 We consider three outcomes (Figure 3): leaving high school before graduating, completing high school with a diploma, and completing high school through a– g course requirements (which makes students eligible for admission to the University of California or California State University).24 22 In their study of the Los Angeles Unified School District, Cannon and Lipscomb found that among the students in grades K -3, 1st grade was the most common retention year (ab out 3% of 1st graders were retained), followed by kindergarten and 2nd grade (about 2% retained in each), with less than 1 percent of 3rd graders being retained. 23 Final high school outcomes are often not reported in CALPADS until the 5th year after a student starts 9th grade. 24 Reasons for leaving school include the following: truant with no known enrollment, student expelled and did not re -enroll, student did not return after academic year, student dropped out and did not enroll in GED program, student dropped out and entered a non -academic institution (such as the job corps or justice system), and student not working toward a high school diploma. 60% 65% 70% 75% 80% 85% 90% 95% 100% 9th 10th 11th12th Percent on-time or better On-time grade IFEP RFEP pre RFEP 2008-09 RFEP post EO EL http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 18 FIGURE 3 Final high school outcomes SOURCE: Authors’ calculations using CALPADS NOTE : Results not tested for statistical significance. Continuing EL students are the most likely students to leave high school without graduating (7%).25 Between 1 and 2 percent of all other students, including all types of RFEPs and IFEPs and EOs, left school. Across the three cohorts of RFEPs, between 84 percent and 89 percent received a diploma. All three groups of RFEPs are much more likely than ELs (56%) to finish high school with a diploma. RFEP students reclassified in the “target” year are as likely as IFEP and more likely than EO students to graduate with a diploma, but students reclassified later are not. RFEPs reclassified prior to high school (44%) are just slightly more likel y than EO students (43%) to have completed their a – g requirements. The RFEP students reclassified in 9t h grade or later are more likely than ELs (7%) to have completed the a –g requirements, but less likely than EO students to have done so. Finally, IFEP students are most likely to have met the hi gher sta ndard of having completed their a – g requirements upon graduation. Summing Up RFEP students outperform continuing EL students on every measure we were able to examine, which suggests that the criteria used to determine when an EL student no longer needs support learning English do separate stronger academic performers from those who are less able. However, not all RFEPs are equal — those who are reclassified at younger grades are more likely to progress on time, have higher test scores, and have positive outcomes in their final years of high school. LTELs who are reclassified do not perform as well, on average, as RFEPs who are reclassified at younger grades. These results suggest that cross -sectional views of RFEPs’ progress over time are complicated by the fact that newly reclassified students, especially those 25 We identify students who leave school by the following exit codes in CALPADS: E140 (no known enro llment, truant), E300 (expelled, no known enrollment), E420 (no show, same school), T270 (trans, drop, adult school), and T380 (trans institution, no HS diploma). 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Leave before graduating Diplomaa-g courses complete Final high school outcomes, 8th grade cohort IFEP RFEP pre RFEP target year (2008-09) RFEP post EO EL http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 19 reclassified after having been LTELs, do not perform as well as ELs reclassified in elementary school. The results also highlight the fact that LTELs who are reclassified still outperform EL students dramatically on multiple measures. We find that RFEP students perform better than EO student in many cases; and in some cases, RFEP students perform better than IFEP students, the group that is often at the top of the performance measures. These last two findings suggest that there is a role for rethinking the goal of EL programs and reclassification standards: What is the right performance level for reclassifying students to insure their academic success after EL support ends ? 26 In the following section, we discuss the variation in district policies for reclassifying students. In later sections we examine how those policies relate to reclassification rates for districts and how they relate to student outcomes (while simultaneously considering the role of student and district characteristics). 26 PPIC research -in-progress that is focusing on longer -term analyses (and using more contextu al data about districts) might help address this question. http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 20 How D oes R eclas sification W ork ? Although state law is very clear on the criteria school districts should use to identi fy English Learners, it is less prescriptive about the policies districts use to reclassify these students. 27 According to state law, school district reclassification decisions must incorporate four criteria: an assessment of English proficiency, an evaluat ion of basic s kills in English, a teacher evaluation of the student, and parental consultation ( Education Code 313(f)). 28 To help districts de velop effective reclassification policies, the State Board of Education established guidelines and recommendations for each of the f our criteria in 2002. 29 For example, the board recommends that districts use the ELA portion of the CST to assess the basic skills of ELs, and set a cut -off score somewhere in the range of Basic to midpoint -Basic (300– 325 points). The guidelines are less specific for the other reclassifica tion criteria. Given the nature of the state guidelines, one might expect large variation in EL reclassification policies and r ates; and in deed, Linquanti (2001), the California State Auditor (2005), and Parrish et al. (2006) found wide variation in reclassification policies across districts in the small sample each used .30 Moreover, the auditor found that of the 180 ELs included in the study, 62 percent met their district’s stated criteria for reclassification but were not reclassified. The auditor recommended clearer statewide reclassification guidelines and regulations to create a more consistent experience for English Learners across the state. Motivated by a desire to better understand and improve EL policy, and with the goal of improving reclassification rates and student outcomes, SB 1108 (Chapter 434, Statutes of 2012) charged the CDE with the task of reviewing school district reclassification policies and practices (with the intent of providing the information necessary for developing new state policy) .31 To help advance the conversation about EL reclassification, the Public Policy Institute of California surveyed local education agencies throughout California. We surveyed school districts about their reclassification policies, rather than collect and review written reclassification policies (such as those contained in English Learner Master Plans), for three reasons. First, we wanted this report to be as timely as possible for current policy discussions. Requesting, reviewing, categorizing, and analyzing di strict plans would have been much more time - and resource- intensive. Second, survey responses afforded the possibility to gather both districts’ reported criteria and additional detail about district reclassification practice that might not be spelled out in district plans. Finally, a survey allowed us to gather detail about both current p olicy and practice as well as 27 Students identified by the Home Language Survey as speaking a language other than English are assessed for their English proficiency through the California English Language Development Test (CELDT). 28 A copy of Education Code 313(f) is available in Appendix F . 29 The SBE approved the “Guidelines for Reclassification of English Learners” on September 11, 2002. On September 6, 2006, the SBE approved modifications to the guidelines. A copy of the guidelines is available in Appendix F . According to the state auditor’s report (2005): “Because th ese are not regulations, school districts are not required to adhere to the department’s guidelines. However, according to the board’s chief legal counsel, the guidelines were based on an analysis of actual test data and developed with public input, so the board expects that school districts will pay great deference to them when making their initial identification and their redesignation decisions. Nevertheless, these are only guidelines and school districts are allowed flexibility in defining their criteria.” (p. 18) 30 Linquanti examined policies from seven districts, prior to SBE issuing its reclassification guidelines. The state auditor reviewed the policies of eight school districts: Anaheim Union High, Long Beach Unified, Los Angeles Unified, Pajaro Valley Unified, Sacramento Unified, San Diego Unified, San Francisco Unified, and Stockton Unified. The auditor also reviewed a total of 180 individual EL cases from these districts. Parrish et al. reported on nine California districts. States, too, vary in their reclassification criteria; see Abedi (2008) for an explanation of how this makes state -to-state comparisons of EL and reclassified EL populations difficult. 31 Contingent upon the availability of funds for the research. http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 21 past policy and practice. In order to complete our longitudinal analysis, we needed to know the policies and practices in place in school districts in the 2008– 09 school year. There are shortcomings with the survey approach, h owever. For examp le, though we attempted to have the most knowledgeable district employee complete the survey, it is possible that the respondent may not always accurat ely report district practi ce. Further, distric t practice could deviate from what was report ed to us. It is worth noting that a review of written district plans would have suffered from the same shortcoming. In the following pages, we present the results of that survey and describe the main differences in reclassification across the state. In this report, when we refer to district p olicy, we are referen cing the policie s reported to us by school district staff in PPIC’s “English Learners Reclassification Survey.” Assessment of English-Language Proficiency: Districts should use the most recent CELDT test data as the primary criterion and consider for reclassification those students whose overall proficiency level is early advanced or higher and each subtest score is intermediate or higher. Students with ove rall proficiency levels in the upper end of intermediate may be considered for reclassification if additional measures determine the likelihood that a student is proficient in English. Teacher Evaluation: Districts should use a student’s academic performan ce and note that incurred deficits in motivation and academic success unrelated to English -language proficiency do not preclude a student from reclassification. Parent Opinion and Consultation: Districts should provide notice to parents or guardians of the ir right s and encourage them to participate in the reclassification process and provide an opportunity for a face -to -face meeting. Comparison of Performance in Basic Skills: A student’s score on the ELA portion of the CST or CMA in the range from the begin ning of Basic up to midpoint Basic suggests that the student may be sufficiently prepared to participate effectively in the curriculum. Districts may select a cut point in this range. Students with scores above the cut point should be considered for reclassification. For students scoring below the cut point, districts should determine whether factors other than English- language proficiency are responsible and whether it is reasonable to reclassify the student. For students in grade 12, the grade 11 CST re sults may be used. For students in grade 1 , districts should base reclassification decision s on the CELDT results, teacher evaluation, parent consultation, and other locally available assessment results. Districts must monitor student performance for two years after reclassification. SOURCE: California Department of Education. 2012, “California English Language Development Test (CELDT), 2012- 2013,” CELDT Information Guide. NOTE : Excerpted from CELDT Information Guide. See Appendix F for complete state guidelines . http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 22 PPIC ’s “ English Learners Reclassif ication Survey ” The PPIC reclassification survey was developed in consultation with EL experts and s everal large school di stricts. The structure of the PPIC reclassification survey reflects the structure of the reclassification guidelines. Districts were asked about the four main components of reclassification decisions: how they assess English proficiency, how they assess basic skills, how teachers’ evaluations are conducted and incorporated, and how parents are consulted and notified of reclassification decisions. A copy of the questions in the survey instrument is in Appendix B. A more complete description of the methods used to validate and distribute the survey is available in Appendix C . The survey was distributed to all local education agencies (LEAs) in existence in June 2013, based on contact information provided by the CDE. For LEAs that receive Title III funding, the primary survey contact was the Title III contact. For all remaining LEAs, the survey was distributed to the superintendent or charter school principal. The LEA contacts were encouraged to forward the survey to the most knowledgeable individual on LEA reclassification policies. Nearly half of the survey respondents were directors of EL services or EL coordinators (Table 3). TABLE 3 Survey respondents Job title Percent of respondents Director of EL services 26.4 EL coordinator/specialist 20.8 Superintendent 18.5 Other 12.9 Director of curriculum 9.9 Assistant superintendent 9.2 Teacher on special assignment 2.3 Total Respondents (N) 303 SOURCE: PPIC’s “English Learners Reclassification Survey.” The final sample used to report district EL policies consists of 303 school districts out of a total of 962 districts in the state .32 The district response rate was 31 percent; however, the districts that responded serve more than 54 percent of the state’s students. Elementary districts were the least likely to respond to the survey, and high school districts were the most likely. Our sample districts are fairly representative of the state as a whole. Of the largest 10 districts in the state, nine responded, which may explain why sample districts have more English Learners and low -income students than non-responders. In Appendix Table A1 , we show that using only districts that responded to our survey does not appreciably change the nature of our sampled 32 In addition to the 3 03 districts that responded, some county offices of education and charter schools or charter management organizations responded. We exclude county offices of education and charter schools from the survey results and analyses. Two districts are included in the survey results section but not in the student outcomes section because of data issues. http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 23 cohorts —indeed, as suggested in Table 4, while it reduces the number of students, it slightly increases the share of EL students in our final analysis. TABLE 4 S urv ey r es pond ents are broadly repr esen tative of the state’s student demographics District characteristics Responded Did not respond Elementary 139 397 Share of elementary students (%) 36 64 Average enrollment 3,037 1,744 High School 33 46 Share of high school students (%) 41 59 Average enrollment 7,439 6,245 Unified 131 208 Share of unified students (%) 64 36 Average enrollment 19,492 6,789 All districts Share of state’s students (%) 54 46 Share of Spanish- speaking ELs (%) 58 42 Share of all other language ELs (%) 62 38 API (average) 780 772 Low -income (average %) 60 56 English Learners (average %) 23 21 Reclassification rate (average %) 12 10 SOURCES: PPIC’s “English Learners Reclassification Survey” and PPIC’s School Finance Model (2013). NOTES: Enrollment is average daily attendance (ADA) from the PPIC School Finance Model (2013); averages are weighted by ADA. Reclassifi cation P ol icies In this section, we report our survey findings on districts’ reported EL programs and reported reclassification policies . 33 We a sked specifically about district policies for non -special -education EL students in structured English immersion (SEI) programs (all districts are required by state law to have an SEI p rogram). Among 33 The survey asks districts about their policies and in some instances their opinions about those policies. In some cases, however, districts may be reporting practi ces rather than policies adopted by their local plans. We did not collect locally adopted plans for comparison with the survey responses. http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 24 districts that also have English language mainstream programs (87%), all but one u se the sam e reclassification criteria as they do for SEI students. Fewer districts (37%) report that they also have alternative EL programs, such as bilingual or dual language programs; and the vast majority of the districts that do ha ve such programs employ the same reclassification criteria in alternative programs as they do for their students in SEI programs. Finally, almost all districts allow special education EL students to be reclassified (these policies are discussed in Appendix C). 34 In the remainder of this section, we consider how the districts have responded to the state guidelines for EL reclassification discussed in the preceding section: Assessment of English Language Proficiency, Evaluation of Basic Skills in English, Teacher Evaluation, and Parental Opinion and Consultation. We asked districts to report their reclassification policies by grade -level groupings: K –2, 3 –5, 6 – 8, and 9 –12. Assessm ent of E nglish Languag e P roficiency All school district s with ELs report using the CELDT to assess English proficiency (Table 5) .35 State guidelines suggest using an overall cut -off score of Early Advanced on the CELDT, a policy adopted by almo st all s cho ol districts across all grade levels. However, about 10 percent of districts report exceeding the state g uidelines and require a score of Advanced for students in grades K–8, and 7 percent require a score of Advanced for high school students (Table 5). Although an overall score of Early Advanced is recommended to demonstrate English proficiency, the CELDT has four subtests (speaking, listening, reading, and writing) , a n d it is possible 34 For each of the 13 state classifications of disabilities, more than 87 percent of districts reported that EL students with that disability could be reclassified. 35 One district reports that no ELs have enrolled for many years. Thus, in practice it has not administered the CELDT in reclassification decisions. EL Pr ogram s English Learners receive instruction in one of three settings. Structured English Immersion (SEI): A classroom setting where English Learners who have not yet acquired reasonable fluency in English, as defined by the school district, receive instruction through an English language acquisition process, in which nearly all classroom instruction is in English but with a curriculum and presentation designed for children who are learning the language. All districts are required to offer an SEI program. English Language Mainstream (ELM): A classroom setting for English Learners who have acquired reason able fluency in English, as defined by the school district. In addition to ELD instruction, English Learners continue to receive additional and appropriate educational services in order to recoup any academic deficits that may have been incurred in other a reas of the core curriculum as a result of language barriers. Alternative Program (Alt): A language acquisition process in which English Learners receive ELD instruction targeted to their English proficiency level and academic subjects are taught in the primary language, as defined by the school district. Placement in an alternative program is triggered by the parents through a parental exception waiver. http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 25 to achieve an overall score of Early Advanced while scoring below Early Advanced on on e ( or more) subtest .36 S tate guidelines recommend that Intermediate or higher in each subtest be considered sufficient as long as the overall score is Early Advanced. About 40 percent of districts exceed state guidelines and require that students achieve Early Advanced on each subtest. Conversely, 60 percent of districts follow t he state guidelines and accept a score of Intermediate on one or two CELDT subtests. TABLE 5 English proficiency criteria, by grade level English proficiency criteria Grades K–2 Grades 3– 5 Grades 6– 8 Grades 9– 12 Using CELDT (%) 99.3 99.6 99.6 100.0 Requiring overall cut -off score of Advanced (%) 9.8 9.6 9.5 7.1 Requiring subtests above Intermediate (%) 43.4 42.6 40.7 39.6 Respondents (N) 144 271 274 168 SOURCE: PPIC’s “English Learners Reclassification Survey.” Eva luation of Basic S kills in E nglish State law requires that the SBE guidelines utilize multiple criteria, including a comparison of the performance of the pupil in basic skills against an empirically established range of performance in basic skills based on the performance of English proficient pupils of the same age. SBE guidelines suggest using the ELA section of the CST and using a cut -off score somewhere in the range of Basic to mid point Basic (a scale score of 300 –325), which the majority of districts follow. However, about 30 percent of districts enrolling s tudents in grades 3 –8 report impo sing a higher cut -off poin t of Proficient (Table 6) . 37 Only one district reports using a cut- off score of Advanced. For high school students, districts typically use a lower cut- off score. About 30 percent of the districts use the CAHSEE (which is thought to be less rigorous than the CST ELA exam) as a supplemental or alternative measure of basic skills for high school students (not shown in table). TABLE 6 Percent of districts imposing basic skills requirements, by grade le vel Basic skills requirement Grade 2 Grades 3– 5 Grades 6– 8 Grades 9– 12 CST ELA 78.2 95.9 95.3 91.0 Basic cut -off point 22.8 27.0 29.9 37.1 Midpoint -Basic cut-off point 40.9 42.3 40.7 40.7 Proficient cut -off point 29.9 30.3 29.1 22.2 CST Math 46.9 47.4 45.8 36.5 36 Subtest weighting for the CELTD varies by gra de, with reading and writing each counting for only 5 percent of the overall score in kindergarten and first grade, and all subtests weighted equally for grades 2 –12 (CDE, 2012). 37 The CST is not administered until grade 2. http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 26 Basic skills requirement Grade 2 Grades 3– 5 Grades 6– 8 Grades 9– 12 Other CST 8.2 8.1 8.4 Writing 8.9 9.2 8.4 Respondents (N) 147 270 273 169 SOURCE: PPIC’s “English Learners Reclassification Survey.” Although state law and state guidelines refer only to basic skills in English, nearly half of all districts report requiring some mastery of math in grades 3 –8 (Table 6). Furthermore, about 8 percent of districts require that students demonstrate some level of basic skills in other subject tests, using either the science or history CST or both. And about 9 percent of districts require that students demonstrate written English proficiency, most commonly through a district writing rubric .38 About 7 percent require all of these additional measures (math CST, writing, and at least one additional CST) in determining a student’s mastery of basic skills. Table 6 does not report basic skills criteria used by districts for students in grades K –1. In the survey, districts were asked to describe the criteria used to assess the basic skills of EL students in grades 2 through 12. Because CST results are not available until late summer after the second grade, only 78 percent of the districts that reported reclassifying second grade students relied on CST results for reclassification . 39 Oth er dis tricts used different assessments. The most frequent responses for measuring the basic skills of stu dents in gra des K –1 included district or teacher rubrics and national assessments of reading development such as the Dynamic Indicators of Basic Early Literacy Skills (DIBELS).40 Teach er E va luation State l aw requires that the teacher evaluation include a review of the EL student’s curriculum mastery ( Education Code 313(f)(2)). State guidelines say that academic performance should be considered as part of the tea cher evaluat ion of EL students. Beyond that, the state offers little guidance about incorporating teachers’ perspectives in reclassification decisions. To capture the range of district practices, we asked districts about a host of potential compon ents of the teacher evaluation. Districts were also allowed to provide more information about a given component or to write in components not included in our list. We then categorized these components as “required” or “considered.” If a district reported that an individual component was required, we prompted th e respondent to enter a grade, GPA, or score, and we determined the component to be a required component of reclassification . 41 If the component was deemed by the district respondent to be “considered” rather than “requ ired,” we did not prompt for a grade, GPA, or score, and we determined the component to be a considered component of reclassification. Most districts reported using required criteria as part of the teacher evaluation (Table 7). Less than 5 percent of districts reported that they do not consider grades, GPA, or assessments across the various grade spans. 38 Responses regarding the testing of writing proficiency were write -in responses, not prompted in the survey questions. 39 Of the districts that serve second grade students, 54 percent responded that they allow students in second grade to be reclassified. See Appendix Table C2 . 40 DIBELS are a set of short (one minute) procedures and measures that can be used to regularly monitor and assess the acquisition of early literacy skills from kindergarten through sixth grade. 