The Impact of English Learner Reclassification on High School Reading and Academic Progress

Abstract

This study estimates the causal impact of 8th grade English learner (EL) reclassification on high school English language arts (ELA) standardized test scores, SAT (Scholastic Aptitude Test) reading, and on-track to graduate status. I apply a regression discontinuity design to rich administrative data from a large district in California. The estimated effects of eighth grade reclassification on 9th, 10th, and 11th grade ELA test scores were statistically insignificant, and I can rule out negative effects less than −0.17 SDs and positive effects larger than 0.09 SDs. The negative effect on SAT reading was fairly large, with a confidence interval of −0.27 SDs to 0.05 SDs. Estimates for on-track status in 10th and 11th grades were positive but imprecise.

Keywords

equity immigration/immigrants policy quasiexperimental analysis regression discontinuity

Despite considerable policy emphasis on improving English learners’ (ELs) academic success, substantial school completion gaps persist between ELs and their peers. Only 67% of ELs graduate from high school in 4 years compared with a national average of 84% (ED Data Express, 2019). Although 81% of native English speakers go to college, only 62% of ELs do; moreover, ELs who attend college are concentrated in 2-year institutions and overrepresented in remediation (Callahan & Humphries, 2016; S. N. Flores & Drake, 2014).

Descriptive and qualitative studies have attributed these school completion gaps to differences in the high school curriculum quality experienced by ELs and non-ELs (e.g., Callahan, 2005; Callahan & Shifrer, 2016; Callahan, Wilkinson, & Muller, 2010; Dabach, 2014; Kanno & Kangas, 2014). Recent papers using quasiexperimental methods also show that EL status impacts students’ achievement and middle school course-taking (e.g., Shin, 2018; Umansky, 2016a, 2016b). However, few studies have identified the causal effects of EL status at high school entry on later outcomes. The quality of high school curriculum and services and their ability to meet student needs play a vital role in students’ postsecondary success. Understanding how EL status shapes high school trajectories is essential to designing policies that promote ELs’ college and career readiness. The Every Student Succeeds Act (ESSA, 2015) prioritizes improving college and career readiness for all students, and many states are working to align K–12 curricula with college entrance requirements. A closer look at ELs’ high school credit earning and reading test scores will inform policymakers of ways to expand ELs’ access to and mastery of college preparatory materials.

This study builds on a developing literature on EL reclassification (e.g., Carlson & Knowles, 2016; Cimpian, Thompson, & Makowski, 2017; Reyes & Hwang, 2019; Robinson-Cimpian & Thompson, 2016) by identifying the effects of eighth grade reclassification on key high school outcomes. Eighth grade is a critical juncture at which students are sorted into various high school tracks and trajectories. Unlike elementary and middle school, students in high school tend to be placed into rigid college preparatory, vocational, or general education tracks (Gamoran, 2009). EL status in high school has been shown to be a strong predictor of track and, as a result, access to college opportunities (Callahan, 2005; Callahan & Shifrer, 2016). Yet, to my knowledge, only one study thus far has focused on the impact of EL reclassification at the moment of middle-to-high-school transition, and it does not address credit completion toward graduation (Reyes & Hwang, 2019).

This study addresses the important need for research in this area by answering the following question:

What are the causal effects of EL reclassification in eighth grade on high school ELA state test scores, SAT (Scholastic Aptitude Test) reading scores, and on-track to graduate status?

Leveraging rich administrative data from a large school district in California, I report the effects of eighth grade EL classification on high school English language arts (ELA) and college entrance reading outcomes, as well as on-track to graduate status. This study extends the literature on EL academic development in several ways. First, it presents estimates on outcomes that previous studies did not investigate (10th and 11th grade ELA standardized test scores, SAT reading scores, and 9th to 11th grade on-track to graduate status). Second, this is one of the first studies to focus on middle-to-high-school transition as the point of EL treatment. By examining EL status assigned during eighth grade, this study includes immigrant students who entered U.S. schools in first grade or later, a substantial student population that is overlooked in most previous quantitative EL research. Finally, whereas previous studies (e.g., Robinson-Cimpian & Thompson, 2016; Umansky & Reardon, 2014) used samples composed mostly of Spanish users, the diverse sample in this study includes a large number of users of Chinese and other Asian languages. Evidence presented in this study on the impact of EL status assigned in eighth grade on high school academic progress will provide new insight to inform curricula and policies.