41 Requiring grades and GPA rather than just considering them does not mean a district is using entirely objective criteria because grades are assigned by teachers and likely include some component unrelated to test scores, such as dis cipline, effort, etc. See Parrish et al. (2006) for a discussion related to using grades in reclassification decisions. TABLE 6 (continued) http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 27 Grades or GPA are used more often in the teacher evaluation of high school students than in grades K –8. Conversely, districts are less likely to require assessments as part of the teacher evaluation in high school than in the lower grade levels. Most districts reported requiring a grade of “C” or better in English -language arts classes, although a few districts require a grade of “B” or better. This matches what districts reported for GPA requirements, which were generally a 2.0 or equivalent, typically associated with a grade of “C.” Districts reported using a variety of assessments that ranged from district rubrics to those developed by testing companies, such as the DIBELS. TABLE 7 Teacher evaluation criteria, by grade level Teacher criteria Grades K–2 Grades 3– 5 Grades 6– 8 Grades 9– 12 Grades/GPA requ ired 64.2 65.3 64.6 70.2 Grades/GPA considered 23.0 25.8 24.0 20.2 Assessments required 50.7 46.1 44.9 34.5 Assessments considered 39.9 43.9 40.5 45.2 None of the above 3.4 2.2 4.4 1.8 Respondents (N) 148 271 274 168 SOURCE: PPIC’s “English Learners Reclassification Survey.” NOTE: Rows do not sum to 100% because respondents could chose more than one category among the first four. Districts also reported using a variety of subjective components in the teacher evaluation criteria (Table 8). Although sta te guidelines appear to discourage the use of students’ behavior and motivation in teacher evaluations, many districts report considering factors such as student attendance, behavior, participation, and discipline. Districts are generally less likely to re port considering these other components for high school students than for students in grades K –8, with one exception: discipline. Disciplinary issues are more likely to be reported as considerations by districts in reclassifying high school ELs than in reclassifying K–8 ELs. Districts that report considering participation tend to be smaller an d to e nroll fewer free and reduced- price lunch students, while districts that report considering discipline an d beha vior tend to enroll fewer English Learners. Among elementary school ELs, elementary districts were more likely than unified districts to report considering behavior and discipline. Among high sch ool ELs, high scho ol districts were more likely to report considering attendance and discipline while unified di stricts were m ore likely to report considering participation and behavior. http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 28 TABLE 8 Subjective evaluation criteria used by teachers, by grade level Subjective criteria Grades K–2 Grades 3– 5 Grades 6– 8 Grades 9– 12 Attendance 30.4 29.5 31.2 23.5 Behavior 20.3 19.6 20.9 13.3 Benchmarks 84.5 80.4 81.4 67.5 Discipline 10.1 9.6 11.1 16.3 Homework 31.8 35.4 37.9 20.5 Participation 61.5 60.2 59.7 46.4 Portfolio 43.9 42.1 43.1 38.8 Projects 45.9 47.2 50.2 37.0 None of the above 13.5 16.2 23.7 27.7 Respondents (N) 148 271 253 166 SOURCE: PPIC’s “English Learners Reclassification Survey.” Parental O pinion and Consult ation State guidelines are equally limited for this element of reclassification policy, stating only that districts should provide parents with a notification of reclassification and the opportunity to participate in the reclassification decision and process, including the opportunity for an in -person meeting. To capture districts’ policies, we first asked how parents are consulted i n the reclassification decision and then how they are notified. Nearly all districts report that they explain reclassification criteria to parents (95%), present the student’s performance data (86%), and solicit parents’ opinions (85%). A majority of districts (62%) report that they provide parents with a comparison of students’ performance data to the reclassification criteria. A handful of districts say they do not consult parents or consult parents in some other manner. Parents are primarily informed o f the district’s reclassification decision by letter (88%) or an in -person meeting (75%). About 58 percent of districts report that they inform parents by phone. Timing State law does not specifically prohibit reclassification in certain grade levels; state guidelines recommend using the CST in grades 2 –12 and an alternative district assessment in first grade (before the CST can be used) as measures of English proficiency. Nearly half or more of all districts report that they do not reclassify in the early grades (K –2): About 30 percent of districts report permitting reclassification in kindergarten, 47 percent in grade 1, and 54 percent in grade 2. Almost all districts report permitting the reclassification of ELs by grade 3 (Figure 4). Given these findings, it may not be surprising that students are more likely to be reclas sified relatively soon after the results of the first CST are available in the summer following se cond grade, many of which count as third grade reclassifications. According to our respondents, ELs are most commo nly reclassified in the middle grades (4 –6), with 5th grade being the modal response http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 29 (56% of di stricts). Parrish et al . (2006) found that reclassification rates were lower in districts that did not permit reclassification until grade 3. 42 FIGURE 4 Percent of districts that permit reclassification, by grade level SOURCE: PPIC’s “English Learners Reclassification Survey.” The reclassification of students is also more common during certain periods of the academic year. The vast majority of districts report reclassifying students in the spring (using the CST that students took nearly 12 months earlier) or in the fall (when the most recent CST scores are available). However, most d istricts reclassify students at multiple points during the year (Figure 5). About 30 percent of districts reclas sify during only one season (fall, winter, spring, or summer). About 2 percent reclassify year -round. Parrish et al . (2006) found that districts reclassifying multiple times durin g the schoo l year were likely to have higher reclassification rates. 42 Future iterations of this research could consider this factor. 0 10 20 30 40 50 60 70 80 90 100 K12345678910 1112 Percent of districts Grade level http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 30 FIGURE 5 Most districts allow for student reclassification at more than one point in time per year The state guidelines were last amended in September of 2006. Very few districts report initiating any changes in their reclassification policies since 2008. About 5 percent of districts changed their English proficiency standards by either using a different overall CELDT score or by changing t heir required CELDT subtest scores. About 15 percent of districts made changes to their basic skills requirements, most often changing the cut -off score required on the ELA portion of the CST or changing requirements for the math CST. Among all districts, 8 percent have made changes to their teacher evaluation standards since 2008. Opinions a bout Re cla ssific ation P olic ies We asked survey respondents about their opinions of their district’s reclassification policies —for example, which aspect of reclassificat ion criteria is the most difficult to achieve by school level (elementary, middle, or high school), allowing just one choice among each of the four criteria. The majority of respondents believe that the basic skills requirement (commonly measured by the CST ELA) is the most difficult requirement across all three grade spans (Table 9), and that it is more difficult for middle - and high school students than for elementary students. Many respondents (40%) consider English proficiency (as measured by the CELDT) to be a more difficult hurdle for elementary school students than for middle- and high school students (27% and 26%, respectively). Very few respondents consider teacher evaluation or parent cons ultation to be difficult barriers (teacher evaluation was selected more often for middle- and high school students than for elem entary students). 0 10 20 30 40 50 60 1 season 2 seasons 3 seasonsYear-round Percent of districts Number of seasons in which reclassification can occur http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 31 TABLE 9 “If you were forced to choose one criteria for each grade level, which of the reclassification criteria would you say is most difficult for EL students to meet in your district?” Elementary Middle High Basic skills 52.8 62.3 67.9 English proficiency 40.1 26.8 25.6 Teacher evaluation 3.3 6.0 8.9 Parent consultation 0.0 0.0 0.5 Don’t know 3.7 4.9 9.5 Respondents (N) 269 265 168 SOURCE: PPIC’s “English Learners Reclassification Survey.” We also asked respondents to rank the importance of the various reclassification criteria for deciding whether to reclassify a student (Table 10). English proficiency (38%) was most likely to rank at the top. Although most respondents believe that basic skills are the most difficult hurdle for students to surmount (Table 9), only 5 percent of respondents ranked this criterion higher than all other factors influencing their decision to reclassify students (Table 10). Another 26 percent of respondents selected basic skills in combination with English proficiency as being most important in the ultimate decision to reclassify a student, and an additional 23 percent ranked basic skills, English proficiency, and teacher evaluation equal ly as th e most important criteria. TABLE 10 “In your opinion, how important are each of the reclassification criteria in the ultimate decision to reclassify a student?” Agree (%) English proficiency is most important 37.6 English proficiency and basic skills are equally important 25.9 English proficiency, basic skills, and teacher evaluation are equally important 22.8 Basic skills are most important 5.2 English proficiency and teacher evaluation are equally important 5.2 Teacher evaluation is most important 4.8 Respondents (N) 290 SOURCE: PPIC’s “English Learners Reclassification Survey.” http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 32 In 2011– 12, reclassification rates ranged from 0 to about 30 percent statewide .43 The vast majority of respondents (94%) believe that their district’s reclassification policies are “about right” (shown in Appendix C), while about 78 percent of respondents believe their reclassification rate is “about right” —with a lmost all of the others believing their reclassification rate is “too low” (shown in Appendix C). Among those who believe the reclassification rate is too low, the vast majority think that their reclassification policies are about right. This would seem to suggest that other factors (such as EL students’ performance or implementation of the reclassification policies) rather than the reclassification polic ies themselves, determine respondents’ satisfaction with district reclassification rates. Summing Up PPIC’s 2013 survey of school district reclassification policies and practices garnered a response rate of 31 percent of districts, representing 54 percent of California’s K –12 students. This report confirms the findings of past research (using small samples of districts) that there is substantial variation in the reclassification practi ces and policies of California school districts. In political and practitioner circles, such variation is often viewed in a negative light because it means that students with the same skills will face different reclassification decisions in different parts of the state. There is also concern statewide about what the “ right” reclassification policy should be. Clearly, based on the survey responses, some districts believe that the state guidelines offer too low a threshold, and they impose many more requirements, particularly with regard to basic skills, including writing samples or content knowledge in other subjects. Do these policy differences affect reclassification rates, which also vary widely across the state (from 0 to 30%)? We address this question in the following section. 43 There are a few districts with substantially higher reclassification rates, but these are districts with very few EL students. The rates were calculated using ELs redesignated during the year in the numerator and the sum of ELs and ELs redesignated in the denomin ator. Data are from DataQuest (http://dq.cde.ca.gov/dataquest/LC/LCOtherDistrict.aspx?dType=all&co=AllCounties&TheYear=2011 -12&sortby=c ). http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 33 Do District Reclas sification P olicie s Infl uence Recla ssifica tion Ra tes? As demonstrated in the previous section, there is a great deal of variation in how and when districts decide that EL students are ready to be reclassified. It is important to understand whether the differences in reclassification policies and practices are associated with differences in district reclassification rates and ultimately, whether the differences in these policies and practices are associated with differences in reclassified students’ outcomes. In this section, we examine those district reclassification policies that are more rigorous than recommended by the SBE guidelines , and using multiple regression analysis, we test w hether districts us ing those more rigorous requirements have lower reclassification rates. Below, we describe in more detail how we categorize the districts’ reclassification policies, then calculate districts’ reclassification r ates and consider the role of other district characteristics. Recall from the section ab ove that the SBE provides guidelines to districts for reclassifying students, most specifically in terms of basic skills and English proficiency. School districts have had more considerable latitude with regard to implementing the criteria for the teacher evaluation and parental opinion and consultation . In modeling the association between district reclassification policies and district reclassification rates, we use reclassification policies and practices from 2012 –13 and reclassification rates from the same period. Because reclassification policies can vary by grade level within a district, we chose the policies for the grade levels in which most reclassification occurs when modeling the overall reclassification rate for the district. For elementary and unified school districts, we used the reclassification policies reported for grades 3 –5.44 Table 11 summarizes how districts reported implementing their own reclassification criteria in 2012 –13. It is stra ightforward to construct a variable to capture whether a district is using a higher threshold than suggested by SBE’s reclassification guidelines in the district’s use of the CELDT and the CST ELA, as well as the effect of requiring additional tests, such as the math CST. State guidelines are less definitive in the case of teacher evaluations, and thus interpreting whether district requirements exceed the state guideline is less clear. If a district reports requiring specific grades (or GPA) or specific assessment results for reclassificati on, we consider that to be a higher threshold than is specified in the state educational code. Districts reporting teacher evaluation components that include considering grades (or GPA) or assessments but not requiring e ither are cons idered to be using the components suggested in the SBE guidelines. If a district reported that it used disciplinary issues as a factor in reclassification decisions, we included that as well. As we noted in the previous section, our survey was not successful in eliciting detailed differences in districts’ parental involvement practices. In addition, district respondents did not rank parental involvement as particularly important in reclassification decisions, so we do not include it in our models that estimate district reclassification rates. 44 Among elementary districts, 97 percent have the same basic skills criteria as for grades 6 -8, and 96 percent have the same CELDT criteria. In unified districts, 94 percent have the same basic skills criteria for grades 6 – 8, and 91 percent have the same criteria for grades 9 –12. CELDT requirements are more likely to be the same across grade levels in unified districts (97% in grades 6 –8 and 96% in grades 9– 12). http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 34 TABLE 11 District reclassification policy Reclassification criteria and requirements Reclassification policies grades 3–5 Reclassification policies grades 9 –12 Elementary districts (%) Unified districts (%) High school districts (%) English Language Proficiency CELDT Overall Performance Level (OPL) OPL is Early Advanced (EA) or higher SBE guideline 88% 91% 97% OPL is Advanced More rigorous 12 7 3 CELDT subtests (reading, writing, speaking, listening) Subtests can be below Early Advanced SBE guideline 51% 65% 50% Subtest scores must all be Early Advanced or higher More rigorous 49 35 50 Basic Skills in English CST ELA Score is Basic or higher SBE guideline 27% 27% 55% Score is mid -Basic or higher More rigorous 39 45 36 Score is Proficient or higher More rigorous 33 27 10 Requiring CST Math More rigorous 57 37 35 Requiring CST history/science More rigorous 9 7 16 Teacher Evaluation Unspecified SBE guideline 24% 22% 13% Require grades/GPA and assessments More rigorous 42 28 13 Require grades/GPA More rigorous 24 37 69 Require assessments More rigorous 9 13 6 Consider disciplinary issues Discouraged 11 9 29 SOURCE : Authors’ tabulations from PPIC’s “English Learners Reclassification Survey.” In addition, we examined the incidence of districts’ use of multiple reclassification requirements that exceed the SBE guidelin es. While none of our surveyed districts employs a ll of the requirements shown in Table 11, most use several of the measures. Of the 294 districts with full reclassification information, only 11 report http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 35 using the measures specified in the SBE guidelines; 42 use one of the more rigorous measures, and 155 (or 53%) use th ree or m ore of these more rigorous measures (See Appendix Table C24 ). However, there are so many different combinations in use that it is, in practice, impossible to test them all. We settled on using the two most common requirement combinations: 1. requiring that all CELDT subscores be Early Advanced or higher and using the CST math test (57 districts), 2. requiring that all CELDT subscores be Early Advanced or higher and requiring both student grades/GPA and teacher assessments (43 districts). In addition to the reclassification policies for each district, we considered the role of district characteristics, such as type of district (elementary, unified, or high school), size of district, 2011– 12 API score, percent of Title I students, average daily attendance, and percent of students that are ELs. District reclassification rates are from CDE’s DataQuest. When 2012–13 reclassification rates were missing but 2011 –12 rates were available (35 districts), we used those and estimated our model for the 291 districts that had complete survey responses and reclassification rate data . 45 The 2012 –13 statewide reclas sification rate was 11.1 percent. The average reclassification rate for surveyed districts with complete responses was 11.2 percent, ranging from 0 to 75 percent, with half of the districts having reclassification rates in the range of 6.2 to 14.5 percent. Just 5 percent of the districts had reclassification rates over 24.5 percent, while 11 percent had reclassification rates lower than 1 percent. E stablishing higher cut -offs on the required state assessments is linked to lower reclassification rates, compared to requiring just the thresholds in the SBE guidelines; in some cases, these differences are statistically significant (see full results in Appendix Table D1 ). Figure 6 shows the reclassification rates for an elementary school district using the SBE reclassification guidelines (first bar, 15%). Each succeeding bar illust rates the predicted reclassification rate of an average unified school district using one of the more stringent requirements, holding other district characteristics constant. 45 Reclassification rates are calculated by dividing the number of RFEP students reclassified since the previous years’ language census by the sum of EL students and RFEP students reclassified since the previous years’ language census. http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 36 FIGURE 6 District reclassification rates under different reclassification requirement scenarios SOURCES: PPIC’s “English Learner s Reclassification Survey” and CDE reclassification rates . NOTE : *** p < .01; ** p<.05, * p<.10 . Regression also included districts’ use of CST other, grades required, and assessments req uired. To e stimate the reclassification rate for the average unified school district (first bar), we multiplied regression coefficients (in Table D1) for district characteristics by their mean values for unified school districts and added the sum to the constant. Sub sequent bars add the coefficient for significant reclassification criteria to the predicted d istrict reclassification rate. Districts that require a CELDT Overall Performance Level of Advanced have lower reclassification rates than those using the state minimum score of Early Advanced. Districts that require all CELDT subtests to be Early Advanced or higher also have lower reclassification rates. The effect of these two requirements is similar — compared to using the minimum criteria in SBE’s guidelines, using the higher cut- offs on the overall CELDT and the CELDT subtests are each associated with district reclassification rates about 3 percentage points lower, a substantial reduction in reclassification. A requirement that CST ELA scores be higher (either mid- Basic or Proficient) is also associated with overall lower reclassification rates —4 percentage points lower for a minimum score of mid -Basic and 3 percentage points lower for a minimum score o f Proficient. Holdi ng other factors constant, requiring the CST math test alone was not linked to a statistically significant difference in reclassification rates. Nor did requiring the results of either the CST history or CST science test indicate statistically significant differences in rates . 46 Reporting the use of disciplinary issues in reclassification was 46 In this report, we did not consider the role of the score required for the CST math, history, or science tests. 15.1% 12.2% 12.2% 11.1% 12.0% 16.0% 12.1% 20.1% 9.6% 0% 5% 10% 15% 20% 25% Percent (%) Reclassification criteria * *** * *** *** * * http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 37 actually associated with higher reclassification rates, suggesting that a lack of disciplinary problem s elevates a student’s chances of reclassification, all else equal . Looking finally at two common combinations of requirements that go beyond the SBE guidelines, we find that districts stipulating both a grade or GPA minimum and required scores on particular assessments have lower reclassif ication rates (about 3 percentage points lower), but this result is only statistically significant at the 10 perc ent level. Holding district characteristics constant, those districts that have a reclassification policy that uses two of the more rigorous reclassification requirements (requiring that the CELDT subt est scores be Early Advanced or higher and requiring the math CST) have reclassification rates 5.4 percentage points lower than districts using neither of these more stringent criteria. We did not find any statistical relationship between other measures of teacher evaluation criteria and district reclassification rates. Summing Up This section has demonstrated a clear association between reported reclassification policies and reclassification rates at the district level. Policies with more- rigorous performance thresholds are associated with lower reclassification rates, holding constant other district characteristics that may also determine reclassification rates. If districts use multiple additional requirements, reclassification rates are further reduced. We tested the most common combination —the CST math requirement and requiring that results on all subtests of the CST ELA be at least Early Advanced—and found that reclassification policies that use those two requirements together redu ce district reclassification rates by 5.4 percentage points relative to the criteria suggested in the SBE guidelines for reclassification. Because fewer students are reclassified in d istricts with higher reclassification thresholds, we might expect to see that those students who succeed in b eing reclassified have better academic outcomes than students who do not. We explore this question in the following section. http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 38 Do Re clas sification P olicie s A ffect Stud ent O utcom es? Available measures indicate that reclassified EL students are among the best performing students in the state. This has been clearly demonstrated using both cross -sectional data (e.g., Hill, 2012; Saunders and Marcelletti, 2013) and longitudinal data (e.g., Flores, Painter, and Pachon 2009), as well as in this report. We have also demonstrated that some or all of the more rigorous reclassification performance thresholds are associated with lower district reclassification rates. Some of the more difficult reclassification performance thresholds may result in keeping EL students classified as ELs for longer durations. In this section, we first examine whether RFEP students are still among the best performing students in comparison to other student language groups when we simultaneously consider the role of student and district characteristics. We then focus more specifically on whether more difficult reclassification standards are associated with better RFEP outcomes. Longitudi nal RFEP Outcomes Earlier in this paper, we noted that cross- sectional research shows that RFEP students outperform EL students in high school but that their advantage appears to shrink at higher grades. Our analysis follows cohorts of EL students who are reclassified at various grade levels and compares them to EO students, to IFEP students, and to students who are never reclassified. This enables us to understand student outcomes for each group of RFEP and EL students without the potential complication of newly arriving ELs or RFEP students who are reclassified at higher grades being indistinguishable from RFEP students reclassified at younger grades. The academic outcomes we consider are CST ELA scores, on -time grade progression, and end -of -high- school outcomes .47 We use multivariate regression analysis to simultaneously consider the role of student characteristics and district characteristics for each of our cohorts of students. Full regression model results are available in Appendix E .48 Figures 7 and 8 indicate that if we use longitudinal data and control for student and district characteristics, RFEP students outperform ELs students on the CST ELA, no matter when they are reclassified (in our target reclassification year, beforehand, or afterward), and no matter what cohort we examine (2nd, 4th, or 7th grade). Again, this is not surprising given that reclassification criteria are intended to separate hi gh-performing ELs from those who could benefit from more English language support. This is true whether we are considering the share of students in each language group who score Basic and higher (Figu re 7) or Proficient and higher (Figure 8) . 49 The gap between CST scores for RFEP students reclassified in a pre -target reclassification year and those reclassified in a post -target reclassification year is larger for older than for younger cohorts. These findings suggest that the cross -sectional view of RFEP performance in high school grades is skewed by the late entrants —those reclassified in grades 8 to 12 generally have lower levels of performance than those reclassified at younger grades. 47 Subsequent extensions of this research could consider other outcomes, such as performance on the CST math and CAHSEE tests. 48 We also estimated models where we did drop all students who transferred districts, including those that moved from their elementa ry to th e appropriate high school or unified dis trict. We found that when we did not allow students to transfer to the appropriated unified or high school district at the end of elementary school, the results were similar. We found that coefficients were of similar sizes and were statically significant in the same cases as in models where students could transfer school districts. The models with no transferring allowed did have sli ghtly hig her R - squared values. These results are available from the authors upon request. 49 Full regression results are available in Appendix E . http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 39 FIGURE 7 Share of cohort, by language group, scoring Basic or higher on CST ELA, controlling for district and student characteristics SOURCES: PPIC’s “English Learners Reclassification Survey” and CALPADS. NOTE: Results for language groups are statistically different than those for EO students with the exception of 7th grade RFEP 2008 –09 students. Predicted values for ELS in 11th grade are slightly negative. Full regression results are available in Appendix Table E1 . RFEP students in the 2nd and 4th grade cohorts also outperform EO students when the performance threshold is scoring Basic or higher on the CST ELA (Figure 7). In the 7th grade cohort, only students reclassified prior to and during the target year are more likely than EO students to score Basic or higher. When the performance threshold is higher (Proficient or higher on the CST ELA, Figure 8), RFEP students reclassified prior to the target year are more likely than EO students to perform above the threshold. However, students reclassified after the target year do not, and only 2nd and 4th grade cohort RFEPs reclassified in the target year exceed EO students. Figures 7 and 8 also show that IFEP students are among the best performing students in the state, although there are many instances in which RFEP students achieve better outcomes. RFEP students reclassified prior to the target reclassification year are more likely than IFEP students to achieve a score above the CST ELA performance measures, with the exception of the 7th grade cohort scoring Proficient or higher. Students reclassified in the target year in the 2nd grade cohort are also more likely than IFEP students to exceed the CST ELA performance measures. 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 2nd grade cohort, 6th grade CST 4th grade cohort, 8th grade CST 7th grade cohort, 11th grade CST Percent scoring Basic or above on 2011-2012 CST ELA Cohort and language group RFEP target year (2008-09) RFEP pre RFEP post IFEP EL EO http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 40 FIGURE 8 Share of cohort, by language group, scoring Proficient or higher on CST ELA, controlling for district and student characteristics SOURCE: PPIC’s “English Learners Reclassification Survey” and CALPADS. NOTE: Results for language groups are statistically different than those for EO students. Predicted values f or ELs in 11th grade are slightly negative. Full regression results are available in Appendix Table E2 . We then follow our 7th grade cohort into their high school years. One importa nt outcome for students is on - time progression to 10th grade. Many students with academic deficits spend more than one year classified as ninth graders. Here, we find that RFEP students reclassified in 8th grade or earlier (target reclassification year or pre -target reclassification year) are much more likely to have made on -time (or better) progress to their 10th grade year than ELs (Figure 9). Those reclassified before 8th grade are just as likely, or even slightly more likely, than IFEP and EO students to be achieving on -time progress to 10th gra de. Students reclassified post -target year are more likely than EL students to have achieved on-time grade pr ogress (similar to EO students, but slightly behind IFEP students). 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 2nd grade cohort, 6th grade CST 4th grade cohort, 8th grade CST 7th grade cohort, 11th grade CST Percent scoring Proficient or above on 2011-2012 CST ELA Cohort and language group RFEP target year (2008-09) RFEP pre RFEP post IFEP EL EO http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 41 FIGURE 9 Percent on -time or better in 10th grade, by language group, 7th grade cohort, controlling for district and student characteristics SOURCE: PPIC’s “English Learners Reclassification Survey” and CALPADS. NOTE: Results for ELs, RFEP pre-target year, and IFEPs are statistically significant from EO students. Full regression results are available in Appendix Table E3 . We can follow our final cohort, those who are 8th graders in 2007– 08, through the end of high school (one year beyond what would have been on -time 12th grade completion). We find that even after controlling for individual and district characteristics, RFEP students have much better high school outcomes than EL and EO students (Figure 10). RFEP students are the least likely of all groups to leave high school without graduating. RFEP students, no matter when reclassified are more likely than any other language group, even IFEP students, to have earned a diploma by 2012– 2013. IFEP stude nts are more likely than other language groups, with the exception of students reclassified prior to the target year, to have completed their a – g requirements before graduation. 76% 78% 80% 82% 84% 86% 88% 90% 92% RFEP target year (2008-09) RFEP pre RFEP post IFEPELEO Percent on-time or better, 10th grade Language group, 7th grade cohort http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 42 FIGURE 10 Final high school outcomes for 8th grade cohort, controlling for district and student characteristics SOURCES: PPIC’s “English Learners Reclassification Survey” and CALPADS. NOTE: Results for language groups are statistically different than those for EO students. Full regression results are available in Technical Appendix Tables E4 –E6 . As this section demonstrates, RFEP students achieve much better academic outcomes than EL students, even after controlling for some systematic differences in student and district characteristics. RFEPs who left the EL group at younger grades (generally before 9th grade) often have better outcomes than EO and sometimes than IFEP students. This finding might prompt policymakers to consider if reclassification thresholds might be set too high, a consideration we return to in the conclusion of this report. In the following section, we examine whether these results might be driven by districts that require high performance from students before they reclassify them. Relationsh ips Be tween More Rigorous Reclassifi cation Policies and RFEP Outcomes In this section, we test the relationship between districts’ reclassification criteria and “ever- classified as EL” (i.e., EL and RFEP) student outcomes. Using ordinary least squares multiple regression, we examine a number of key academic outcomes for our EL and RFEP student cohorts: 2011– 12 CST ELA scores, on -time grade progression, leaving high school without graduating, graduating, and graduating having completed a– g requirements. We simultaneously consider the same district - and student -level characteristics used to predict student outcomes in the previous section. We also include the same reclassification policy requirements that 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Leaving HS before graduating DiplomaA-g courses complete Percent with high school outcome RFEP target year (2008-09) RFEP pre RFEP post IFEP EL EO http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 43 we used in our examination of the association between the criteria and district reclassification rates, testing the requirements s hown in the box below. Outcomes for Students Reclassified in Elementary or Middle School We use reclassification policy variables from grades 3 to 5 for the 2nd and 4th grade cohorts, from grades 6 to 8 for the 7th grade cohort, and from grades 9 to 12 for the 8th grade cohort. When simultaneously considering the role of student and district characteristics, we find a statistically significant role of district reclassification policies in predicting EL and RFEP student outcomes. In general, district policies appear to matter somewhat more for the outcomes of younger students, and in the direction we would expect. More of the variables we considered are statistically significant for our 2nd grade cohort, and the association between the reclassification category and the outcome is almost always stronger for the younger students —i.e., the 2nd and 4th grade cohorts versus the 7th grade cohort (see Table 12; full regression results are available in Appendix Tables E7, E8, and E9 ). Not surprisingly, students in districts requiring higher CST ELA scores for reclassification have a greater proportion of students scoring at least Basic (columns 1 –3) or at least Proficient (columns 4 –6) on the CST ELA in 2011 –12 than students in districts that do not. Higher CELDT score requirements are associated with being more likely to score Proficient or higher on the CST outcomes for the cohort of 2nd grade RFEP students and a Basic or higher for the cohort of 7th grade students. If a district’s teacher evaluation reclassification standard requires some use of grades/GPA and/or assessments, RFEP students have higher CST scores than those in districts without teacher evaluation requirements. Increases over the Basic threshold range from 4 to 6 percentage points for the 2nd and 4th grade cohorts and from 0 to 5 percentage points for the 7th grade cohort. Results are similar for the association with teacher evaluation requirements and increases over the Proficient threshold, with 2nd and 4th grade cohorts CELDT Overall Performance Level of Advanced (versus Early Advanced or higher) CELDT subtest (reading, writing, speaking, listening) scores must all be Early Advanced or higher (versus allowing some intermediate scores) CST ELA score of mid-Basic or higher (versus Basic or higher) CST ELA score of Proficient or higher (versus Basic or higher) CST math test versus no CST test requirement beyond CST ELA CST history/science test versus no CST test requirement beyond CST ELA Require grades/GPA and assessments versus just considering grades/GPA and assessments Require grades/GPA versus just considering grades/GPA and assessments Require assessments versus just considering grades/GPA and assessments Consider disciplinary issues versus not considering disciplinary issues CELDT subtest scores must all be Early Advanced or higher AND require CST math test CELDT subtest scores must all be Early Advanced or higher AND require grades/GPA and assessments http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 44 achieving increases around 4 to 6 percentage points. For the 7th grade cohort, effects are slightly negative. If districts consider a student’s disciplinary record, it has a positive effect on CST scores for youngest cohort (2nd grade) but a negative effect for the 7th grade cohort. This mixed finding suggests that considering a student’s lack of disciplinary problems helps students in younger grades but hinders students in older grades. Recall that overall, it is associated with higher, not lower, reclassification rates. http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 45 TABLE 12 Per centage point change for student outcomes and district recl assification rat es associated with specific reclassification requirements Percentage Point Change Attaining Basic CST ELA Attaining Proficient CST ELA 10th grade on time Reclassification rate 2nd grade cohort 4th grade cohort 7th grade cohort 2nd grade cohort 4th grade cohort 7th grade cohort 7th grade cohort District English Proficiency CELDT Advanced 0.002 -0.014 0.067 *** 0.068 *** 0.024 0.010 -0.060 *** -0.029 * CELDT no intermediate 0.007 0.001 -0.181 *** 0.037 *** 0.012 -0.140 *** -0.045 *** -0.028 * Basic Skills CST ELA mid- Basic 0.028 *** 0.049 *** 0.027 *** 0.020 *** 0.041 *** 0.017 ** 0.036 *** -0.040 ** CST Proficient 0.042 *** 0.054 *** 0.011 0.044 *** 0.056 *** 0.030 *** 0.051 *** -0.030 ** CST Math 0.001 -0.008 -0.016 * -0.037 *** -0.058 *** -0.004 -0.013 * 0.009 CST Other -0.005 0.030 ** 0.044 *** 0.037 *** 0.056 *** 0.034 ** 0.010 -0.005 Teacher Evaluation Grades/GPA and Assessments required 0.042 *** 0.042 *** 0.011 0.036 *** 0.037 *** -0.052 *** 0.012 -0.030 * Grades (only) required 0.040 *** 0.042 *** 0.050 *** 0.055 *** 0.046 *** 0.004 0.026 *** -0.015 Assessments (only) required 0.055 *** 0.034 *** 0.038 *** 0.041 *** 0.014 0.009 0.028 *** -0.002 Disciplinary record 0.032 ** 0.022 -0.089 *** 0.007 0.010 -0.063 *** -0.011 * 0.050 ** Combination interaction terms CELDT no intermediate and CST Math 0.039 * 0.040 * 0.198 *** 0.032 * 0.066 *** 0.114 *** 0.061 *** -0.035 * CELDT no intermediate and Grades & Assessment 0.024 * 0.046 * 0.098 *** 0.031 * 0.031 0.122 *** 0.060 *** 0.014 Combination net effects CELDT no intermediate and CST Math 0.047 0.033 0.001 0.031 0.020 -0.030 0.002 -0.054 CELDT no intermediate and Grades & Assessment 0.073 0.088 -0.072 0.103 0.080 -0.070 0.027 -0.044 SOURCES: PPIC’s “English Learners Reclassification Survey” and CALPADS. NOTES: *** p<0.01, ** p< 0.05, * p<0.1, ** p<0.05 and * p <0.10 for district reclassification rates. Full regression results for student outcomes are available in Appendix Tables E7 –E9 and for district reclassification rates in Appendix Table E1 . CST scores are measured in 2011–2012 which is the 6th grade for the 2nd grade cohort, 8th grade for the 4th grade coho rt, and 11th grade for the 7th grade cohort. http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 46 In a district that uses common combinations of more rigorous reclassification standards, CST scores are higher, especially among younger students. For example, in a district that requires all CELDT scores to be at least Early Advanced and that uses the mat h CST, the share of RFEP students from the 2nd grade cohort with scores on the CST ELA that are above the Basic threshold is 4.7 percentage points higher than a district that requires neither. The effect is smaller for older cohorts. Districts using reclas sification policies with these performance thresholds have a 5 percentage point lower reclassification rate. More rigorous reclassification policies do appear linked to greater percentages of our 7th grade RFEP cohort being “on -time” 10th graders. Requirin g higher CST thresholds increases the share of RFEP students being “on -time” 10th graders by 4 to 5 percentage points. However, doing so is associated with a 3 to 4 percentage point lower reclassification rate. Similarly, teacher evaluation standards that are more demanding are associated with more on -time 10th grade RFEPs (3 percentage points more), but with reclassification rate reductions as well, although these are not statistically significant, likely due to small sample sizes in our model (see Appendix Table D1 ). Elevated CELDT score requirements, on their own, are associated with lower shares of RFEPs being on time for 10th grade, but districts requiring both CELDT subtest scores to be Early Advanced or higher and requiring grades/GPA and assessments raises the share of RFEP students being on time in 10th grade by 3 percentage points. Outcomes for Students Reclassified in High School The common combinations of reclassification requirements we tested are statistically significant (Table 1 3). Our estimate of their net effects suggest they make little difference in reducing the share of RFEPs who leave high school without graduating (a subset of the dropouts as defined by CDE) , 50 no r do they make much difference in the share that earn a diploma. However, using the combined criteria is associated with reduced shares of RFEP students completing their a –g requirements. Recall that using these standards is associated with lower district reclassification rates of between 4 and 5 percentage points. More rigorous CST ELA requirements are associated with reduced shares of RFEP students earning diplomas and meeting their a –g requirements and increased shares leaving high school before graduati ng, a counter-intuitive finding. Perhaps remaining in EL status longer in districts with more rigorous reclassification standards means losing access to some academic instruction that would have helped with a –g requirements and high school completion. Howe ver, we do observe somewhat improved “end of high school” outcomes for RFEP students in districts where the history and/or social science CST is required. The teacher evalua tion requirements are linked to higher percent ages of students persisting in high school (i.e., not leaving before graduation) and earning a diploma. There are no statistically s ignificant associations between district’s use of disciplinary issues in reclassification decisions for their high school students and outcomes for those studen ts. Full regression results are available in Appendix Tables E10, E11, and E12 . 50 Recall from Figures 3 and 8 that relatively few RFEP students leave high school without a diploma: Only 100 RFEPs who were reclassified between 2008– 09 and 2012–13 dropped out of school. http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 47 TABLE 13 End of high school o utcomes, cohort 4 High School Outcomes, Percentage Point Change Reclassification Rate, Percentage Point Change Leaving HS before graduating Diploma a –g requirements District English Proficiency CELDT Advanced -0.012 0.059 ** 0.095 *** -0.029 * CELDT no Intermediate 0.028 *** -0.110 *** -0.191 *** -0.028 * Basic Skills CST ELA m id-Basic 0.010 *** -0.012 -0.050 *** -0.040 ** CST Proficient 0.004 -0.049 *** 0.006 -0.030 ** CST Math 0.002 0.021 * -0.026 * 0.009 CST Other -0.021 ** 0.060 *** 0.110 *** -0.005 Teacher Evaluation Grades/GPA and assessments required 0.005 0.029 ** -0.102 *** -0.030 * Grades (only) required -0.015 *** 0.030 *** -0.005 -0.015 Assessments (only) required -0.007 0.021 0.008 -0.002 Disciplinary record 0.008 -0.012 -0.008 0.050 ** Combination CELDT no intermediate and CST Math -0.019 * 0.100 *** 0.194 *** -0.035 * CELDT no intermediate and require Grades/GPA and Assessments -0.038 *** 0.073 *** 0.184 *** 0.014 Combination Net Effects CELDT no intermediate and CST Math 0.011 0.011 -0.023 -0.054 CELDT no intermediate and require Grades /GPA and Assessment -0.005 -0.008 -0.109 -0.044 SOURCES PPIC’s “English Learners Reclassification Survey” and CALPADS. NOTE: *** p<0.01, ** p< 0.05, * p<0.1, ** p<0.05 and * p <0.10 for district reclassification rates. Full regression results for student outcomes are available in Appendix Tables E10 – E12 and for district r eclassification rates in Appendix Table E1 . Looking across all outcomes, we can attempt to generalize about the benefits and costs of different reclassification policies. Requiring a CELDT score of Advanced improves academic outcomes for students in the 7th and 8th grade cohorts, but not for younger cohorts. Not allowing CELDT subtest scores to fall below Early Advanced is associated with better outcomes for the 2nd grade cohort, but not for 7th and 8th grade cohorts. Having higher CST ELA cut -offs improves outcomes for the three youngest cohorts (2n d, 4th, and 7th grade) but does not improve outcomes for the oldest cohort (8th grade). The gains for the younger cohorts are relatively small, especially relative to the reduction in district reclassification rates. For example, http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 48 students reclassified in 5 th grade in districts requiring a score of Proficient or above on the CST ELA are 5 percentage points more likely to score Proficient or above on their 8th grade CSTs than similar students in districts without that requireme nt, translating to 66 percent r ather than 61 percent scoring Proficient or above. If the average annual reclassification rate among districts using criteria suggested in SBE guidelines were 10 percent, districts that add this more rigorous standard would have reclassification rates of 7 percent because the requirement is associated with a 3 percentage point reduction in reclassification rates. Using an additional CST assessment (i.e., history or social science) is associated with better academic outcomes for all students. Recall that we did not find that using this requirement was associated with lower reclassification rates overall, but that this is probably because this requirement is used relatively infrequently (in 9% of elementary, 7% of unified, and 16% of high school districts). Requiring, rather than considering, some form of teacher evaluation criteria to reclassify EL students is linked to improved academic outcomes for all cohorts, with one exception —a–g requirements for the 8th grade cohort. While there are clear associations between more rigorous reclassification policies and outcomes for RFEP students, they often vary by outcome and by grade level. In the cases where more rigorous standards are associated with better RFEP outcomes, it is important to question whether the outcomes are sufficiently improved to justify the much lower district reclassification rates that result. Recall that most districts report using more than one category of more rigorous reclassification policy. As described above, even districts that increase the rigor of just one or two more reclassification requirements can significantly lower their reclassification rate. http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 49 How D o We E valua te t he Trad e-offs Betw een Low er R ecla ssification Rat es but Better O utcom es? We have demonstrated that RFEP students not only outperform EL students, but often do better than EO and IFEP students. RFEP students reclassified at early grades do well for all six years we observe them. RFEP students reclassified at older grades perform well on academic m easures but generally not as well as those reclassified in elementary school. There may be room to lower the reclassi fication standards in districts us ing more rigorous thresholds and still ensure that RFEP students perform well enough in academ ic instruction without additional English language support. However, it should be noted that because districts choose their reclassification policies and because their choice of policies may be somehow related to factors we cannot measure (about their students, their EL instruction, or other factors), we may have under- or over- estimated the true effect o f the policy on the outcomes of interest. Future extensions of this research could explore this issue further. We find that there is wide variation in school districts’ EL reclassification policies, with most districts adopting at least some criteria more rigorous than suggested by the SBE’s 2006 reclassification guidelines. Having different reclassification standards across the state’s districts makes it difficult to compare EL and RFEP outcomes across districts. More troublesome, however, is that we find that districts using more rigorous reclassification standards have lower reclassification rates. Districts using one the most common combinations of requirements (requiring the math CST and requiring that all CELDT subtest scores be Early Advanced or highe r) reduce their reclassification rates by 5.4 percentage points. If a district’s reclassification rate is 10 percent, using that combination of requirements reduces the number of new RFEP students each year by more than half in that district compared to a district that did not use the more rigorous policy. We find RFEP students’ outcomes are better in school districts with more rigorous performance thresholds, but not by much. Is the trade -off between improved RFEP outcomes and fewer RFEP students worth it? It would seem that the costs of these requirements —in terms of lower reclassification rates—are disproportionate to their benefit. Given that we also find that RFEP students usually out -perform EO students (and sometimes IFEP students as well) and general ly perform at very high levels even in districts with more relaxed reclassification standards, we recommend that reclassification standards be lowered in districts using standards more rigorous than those suggested in the SBE guidelines. This report tested the association between more rigorous reclassification standards and reclassified students’ outcomes, but could not test for the possibility that even the thresholds suggested in SBE guidelines might be too rigorous. We recommend that there be additional research to define ideal reclassification standards. If ELs have access to the same academic content as other students (RFEPs, EOs, and IFEPs), then the status itself may not be problematic in terms of a student’s ability to succeed in school (although the re is some evidence that Long-T erm EL students are subject to stigma (Dabach and Callahan, 2011)). But if an EL student is not reclassified because the standards to allow him or her to transition to RFEP status are too rigorous, then it is possible that di stricts with high reclassification thresholds are restricting the access of EL students to the full range of academic instruction that non -EL students receive, which could suppress the academic achievement of high -performing EL students. http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 50 Establishing ideal reclassification standards will not be easy, but we have the time and opportunity to do so. Saunders and Marcelletti (2013) suggest that it is a reasonable goal to expect RFEP students to achieve at least the same level of outcomes as EO students. However, many EO students are not meeting Annual Yearly Progress (AYP) goals or even achieving a Basic score on the state’s standardized ELA tests (Abedi, 2008). In addition, EO students vary in their demographic composition across districts. Robinson (2011) arrives at the conclusion that ideal reclassification thresholds probably depend on high school students’ learning environments after reclassification. Given the current policy environment, this is an ideal time to examine the components of reclassification decisions that are associated with getting the greatest number of EL students reclassified while achieving a minimum agreed- upon standard of success. Two major policy changes will affect EL and RFEP students in the coming years. The first relates to the Common Core State Standards currently being implemented in district curricula. Once the new assessments for the standards are in place, there will no longer be state standardized assessments in ELA for 9th and 10th graders, and districts will need to rethink their practices for reclassifying students during the high school years. We see evidence in the data that even Long -Term English Leaners (LTELs) are being reclassified in high school, and districts certainly reported that to be the case in the survey. One possibility is that districts will increasingly use the CAHSEE instead of the CST ELA for grades 10, and 12, but a passing score on the ELA portion of the CAHSEE is thought to reflect a lower standard than a score of Basic on the CST ELA. Using the CAHSEE would still mean that the 9th grade year has no ELA standardized test, which is currently the type of assessment used by districts to evaluate an EL students’ basic skill level in ELA. In addition, districts will to stop using the CST this year in order to prepare for the Smarter Balanced assessments. A second issue involves the interplay between reclassification and the new Local Control Funding Formula (LCFF). It is important that funding directed to districts for EL students not discourage reclassification. While the LCFF provides funding only for EL students and not for RFEPs, many RFEP students will generate funding for their district based on their poverty status. We found that 79 percent of the ELs in our cohorts of students were low -income students .51 We also found that RFEP students reclassified by 2008– 09 have poverty rates of 63 percent, and those reclassified after 2008– 09 have poverty rates of 77 percent. Still, the instruction or supplemental services that might help low -income RFEPs stay on track could be very different from what low -income EO students need. Another important issue is the flexibility involved in monitoring academic outcomes for RFEP students. Most districts reported that they monitor RFEPs for two years; but during this time of dramatic changes for RFEP students, we suggest that the role for monitoring, as well as adapting services to RFEP students, should be expanded. 51 Our measure of low -income students was whether they participated in the free or reduced price lunch programs, which include income cut-offs of 130 percent and 185 percent of the federal poverty level, respectively. http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 51 References Abedi, J. 2008. “Classification System for English Language Learners: Issues and Recommendations,” Educational Measurement: Issues and Practices, pp. 17 –31. Assembly Bill 2193. Chapter 427, “An act to add Sections 313.1 and 313.2 to the Education Code , relating to English Learners.” Accessed from http://leginfo.legislature.ca.gov/faces/billNavClient.xhtml?bill_id=201120120AB2193 . California Department of Education. 