The Effects of EL Status

EL status is not simply an indicator of a student’s true English language proficiency. Rather, EL status is a category constructed and interpreted by education agencies. In many U.S. states and school districts, students who use a language other than English at home are classified as ELs or non-ELs based on performance on an English proficiency test. ELs are eligible to receive language support services, whereas non-ELs are not. As schools administer specialized instructional services to ELs, EL status itself directly impacts the curriculum to which students have access, which in turn affects their academic development and school completion.

Although policies often vary between states and districts, students whose primary home language is not English typically enter and exit EL status through the process of classification and reclassification (Abedi, 2008; Linquanti, Cook, Bailey, & MacDonald, 2016). Those who pass the initial English test upon entering school are classified as Initially Fluent English Proficient (IFEP) and placed in monolingual English classrooms without language support. Those who do not score high enough to pass the initial test are classified as ELs and become eligible to receive language services such as designated English language development (ELD) and sheltered content courses. ELs are then assessed annually until they demonstrate proficiency to reclassify. In some contexts, other factors such as stakeholders’ evaluations and educators’ policy interpretation affect the reclassification decision (Mavrogordato & White, 2017). That is, students who have met the test score thresholds for reclassification stay longer in EL status because of parent, teacher, or district recommendation (Estrada & Wang, 2018). Once reclassified, an EL student exits EL status and is no longer eligible for language services.

The above description generalizes the EL classification, service, and exit process. However, there are exceptions and ongoing developments. In recent years, innovative language and teaching models have introduced variations in the curriculum and services available to IFEPs and ELs. Some schools offer dual-language immersion programs, which enroll IFEPs and native English users alongside ELs. Coteaching models that leverage the expertise of a mainstream subject content teacher and an ELD teacher are also becoming increasingly common (Martin-Beltran & Peercy, 2014).

EL services are intended to help students develop English proficiency and access academic content. Examples of support services at the high school level include designated ELD classes, sheltered academic content courses, and, in some schools, EL coordinators who provide general academic as well as college application support. When implemented well, these services provide a supportive environment in which ELs can develop competencies in English and across academic subjects. However, EL status and the accompanying EL services can also have unintended and potentially harmful effects. A landmark study shows that at the middle school level EL status causes students to be excluded from accessing certain curricula (e.g., math, science) altogether (Umansky, 2016a). These negative effects can be attributed to logistical and social factors. ELD often takes up two or more periods in ELs’ daily schedules, crowding out academic subject courses (e.g., Lillie, Markos, Arias, & Wiley, 2012). The EL label can also harm students through the stigma it creates, causing teachers, counselors, and even the ELs themselves to lower expectations for their academic achievement (Kanno & Kangas, 2014).

Just like entering EL status, exiting, or reclassification, can affect students’ outcomes. Ideally, students would stay in EL status for precisely as long as they need to receive support services. But language proficiency is multidimensional and difficult to fully capture. So the key is not identifying the perfect moment for reclassification, but matching support services to individual students’ needs. Removing a necessary service too soon places EL students into an instructional environment before they are ready, whereas removing services too late may deny these students access to rigorous academic materials and classroom peers with native or fluent English proficiency (Robinson, 2011). Not surprisingly, reclassification has been shown to influence downstream outcomes such as high school graduation and college choice (e.g., Johnson, 2019a; Robinson-Cimpian & Thompson, 2016).

As demonstrated in extant studies, reclassification effects observed in an educational context can be a powerful lever used to modify policy to better suit student needs. For example, Robinson-Cimpian and Thompson (2016) used longitudinal data from Los Angeles Unified School District and found that a negative reclassification effect on graduation disappeared after making reclassification criteria more stringent. As the setting and implementation of reclassification policies vary considerably across districts and states, the effects of those policies are also likely to vary from context to context (Cimpian et al., 2017). Therefore, it is crucial to study the impact of EL reclassification within local educational contexts to inform policy and practice. If we observe that reclassification has a positive effect in a district or state, ELs are harmed by maintaining their EL status and by EL curricula and services. On the contrary, if reclassification has a negative effect, services are being removed too soon while the students still need them.