2012, “California English Language Development Test (CELDT), 2012– 2013,” CELDT Information Guide . California Department of Education. 2011. A Comparison Study of Kindergarteners and Grade One English -Fluent Students and English Learners on the 2010– 2011 Edition of the CELDT . California Legislative Analyst’s Office. 2007. Analysis of the 2007– 2008 Budget Bill: Education . California State Auditor. 2005. Department of Education: School Districts’ Inconsistent Identification of Redesignation of English Learners Cause Funding Variances and Make Comparisons of Performance Outcomes Difficult , Bureau of State Audits. Cannon, J. S., and S. Lipscomb. 2011. Early Grade Retention and Student Success: Evidence from Los Angeles, Public Policy Institute of California, San Francisco, CA. Dabach, Dafney B., and Rebecca M. Callahan. 2011. Rights versus Reality: The Gap between Civil Rights and English Learners’ High School Educational Opportunities , Teachers College Record. EdSource. 2008. English Learners in California: What the Numbers Say , Mountain View, CA. Flores, Edward, Gary Painter, and Harry Pachon. 2009. ¿Qué Pasa?: Are English Language Learning Students Remaining in English Learning Classes Too Long? Tomás Rivera Policy Institute Full Report, Los Angeles, CA. Gándara, P., and R. Rumberger. 2006. Resource Needs for California’s English Learners , Linguistic Minority Research Institute, University of California. Hill, L.E. 2012. California’s English Learner Students, Public Policy Institute of California, San Francisco, CA. Hopkins, M., K.D. Thompson, R. Linquanti, K. Hakuta, and D. August. 2012. “Fully Accounting for English Learner Performance: A Key Issue in ESEA Reauthorization,“ Educational Researcher , Vol. 42, No. 2., pp. 101– 108. Linquanti, R. 2001. The Reclassification Dilemma: Challenges and Choices in Fostering Meaningful Accountability for English Learners , UC Language Minority Research Institute Policy Report, University of California, Davis, CA. Mechelli, A., J.T. Crinion, U. Noppeney, J. O’Doherty, J. Ashburner, R.S. Frackowiak, and C.J. Price. 2004. “Structural pasticity in the bilingual brain: proficiency in a second language and age at acquisition affect grey -matter density,” Nature , Vol. 431. Parrish, T., A. Merickel, M. Perez, R. Linquanti, M. Socia, A. Spain, C. Speroni, P. Esra, L. Brock, and D. Delancey. 2006. Effects of the Implementation of Proposition 227 on the education of English Learners, K–12, Findings from a Five -Year Evaluation , American Institutes for Research and WestEd, Palo Alto, CA. PPIC School Finance Model. 2013. Available at www.ppic.org/main/dataSet.asp?i=1229. Robinson, J.P. 2011. “Evaluating Criteria for EL Reclassification: A Casual -Effects Approach Using a Binding -Score Regression Discontinuity Design with Instrumental Variables,” Education Evaluation and Policy Analysis . Rose, H., J. Sonstelie, and M. Weston. 2012. Funding Formulas for California’s Schools III: An Analysis of Governor Brown’s Weighted Pupil Funding Formula, Public Policy Institute of California, San Francisco, CA. Saunders, W. M., and D. J. Marcelletti. 2013. “The Gap That Can’t Go Away: The Catch -22 of Reclassification in Monitoring the Progress of English Learners,” Educational Evaluation and Policy Analysis , Vol. 35, No. 139. Warren, P., and H. Hough. 2013. Increasing the Usefulness of California’s Educational Data, Public Policy Institute of California, San Francisco, CA. http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 52 About the Authors Laura E. Hill is a research fellow at PPIC, where her research interests include immigrant well -being, English Learners in K –12 schools, race and ethnicity, and youth. She has been a research associate at the SPHERE Institute and a National Institute of Aging postdocto ral fellow. Laura holds a Ph.D. in demography from the University of California, Berkeley. Margaret Weston is a research fellow at PPIC’s Sacramento Center, where she focuses on K –12 school finance. Before joining PPIC, she taught high school English and d rama in Baltimore City Public Schools through Teach For America. She holds a master’s degree in teaching from Johns Hopkins University and a master of public policy degree from the University of Michigan. She is pursuing a Ph.D. in school organization and education policy at the University of California, Davis. Joseph M. Hayes is a research associate at PPIC, where he studies population change and corrections issues. Recent projects have focused on inter- regional migration, estimates of the undocumented imm igrant population, and the changing composition of the prison population. He holds an M.S. in agricultural economics from the University of Wisconsin, Madison. Acknow ledgm ents The authors wish to thank Dave Lesher, Hans Johnson, Patrick Murphy, Athena Brek ke, Abby Cook, Lisa Cole, and Jennifer Paluch at PPIC for their support in developing, producing, and publishing this r eport. W e also thank Caitrin Phillips Chappelle, Sonja Petek, Jan Mayer, Kris Nichols, Karen Bachofer, and Mary Estelle Anderson for their help with the development of the reclas sification survey. We t hank the California Department of Education (CDE), for providing student data. We also wish to thank Robert Linquanti, Leonor Ehling, Ellen Hanak, Lynette Ubois, and Caroline Danielson for comments on earlier versions of the draft and Gary Bjork, Jenny Miyasaki, and Kate Reber for their editorial support. Any errors are our own. PUBLIC POLICY INSTITUTE OF CALIFORNIA Board of Directors Donna Lucas, Chair Chief Executive Officer Lucas Public Affairs Mark Baldassare President and CEO Public Policy Institute of California Ruben Barrales President and CEO GROW Elect María Blanco Vice President, Civic Engagement California Community Foundation Brigitte Bren Attorney Walter B. Hewlett Chair, Board of Directors The William and Flora Hewlett Foundation Phil Isenberg Chair Delta Stewardship Council Mas Masumoto Author and F armer Steven A. Merksamer Senior Partner Nielsen, Merksamer, Parrinello, Gross & Leoni, LLP Kim Polese Chairman ClearStreet, Inc. Thomas C. Sutton Retired Chairman and CEO Pacific Life Insurance Company The Public Policy Institute of California is dedicated to informing and improving public policy in California through independent, objective, nonpartisan research on major economic, social, and political issues. The institute’s goal is to raise public awar eness and to give elected representatives and other decisionmakers a more informed basis for developing policies and programs. The institute’s research focuses on the underlying forces shaping California’s future, cutting across a wide range of public poli cy concerns, including economic development, education, environment and resources, governance, population, public finance, and social and health policy. 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. PPIC was established in 1994 with an endowment from William R. Hewlett. Mark Baldassare is President and Chief Executive Officer of PPIC. Donna Lucas is Chair of the Board of Directors. Short sections of text, not to exceed three paragraphs, may be quoted without written permission provided that full attribution is given to the source. Research publications reflect the views of t he authors and do not necessarily reflect the views of the staff, officers, or Board of Directors of the Public Policy Institute of California. Copyright © 201 4 Public Policy Institute of California All rights reserved. San Francisco, CA PUBLIC POLICY INSTITUTE OF CALIFORNIA 500 Washington Street, Suite 600 San Francisco, California 94111 phone: 415.291.4400 fax: 415.291.4401 www.ppic.org PPIC SACRAMENTO CENT ER Senator Office Building 1121 L Street, Suite 801 Sacramento, California 95814 phone: 916.440.1120 fax: 916.440.1121" } ["___content":protected]=> string(102) "

R 114LHR

" ["_permalink":protected]=> string(101) "https://www.ppic.org/publication/reclassification-of-english-learner-students-in-california/r_114lhr/" ["_next":protected]=> array(0) { } ["_prev":protected]=> array(0) { } ["_css_class":protected]=> NULL ["id"]=> int(8892) ["ID"]=> int(8892) ["post_author"]=> string(1) "1" ["post_content"]=> string(0) "" ["post_date"]=> string(19) "2017-05-20 02:41:51" ["post_excerpt"]=> string(0) "" ["post_parent"]=> int(4324) ["post_status"]=> string(7) "inherit" ["post_title"]=> string(8) "R 114LHR" ["post_type"]=> string(10) "attachment" ["slug"]=> string(8) "r_114lhr" ["__type":protected]=> NULL ["_wp_attached_file"]=> string(12) "R_114LHR.pdf" ["wpmf_size"]=> string(6) "702148" ["wpmf_filetype"]=> string(3) "pdf" ["wpmf_order"]=> string(1) "0" ["searchwp_content"]=> string(135424) "Reclassification of English Learner Students in California January 2014 Laura E. Hill, Margaret Weston, an d Joseph M. Hayes with research support from Da vid Ez ekiel and B elen Chavez http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 2 Summary Former English Learner students who have improved their facility with English to such a degree that they have been reclassified by their school districts as fluent in the English language are among the best performing students in the state. Because these Reclassified Fluent English Proficient (RFEP) students have much better academic outcomes than English Learner (EL) students, policymakers conjecture that reclassifying ELs more quickly might help close the state’s persistent achievement gap between EL and non- EL or English only (EO) students. To substantiate this conjecture —and noting that the standards for reclassification currently vary greatly among school districts—policymakers are interested in assessing whether district s with more rigorous reclassification standards have systematically lower reclassification rates, but also better student outcomes, than districts with less rigorous standards. Because districts determine their own reclassification criteria, it is difficul t to compare reclassification rates, the progress of ELs, and the outcomes for ELs and RFEPs across school districts throughout the state. We hope this report is informative to policymakers interested in Senate Bill 1108, which has as its goal documenting reclassification policies in California’s school districts and their link to student outcomes. In this report, we are able to overcome the key limitations of previous research through the use of California L ongitudinal Pupil Achievement D ata System (CALPAD S) provided under an arrangement with the California Department of Education (CDE). The data enabled us to track students in each California school district from 2007 –08 through 2012– 13, excluding students in charter schools. This report provides the first longitudinal analysis of the transition from EL to RFEP status for all California school districts. Our analysis indicates that RFEP students not only outperform EL students, but also often do as well as native English speakers when it comes to measures of academic outcomes, such as standardized tests and on-time grade progression. We also conducted a survey of school districts, asking detailed questions about their current and former reclassification policies and practices. We found that more than 90 percent of responding districts report using more demanding criteria than are suggested by the State Board of Education (SBE) guidelines for the four reclassification criteria specified in California Education Code Section 313(f). Our analysis of student -level longitudinal data in conjunction with survey responses reveals that districts using more stringent reclassification criteria have lower reclassification rates. For example, the roughly 30 percent of districts that require an EL student to score Proficient or higher on the English Language Arts portion of the California Standards T est have district reclassification rates that are 3 percentage points lower than in districts that require a score of Basic or higher. In other words , if the average annual reclassification rate among districts using Basic or higher (as suggested by the SBE guidelines) were 10 percent, districts that use Proficient or higher (a more rigorous criterion) would have reclassification rates of 7 percent. Th is translates to a 30 percent reduction in the number of students reclassified in districts using the more rigorous criterion. If districts require higher standards on more than one criterion, their reclassification rates are even lower. However, using stricter criteria than suggested in the state guidelines is also associated with slightly better outcomes for RFEP students. For example, requiring a score of Proficient or higher for students reclassified in 3rd grade is associated with larger percentages of these students scoring Proficient or higher on 6th grade standardized tests (a 4 percentage point increase, which would increase the share scoring Proficient or higher from 78 to 82 percent). Stricter criteria are also associated with a greater likelihood of on-time grade progress among students reclassified in the 8th grade (a 5 percentage point increase, increasing the http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 3 likelihood from 90 to 95 percent) and a 3 percentage point increase in the share scoring Proficient or higher on standardized tests (incr easing the likelihood from 14 to 17 percent). Among students reclassified in 9th grade, the Proficient criterion is not related to completing a –g requirements or leaving high school without graduating, but it is negatively associated with getting a high sc hool diploma, reducing the chances by 5 percentage points. Districts and policymakers should decide if these mostly positive outcomes are sufficiently large to justify the more demanding reclassification criteria that hinder so many ELs from transitioning to the RFEP population. In addition to these considerations, two major policy shifts occurring in K –12 education are also relevant to the well -being of the EL population. Under the Common Core State Standards and the new assessments being developed to test them, the assessments and criteria for EL reclassification will necessarily change in the coming years. Indeed, the California Standards Test (CST) will no longer be used, even in the current school year. The new reclassification criteria using th e new assessments will need to be crafted carefully. Current law requires analysis to determine the new reclassification criteria and we agree that such research and analysis are needed. The second relevant policy shift is embodied in the Local Control Funding Formula (LCFF), which increases funding for districts with large populations of EL students but does not provide funding for RFEP students. This new funding formula may reduce districts’ incentives to reclassify students.1 However, a statewide reclassification policy would help deter the likelihood of a district restricting EL students from being reclassified when they are ready to transition to RFEP status. Based on the findings of this report, we recommend that there be one standard for reclas sifying EL students statewide, and that the standard be set, for now, using the assessments and levels recommended in the guidelines provided by the State Board of Education. In most cases, this means districts will need to lower their reclassification sta ndards. The implementation of the Common Core State Standards and the Local Control Funding Formula represent new opportunities for success in K –12 education, and the progress and outcomes of both EL and RFEP student s should be closely monitored to insure that these students participate in these opportunities. 1 The new funding formula goes into effect this school year (2013 -2014), but not all of the funding will be available in the first year. Because many RFEP students are also members of low -income families (and low-income students generate funding for district s as well), districts should not face as large a disincentive to reclassify students as they otherwise might. Contents Summary 2 Figures 5 Tables 6 Acronyms 7 Introduction 8 Recl assification and S tudent Outcom es 10 Student C ohort Sel ect ion an d Defin ition 10 Student O utco mes 13 How Does R ecl assification W ork ? 20 PPIC’s “ English Learners Reclassif ication Survey” 22 Reclassification P ol icies 23 Do District R ecl assification Polici es Infl uen ce Recl assific ation Ra tes? 33 Do Recl assification Polici es A ffect Stud ent Outcom es? 38 Longitudi nal RFEP Outcomes 38 Relationships Be tween More Rigorous Reclassifi cation Policies an d RFEP Outcomes 42 How Do We E valuat e t he T rad e-offs Betw een Low er R ecl assification R at es but Bett er Outcom es? 49 References 51 About the A uthors 52 Acknow ledgm ents 52 Technical appendices to this paper are available on the PPIC website: www.ppic.org/content/pubs/other/114LHR_appendix.pdf http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 5 Figures 1. Pe rcent of stud ents scoring B asic or high er on the CST ELA 15 2. On -time or better grade progression, 7th grade cohort 17 3. Final high school outcomes 18 4. Percent of districts that permit reclassification, by grade level 29 5. Most districts allow for student reclassification at more than one point in time per year 30 6. District reclassification rates under different reclassification requirement scenarios 36 7. Share of cohort, by language group, scoring Basic or higher on CST ELA, controlling for district and student cha racteristics 39 8. Share of cohort, by language group, scoring Proficient or higher on CST ELA, controlling for district and student characteristics 40 9. Percent on -time or better in 10th grade, by language group, 7th grade cohort, controlling for district and student characteristics 41 10. Final high school outcomes for 8th grade cohort, controlling for district and student characteristics 42 http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 6 Tables 1. N umb er and cl assific ation st atus of students included in our a n alysis , by cohort 12 2. Re classific ation r a te s for four cohorts of E Ls 13 3. Survey respondents 22 4. S urv ey r es pond ents are broadly repr esen tative of the state’s student demographics 23 5. English proficiency criteria , by grade level 25 6. Percent of districts imposing basic skills requirements, by grade level 25 7. Teacher evaluation criteria, by grade level 27 8. Subjective evaluation criteria used by teachers, by grade level 28 9. “If you were forced to choose one criteria for each grade level, which of the reclassification criteria would you say is most difficult for EL students to meet in your district?” 31 10. “In your opinion, how important are each of the reclassification criteria in the ultimate decision to reclassify a student?” 31 11. District reclassification policy 34 12. Per centag e point change for stud ent outcomes and district recl assification rat es associated with specific recl assification requirements 45 13. End of high school outcomes, cohort 4 47 http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 7 Acronyms ADA API AYP CAHSEE CALPADS CDE CELDT CST DIBELS EL Average Daily Attendance Academic Performance Index Annual Yearly Progress California High School Exit Exam California Longitudinal Pupil Achievement Data System California Department of Education California English Language Development Test California Standards Test Dynamic Indicators of Basic Early Literacy Skills English Learner ELA ELD EO IFEP LCFF LEA LTEL OPL RFEP SBE SEI English Language Arts English Language Development English Only Initially Fluent English Proficient Local Control Funding Formula Local Education Agency Long -Term English Learner Overall Performance Level Reclassified Fluent English Proficient State Board of Education Structured English Immersion http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 8 Introduction California’s K –12 public schools include some 1.4 million English Learners (ELs), representing about 22 percent of the student body. Researchers have found a persistent achievement gap between these ELs and their native -English -speaking peers. However, the EL designation is intended to be temporary. Once ELs become proficient in English, they are reclassif ied and analysts have found that the achievement gap between former EL students —i.e., Reclassified Fluent English Proficient (RFEP) students— and native English speakers is small or even positive, with RFEP students sometimes outperforming native English sp eakers (Saunders and Marcelletti, 2013; Hill, 2012; EdSource, 2008; Gándara and Rumberger, 2006). Because reclassified ELs perform so much better than ELs on a host of academic outcomes, policymakers are avidly interested in designing policies that help EL students transition quickly into unsupported academic instruction. This endeavor assumes that once the students are reclassified, their increased access to academic instruction will increase their academic performance and, correspondingly, that bringing m ore students into the RFEP student group will reduce the achievement gap. However, one must consider the possibility that the gap between RFEP and EL student performance may be the result of reclassifying only the very strongest students among the EL group . Currently, California school districts are allowed to determine their own reclassification policies. The State Board of Education has issued guidelines for the four reclassification criteria specified in California’s Education Code Section 313(f).2 In October 2012, Governor Jerry Brown signed SB 1108 (Chapter 434, Statutes of 2012) , which reflects an interest among policymakers in understanding the variance in reclassification policies across school districts and their association with student outcomes. Until now, there has been a lack of general knowledge about how and when districts reclassify their EL students. In addition, there is no consensus about the ideal reclassification policy for ensuring the success of EL and RFEP students. At the same time, the governor’s 2013 –2014 budget has seriously changed the way school districts are funded, significantly increasing per -pupil funding for EL and low -income students while allowing districts to exert their own control over how those dollars are spent.3 Some are concerned that the extra funding will not be spent on EL students and that districts will have incentives to keep EL students from being reclassified (extra funding is not available for RFEP students).4 However, most EL students (estimates range from 74 to 85 percent) are also members of low -income families (Rose, Sonstelie, and Weston, 2012; California Legislative Analyst’s Office, 2007), which means that even if they are reclassified, these students will still generate extra funding for their school districts. Would lowering reclassification standards reduce the achievement gap by increasing the number of RFEP students, or would it increase the number of RFEP students but lower their performance as a group? 2 Section 313(f) reads as follows, “The reclassification procedures developed by the department shall utilize multiple criteria in determining whe ther to reclassify a pupil as proficient in English, including, but not limited to, all of the following: (1) Assessment of l anguage proficiency using an objective assessment instrument, including, but not limited to, the English language development test that is developed or acquired pursuant to Section 60810. (2) Teacher evaluation, including, but not limited to, a review of the pupil’s curriculum mastery. (3) Parental opinion and consultation. (4) Comparison of the performance of the pupil in basic skill s against an empirically established range of performance in basic skills based upon the performance of English-proficient pupils of the same age, that demonstrates whether the pupil is sufficiently proficient in English to participate effectively in a cur riculum designed for pupils of the same age whose native language is English.” 3 Draft regulations were released by the SBE in October, 2013. Emergency regulations were adopted January 16, 2014, and final r egulations are expected later in 2014 . 4 Students classified as both EL and low -income students do not generate additional funding, compared to those who are classified only as low- income students. http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 9 Understanding the relationships between reclassification policies, reclassification rates, and reclassified student outcomes is a necessary prerequisite before considering recommendations about whether more EL students should be reclassified, and more quickly, than is currently the norm. In summer 2013, PPIC conducted a survey of school districts, documenting reclassification policies and practices across a variety of the state’s school districts. The researchers also analyzed six years of longitudinal student-level data captured in CDE’s California Longitudinal Pupil Achievement Data System (CALPADS). This report presents the results of this research. In the following pages, we  Explore differences in academic outcomes for ELs, RFEPs, and other language groups, tracking these students as they progress though school.  Report reclassification policies and practices in school districts throughout the state.  Examine whether school districts who report using more rigorous reclassification policies and practices have lower reclassification rates than districts who report using less rigorous criteria.  Consider whether reclassified students in school districts who report more rigorous reclassification policies and practices have better outcomes than districts who report more relaxed policies.  Conclu de with preliminary recommendations for reclassification policies. This researc h is an important first step in understanding the relationship between reclassification policies and student outcomes in California’s school districts. We hope this report is helpful to legislators, CDE, and the State Board of Education as they contemplate a response to SB 1108. We should note that this report investigates just one policy lever —reclassification policies—and does not consider other important issues. For example, it does not contribute to valuable discussions about the validity and reliability of the standardized tests used to establish English proficiency. 5 Nor does it provide analysis related to EL services and program delivery or teacher preparation. However, given that decisions about how ELs are taught and served are left to the local level, setting uniform reclassification policy is one of the few levers currently available to the state. 5 For a discussion of this issue, see Abedi (2008). http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 10 Recla ssification and S tudent O utcom es Research on outcomes among California’s EL and RFEP students usually examines cross- sectional differences among student groups—RFEPs compared to ELs or English only (EO) students —(e.g., Saunders and Marcelletti, 2013; Hill, 2012; EdSource, 2008; and Gándara and Rumberger, 2006). Other studies track students through longitudinal data, although these studies generally focus on only one or two districts at a time (e.g., Flores, Painter, and Pachon, 2009; Hopkins et al., 2012; Robinson, 2011). The cross- sectional research has found large differences between EO students and ELs, but this work often overlooks the confounding factor that the most successful EL students are reclassified as RFEPs and are not included in either comparison group. When RFEP and EL students are combined into an “ever -EL” student group, the gap between ever -EL and EO students is considerably smaller and has declined somewhat over time (Saunders and Marcelletti, 2013). However, even cross -sectional research that refines comparison groups for ELs cannot account for “time since reclassification” or new entrants to the EL population. Indeed, the inability to control for those potentially confounding factors in cross -section al research may explain why RFEP students outperform EO students in elementary grades, but have worse outcomes by the end of high school (Gándara and Rumberger, 2006; Hill 2012). In 4th grade, for example, RFEP students have much higher scores on standard ized tests than native English speakers, but among 10th graders, native English speakers have slightly higher scores than RFEP students. It is important that we understand whether the apparent decline of 10th grade RFEP students’ scores is real or whether 10th grade scores are lower because more recently reclassified students in the cohort have lower scores than students reclassified at earlier grades. In our research, we follow the outcomes of individual students over time; thus we can consider the role of “time since reclassification” in the academic outcomes of RFEP students. We begin this section by describing our longitudinal student- level data and the cohorts of students we construct for analysis. We then examine standardized test scores, on- time grade progression, and high school outcomes across our student cohorts and language groups. Student C ohort Selection and Defin ition In this report, we are able to overcome the key limitations of previous research through the use of the state’s longitudinal student-level data system (CALPADS). The data enabled us to track students in each of the state’s school districts from 2007 –08 through 2012– 13, excluding students in charter schools. This report provides the first longitudinal analysis of the transition from EL to RFEP status for all California school districts. Because we can use data for all of the state’s school districts and can follow students across many years of data, our research is uniquely situated in its ability to compare the outcomes of RFEPs with the outcomes of ELs, EOs, and students who have a primary language other than English but are designated as Initially Fluent English Proficient (IFEP). We can also simultaneously consider a student’s ba ckground characteristics, the grade in which a student was reclassified, the role of district characteristics (such as demographic composition and performance on key measures), and finally, the association between district reclassification policies and rec lassified students’ outcomes. http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 11 Using CALPADS, we constructed four cohorts of students based on the grade they were in during the 2007 –08 school year (grades 2, 4, 7, or 8) and then divided each cohort into groups based on language status: EL, IFE P, RFEP, o r EO. For our purposes, ELs must have been ELs in kindergarten and must not have been reclassified while we observe them (school years 2007– 08 through 2012– 13). We have divided the RFEP students in each cohort into three groups: those reclassified in 2008 –09 (our “target” reclassification year), those reclassified beforehand (“pre-target”), and those reclassified afterward (“post-target”). Because reclassification policies can change over time, examining student progress relative to policies in place at one point in time (2008– 09) allows us to draw connections between reclassification criteria and later student outcomes. EL or RFEP students who started school in California after kindergarten are not included in our cohorts. Most ELs in California started school in the United States in kindergarten, rather than arriving from abroad as older students in later grades.6 Table 1 summarizes the number of students and their classification status in each of our four cohorts. The 2nd grade cohort allows us to observe a large cohort of students who were reclassified at the first opportunity in some districts —at the beginning of 3rd grade. Many districts do not reclassify before this point. The 4th grade cohort (students reclassi fied in 5th grade) represents another important reclassification year —this is commonly understood to be the modal reclassification grade . 7 The 7th grade cohort represents students who are L ong-T erm English Learners (LTELs), and we can observe their outcomes to the beginning of their 12th grade year. 8 The final cohort, the 8th grade cohort, also includes LTELs, enabling us to observe their end -of -high- school outcomes. 9 6 Published CDE data suggest that the share of EL students who arrived from abroad within the last year is relatively small by grade. The share peaks in number in grade 2 —representing approximately 4,000 (or 2%) of 2nd grade ELs —and peaks in percentage in grade 9, representing about 6% of 9th grade ELs (Hill, 2012). Their cumulative percentage by grade depends on their reclassification rates. Future extensions of this research could include EL students who arrived in the United States after the kindergarten year, but we do not include them h ere. 7 Ther e are at least two reasons why this is considered the modal reclassification grade: 1) elementary schools wish to transition students to middle school as RFEPs in order to avoid having students become Long -Term ELs, and 2) one component of the redesignation criteria is easiest to achieve in the prior year (the CST ELA, administered in the 4th grade). 8 AB 2193 (Chapter 47, Statutes of 2012) defines an LTEL as “an English Learner who is enrolled in any of grades 6 to 12, inclu sive, has been enrol led in schools in the United States for more than six years, has remained at the same English language proficiency level for two or more consecutive years as determined by the English language development test identified or developed pursuant to Section 60 810, or any successor test, and scores far below Basic or below Basic on the English language arts standards-based achievement test administered pursuant to Section 60640, or any successor test.” (Accessed from http://leginfo.legislature.ca.gov/faces/billNavClient.xhtml?bill_id=201120120AB2193 .) Because we do not use the English language development test or the English language arts standards-based achieve ment test to create the cohort, our LTELs are best thought of as an approximation of official LTELs. 9 Most end -of-high school outcomes such as graduation, finishing a –g course requirements, and reasons for leaving without a diploma are not recorded for al l students until five years after beginning 9th grade. http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 12 TABLE 1 N umb er and cl assific ation st atus of students included in our a n alysis , by cohort Grade 2 Grade 4 Grade 7 Grade 8 EO 85,167 77,337 112,293 67,131 IFEP 12,568 13,642 18,967 11,215 RFEP, pre-target year 7,692 18,038 48,267 33,871 Long-Term English Learners RFEP, target year (2008– 09) 3,246 2,623 1,404 997 RFEP, post -target year 22,583 9,898 4,544 2,363 EL never reclassified (2008– 09 –2012– 13) 19,538 10,400 7,516 4,628 Total 150,794 131,938 192,991 120,205 Because we are interested in understanding the relationship between district reclassification policies and outcomes for students, we generally restrict our sample to students who do not transfer across districts — with one important exception.10 We do include students who transfer from their elementary school district to the appropriate high school or unified school district .11, 12 Similarly, we require that students be present in the data for all six years of our study, with exceptions made for our two oldest cohorts. In our 7th grade cohort, we require that students remain in the data until the third year of the study, when they should be in 9th grade if they are progressing on time. After 9th grade, we do allow the students to be missing, because dropping out (or remaining in school) is an outcome of interest for high school students. We allow an analogous exception for our 8th grade cohort, requiring them to remain in the data only until the 3rd year of the study or 10th grade. Of all of our data cleaning and cohort construction restrictions, the one with the most substantial impact on the representativeness of our sample is our decision to exclude ELs who arrive in the California school system after kindergarten. Excluding these later -arriving ELs significantly affects the share of ELs in the 7th and 8th grade cohorts. In 2007– 08 ELs represented 21 percent of 7th graders an d 18 percent of 8th graders (the first year of our data; see Appendix Table A1 ). When we drop students who are missing and ELs who did not start in kindergarten, the share of ELs falls to 8 percent in both cohorts. Requiring that ELs be in California schools since kindergarten also substantially affects the share of Latinos in our sample, since most ELs speak Spanish, but not the share of economically disadvantaged students. We also exclude special education students.13 Districts may vary in the extent to which ELs are classified as 10 See Appendix Table A1 for details on the number of students not included in the study. 11 In practice, we drop students from the study who transfer out of their district before the appropriate year to complete a transfer from elementary districts to high school districts. For example, a student enrolled in a K –8 district who transfers districts is dropped from the analysis unles s that transfer occurs in the 9th grade. For students who transfer in the appropriate year, we require that their transfer be within the county in order to stay in the data. Transfers are not permitted for students in unified school districts. 12 Future res earch could assess the effect of excluding more mobile students from our analysis. 13 Special education students who are ELs can be reclassified, but often through different assessment instruments. Our survey of district reclassification policies did ask ab out reclassification policies for special education students (see the following section and Appendix B ), but our http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 13 special education students, and perhaps even more so in the case of LTELs who are also special education students. We do find that excluding special education students has a somewhat larger effect on the characteristics of the 7th and 8th grade cohorts.14 Due to these data cleaning and cohort construction restrictions, we believe that this report best describes outcomes and predictors of outcomes for ELs who start in California schools in kindergarten. Additional research should address late -arriving ELs and the role of special education status in EL classification and reclassification. Student O utco mes Our analyses focus on the following student outcomes: scoring Basic or higher on the California Standards English Language Arts (CST ELA) test, scoring Proficient or higher on the CST ELA, and on -time grade progression. For our oldest cohort of students, we also consider earning a high school diploma, meeting a – g requirements, and leaving prior to completing high school.15 Before discussing these outcomes for EL and reclassified students, we consider differences in reclassification rates by grade among our cohorts. Among our four cohorts, across all districts, reclassification rates in our target reclassification years are highest for students in the 8th grade cohort, who are reclassified in the 9th grade (12.5%), although this rate is similar to the reclassification rates in the 5th grade.16 TABLE 2 Re classific ation r at es for four cohorts of E Ls Reclassification rates observed in: 2nd grade cohort 4th grade cohort 7th grade cohort 8th grade cohort 3rd grade 7.1% 5th grade 11.4% 8th grade 10.4% 9th grade 12.5% Rates of reclassification, 2008– 09 – 2012– 13 56.9% 54.6% 44.2% 42.1% Note: Students in all cohorts h ave been EL students since k ind erg art en. Looking at longer term outcomes for ELs, the 2nd and 4th grade cohorts have roughly equal chances of being reclassified during the next five years: 57 percent of 2nd grade ELs are reclassified as RFEP students by 7th grade, and 55 percent of 4th grade ELs are reclassified as RFEP students by the 9th grade. In the 7th and 8th grade cohorts, students who were reclassified in the target year or later are considered Long -T erm English analyses in this report exclude special education students. Of course, districts can also differ is their classification of EL students into special education in the first place. We leave these considerations for later studies. 14 When special education students are excluded, the share of ELs drops by 1 to 2 percentage points across cohorts, but since the share of ELs remaining is already substantially lower in the 7th and 8th grade cohorts, this exclusion has a larger impact on the older cohorts. 15 We hope in future extensions of this research to incorporate other test results, such as CST math and CAHSEE scores and pass rates. 16 Reclassification rates are calculated by dividing the number of new RFEP students in a calendar year by the sum of ELs and new RFEPs. http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 14 Learners. Among students who are still ELs in 7th grade, 44 percent become RFEPs by the end of high school. Rates for the 8th grade cohort are similar (42%). 17 Although these reclassification rates are lower than the reclassification rates in the younger grades, it is notable that sizeable percentages of LTELs ar e ultimately reclassified by the end of high school. In the following sections, we discuss outcomes for reclassified and non- reclassified EL students, as measured by CST ELA scores, on -time grade progression, and final high school outcomes. Student O utcom es: C ST E LA Scor es We examine ELA scores for the youngest three cohorts.18 Descriptively, across all three cohorts, RFEP students, no matter w hen they are reclassified, are more likely than EL students to achieve a score of Basic or higher on the CST ELA, and their performance is consistently better across the years following their reclassification. Recall that we are following individual students over time, and EL students are those who have not been reclassified by the last year in which we observe them (2012 –2013). That RFEP students outperform EL students is not surprising, given that reclassified students exit the EL student population because they meet reclassification criteria that include the CST ELA . 19 However, previous research based on cross- sectional data has questioned whether these differences pers ist over time. More notable is the finding that RFEP students often outperform native English speakers, although the differences are generally small (Figure 1). RFEP students in our 2nd grade cohort are a little more likely than EO students to score Basic or higher on the 6th grade ELA, no matter when they were reclassified. The scores of IFEP students are very similar to those of RFEP students .20 In the 4th grade cohort, RFEP students reclassified before our “target” year perform better than EO and IFEP students when they take the CST ELA in 8th grade, but those reclassified during or after the “target” year do not. In the 7th grade cohort, RFEP students reclassified by the target year (i.e., pre-target students) or during the target year have higher ELA scores in the 11th grade —78 percent and 73 percent, respectively —than EO students (71%). The results are somewhat different if we examine the share of students who score Proficient or higher on the CST ELA. 21 Very few EL students are able to reach the Profi cient level (results are shown in Appendix Figure A1 ). In our 2nd and 4th grade cohorts, RFEP students reclassified prior to our target year tested Proficient in a roughly equal pro portion to IFEP students and higher than EO students. In the 2nd, 4th and 7th grade cohorts, students reclassified in the target year or later do not perform as well as EO or IFEP students. There is little evidence that RFEP students lose ground over time. They do not exhibit steeper declines in test scores than other student language groups (e.g., IFEPs or EOs). Turning to outcomes for ELs, younger EL 17 Of course, many ELs might have dropped out of school or moved out of state between the 10th and 12th grades, so the reclassification rate may actually be lower. As noted in the text and further discussed in the Technical Appendix, we do not allow students in the younger cohorts to exit from the data. 18 In a later section, we analyze factors (including district reclassification criteria) that may predict differences across groups. Hence, we do not in this section test the d ifferences for statistical significance. 19 Future research could compare outcomes in later grades for EO, IFEP, and RFEP students with the same 3rd grade ELA scores. 20 IFEP students are likely to speak a language other than English at home but are designated fluent in English if they pass an English proficiency test in kindergarten. Being bilingual, especially at a young age and with a high degree of fluency, is associated with increased density of grey matter in the brain (Mechelli et al., 2004). T his, along with higher average socioeconomic status (52 percent of IFEP students are low income, in comparison to 79 percent of EL and 63 percent of RFEP students reclassified in 2008 -09), may explain the very high performance of IFEP students. 21 Federal guidelines require that all students score Proficient or higher by 2014. http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 15 cohorts appear to make gains in grades 3 through 6 (2nd grade cohort) and grades 7 and 8 (4th grade cohort), while other student language groups do not. FIGU RE 1 Pe rcent of stude nts scoring B asic or high er on the CST ELA 2nd Grade Cohort 4th G rade Cohort 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 2nd 3rd 4th5th6th Grade IFEP RFEP pre RFEP 2008-09 RFEP post EO EL 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 4th 5th6th7th8th Grade IFEP RFEP pre RFEP 2008-09 RFEP post EO EL http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 16 7th G rade Cohort SOURCE: Authors’ calculations using CALPADS NOTE: Results are not teste d for statistical significance. Student O utcomes: O n-T ime Gra de P rog ression Descriptively, EL students are much less likely than other types of students to advance one grade per school year (Figure 2). RFEP students, regardless of when they were reclassified, are the most successful students in terms of on -time (or better) grade progression to 12th grade: Over 82 percent progressed on time to their final year in high school. (Recall that we do allow students to leave the sample starting with the 10th grade, so these on -time rates are just of those students who remain in the appropriate district each year we observe them.) About 81 percent of IFEP students make on -time progress to the 12th grade. The two groups with the worst outcomes on this measu re are EO and continuing EL students. EO students (80%) are somewhat more likely than continuing EL students (74%) to reach the 12th grade on time. 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 7th 8th9th10th 11th Grade IFEP RFEP pre RFEP 2008-09 RFEP post EO EL http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 17 FIGURE 2 On -time or better grade progression, 7th grade cohort SOURCE: Authors’ calculations using CALPADS NOTE : All students in Figure 2 are in the 7th grade cohort. Results are not tested for statistical significance. The share making on -time grade progression is slightly higher in the 11th than 10th grade for EL, RFEP target -year, and RFEP post targe t- year groups, which is likely due to students dropping out of school between the 10th and 11th grade. CALPADS lacks information on grade retention that may have occurred prior to 2007– 08. This data limitation likely results in an underestimate of the retention patterns for ever -EL students, who are m ore likely to be retained in grades K –3 than non- EL students (Cannon and Lipscomb, 2011) . 22 If we cou ld observe early grade retention, the lines in Figure 2 for the EL and RFEP groups would likely be lower, but they would still show the same rate of on -time progression in the high school years. High S chool O utcom es We constructed our 8th grade cohort in such a way as to observe differences in final high school outcomes .23 We consider three outcomes (Figure 3): leaving high school before graduating, completing high school with a diploma, and completing high school through a– g course requirements (which makes students eligible for admission to the University of California or California State University).24 22 In their study of the Los Angeles Unified School District, Cannon and Lipscomb found that among the students in grades K -3, 1st grade was the most common retention year (ab out 3% of 1st graders were retained), followed by kindergarten and 2nd grade (about 2% retained in each), with less than 1 percent of 3rd graders being retained. 23 Final high school outcomes are often not reported in CALPADS until the 5th year after a student starts 9th grade. 24 Reasons for leaving school include the following: truant with no known enrollment, student expelled and did not re -enroll, student did not return after academic year, student dropped out and did not enroll in GED program, student dropped out and entered a non -academic institution (such as the job corps or justice system), and student not working toward a high school diploma. 60% 65% 70% 75% 80% 85% 90% 95% 100% 9th 10th 11th12th Percent on-time or better On-time grade IFEP RFEP pre RFEP 2008-09 RFEP post EO EL http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 18 FIGURE 3 Final high school outcomes SOURCE: Authors’ calculations using CALPADS NOTE : Results not tested for statistical significance. Continuing EL students are the most likely students to leave high school without graduating (7%).25 Between 1 and 2 percent of all other students, including all types of RFEPs and IFEPs and EOs, left school. Across the three cohorts of RFEPs, between 84 percent and 89 percent received a diploma. All three groups of RFEPs are much more likely than ELs (56%) to finish high school with a diploma. RFEP students reclassified in the “target” year are as likely as IFEP and more likely than EO students to graduate with a diploma, but students reclassified later are not. RFEPs reclassified prior to high school (44%) are just slightly more likel y than EO students (43%) to have completed their a – g requirements. The RFEP students reclassified in 9t h grade or later are more likely than ELs (7%) to have completed the a –g requirements, but less likely than EO students to have done so. Finally, IFEP students are most likely to have met the hi gher sta ndard of having completed their a – g requirements upon graduation. Summing Up RFEP students outperform continuing EL students on every measure we were able to examine, which suggests that the criteria used to determine when an EL student no longer needs support learning English do separate stronger academic performers from those who are less able. However, not all RFEPs are equal — those who are reclassified at younger grades are more likely to progress on time, have higher test scores, and have positive outcomes in their final years of high school. LTELs who are reclassified do not perform as well, on average, as RFEPs who are reclassified at younger grades. These results suggest that cross -sectional views of RFEPs’ progress over time are complicated by the fact that newly reclassified students, especially those 25 We identify students who leave school by the following exit codes in CALPADS: E140 (no known enro llment, truant), E300 (expelled, no known enrollment), E420 (no show, same school), T270 (trans, drop, adult school), and T380 (trans institution, no HS diploma). 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Leave before graduating Diplomaa-g courses complete Final high school outcomes, 8th grade cohort IFEP RFEP pre RFEP target year (2008-09) RFEP post EO EL http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 19 reclassified after having been LTELs, do not perform as well as ELs reclassified in elementary school. The results also highlight the fact that LTELs who are reclassified still outperform EL students dramatically on multiple measures. We find that RFEP students perform better than EO student in many cases; and in some cases, RFEP students perform better than IFEP students, the group that is often at the top of the performance measures. These last two findings suggest that there is a role for rethinking the goal of EL programs and reclassification standards: What is the right performance level for reclassifying students to insure their academic success after EL support ends ? 26 In the following section, we discuss the variation in district policies for reclassifying students. In later sections we examine how those policies relate to reclassification rates for districts and how they relate to student outcomes (while simultaneously considering the role of student and district characteristics). 26 PPIC research -in-progress that is focusing on longer -term analyses (and using more contextu al data about districts) might help address this question. http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 20 How D oes R eclas sification W ork ? Although state law is very clear on the criteria school districts should use to identi fy English Learners, it is less prescriptive about the policies districts use to reclassify these students. 27 According to state law, school district reclassification decisions must incorporate four criteria: an assessment of English proficiency, an evaluat ion of basic s kills in English, a teacher evaluation of the student, and parental consultation ( Education Code 313(f)). 28 To help districts de velop effective reclassification policies, the State Board of Education established guidelines and recommendations for each of the f our criteria in 2002. 29 For example, the board recommends that districts use the ELA portion of the CST to assess the basic skills of ELs, and set a cut -off score somewhere in the range of Basic to midpoint -Basic (300– 325 points). The guidelines are less specific for the other reclassifica tion criteria. Given the nature of the state guidelines, one might expect large variation in EL reclassification policies and r ates; and in deed, Linquanti (2001), the California State Auditor (2005), and Parrish et al. (2006) found wide variation in reclassification policies across districts in the small sample each used .30 Moreover, the auditor found that of the 180 ELs included in the study, 62 percent met their district’s stated criteria for reclassification but were not reclassified. The auditor recommended clearer statewide reclassification guidelines and regulations to create a more consistent experience for English Learners across the state. Motivated by a desire to better understand and improve EL policy, and with the goal of improving reclassification rates and student outcomes, SB 1108 (Chapter 434, Statutes of 2012) charged the CDE with the task of reviewing school district reclassification policies and practices (with the intent of providing the information necessary for developing new state policy) .31 To help advance the conversation about EL reclassification, the Public Policy Institute of California surveyed local education agencies throughout California. We surveyed school districts about their reclassification policies, rather than collect and review written reclassification policies (such as those contained in English Learner Master Plans), for three reasons. First, we wanted this report to be as timely as possible for current policy discussions. Requesting, reviewing, categorizing, and analyzing di strict plans would have been much more time - and resource- intensive. Second, survey responses afforded the possibility to gather both districts’ reported criteria and additional detail about district reclassification practice that might not be spelled out in district plans. Finally, a survey allowed us to gather detail about both current p olicy and practice as well as 27 Students identified by the Home Language Survey as speaking a language other than English are assessed for their English proficiency through the California English Language Development Test (CELDT). 28 A copy of Education Code 313(f) is available in Appendix F . 