Two more recent studies have reported the causal link between EL reclassification and late high school outcomes. One used data from a district in California with large Spanish and Chinese speaker populations and found that conditional on attending college, reclassifying in 8th grade decreased the probability of choosing a 4-year college or enrolling full-time (Johnson, 2019a). The other examined students in Wisconsin, the majority of whom were Spanish or Hmong speakers, and found that those who exited EL services in 10th grade had higher ACT (American College Testing) scores than their peers who remained EL (Carlson & Knowles, 2016). These mixed findings also point to the need for further investigation of the effects of reclassification policies across policy contexts and subgroups of EL students.

EL Status and High School Achievement

For students with relatively high English proficiency who score close to the reclassification eligibility threshold, answering one additional question correctly or incorrectly on the annual test could determine whether or not they are classified as ELs for the next academic year and impact their subsequent schooling experiences. Leveraging test score thresholds for EL assignment, two studies have applied the regression discontinuity (RD) design to data from large school districts in California to estimate the impact of initial EL classification upon district entry on ELA achievement. Umansky (2016b) found that students barely classified as ELs in kindergarten had significantly lower ELA test scores in second through 10th grades compared with students who barely passed the initial English proficiency test. In contrast, Shin (2018) found that EL classification in kindergarten increased ELA test scores in the early grades, but the positive effects were weaker in middle and high school.

In many districts that serve large EL populations, the EL students in middle and high school differ from those in early grades. Typically, a large proportion of kindergarten ELs have reclassified by the end of elementary school and are eligible for the same types of academic services as monolingual and native English users (Shin, 2018). Thus, ELs in high school are either (a) immigrants who arrived in the United States during late elementary or middle school or (b) long-term ELs who have maintained EL status for 5 or more years, often because of low scores on reading and writing sections of the annual assessments. These two groups differ in important ways.

Immigrant students who arrive in late elementary grades or middle school bring to U.S. schools a wide array of cultural, linguistic, and academic experiences (Umansky et al., 2018). Many are literate in their home language. Some even have fluent or high English proficiency. Others who have experienced interrupted formal education may not be literate in any language and, as a result, struggle to acquire academic content in other subjects (DeCapua & Marshall, 2010; Olsen & Jaramillo, 1999). During their first 3 years in the United States, immigrant students are eligible for participation in the Title III Immigrant Student Subgrant Program (ESSA, 2015).

Long-term ELs, by definition, have a history of struggling with English language and literacy. Many states and districts assign long-term status to students who have not been reclassified after 5 or more years, others use 6 or 7 years as the threshold (N. Flores, Kleyn, & Menken, 2015; Freeman, Freeman, & Mercuri, 2002; Menken & Kleyn, 2009; Olsen, 2010). Educators often express concern about long-term ELs’ stagnant academic English proficiency and low grades and test scores in other academic subjects (Callahan & Shifrer, 2016; Olsen, 2014). Research points out that the problem is with the system and not the students. Many ELs end up in long-term status not because of their own effort or ability but because schools fail to provide them with appropriate support (Olsen, 2010). Students who maintain EL status into secondary school tend to be perceived by teachers and counselors not as bilingual but as semilingual, and, as a result, are not given access to rigorous academic courses and materials (Brooks, 2017; Kim & Garcia, 2014; Thompson, 2015). In addition, a disproportionately large fraction (e.g., 17.4% in my sample) of long-term ELs are identified as having a disability (Umansky, Thompson, & Diaz, 2017). The need to prevent ELs from entering long-term status and to provide suitable accommodation to dually identified ELs has been brought to the forefront of education policy. Under ESSA (2015), schools and districts are now required to disaggregate EL data and report long-term EL performance separately from other subgroups.

In this article, I use the term long-term ELs to refer to students who entered U.S. schools in early elementary school grades and have not reclassified prior to high school. My intention is to highlight the urgency to identify and address the needs of this subgroup, not to reinforce the label itself.

Eighth Grade Reclassification

Given the distinct needs of more recent immigrants and long-term ELs, using only kindergarten test scores to estimate the effects of EL treatment on their high school outcomes can be problematic. Kindergarten tests reflect the English proficiency of neither group as they enter high school. Immigrant students are not tested in kindergarten because they arrive in the United States in later grades. For long-term ELs, 9 years would have elapsed between the kindergarten test and high school. Kindergarten test scores do not inform well on the language proficiencies of students entering high school. In addition, the English proficiency exam given to kindergarten students in some states does not test the academic reading and writing domains. Thus, kindergarten test scores do not provide useful information about English proficiency that would be relevant to high school performance, which is mostly assessed through reading and writing. Scores and EL status assigned around the time of high school entry would be more likely to reflect long-term ELs’ English proficiency, especially reading and writing.