29 The SBE approved the “Guidelines for Reclassification of English Learners” on September 11, 2002. On September 6, 2006, the SBE approved modifications to the guidelines. A copy of the guidelines is available in Appendix F . According to the state auditor’s report (2005): “Because th ese are not regulations, school districts are not required to adhere to the department’s guidelines. However, according to the board’s chief legal counsel, the guidelines were based on an analysis of actual test data and developed with public input, so the board expects that school districts will pay great deference to them when making their initial identification and their redesignation decisions. Nevertheless, these are only guidelines and school districts are allowed flexibility in defining their criteria.” (p. 18) 30 Linquanti examined policies from seven districts, prior to SBE issuing its reclassification guidelines. The state auditor reviewed the policies of eight school districts: Anaheim Union High, Long Beach Unified, Los Angeles Unified, Pajaro Valley Unified, Sacramento Unified, San Diego Unified, San Francisco Unified, and Stockton Unified. The auditor also reviewed a total of 180 individual EL cases from these districts. Parrish et al. reported on nine California districts. States, too, vary in their reclassification criteria; see Abedi (2008) for an explanation of how this makes state -to-state comparisons of EL and reclassified EL populations difficult. 31 Contingent upon the availability of funds for the research. http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 21 past policy and practice. In order to complete our longitudinal analysis, we needed to know the policies and practices in place in school districts in the 2008– 09 school year. There are shortcomings with the survey approach, h owever. For examp le, though we attempted to have the most knowledgeable district employee complete the survey, it is possible that the respondent may not always accurat ely report district practi ce. Further, distric t practice could deviate from what was report ed to us. It is worth noting that a review of written district plans would have suffered from the same shortcoming. In the following pages, we present the results of that survey and describe the main differences in reclassification across the state. In this report, when we refer to district p olicy, we are referen cing the policie s reported to us by school district staff in PPIC’s “English Learners Reclassification Survey.” Assessment of English-Language Proficiency: Districts should use the most recent CELDT test data as the primary criterion and consider for reclassification those students whose overall proficiency level is early advanced or higher and each subtest score is intermediate or higher. Students with ove rall proficiency levels in the upper end of intermediate may be considered for reclassification if additional measures determine the likelihood that a student is proficient in English. Teacher Evaluation: Districts should use a student’s academic performan ce and note that incurred deficits in motivation and academic success unrelated to English -language proficiency do not preclude a student from reclassification. Parent Opinion and Consultation: Districts should provide notice to parents or guardians of the ir right s and encourage them to participate in the reclassification process and provide an opportunity for a face -to -face meeting. Comparison of Performance in Basic Skills: A student’s score on the ELA portion of the CST or CMA in the range from the begin ning of Basic up to midpoint Basic suggests that the student may be sufficiently prepared to participate effectively in the curriculum. Districts may select a cut point in this range. Students with scores above the cut point should be considered for reclassification. For students scoring below the cut point, districts should determine whether factors other than English- language proficiency are responsible and whether it is reasonable to reclassify the student. For students in grade 12, the grade 11 CST re sults may be used. For students in grade 1 , districts should base reclassification decision s on the CELDT results, teacher evaluation, parent consultation, and other locally available assessment results. Districts must monitor student performance for two years after reclassification. SOURCE: California Department of Education. 2012, “California English Language Development Test (CELDT), 2012- 2013,” CELDT Information Guide. NOTE : Excerpted from CELDT Information Guide. See Appendix F for complete state guidelines . http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 22 PPIC ’s “ English Learners Reclassif ication Survey ” The PPIC reclassification survey was developed in consultation with EL experts and s everal large school di stricts. The structure of the PPIC reclassification survey reflects the structure of the reclassification guidelines. Districts were asked about the four main components of reclassification decisions: how they assess English proficiency, how they assess basic skills, how teachers’ evaluations are conducted and incorporated, and how parents are consulted and notified of reclassification decisions. A copy of the questions in the survey instrument is in Appendix B. A more complete description of the methods used to validate and distribute the survey is available in Appendix C . The survey was distributed to all local education agencies (LEAs) in existence in June 2013, based on contact information provided by the CDE. For LEAs that receive Title III funding, the primary survey contact was the Title III contact. For all remaining LEAs, the survey was distributed to the superintendent or charter school principal. The LEA contacts were encouraged to forward the survey to the most knowledgeable individual on LEA reclassification policies. Nearly half of the survey respondents were directors of EL services or EL coordinators (Table 3). TABLE 3 Survey respondents Job title Percent of respondents Director of EL services 26.4 EL coordinator/specialist 20.8 Superintendent 18.5 Other 12.9 Director of curriculum 9.9 Assistant superintendent 9.2 Teacher on special assignment 2.3 Total Respondents (N) 303 SOURCE: PPIC’s “English Learners Reclassification Survey.” The final sample used to report district EL policies consists of 303 school districts out of a total of 962 districts in the state .32 The district response rate was 31 percent; however, the districts that responded serve more than 54 percent of the state’s students. Elementary districts were the least likely to respond to the survey, and high school districts were the most likely. Our sample districts are fairly representative of the state as a whole. Of the largest 10 districts in the state, nine responded, which may explain why sample districts have more English Learners and low -income students than non-responders. In Appendix Table A1 , we show that using only districts that responded to our survey does not appreciably change the nature of our sampled 32 In addition to the 3 03 districts that responded, some county offices of education and charter schools or charter management organizations responded. We exclude county offices of education and charter schools from the survey results and analyses. Two districts are included in the survey results section but not in the student outcomes section because of data issues. http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 23 cohorts —indeed, as suggested in Table 4, while it reduces the number of students, it slightly increases the share of EL students in our final analysis. TABLE 4 S urv ey r es pond ents are broadly repr esen tative of the state’s student demographics District characteristics Responded Did not respond Elementary 139 397 Share of elementary students (%) 36 64 Average enrollment 3,037 1,744 High School 33 46 Share of high school students (%) 41 59 Average enrollment 7,439 6,245 Unified 131 208 Share of unified students (%) 64 36 Average enrollment 19,492 6,789 All districts Share of state’s students (%) 54 46 Share of Spanish- speaking ELs (%) 58 42 Share of all other language ELs (%) 62 38 API (average) 780 772 Low -income (average %) 60 56 English Learners (average %) 23 21 Reclassification rate (average %) 12 10 SOURCES: PPIC’s “English Learners Reclassification Survey” and PPIC’s School Finance Model (2013). NOTES: Enrollment is average daily attendance (ADA) from the PPIC School Finance Model (2013); averages are weighted by ADA. Reclassifi cation P ol icies In this section, we report our survey findings on districts’ reported EL programs and reported reclassification policies . 33 We a sked specifically about district policies for non -special -education EL students in structured English immersion (SEI) programs (all districts are required by state law to have an SEI p rogram). Among 33 The survey asks districts about their policies and in some instances their opinions about those policies. In some cases, however, districts may be reporting practi ces rather than policies adopted by their local plans. We did not collect locally adopted plans for comparison with the survey responses. http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 24 districts that also have English language mainstream programs (87%), all but one u se the sam e reclassification criteria as they do for SEI students. Fewer districts (37%) report that they also have alternative EL programs, such as bilingual or dual language programs; and the vast majority of the districts that do ha ve such programs employ the same reclassification criteria in alternative programs as they do for their students in SEI programs. Finally, almost all districts allow special education EL students to be reclassified (these policies are discussed in Appendix C). 34 In the remainder of this section, we consider how the districts have responded to the state guidelines for EL reclassification discussed in the preceding section: Assessment of English Language Proficiency, Evaluation of Basic Skills in English, Teacher Evaluation, and Parental Opinion and Consultation. We asked districts to report their reclassification policies by grade -level groupings: K –2, 3 –5, 6 – 8, and 9 –12. Assessm ent of E nglish Languag e P roficiency All school district s with ELs report using the CELDT to assess English proficiency (Table 5) .35 State guidelines suggest using an overall cut -off score of Early Advanced on the CELDT, a policy adopted by almo st all s cho ol districts across all grade levels. However, about 10 percent of districts report exceeding the state g uidelines and require a score of Advanced for students in grades K–8, and 7 percent require a score of Advanced for high school students (Table 5). Although an overall score of Early Advanced is recommended to demonstrate English proficiency, the CELDT has four subtests (speaking, listening, reading, and writing) , a n d it is possible 34 For each of the 13 state classifications of disabilities, more than 87 percent of districts reported that EL students with that disability could be reclassified. 35 One district reports that no ELs have enrolled for many years. Thus, in practice it has not administered the CELDT in reclassification decisions. EL Pr ogram s English Learners receive instruction in one of three settings. Structured English Immersion (SEI): A classroom setting where English Learners who have not yet acquired reasonable fluency in English, as defined by the school district, receive instruction through an English language acquisition process, in which nearly all classroom instruction is in English but with a curriculum and presentation designed for children who are learning the language. All districts are required to offer an SEI program. English Language Mainstream (ELM): A classroom setting for English Learners who have acquired reason able fluency in English, as defined by the school district. In addition to ELD instruction, English Learners continue to receive additional and appropriate educational services in order to recoup any academic deficits that may have been incurred in other a reas of the core curriculum as a result of language barriers. Alternative Program (Alt): A language acquisition process in which English Learners receive ELD instruction targeted to their English proficiency level and academic subjects are taught in the primary language, as defined by the school district. Placement in an alternative program is triggered by the parents through a parental exception waiver. http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 25 to achieve an overall score of Early Advanced while scoring below Early Advanced on on e ( or more) subtest .36 S tate guidelines recommend that Intermediate or higher in each subtest be considered sufficient as long as the overall score is Early Advanced. About 40 percent of districts exceed state guidelines and require that students achieve Early Advanced on each subtest. Conversely, 60 percent of districts follow t he state guidelines and accept a score of Intermediate on one or two CELDT subtests. TABLE 5 English proficiency criteria, by grade level English proficiency criteria Grades K–2 Grades 3– 5 Grades 6– 8 Grades 9– 12 Using CELDT (%) 99.3 99.6 99.6 100.0 Requiring overall cut -off score of Advanced (%) 9.8 9.6 9.5 7.1 Requiring subtests above Intermediate (%) 43.4 42.6 40.7 39.6 Respondents (N) 144 271 274 168 SOURCE: PPIC’s “English Learners Reclassification Survey.” Eva luation of Basic S kills in E nglish State law requires that the SBE guidelines utilize multiple criteria, including a comparison of the performance of the pupil in basic skills against an empirically established range of performance in basic skills based on the performance of English proficient pupils of the same age. SBE guidelines suggest using the ELA section of the CST and using a cut -off score somewhere in the range of Basic to mid point Basic (a scale score of 300 –325), which the majority of districts follow. However, about 30 percent of districts enrolling s tudents in grades 3 –8 report impo sing a higher cut -off poin t of Proficient (Table 6) . 37 Only one district reports using a cut- off score of Advanced. For high school students, districts typically use a lower cut- off score. About 30 percent of the districts use the CAHSEE (which is thought to be less rigorous than the CST ELA exam) as a supplemental or alternative measure of basic skills for high school students (not shown in table). TABLE 6 Percent of districts imposing basic skills requirements, by grade le vel Basic skills requirement Grade 2 Grades 3– 5 Grades 6– 8 Grades 9– 12 CST ELA 78.2 95.9 95.3 91.0 Basic cut -off point 22.8 27.0 29.9 37.1 Midpoint -Basic cut-off point 40.9 42.3 40.7 40.7 Proficient cut -off point 29.9 30.3 29.1 22.2 CST Math 46.9 47.4 45.8 36.5 36 Subtest weighting for the CELTD varies by gra de, with reading and writing each counting for only 5 percent of the overall score in kindergarten and first grade, and all subtests weighted equally for grades 2 –12 (CDE, 2012). 37 The CST is not administered until grade 2. http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 26 Basic skills requirement Grade 2 Grades 3– 5 Grades 6– 8 Grades 9– 12 Other CST 8.2 8.1 8.4 Writing 8.9 9.2 8.4 Respondents (N) 147 270 273 169 SOURCE: PPIC’s “English Learners Reclassification Survey.” Although state law and state guidelines refer only to basic skills in English, nearly half of all districts report requiring some mastery of math in grades 3 –8 (Table 6). Furthermore, about 8 percent of districts require that students demonstrate some level of basic skills in other subject tests, using either the science or history CST or both. And about 9 percent of districts require that students demonstrate written English proficiency, most commonly through a district writing rubric .38 About 7 percent require all of these additional measures (math CST, writing, and at least one additional CST) in determining a student’s mastery of basic skills. Table 6 does not report basic skills criteria used by districts for students in grades K –1. In the survey, districts were asked to describe the criteria used to assess the basic skills of EL students in grades 2 through 12. Because CST results are not available until late summer after the second grade, only 78 percent of the districts that reported reclassifying second grade students relied on CST results for reclassification . 39 Oth er dis tricts used different assessments. The most frequent responses for measuring the basic skills of stu dents in gra des K –1 included district or teacher rubrics and national assessments of reading development such as the Dynamic Indicators of Basic Early Literacy Skills (DIBELS).40 Teach er E va luation State l aw requires that the teacher evaluation include a review of the EL student’s curriculum mastery ( Education Code 313(f)(2)). State guidelines say that academic performance should be considered as part of the tea cher evaluat ion of EL students. Beyond that, the state offers little guidance about incorporating teachers’ perspectives in reclassification decisions. To capture the range of district practices, we asked districts about a host of potential compon ents of the teacher evaluation. Districts were also allowed to provide more information about a given component or to write in components not included in our list. We then categorized these components as “required” or “considered.” If a district reported that an individual component was required, we prompted th e respondent to enter a grade, GPA, or score, and we determined the component to be a required component of reclassification . 41 If the component was deemed by the district respondent to be “considered” rather than “requ ired,” we did not prompt for a grade, GPA, or score, and we determined the component to be a considered component of reclassification. Most districts reported using required criteria as part of the teacher evaluation (Table 7). Less than 5 percent of districts reported that they do not consider grades, GPA, or assessments across the various grade spans. 38 Responses regarding the testing of writing proficiency were write -in responses, not prompted in the survey questions. 39 Of the districts that serve second grade students, 54 percent responded that they allow students in second grade to be reclassified. See Appendix Table C2 . 40 DIBELS are a set of short (one minute) procedures and measures that can be used to regularly monitor and assess the acquisition of early literacy skills from kindergarten through sixth grade. 41 Requiring grades and GPA rather than just considering them does not mean a district is using entirely objective criteria because grades are assigned by teachers and likely include some component unrelated to test scores, such as dis cipline, effort, etc. See Parrish et al. (2006) for a discussion related to using grades in reclassification decisions. TABLE 6 (continued) http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 27 Grades or GPA are used more often in the teacher evaluation of high school students than in grades K –8. Conversely, districts are less likely to require assessments as part of the teacher evaluation in high school than in the lower grade levels. Most districts reported requiring a grade of “C” or better in English -language arts classes, although a few districts require a grade of “B” or better. This matches what districts reported for GPA requirements, which were generally a 2.0 or equivalent, typically associated with a grade of “C.” Districts reported using a variety of assessments that ranged from district rubrics to those developed by testing companies, such as the DIBELS. TABLE 7 Teacher evaluation criteria, by grade level Teacher criteria Grades K–2 Grades 3– 5 Grades 6– 8 Grades 9– 12 Grades/GPA requ ired 64.2 65.3 64.6 70.2 Grades/GPA considered 23.0 25.8 24.0 20.2 Assessments required 50.7 46.1 44.9 34.5 Assessments considered 39.9 43.9 40.5 45.2 None of the above 3.4 2.2 4.4 1.8 Respondents (N) 148 271 274 168 SOURCE: PPIC’s “English Learners Reclassification Survey.” NOTE: Rows do not sum to 100% because respondents could chose more than one category among the first four. Districts also reported using a variety of subjective components in the teacher evaluation criteria (Table 8). Although sta te guidelines appear to discourage the use of students’ behavior and motivation in teacher evaluations, many districts report considering factors such as student attendance, behavior, participation, and discipline. Districts are generally less likely to re port considering these other components for high school students than for students in grades K –8, with one exception: discipline. Disciplinary issues are more likely to be reported as considerations by districts in reclassifying high school ELs than in reclassifying K–8 ELs. Districts that report considering participation tend to be smaller an d to e nroll fewer free and reduced- price lunch students, while districts that report considering discipline an d beha vior tend to enroll fewer English Learners. Among elementary school ELs, elementary districts were more likely than unified districts to report considering behavior and discipline. Among high sch ool ELs, high scho ol districts were more likely to report considering attendance and discipline while unified di stricts were m ore likely to report considering participation and behavior. http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 28 TABLE 8 Subjective evaluation criteria used by teachers, by grade level Subjective criteria Grades K–2 Grades 3– 5 Grades 6– 8 Grades 9– 12 Attendance 30.4 29.5 31.2 23.5 Behavior 20.3 19.6 20.9 13.3 Benchmarks 84.5 80.4 81.4 67.5 Discipline 10.1 9.6 11.1 16.3 Homework 31.8 35.4 37.9 20.5 Participation 61.5 60.2 59.7 46.4 Portfolio 43.9 42.1 43.1 38.8 Projects 45.9 47.2 50.2 37.0 None of the above 13.5 16.2 23.7 27.7 Respondents (N) 148 271 253 166 SOURCE: PPIC’s “English Learners Reclassification Survey.” Parental O pinion and Consult ation State guidelines are equally limited for this element of reclassification policy, stating only that districts should provide parents with a notification of reclassification and the opportunity to participate in the reclassification decision and process, including the opportunity for an in -person meeting. To capture districts’ policies, we first asked how parents are consulted i n the reclassification decision and then how they are notified. Nearly all districts report that they explain reclassification criteria to parents (95%), present the student’s performance data (86%), and solicit parents’ opinions (85%). A majority of districts (62%) report that they provide parents with a comparison of students’ performance data to the reclassification criteria. A handful of districts say they do not consult parents or consult parents in some other manner. Parents are primarily informed o f the district’s reclassification decision by letter (88%) or an in -person meeting (75%). About 58 percent of districts report that they inform parents by phone. Timing State law does not specifically prohibit reclassification in certain grade levels; state guidelines recommend using the CST in grades 2 –12 and an alternative district assessment in first grade (before the CST can be used) as measures of English proficiency. Nearly half or more of all districts report that they do not reclassify in the early grades (K –2): About 30 percent of districts report permitting reclassification in kindergarten, 47 percent in grade 1, and 54 percent in grade 2. Almost all districts report permitting the reclassification of ELs by grade 3 (Figure 4). Given these findings, it may not be surprising that students are more likely to be reclas sified relatively soon after the results of the first CST are available in the summer following se cond grade, many of which count as third grade reclassifications. According to our respondents, ELs are most commo nly reclassified in the middle grades (4 –6), with 5th grade being the modal response http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 29 (56% of di stricts). Parrish et al . (2006) found that reclassification rates were lower in districts that did not permit reclassification until grade 3. 42 FIGURE 4 Percent of districts that permit reclassification, by grade level SOURCE: PPIC’s “English Learners Reclassification Survey.” The reclassification of students is also more common during certain periods of the academic year. The vast majority of districts report reclassifying students in the spring (using the CST that students took nearly 12 months earlier) or in the fall (when the most recent CST scores are available). However, most d istricts reclassify students at multiple points during the year (Figure 5). About 30 percent of districts reclas sify during only one season (fall, winter, spring, or summer). About 2 percent reclassify year -round. Parrish et al . (2006) found that districts reclassifying multiple times durin g the schoo l year were likely to have higher reclassification rates. 42 Future iterations of this research could consider this factor. 0 10 20 30 40 50 60 70 80 90 100 K12345678910 1112 Percent of districts Grade level http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 30 FIGURE 5 Most districts allow for student reclassification at more than one point in time per year The state guidelines were last amended in September of 2006. Very few districts report initiating any changes in their reclassification policies since 2008. About 5 percent of districts changed their English proficiency standards by either using a different overall CELDT score or by changing t heir required CELDT subtest scores. About 15 percent of districts made changes to their basic skills requirements, most often changing the cut -off score required on the ELA portion of the CST or changing requirements for the math CST. Among all districts, 8 percent have made changes to their teacher evaluation standards since 2008. Opinions a bout Re cla ssific ation P olic ies We asked survey respondents about their opinions of their district’s reclassification policies —for example, which aspect of reclassificat ion criteria is the most difficult to achieve by school level (elementary, middle, or high school), allowing just one choice among each of the four criteria. The majority of respondents believe that the basic skills requirement (commonly measured by the CST ELA) is the most difficult requirement across all three grade spans (Table 9), and that it is more difficult for middle - and high school students than for elementary students. Many respondents (40%) consider English proficiency (as measured by the CELDT) to be a more difficult hurdle for elementary school students than for middle- and high school students (27% and 26%, respectively). Very few respondents consider teacher evaluation or parent cons ultation to be difficult barriers (teacher evaluation was selected more often for middle- and high school students than for elem entary students). 0 10 20 30 40 50 60 1 season 2 seasons 3 seasonsYear-round Percent of districts Number of seasons in which reclassification can occur http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 31 TABLE 9 “If you were forced to choose one criteria for each grade level, which of the reclassification criteria would you say is most difficult for EL students to meet in your district?” Elementary Middle High Basic skills 52.8 62.3 67.9 English proficiency 40.1 26.8 25.6 Teacher evaluation 3.3 6.0 8.9 Parent consultation 0.0 0.0 0.5 Don’t know 3.7 4.9 9.5 Respondents (N) 269 265 168 SOURCE: PPIC’s “English Learners Reclassification Survey.” We also asked respondents to rank the importance of the various reclassification criteria for deciding whether to reclassify a student (Table 10). English proficiency (38%) was most likely to rank at the top. Although most respondents believe that basic skills are the most difficult hurdle for students to surmount (Table 9), only 5 percent of respondents ranked this criterion higher than all other factors influencing their decision to reclassify students (Table 10). Another 26 percent of respondents selected basic skills in combination with English proficiency as being most important in the ultimate decision to reclassify a student, and an additional 23 percent ranked basic skills, English proficiency, and teacher evaluation equal ly as th e most important criteria. TABLE 10 “In your opinion, how important are each of the reclassification criteria in the ultimate decision to reclassify a student?” Agree (%) English proficiency is most important 37.6 English proficiency and basic skills are equally important 25.9 English proficiency, basic skills, and teacher evaluation are equally important 22.8 Basic skills are most important 5.2 English proficiency and teacher evaluation are equally important 5.2 Teacher evaluation is most important 4.8 Respondents (N) 290 SOURCE: PPIC’s “English Learners Reclassification Survey.” http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 32 In 2011– 12, reclassification rates ranged from 0 to about 30 percent statewide .43 The vast majority of respondents (94%) believe that their district’s reclassification policies are “about right” (shown in Appendix C), while about 78 percent of respondents believe their reclassification rate is “about right” —with a lmost all of the others believing their reclassification rate is “too low” (shown in Appendix C). Among those who believe the reclassification rate is too low, the vast majority think that their reclassification policies are about right. This would seem to suggest that other factors (such as EL students’ performance or implementation of the reclassification policies) rather than the reclassification polic ies themselves, determine respondents’ satisfaction with district reclassification rates. Summing Up PPIC’s 2013 survey of school district reclassification policies and practices garnered a response rate of 31 percent of districts, representing 54 percent of California’s K –12 students. This report confirms the findings of past research (using small samples of districts) that there is substantial variation in the reclassification practi ces and policies of California school districts. In political and practitioner circles, such variation is often viewed in a negative light because it means that students with the same skills will face different reclassification decisions in different parts of the state. There is also concern statewide about what the “ right” reclassification policy should be. Clearly, based on the survey responses, some districts believe that the state guidelines offer too low a threshold, and they impose many more requirements, particularly with regard to basic skills, including writing samples or content knowledge in other subjects. Do these policy differences affect reclassification rates, which also vary widely across the state (from 0 to 30%)? We address this question in the following section. 43 There are a few districts with substantially higher reclassification rates, but these are districts with very few EL students. The rates were calculated using ELs redesignated during the year in the numerator and the sum of ELs and ELs redesignated in the denomin ator. Data are from DataQuest (http://dq.cde.ca.gov/dataquest/LC/LCOtherDistrict.aspx?dType=all&co=AllCounties&TheYear=2011 -12&sortby=c ). http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 33 Do District Reclas sification P olicie s Infl uence Recla ssifica tion Ra tes? As demonstrated in the previous section, there is a great deal of variation in how and when districts decide that EL students are ready to be reclassified. It is important to understand whether the differences in reclassification policies and practices are associated with differences in district reclassification rates and ultimately, whether the differences in these policies and practices are associated with differences in reclassified students’ outcomes. In this section, we examine those district reclassification policies that are more rigorous than recommended by the SBE guidelines , and using multiple regression analysis, we test w hether districts us ing those more rigorous requirements have lower reclassification rates. Below, we describe in more detail how we categorize the districts’ reclassification policies, then calculate districts’ reclassification r ates and consider the role of other district characteristics. Recall from the section ab ove that the SBE provides guidelines to districts for reclassifying students, most specifically in terms of basic skills and English proficiency. School districts have had more considerable latitude with regard to implementing the criteria for the teacher evaluation and parental opinion and consultation . In modeling the association between district reclassification policies and district reclassification rates, we use reclassification policies and practices from 2012 –13 and reclassification rates from the same period. Because reclassification policies can vary by grade level within a district, we chose the policies for the grade levels in which most reclassification occurs when modeling the overall reclassification rate for the district. For elementary and unified school districts, we used the reclassification policies reported for grades 3 –5.44 Table 11 summarizes how districts reported implementing their own reclassification criteria in 2012 –13. It is stra ightforward to construct a variable to capture whether a district is using a higher threshold than suggested by SBE’s reclassification guidelines in the district’s use of the CELDT and the CST ELA, as well as the effect of requiring additional tests, such as the math CST. State guidelines are less definitive in the case of teacher evaluations, and thus interpreting whether district requirements exceed the state guideline is less clear. If a district reports requiring specific grades (or GPA) or specific assessment results for reclassificati on, we consider that to be a higher threshold than is specified in the state educational code. Districts reporting teacher evaluation components that include considering grades (or GPA) or assessments but not requiring e ither are cons idered to be using the components suggested in the SBE guidelines. If a district reported that it used disciplinary issues as a factor in reclassification decisions, we included that as well. As we noted in the previous section, our survey was not successful in eliciting detailed differences in districts’ parental involvement practices. In addition, district respondents did not rank parental involvement as particularly important in reclassification decisions, so we do not include it in our models that estimate district reclassification rates. 44 Among elementary districts, 97 percent have the same basic skills criteria as for grades 6 -8, and 96 percent have the same CELDT criteria. In unified districts, 94 percent have the same basic skills criteria for grades 6 – 8, and 91 percent have the same criteria for grades 9 –12. CELDT requirements are more likely to be the same across grade levels in unified districts (97% in grades 6 –8 and 96% in grades 9– 12). http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 34 TABLE 11 District reclassification policy Reclassification criteria and requirements Reclassification policies grades 3–5 Reclassification policies grades 9 –12 Elementary districts (%) Unified districts (%) High school districts (%) English Language Proficiency CELDT Overall Performance Level (OPL) OPL is Early Advanced (EA) or higher SBE guideline 88% 91% 97% OPL is Advanced More rigorous 12 7 3 CELDT subtests (reading, writing, speaking, listening) Subtests can be below Early Advanced SBE guideline 51% 65% 50% Subtest scores must all be Early Advanced or higher More rigorous 49 35 50 Basic Skills in English CST ELA Score is Basic or higher SBE guideline 27% 27% 55% Score is mid -Basic or higher More rigorous 39 45 36 Score is Proficient or higher More rigorous 33 27 10 Requiring CST Math More rigorous 57 37 35 Requiring CST history/science More rigorous 9 7 16 Teacher Evaluation Unspecified SBE guideline 24% 22% 13% Require grades/GPA and assessments More rigorous 42 28 13 Require grades/GPA More rigorous 24 37 69 Require assessments More rigorous 9 13 6 Consider disciplinary issues Discouraged 11 9 29 SOURCE : Authors’ tabulations from PPIC’s “English Learners Reclassification Survey.” In addition, we examined the incidence of districts’ use of multiple reclassification requirements that exceed the SBE guidelin es. While none of our surveyed districts employs a ll of the requirements shown in Table 11, most use several of the measures. Of the 294 districts with full reclassification information, only 11 report http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 35 using the measures specified in the SBE guidelines; 42 use one of the more rigorous measures, and 155 (or 53%) use th ree or m ore of these more rigorous measures (See Appendix Table C24 ). However, there are so many different combinations in use that it is, in practice, impossible to test them all. We settled on using the two most common requirement combinations: 1. requiring that all CELDT subscores be Early Advanced or higher and using the CST math test (57 districts), 2. requiring that all CELDT subscores be Early Advanced or higher and requiring both student grades/GPA and teacher assessments (43 districts). In addition to the reclassification policies for each district, we considered the role of district characteristics, such as type of district (elementary, unified, or high school), size of district, 2011– 12 API score, percent of Title I students, average daily attendance, and percent of students that are ELs. District reclassification rates are from CDE’s DataQuest. When 2012–13 reclassification rates were missing but 2011 –12 rates were available (35 districts), we used those and estimated our model for the 291 districts that had complete survey responses and reclassification rate data . 45 The 2012 –13 statewide reclas sification rate was 11.1 percent. The average reclassification rate for surveyed districts with complete responses was 11.2 percent, ranging from 0 to 75 percent, with half of the districts having reclassification rates in the range of 6.2 to 14.5 percent. Just 5 percent of the districts had reclassification rates over 24.5 percent, while 11 percent had reclassification rates lower than 1 percent. E stablishing higher cut -offs on the required state assessments is linked to lower reclassification rates, compared to requiring just the thresholds in the SBE guidelines; in some cases, these differences are statistically significant (see full results in Appendix Table D1 ). Figure 6 shows the reclassification rates for an elementary school district using the SBE reclassification guidelines (first bar, 15%). Each succeeding bar illust rates the predicted reclassification rate of an average unified school district using one of the more stringent requirements, holding other district characteristics constant. 45 Reclassification rates are calculated by dividing the number of RFEP students reclassified since the previous years’ language census by the sum of EL students and RFEP students reclassified since the previous years’ language census. http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 36 FIGURE 6 District reclassification rates under different reclassification requirement scenarios SOURCES: PPIC’s “English Learner s Reclassification Survey” and CDE reclassification rates . NOTE : *** p < .01; ** p<.05, * p<.10 . Regression also included districts’ use of CST other, grades required, and assessments req uired. To e stimate the reclassification rate for the average unified school district (first bar), we multiplied regression coefficients (in Table D1) for district characteristics by their mean values for unified school districts and added the sum to the constant. Sub sequent bars add the coefficient for significant reclassification criteria to the predicted d istrict reclassification rate. Districts that require a CELDT Overall Performance Level of Advanced have lower reclassification rates than those using the state minimum score of Early Advanced. Districts that require all CELDT subtests to be Early Advanced or higher also have lower reclassification rates. The effect of these two requirements is similar — compared to using the minimum criteria in SBE’s guidelines, using the higher cut- offs on the overall CELDT and the CELDT subtests are each associated with district reclassification rates about 3 percentage points lower, a substantial reduction in reclassification. A requirement that CST ELA scores be higher (either mid- Basic or Proficient) is also associated with overall lower reclassification rates —4 percentage points lower for a minimum score of mid -Basic and 3 percentage points lower for a minimum score o f Proficient. Holdi ng other factors constant, requiring the CST math test alone was not linked to a statistically significant difference in reclassification rates. Nor did requiring the results of either the CST history or CST science test indicate statistically significant differences in rates . 46 Reporting the use of disciplinary issues in reclassification was 46 In this report, we did not consider the role of the score required for the CST math, history, or science tests. 15.1% 12.2% 12.2% 11.1% 12.0% 16.0% 12.1% 20.1% 9.6% 0% 5% 10% 15% 20% 25% Percent (%) Reclassification criteria * *** * *** *** * * http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 37 actually associated with higher reclassification rates, suggesting that a lack of disciplinary problem s elevates a student’s chances of reclassification, all else equal . Looking finally at two common combinations of requirements that go beyond the SBE guidelines, we find that districts stipulating both a grade or GPA minimum and required scores on particular assessments have lower reclassif ication rates (about 3 percentage points lower), but this result is only statistically significant at the 10 perc ent level. Holding district characteristics constant, those districts that have a reclassification policy that uses two of the more rigorous reclassification requirements (requiring that the CELDT subt est scores be Early Advanced or higher and requiring the math CST) have reclassification rates 5.4 percentage points lower than districts using neither of these more stringent criteria. We did not find any statistical relationship between other measures of teacher evaluation criteria and district reclassification rates. Summing Up This section has demonstrated a clear association between reported reclassification policies and reclassification rates at the district level. Policies with more- rigorous performance thresholds are associated with lower reclassification rates, holding constant other district characteristics that may also determine reclassification rates. If districts use multiple additional requirements, reclassification rates are further reduced. We tested the most common combination —the CST math requirement and requiring that results on all subtests of the CST ELA be at least Early Advanced—and found that reclassification policies that use those two requirements together redu ce district reclassification rates by 5.4 percentage points relative to the criteria suggested in the SBE guidelines for reclassification. Because fewer students are reclassified in d istricts with higher reclassification thresholds, we might expect to see that those students who succeed in b eing reclassified have better academic outcomes than students who do not. We explore this question in the following section. http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 38 Do Re clas sification P olicie s A ffect Stud ent O utcom es? Available measures indicate that reclassified EL students are among the best performing students in the state. This has been clearly demonstrated using both cross -sectional data (e.g., Hill, 2012; Saunders and Marcelletti, 2013) and longitudinal data (e.g., Flores, Painter, and Pachon 2009), as well as in this report. We have also demonstrated that some or all of the more rigorous reclassification performance thresholds are associated with lower district reclassification rates. Some of the more difficult reclassification performance thresholds may result in keeping EL students classified as ELs for longer durations. In this section, we first examine whether RFEP students are still among the best performing students in comparison to other student language groups when we simultaneously consider the role of student and district characteristics. We then focus more specifically on whether more difficult reclassification standards are associated with better RFEP outcomes. Longitudi nal RFEP Outcomes Earlier in this paper, we noted that cross- sectional research shows that RFEP students outperform EL students in high school but that their advantage appears to shrink at higher grades. Our analysis follows cohorts of EL students who are reclassified at various grade levels and compares them to EO students, to IFEP students, and to students who are never reclassified. This enables us to understand student outcomes for each group of RFEP and EL students without the potential complication of newly arriving ELs or RFEP students who are reclassified at higher grades being indistinguishable from RFEP students reclassified at younger grades. The academic outcomes we consider are CST ELA scores, on -time grade progression, and end -of -high- school outcomes .47 We use multivariate regression analysis to simultaneously consider the role of student characteristics and district characteristics for each of our cohorts of students. Full regression model results are available in Appendix E .48 Figures 7 and 8 indicate that if we use longitudinal data and control for student and district characteristics, RFEP students outperform ELs students on the CST ELA, no matter when they are reclassified (in our target reclassification year, beforehand, or afterward), and no matter what cohort we examine (2nd, 4th, or 7th grade). Again, this is not surprising given that reclassification criteria are intended to separate hi gh-performing ELs from those who could benefit from more English language support. This is true whether we are considering the share of students in each language group who score Basic and higher (Figu re 7) or Proficient and higher (Figure 8) . 49 The gap between CST scores for RFEP students reclassified in a pre -target reclassification year and those reclassified in a post -target reclassification year is larger for older than for younger cohorts. These findings suggest that the cross -sectional view of RFEP performance in high school grades is skewed by the late entrants —those reclassified in grades 8 to 12 generally have lower levels of performance than those reclassified at younger grades. 47 Subsequent extensions of this research could consider other outcomes, such as performance on the CST math and CAHSEE tests. 48 We also estimated models where we did drop all students who transferred districts, including those that moved from their elementa ry to th e appropriate high school or unified dis trict. We found that when we did not allow students to transfer to the appropriated unified or high school district at the end of elementary school, the results were similar. We found that coefficients were of similar sizes and were statically significant in the same cases as in models where students could transfer school districts. The models with no transferring allowed did have sli ghtly hig her R - squared values. These results are available from the authors upon request. 49 Full regression results are available in Appendix E . http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 39 FIGURE 7 Share of cohort, by language group, scoring Basic or higher on CST ELA, controlling for district and student characteristics SOURCES: PPIC’s “English Learners Reclassification Survey” and CALPADS. NOTE: Results for language groups are statistically different than those for EO students with the exception of 7th grade RFEP 2008 –09 students. Predicted values for ELS in 11th grade are slightly negative. Full regression results are available in Appendix Table E1 . RFEP students in the 2nd and 4th grade cohorts also outperform EO students when the performance threshold is scoring Basic or higher on the CST ELA (Figure 7). In the 7th grade cohort, only students reclassified prior to and during the target year are more likely than EO students to score Basic or higher. When the performance threshold is higher (Proficient or higher on the CST ELA, Figure 8), RFEP students reclassified prior to the target year are more likely than EO students to perform above the threshold. However, students reclassified after the target year do not, and only 2nd and 4th grade cohort RFEPs reclassified in the target year exceed EO students. Figures 7 and 8 also show that IFEP students are among the best performing students in the state, although there are many instances in which RFEP students achieve better outcomes. RFEP students reclassified prior to the target reclassification year are more likely than IFEP students to achieve a score above the CST ELA performance measures, with the exception of the 7th grade cohort scoring Proficient or higher. Students reclassified in the target year in the 2nd grade cohort are also more likely than IFEP students to exceed the CST ELA performance measures. 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 2nd grade cohort, 6th grade CST 4th grade cohort, 8th grade CST 7th grade cohort, 11th grade CST Percent scoring Basic or above on 2011-2012 CST ELA Cohort and language group RFEP target year (2008-09) RFEP pre RFEP post IFEP EL EO http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 40 FIGURE 8 Share of cohort, by language group, scoring Proficient or higher on CST ELA, controlling for district and student characteristics SOURCE: PPIC’s “English Learners Reclassification Survey” and CALPADS. NOTE: Results for language groups are statistically different than those for EO students. Predicted values f or ELs in 11th grade are slightly negative. Full regression results are available in Appendix Table E2 . We then follow our 7th grade cohort into their high school years. One importa nt outcome for students is on - time progression to 10th grade. Many students with academic deficits spend more than one year classified as ninth graders. Here, we find that RFEP students reclassified in 8th grade or earlier (target reclassification year or pre -target reclassification year) are much more likely to have made on -time (or better) progress to their 10th grade year than ELs (Figure 9). Those reclassified before 8th grade are just as likely, or even slightly more likely, than IFEP and EO students to be achieving on -time progress to 10th gra de. Students reclassified post -target year are more likely than EL students to have achieved on-time grade pr ogress (similar to EO students, but slightly behind IFEP students). 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 2nd grade cohort, 6th grade CST 4th grade cohort, 8th grade CST 7th grade cohort, 11th grade CST Percent scoring Proficient or above on 2011-2012 CST ELA Cohort and language group RFEP target year (2008-09) RFEP pre RFEP post IFEP EL EO http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 41 FIGURE 9 Percent on -time or better in 10th grade, by language group, 7th grade cohort, controlling for district and student characteristics SOURCE: PPIC’s “English Learners Reclassification Survey” and CALPADS. NOTE: Results for ELs, RFEP pre-target year, and IFEPs are statistically significant from EO students. Full regression results are available in Appendix Table E3 . We can follow our final cohort, those who are 8th graders in 2007– 08, through the end of high school (one year beyond what would have been on -time 12th grade completion). We find that even after controlling for individual and district characteristics, RFEP students have much better high school outcomes than EL and EO students (Figure 10). RFEP students are the least likely of all groups to leave high school without graduating. RFEP students, no matter when reclassified are more likely than any other language group, even IFEP students, to have earned a diploma by 2012– 2013. IFEP stude nts are more likely than other language groups, with the exception of students reclassified prior to the target year, to have completed their a – g requirements before graduation. 