We want to know how ELs are doing academically compared with non-ELs. Using kindergarten test scores to compare the high school outcomes of students initially classified as ELs and students never classified as ELs only gives us half the picture. The other half requires a comparison of students who have exited EL status before high school with those who remain. By examining students at the cusp of reclassification in eighth grade, I compare students at the same starting line at the moment of high school entrance. Any subsequent differences in outcomes can be attributed to how well high school programs and services match students’ needs.

Recent studies on high school curriculum access also show the need to focus on the 8th grade juncture. As shown by a mature body of course-taking research, 8th grade achievement explains a large part of high school curriculum access (e.g., Conger, Long, & Iatarola, 2009). Recent descriptive research demonstrates that course-taking and credit completion gaps between EL subgroups (e.g., newcomers, long-term ELs) and between ELs and non-ELs develop in the first two semesters of high school and grow larger in later years (Johnson, 2019b). Studies focusing on kindergarten entry or 10th grade reclassification (e.g., Carlson & Knowles, 2016) are unable to account for this pattern as newcomers would not have been assessed in kindergarten and course-taking gaps develop before 10th grade.

To my best knowledge, only one paper has focused on the effects of reclassification at the moment of middle-to-high-school transition (Reyes & Hwang, 2019). Using data from a district in Southern California with eight middle schools, Reyes and Hwang (2019) examined the effects of reclassification in seventh or eighth grade on ninth grade ELA standardized test scores, high school exit exam scores, math course placement, attendance, and suspensions. They found that although reclassification is associated with better academic and behavioral outcomes in high school, most causal estimates were small and nonsignificant. These results highlight the need for robust evidence derived from experimental and quasiexperimental research designs. The associations estimated using descriptive approaches may not reflect the true impact of EL reclassification, as they are prone to confounding reclassification with other factors that influence students’ downstream outcomes.

Contributions of this Study

This study presents new evidence on EL academic development by examining reclassification at middle-to-high-school transition, a critical period that has largely been overlooked by previous causal studies. Leveraging rich administrative data from a large district in California, I estimate the impact of EL reclassification based on 8th grade annual assessments on 9th to 11th grade ELA state test scores, SAT reading scores, and on-track to graduate status. To identify these causal effects, I employ a binding RD design, using centered standardized eighth grade English proficiency and ELA state test scores as the running variable.

This study is similar to Reyes and Hwang (2019) in a few ways. Both use an RD research design; both samples include a large number of Hispanic and Asian students in California; and the outcomes of interest are proxies for high school performance and include 9th grade ELA state test scores. Extending from Reyes and Hwang (2019), I also examine 10th and 11th grade ELA test scores, as well as SAT reading scores, to identify reclassification effects on academic English reading over time. This study also builds on the analysis of math placement by Reyes and Hwang (2019) by looking at on-track to graduate status in 9th, 10th, and 11th grades. On-track status is determined by comparing students’ actual credit earning overall, as well as in specific subjects, to expected credit earning toward state college entrance requirements. This provides a more complete picture of high school academic progress than looking at placement in a single subject. The policy context in this study also differs from the anonymous district in Reyes and Hwang (2019). The district in this study is much larger, with 21 schools serving 6th to 8th grades and 17 high schools. It also has one of the longest and richest histories among districts in the nation of serving a diverse population of immigrants. Currently, more than 20% of students enrolled in the district are ELs. In response to the wide array of assets and needs that students bring, the district offers a variety of language programs tailored to EL subgroups at the middle and high school levels. The vast majority of high school ELs take designated ELD classes and sheltered academic subject courses, where content is made accessible through modified language. Because of these similarities and differences, the findings from this study offer an interesting comparison to Reyes and Hwang (2019) and earlier RD studies.

This article also explores heterogeneity in the effects of EL reclassification by home language and time spent as an EL prior to high school. Subgroups of ELs reclassify at very different rates (Thompson, 2017; Umansky & Reardon, 2014), so we might expect them to respond to reclassification in different ways as well. For this reason, I look at the differential effect of eighth grade reclassification on Spanish and Chinese speakers and on long-term ELs and more recent immigrants. These interaction analyses were unfortunately underpowered due to the small sample size. I report the results in Supplemental Tables A3 to A8 in the online version of the journal.