76% 78% 80% 82% 84% 86% 88% 90% 92% RFEP target year (2008-09) RFEP pre RFEP post IFEPELEO Percent on-time or better, 10th grade Language group, 7th grade cohort http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 42 FIGURE 10 Final high school outcomes for 8th grade cohort, controlling for district and student characteristics SOURCES: PPIC’s “English Learners Reclassification Survey” and CALPADS. NOTE: Results for language groups are statistically different than those for EO students. Full regression results are available in Technical Appendix Tables E4 –E6 . As this section demonstrates, RFEP students achieve much better academic outcomes than EL students, even after controlling for some systematic differences in student and district characteristics. RFEPs who left the EL group at younger grades (generally before 9th grade) often have better outcomes than EO and sometimes than IFEP students. This finding might prompt policymakers to consider if reclassification thresholds might be set too high, a consideration we return to in the conclusion of this report. In the following section, we examine whether these results might be driven by districts that require high performance from students before they reclassify them. Relationsh ips Be tween More Rigorous Reclassifi cation Policies and RFEP Outcomes In this section, we test the relationship between districts’ reclassification criteria and “ever- classified as EL” (i.e., EL and RFEP) student outcomes. Using ordinary least squares multiple regression, we examine a number of key academic outcomes for our EL and RFEP student cohorts: 2011– 12 CST ELA scores, on -time grade progression, leaving high school without graduating, graduating, and graduating having completed a– g requirements. We simultaneously consider the same district - and student -level characteristics used to predict student outcomes in the previous section. We also include the same reclassification policy requirements that 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Leaving HS before graduating DiplomaA-g courses complete Percent with high school outcome RFEP target year (2008-09) RFEP pre RFEP post IFEP EL EO http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 43 we used in our examination of the association between the criteria and district reclassification rates, testing the requirements s hown in the box below. Outcomes for Students Reclassified in Elementary or Middle School We use reclassification policy variables from grades 3 to 5 for the 2nd and 4th grade cohorts, from grades 6 to 8 for the 7th grade cohort, and from grades 9 to 12 for the 8th grade cohort. When simultaneously considering the role of student and district characteristics, we find a statistically significant role of district reclassification policies in predicting EL and RFEP student outcomes. In general, district policies appear to matter somewhat more for the outcomes of younger students, and in the direction we would expect. More of the variables we considered are statistically significant for our 2nd grade cohort, and the association between the reclassification category and the outcome is almost always stronger for the younger students —i.e., the 2nd and 4th grade cohorts versus the 7th grade cohort (see Table 12; full regression results are available in Appendix Tables E7, E8, and E9 ). Not surprisingly, students in districts requiring higher CST ELA scores for reclassification have a greater proportion of students scoring at least Basic (columns 1 –3) or at least Proficient (columns 4 –6) on the CST ELA in 2011 –12 than students in districts that do not. Higher CELDT score requirements are associated with being more likely to score Proficient or higher on the CST outcomes for the cohort of 2nd grade RFEP students and a Basic or higher for the cohort of 7th grade students. If a district’s teacher evaluation reclassification standard requires some use of grades/GPA and/or assessments, RFEP students have higher CST scores than those in districts without teacher evaluation requirements. Increases over the Basic threshold range from 4 to 6 percentage points for the 2nd and 4th grade cohorts and from 0 to 5 percentage points for the 7th grade cohort. Results are similar for the association with teacher evaluation requirements and increases over the Proficient threshold, with 2nd and 4th grade cohorts CELDT Overall Performance Level of Advanced (versus Early Advanced or higher) CELDT subtest (reading, writing, speaking, listening) scores must all be Early Advanced or higher (versus allowing some intermediate scores) CST ELA score of mid-Basic or higher (versus Basic or higher) CST ELA score of Proficient or higher (versus Basic or higher) CST math test versus no CST test requirement beyond CST ELA CST history/science test versus no CST test requirement beyond CST ELA Require grades/GPA and assessments versus just considering grades/GPA and assessments Require grades/GPA versus just considering grades/GPA and assessments Require assessments versus just considering grades/GPA and assessments Consider disciplinary issues versus not considering disciplinary issues CELDT subtest scores must all be Early Advanced or higher AND require CST math test CELDT subtest scores must all be Early Advanced or higher AND require grades/GPA and assessments http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 44 achieving increases around 4 to 6 percentage points. For the 7th grade cohort, effects are slightly negative. If districts consider a student’s disciplinary record, it has a positive effect on CST scores for youngest cohort (2nd grade) but a negative effect for the 7th grade cohort. This mixed finding suggests that considering a student’s lack of disciplinary problems helps students in younger grades but hinders students in older grades. Recall that overall, it is associated with higher, not lower, reclassification rates. http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 45 TABLE 12 Per centage point change for student outcomes and district recl assification rat es associated with specific reclassification requirements Percentage Point Change Attaining Basic CST ELA Attaining Proficient CST ELA 10th grade on time Reclassification rate 2nd grade cohort 4th grade cohort 7th grade cohort 2nd grade cohort 4th grade cohort 7th grade cohort 7th grade cohort District English Proficiency CELDT Advanced 0.002 -0.014 0.067 *** 0.068 *** 0.024 0.010 -0.060 *** -0.029 * CELDT no intermediate 0.007 0.001 -0.181 *** 0.037 *** 0.012 -0.140 *** -0.045 *** -0.028 * Basic Skills CST ELA mid- Basic 0.028 *** 0.049 *** 0.027 *** 0.020 *** 0.041 *** 0.017 ** 0.036 *** -0.040 ** CST Proficient 0.042 *** 0.054 *** 0.011 0.044 *** 0.056 *** 0.030 *** 0.051 *** -0.030 ** CST Math 0.001 -0.008 -0.016 * -0.037 *** -0.058 *** -0.004 -0.013 * 0.009 CST Other -0.005 0.030 ** 0.044 *** 0.037 *** 0.056 *** 0.034 ** 0.010 -0.005 Teacher Evaluation Grades/GPA and Assessments required 0.042 *** 0.042 *** 0.011 0.036 *** 0.037 *** -0.052 *** 0.012 -0.030 * Grades (only) required 0.040 *** 0.042 *** 0.050 *** 0.055 *** 0.046 *** 0.004 0.026 *** -0.015 Assessments (only) required 0.055 *** 0.034 *** 0.038 *** 0.041 *** 0.014 0.009 0.028 *** -0.002 Disciplinary record 0.032 ** 0.022 -0.089 *** 0.007 0.010 -0.063 *** -0.011 * 0.050 ** Combination interaction terms CELDT no intermediate and CST Math 0.039 * 0.040 * 0.198 *** 0.032 * 0.066 *** 0.114 *** 0.061 *** -0.035 * CELDT no intermediate and Grades & Assessment 0.024 * 0.046 * 0.098 *** 0.031 * 0.031 0.122 *** 0.060 *** 0.014 Combination net effects CELDT no intermediate and CST Math 0.047 0.033 0.001 0.031 0.020 -0.030 0.002 -0.054 CELDT no intermediate and Grades & Assessment 0.073 0.088 -0.072 0.103 0.080 -0.070 0.027 -0.044 SOURCES: PPIC’s “English Learners Reclassification Survey” and CALPADS. NOTES: *** p<0.01, ** p< 0.05, * p<0.1, ** p<0.05 and * p <0.10 for district reclassification rates. Full regression results for student outcomes are available in Appendix Tables E7 –E9 and for district reclassification rates in Appendix Table E1 . CST scores are measured in 2011–2012 which is the 6th grade for the 2nd grade cohort, 8th grade for the 4th grade coho rt, and 11th grade for the 7th grade cohort. http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 46 In a district that uses common combinations of more rigorous reclassification standards, CST scores are higher, especially among younger students. For example, in a district that requires all CELDT scores to be at least Early Advanced and that uses the mat h CST, the share of RFEP students from the 2nd grade cohort with scores on the CST ELA that are above the Basic threshold is 4.7 percentage points higher than a district that requires neither. The effect is smaller for older cohorts. Districts using reclas sification policies with these performance thresholds have a 5 percentage point lower reclassification rate. More rigorous reclassification policies do appear linked to greater percentages of our 7th grade RFEP cohort being “on -time” 10th graders. Requirin g higher CST thresholds increases the share of RFEP students being “on -time” 10th graders by 4 to 5 percentage points. However, doing so is associated with a 3 to 4 percentage point lower reclassification rate. Similarly, teacher evaluation standards that are more demanding are associated with more on -time 10th grade RFEPs (3 percentage points more), but with reclassification rate reductions as well, although these are not statistically significant, likely due to small sample sizes in our model (see Appendix Table D1 ). Elevated CELDT score requirements, on their own, are associated with lower shares of RFEPs being on time for 10th grade, but districts requiring both CELDT subtest scores to be Early Advanced or higher and requiring grades/GPA and assessments raises the share of RFEP students being on time in 10th grade by 3 percentage points. Outcomes for Students Reclassified in High School The common combinations of reclassification requirements we tested are statistically significant (Table 1 3). Our estimate of their net effects suggest they make little difference in reducing the share of RFEPs who leave high school without graduating (a subset of the dropouts as defined by CDE) , 50 no r do they make much difference in the share that earn a diploma. However, using the combined criteria is associated with reduced shares of RFEP students completing their a –g requirements. Recall that using these standards is associated with lower district reclassification rates of between 4 and 5 percentage points. More rigorous CST ELA requirements are associated with reduced shares of RFEP students earning diplomas and meeting their a –g requirements and increased shares leaving high school before graduati ng, a counter-intuitive finding. Perhaps remaining in EL status longer in districts with more rigorous reclassification standards means losing access to some academic instruction that would have helped with a –g requirements and high school completion. Howe ver, we do observe somewhat improved “end of high school” outcomes for RFEP students in districts where the history and/or social science CST is required. The teacher evalua tion requirements are linked to higher percent ages of students persisting in high school (i.e., not leaving before graduation) and earning a diploma. There are no statistically s ignificant associations between district’s use of disciplinary issues in reclassification decisions for their high school students and outcomes for those studen ts. Full regression results are available in Appendix Tables E10, E11, and E12 . 50 Recall from Figures 3 and 8 that relatively few RFEP students leave high school without a diploma: Only 100 RFEPs who were reclassified between 2008– 09 and 2012–13 dropped out of school. http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 47 TABLE 13 End of high school o utcomes, cohort 4 High School Outcomes, Percentage Point Change Reclassification Rate, Percentage Point Change Leaving HS before graduating Diploma a –g requirements District English Proficiency CELDT Advanced -0.012 0.059 ** 0.095 *** -0.029 * CELDT no Intermediate 0.028 *** -0.110 *** -0.191 *** -0.028 * Basic Skills CST ELA m id-Basic 0.010 *** -0.012 -0.050 *** -0.040 ** CST Proficient 0.004 -0.049 *** 0.006 -0.030 ** CST Math 0.002 0.021 * -0.026 * 0.009 CST Other -0.021 ** 0.060 *** 0.110 *** -0.005 Teacher Evaluation Grades/GPA and assessments required 0.005 0.029 ** -0.102 *** -0.030 * Grades (only) required -0.015 *** 0.030 *** -0.005 -0.015 Assessments (only) required -0.007 0.021 0.008 -0.002 Disciplinary record 0.008 -0.012 -0.008 0.050 ** Combination CELDT no intermediate and CST Math -0.019 * 0.100 *** 0.194 *** -0.035 * CELDT no intermediate and require Grades/GPA and Assessments -0.038 *** 0.073 *** 0.184 *** 0.014 Combination Net Effects CELDT no intermediate and CST Math 0.011 0.011 -0.023 -0.054 CELDT no intermediate and require Grades /GPA and Assessment -0.005 -0.008 -0.109 -0.044 SOURCES PPIC’s “English Learners Reclassification Survey” and CALPADS. NOTE: *** p<0.01, ** p< 0.05, * p<0.1, ** p<0.05 and * p <0.10 for district reclassification rates. Full regression results for student outcomes are available in Appendix Tables E10 – E12 and for district r eclassification rates in Appendix Table E1 . Looking across all outcomes, we can attempt to generalize about the benefits and costs of different reclassification policies. Requiring a CELDT score of Advanced improves academic outcomes for students in the 7th and 8th grade cohorts, but not for younger cohorts. Not allowing CELDT subtest scores to fall below Early Advanced is associated with better outcomes for the 2nd grade cohort, but not for 7th and 8th grade cohorts. Having higher CST ELA cut -offs improves outcomes for the three youngest cohorts (2n d, 4th, and 7th grade) but does not improve outcomes for the oldest cohort (8th grade). The gains for the younger cohorts are relatively small, especially relative to the reduction in district reclassification rates. For example, http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 48 students reclassified in 5 th grade in districts requiring a score of Proficient or above on the CST ELA are 5 percentage points more likely to score Proficient or above on their 8th grade CSTs than similar students in districts without that requireme nt, translating to 66 percent r ather than 61 percent scoring Proficient or above. If the average annual reclassification rate among districts using criteria suggested in SBE guidelines were 10 percent, districts that add this more rigorous standard would have reclassification rates of 7 percent because the requirement is associated with a 3 percentage point reduction in reclassification rates. Using an additional CST assessment (i.e., history or social science) is associated with better academic outcomes for all students. Recall that we did not find that using this requirement was associated with lower reclassification rates overall, but that this is probably because this requirement is used relatively infrequently (in 9% of elementary, 7% of unified, and 16% of high school districts). Requiring, rather than considering, some form of teacher evaluation criteria to reclassify EL students is linked to improved academic outcomes for all cohorts, with one exception —a–g requirements for the 8th grade cohort. While there are clear associations between more rigorous reclassification policies and outcomes for RFEP students, they often vary by outcome and by grade level. In the cases where more rigorous standards are associated with better RFEP outcomes, it is important to question whether the outcomes are sufficiently improved to justify the much lower district reclassification rates that result. Recall that most districts report using more than one category of more rigorous reclassification policy. As described above, even districts that increase the rigor of just one or two more reclassification requirements can significantly lower their reclassification rate. http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 49 How D o We E valua te t he Trad e-offs Betw een Low er R ecla ssification Rat es but Better O utcom es? We have demonstrated that RFEP students not only outperform EL students, but often do better than EO and IFEP students. RFEP students reclassified at early grades do well for all six years we observe them. RFEP students reclassified at older grades perform well on academic m easures but generally not as well as those reclassified in elementary school. There may be room to lower the reclassi fication standards in districts us ing more rigorous thresholds and still ensure that RFEP students perform well enough in academ ic instruction without additional English language support. However, it should be noted that because districts choose their reclassification policies and because their choice of policies may be somehow related to factors we cannot measure (about their students, their EL instruction, or other factors), we may have under- or over- estimated the true effect o f the policy on the outcomes of interest. Future extensions of this research could explore this issue further. We find that there is wide variation in school districts’ EL reclassification policies, with most districts adopting at least some criteria more rigorous than suggested by the SBE’s 2006 reclassification guidelines. Having different reclassification standards across the state’s districts makes it difficult to compare EL and RFEP outcomes across districts. More troublesome, however, is that we find that districts using more rigorous reclassification standards have lower reclassification rates. Districts using one the most common combinations of requirements (requiring the math CST and requiring that all CELDT subtest scores be Early Advanced or highe r) reduce their reclassification rates by 5.4 percentage points. If a district’s reclassification rate is 10 percent, using that combination of requirements reduces the number of new RFEP students each year by more than half in that district compared to a district that did not use the more rigorous policy. We find RFEP students’ outcomes are better in school districts with more rigorous performance thresholds, but not by much. Is the trade -off between improved RFEP outcomes and fewer RFEP students worth it? It would seem that the costs of these requirements —in terms of lower reclassification rates—are disproportionate to their benefit. Given that we also find that RFEP students usually out -perform EO students (and sometimes IFEP students as well) and general ly perform at very high levels even in districts with more relaxed reclassification standards, we recommend that reclassification standards be lowered in districts using standards more rigorous than those suggested in the SBE guidelines. This report tested the association between more rigorous reclassification standards and reclassified students’ outcomes, but could not test for the possibility that even the thresholds suggested in SBE guidelines might be too rigorous. We recommend that there be additional research to define ideal reclassification standards. If ELs have access to the same academic content as other students (RFEPs, EOs, and IFEPs), then the status itself may not be problematic in terms of a student’s ability to succeed in school (although the re is some evidence that Long-T erm EL students are subject to stigma (Dabach and Callahan, 2011)). But if an EL student is not reclassified because the standards to allow him or her to transition to RFEP status are too rigorous, then it is possible that di stricts with high reclassification thresholds are restricting the access of EL students to the full range of academic instruction that non -EL students receive, which could suppress the academic achievement of high -performing EL students. http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 50 Establishing ideal reclassification standards will not be easy, but we have the time and opportunity to do so. Saunders and Marcelletti (2013) suggest that it is a reasonable goal to expect RFEP students to achieve at least the same level of outcomes as EO students. However, many EO students are not meeting Annual Yearly Progress (AYP) goals or even achieving a Basic score on the state’s standardized ELA tests (Abedi, 2008). In addition, EO students vary in their demographic composition across districts. Robinson (2011) arrives at the conclusion that ideal reclassification thresholds probably depend on high school students’ learning environments after reclassification. Given the current policy environment, this is an ideal time to examine the components of reclassification decisions that are associated with getting the greatest number of EL students reclassified while achieving a minimum agreed- upon standard of success. Two major policy changes will affect EL and RFEP students in the coming years. The first relates to the Common Core State Standards currently being implemented in district curricula. Once the new assessments for the standards are in place, there will no longer be state standardized assessments in ELA for 9th and 10th graders, and districts will need to rethink their practices for reclassifying students during the high school years. We see evidence in the data that even Long -Term English Leaners (LTELs) are being reclassified in high school, and districts certainly reported that to be the case in the survey. One possibility is that districts will increasingly use the CAHSEE instead of the CST ELA for grades 10, and 12, but a passing score on the ELA portion of the CAHSEE is thought to reflect a lower standard than a score of Basic on the CST ELA. Using the CAHSEE would still mean that the 9th grade year has no ELA standardized test, which is currently the type of assessment used by districts to evaluate an EL students’ basic skill level in ELA. In addition, districts will to stop using the CST this year in order to prepare for the Smarter Balanced assessments. A second issue involves the interplay between reclassification and the new Local Control Funding Formula (LCFF). It is important that funding directed to districts for EL students not discourage reclassification. While the LCFF provides funding only for EL students and not for RFEPs, many RFEP students will generate funding for their district based on their poverty status. We found that 79 percent of the ELs in our cohorts of students were low -income students .51 We also found that RFEP students reclassified by 2008– 09 have poverty rates of 63 percent, and those reclassified after 2008– 09 have poverty rates of 77 percent. Still, the instruction or supplemental services that might help low -income RFEPs stay on track could be very different from what low -income EO students need. Another important issue is the flexibility involved in monitoring academic outcomes for RFEP students. Most districts reported that they monitor RFEPs for two years; but during this time of dramatic changes for RFEP students, we suggest that the role for monitoring, as well as adapting services to RFEP students, should be expanded. 51 Our measure of low -income students was whether they participated in the free or reduced price lunch programs, which include income cut-offs of 130 percent and 185 percent of the federal poverty level, respectively. http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 51 References Abedi, J. 2008. “Classification System for English Language Learners: Issues and Recommendations,” Educational Measurement: Issues and Practices, pp. 17 –31. Assembly Bill 2193. Chapter 427, “An act to add Sections 313.1 and 313.2 to the Education Code , relating to English Learners.” Accessed from http://leginfo.legislature.ca.gov/faces/billNavClient.xhtml?bill_id=201120120AB2193 . California Department of Education. 2012, “California English Language Development Test (CELDT), 2012– 2013,” CELDT Information Guide . California Department of Education. 2011. A Comparison Study of Kindergarteners and Grade One English -Fluent Students and English Learners on the 2010– 2011 Edition of the CELDT . California Legislative Analyst’s Office. 2007. Analysis of the 2007– 2008 Budget Bill: Education . California State Auditor. 2005. Department of Education: School Districts’ Inconsistent Identification of Redesignation of English Learners Cause Funding Variances and Make Comparisons of Performance Outcomes Difficult , Bureau of State Audits. Cannon, J. S., and S. Lipscomb. 2011. Early Grade Retention and Student Success: Evidence from Los Angeles, Public Policy Institute of California, San Francisco, CA. Dabach, Dafney B., and Rebecca M. Callahan. 2011. Rights versus Reality: The Gap between Civil Rights and English Learners’ High School Educational Opportunities , Teachers College Record. EdSource. 2008. English Learners in California: What the Numbers Say , Mountain View, CA. Flores, Edward, Gary Painter, and Harry Pachon. 2009. ¿Qué Pasa?: Are English Language Learning Students Remaining in English Learning Classes Too Long? Tomás Rivera Policy Institute Full Report, Los Angeles, CA. Gándara, P., and R. Rumberger. 2006. Resource Needs for California’s English Learners , Linguistic Minority Research Institute, University of California. Hill, L.E. 2012. California’s English Learner Students, Public Policy Institute of California, San Francisco, CA. Hopkins, M., K.D. Thompson, R. Linquanti, K. Hakuta, and D. August. 2012. “Fully Accounting for English Learner Performance: A Key Issue in ESEA Reauthorization,“ Educational Researcher , Vol. 42, No. 2., pp. 101– 108. Linquanti, R. 2001. The Reclassification Dilemma: Challenges and Choices in Fostering Meaningful Accountability for English Learners , UC Language Minority Research Institute Policy Report, University of California, Davis, CA. Mechelli, A., J.T. Crinion, U. Noppeney, J. O’Doherty, J. Ashburner, R.S. Frackowiak, and C.J. Price. 2004. “Structural pasticity in the bilingual brain: proficiency in a second language and age at acquisition affect grey -matter density,” Nature , Vol. 431. Parrish, T., A. Merickel, M. Perez, R. Linquanti, M. Socia, A. Spain, C. Speroni, P. Esra, L. Brock, and D. Delancey. 2006. Effects of the Implementation of Proposition 227 on the education of English Learners, K–12, Findings from a Five -Year Evaluation , American Institutes for Research and WestEd, Palo Alto, CA. PPIC School Finance Model. 2013. Available at www.ppic.org/main/dataSet.asp?i=1229. Robinson, J.P. 2011. “Evaluating Criteria for EL Reclassification: A Casual -Effects Approach Using a Binding -Score Regression Discontinuity Design with Instrumental Variables,” Education Evaluation and Policy Analysis . Rose, H., J. Sonstelie, and M. Weston. 2012. Funding Formulas for California’s Schools III: An Analysis of Governor Brown’s Weighted Pupil Funding Formula, Public Policy Institute of California, San Francisco, CA. Saunders, W. M., and D. J. Marcelletti. 2013. “The Gap That Can’t Go Away: The Catch -22 of Reclassification in Monitoring the Progress of English Learners,” Educational Evaluation and Policy Analysis , Vol. 35, No. 139. Warren, P., and H. Hough. 2013. Increasing the Usefulness of California’s Educational Data, Public Policy Institute of California, San Francisco, CA. http://www.ppic.org /main/home.asp Reclassification of English Learner Students in California 52 About the Authors Laura E. Hill is a research fellow at PPIC, where her research interests include immigrant well -being, English Learners in K –12 schools, race and ethnicity, and youth. She has been a research associate at the SPHERE Institute and a National Institute of Aging postdocto ral fellow. Laura holds a Ph.D. in demography from the University of California, Berkeley. Margaret Weston is a research fellow at PPIC’s Sacramento Center, where she focuses on K –12 school finance. Before joining PPIC, she taught high school English and d rama in Baltimore City Public Schools through Teach For America. She holds a master’s degree in teaching from Johns Hopkins University and a master of public policy degree from the University of Michigan. She is pursuing a Ph.D. in school organization and education policy at the University of California, Davis. Joseph M. Hayes is a research associate at PPIC, where he studies population change and corrections issues. Recent projects have focused on inter- regional migration, estimates of the undocumented imm igrant population, and the changing composition of the prison population. He holds an M.S. in agricultural economics from the University of Wisconsin, Madison. Acknow ledgm ents The authors wish to thank Dave Lesher, Hans Johnson, Patrick Murphy, Athena Brek ke, Abby Cook, Lisa Cole, and Jennifer Paluch at PPIC for their support in developing, producing, and publishing this r eport. W e also thank Caitrin Phillips Chappelle, Sonja Petek, Jan Mayer, Kris Nichols, Karen Bachofer, and Mary Estelle Anderson for their help with the development of the reclas sification survey. We t hank the California Department of Education (CDE), for providing student data. We also wish to thank Robert Linquanti, Leonor Ehling, Ellen Hanak, Lynette Ubois, and Caroline Danielson for comments on earlier versions of the draft and Gary Bjork, Jenny Miyasaki, and Kate Reber for their editorial support. Any errors are our own. PUBLIC POLICY INSTITUTE OF CALIFORNIA Board of Directors Donna Lucas, Chair Chief Executive Officer Lucas Public Affairs Mark Baldassare President and CEO Public Policy Institute of California Ruben Barrales President and CEO GROW Elect María Blanco Vice President, Civic Engagement California Community Foundation Brigitte Bren Attorney Walter B. Hewlett Chair, Board of Directors The William and Flora Hewlett Foundation Phil Isenberg Chair Delta Stewardship Council Mas Masumoto Author and F armer Steven A. Merksamer Senior Partner Nielsen, Merksamer, Parrinello, Gross & Leoni, LLP Kim Polese Chairman ClearStreet, Inc. Thomas C. Sutton Retired Chairman and CEO Pacific Life Insurance Company The Public Policy Institute of California is dedicated to informing and improving public policy in California through independent, objective, nonpartisan research on major economic, social, and political issues. The institute’s goal is to raise public awar eness and to give elected representatives and other decisionmakers a more informed basis for developing policies and programs. The institute’s research focuses on the underlying forces shaping California’s future, cutting across a wide range of public poli cy concerns, including economic development, education, environment and resources, governance, population, public finance, and social and health policy. 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. PPIC was established in 1994 with an endowment from William R. Hewlett. Mark Baldassare is President and Chief Executive Officer of PPIC. Donna Lucas is Chair of the Board of Directors. Short sections of text, not to exceed three paragraphs, may be quoted without written permission provided that full attribution is given to the source. Research publications reflect the views of t he authors and do not necessarily reflect the views of the staff, officers, or Board of Directors of the Public Policy Institute of California. Copyright © 201 4 Public Policy Institute of California All rights reserved. San Francisco, CA PUBLIC POLICY INSTITUTE OF CALIFORNIA 500 Washington Street, Suite 600 San Francisco, California 94111 phone: 415.291.4400 fax: 415.291.4401 www.ppic.org PPIC SACRAMENTO CENT ER Senator Office Building 1121 L Street, Suite 801 Sacramento, California 95814 phone: 916.440.1120 fax: 916.440.1121" ["post_date_gmt"]=> string(19) "2017-05-20 09:41:51" ["comment_status"]=> string(4) "open" ["ping_status"]=> string(6) "closed" ["post_password"]=> string(0) "" ["post_name"]=> string(8) "r_114lhr" ["to_ping"]=> string(0) "" ["pinged"]=> string(0) "" ["post_modified"]=> string(19) "2017-05-20 02:41:51" ["post_modified_gmt"]=> string(19) "2017-05-20 09:41:51" ["post_content_filtered"]=> string(0) "" ["guid"]=> string(50) "http://148.62.4.17/wp-content/uploads/R_114LHR.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) }