Data

My sample includes a total of 5,211 students and is composed of students previously classified as ELs between kindergarten and seventh grade who took the English proficiency and ELA state standardized tests in eighth grade. Eighth grade test scores are from academic years 2004–2005 to 2014–2015 and available for students who were expected to graduate from high school between 2009 and 2017. Table 1, Panel A presents the summary statistics for the sample. About 42% of the sample were female and approximately 42% Spanish users. Another 38% of students reported varieties of Chinese (e.g., Mandarin, Cantonese) as their home language.

Table 1

Sample Summary Statistics and Outcome Availability

Panel A: Summary Statistics

	ELs (N = 4,323)	Non-ELs (N = 888)	Pooled (N = 5,211)
Covariate means
Female	0.415	0.457	0.422
Hispanic	0.455	0.361	0.439
Chinese (language)	0.376	0.411	0.382
Spanish (language)	0.430	0.340	0.415
SPED	0.178	0.082	0.161
English test age (years)	13.619	13.503	13.600
Eighth grade unexcused absences (classes)	48.557	27.709	45.062
Eighth grade days suspended	0.367	0.215	0.341
Number of students by high school graduation cohort
2009	555	109	664
2010	489	95	584
2011	550	100	650
2012	539	136	675
2013	550	98	648
2014	480	120	600
2015	411	80	491
2016	402	102	504
2017	347	48	395
Number of students by grade of initial EL classification
K	389	92	481
1	91	30	121
2	111	28	139
3	180	47	227
4	315	66	381
5	395	78	473
6	619	73	692
7	2,223	474	2,697

Note. ELs refer to students who scored below the cut score on the eighth grade test. Non-ELs refer to students who scored above the cut score on the eighth grade test. English test age refers to students’ age on the day of the eighth grade English proficiency test. EL = English learner; SPED = special education.

Panel B: Outcome Availability

Outcome	EL mean	EL No.	Non-EL mean	Non-EL No.	Pooled mean	Pooled No.
Test scores (std)
9th grade ELA	−0.855	3,703	−0.032	790	−0.710	4,493
10th grade ELA	−0.746	3,032	0.043	675	−0.602	3,707
11th grade ELA	−0.673	2,823	0.086	658	−0.530	3,481
SAT reading	−1.423	1,322	−0.738	427	−1.256	1,749
On-track to graduate
9th grade	0.477	4,029	0.643	815	0.505	4,844
10th grade	0.391	3,942	0.597	796	0.426	4,738
11th grade	0.380	3,811	0.624	766	0.421	4,577

Note. ELs refer to students who scored below the cut score on the eighth grade test. Non-ELs refer to students who scored above the cut score on the eighth grade test. EL = English learner; ELA = English language arts; SAT = Scholastic Aptitude Test.

Table 2 shows the background characteristics, eighth grade test scores, and outcomes by students’ home language and grade of initial EL classification subgroup. Compared with Chinese students, a much larger fraction of Spanish students were also identified for special education. Spanish users also had lower test scores and on-track to graduate probabilities relative to Chinese users. However, as mentioned earlier, these differences should not be attributed to home language alone because they may reflect socioeconomic status and other unobserved factors. Compared with students initially classified as ELs between fourth and seventh grades (“4–7 entry”), K–3 entry students had a higher average eighth grade binding score but lower standardized ELA test scores and lower probabilities to be on track in 9th to 11th grades.

Table 2

Covariate and Outcome Means by Home Language and Grade-of-Entry Subgroup

	Spanish (N = 2,162)	Chinese (N = 2,052)	K–3 entry (N = 968)	Grade 4–7 entry (N = 4,243)
Covariate means
Female	0.434	0.424	0.412	0.424
SPED	0.206	0.116	0.174	0.159
English test age (years)	13.608	13.608	13.577	13.605
8th grade unexcused absences (classes)	69.158	21.271	51.402	43.643
8th grade days suspended	0.531	0.143	0.392	0.330
8th grade binding scores
Binding score	−0.810	−0.787	−0.528	−0.819
Domain that became binding
CELDT overall (percentage of subgroup)	19.75	41.58	14.26	32.97
ELA (percentage of subgroup)	70.58	49.87	77.17	57.91
Test scores (std)
9th grade ELA	−0.867	−0.580	−0.802	−0.693
10th grade ELA	−0.842	−0.399	−0.673	−0.591
11th grade ELA	−0.809	−0.327	−0.610	−0.516
SAT reading	−1.367	−1.242	−1.238	−1.259
On-track to graduate
9th grade	0.353	0.642	0.434	0.520
10th grade	0.241	0.592	0.365	0.439
11th grade	0.223	0.604	0.407	0.424

Note. Spanish and Chinese refer to students’ reported home language. K–3 entry refers to students who entered the district between kindergarten and third grade. Grade 4–7 entry refers to students who entered the district between fourth and seventh grades. Binding score is the lowest centered standardized score among overall, listening, speaking, reading, and writing sections of the CELDT, and state standardized ELA test. On-track to graduate status is calculated based on overall and by-subject credit accumulation at the end of each grade level. SPED = special education; CELDT = California English Language Development Test; ELA = English language arts; SAT = Scholastic Aptitude Test.

Test Scores Used for EL Reclassification

The sample of eighth grade students includes all students in the district who took the annual English and ELA exams to determine eligibility for reclassification, or exit from EL services. The English proficiency test administered during the data time frame was the California English Language Development Test (CELDT). Students who scored “early advanced” or higher overall and “intermediate” in all constructs (listening, speaking, reading, writing) met the CELDT criteria for reclassification. In addition, students had to get a scale score of mid-basic level on the ELA standardized test. To make scores comparable across tests and cohorts, I center and standardize all section scores using the standard deviation (SD) and level cut scores for eighth grade students in each year. A standardized and centered score of 0 or higher indicates that the student passed a section of the test. Thus, the lowest section score becomes the binding score for each student. A negative binding score indicates that the student did not meet the required score for at least one section, did not become eligible for reclassification, and should remain EL for the following year.

Outcomes

My main outcomes of interest are ELA state test scores, SAT reading scores, and on-track to graduate status. These are important indicators for high school academic progress and readiness for college-level English reading. The availability of each outcome is presented in Table 1, Panel B. I test the balance of outcome missingness across the EL reclassification threshold because missing test scores or on-track status in later grades may be an indication of early departure from school.

I use standardized ELA state test scores for measures of academic English and reading. The availability of ELA test score data varies by cohort. The State of California ended the administration of the California Standards Test (CST) in 2012–2013 and transitioned to Smarter Balanced starting in 2014–2015. Each cohort was tested only once or twice during 9th to 11th grades. To make test scores across years and across the two different assessments comparable, I center and standardize the scores using cut scores and state SDs for each test year and grade combination. This way, I am able to capture students’ performance relative to their grade-level peers across years and assessments.

SAT critical reading scores (possible score range 200–800) are available for students who took the SAT test between 2008 and 2017. If a student took the same test more than once, I use their first score because later test-taking and scores may be endogenous. I standardize scores using the mean (500) and SD (100).

On-track to graduate status is an indicator summarizing students’ credit accumulation in high school. Since the 1990s, states and districts have developed and utilized on-track indicators to monitor students’ progress through high school. Extant research shows that on-track status provides valuable information that supplements standardized test scores and grade point averages (e.g., Allensworth & Easton, 2007). More specifically, being on track in ninth grade strongly predicts on-time graduation (e.g., Hartman, Wilkins, Gregory, Gould, & D’Souza, 2011; Norbury et al., 2012; Stuit et al., 2016). The formulae used to calculate on-track status have varied by context. Many include a credit count component and a course grade component (e.g., Allensworth & Easton, 2007); research also suggests taking into account proxies for student engagement, such as attendance and behavior (Balfanz & Fox, 2011).

The district in this study calculates on-track status at the end of each semester and evaluates credit accumulation in terms of both overall credits and subject-specific credits. For instance, at the end of ninth grade, students who earned credit toward one year-long math class, one year-long ELA class, and a total of five year-long courses are considered on track. If the student failed to complete either math or ELA, or passed fewer than five year-long courses, she is considered off track. The indicator is particularly informative in this context because the district’s graduation requirements are aligned with the 4-year college entrance requirements in the State of California. In other words, being on track to graduate is a proxy of being on track for college readiness. The on-track indicator is not available for the whole sample in each grade level due to enrollment stoppage and transfers outside of the district. ELs and other immigrant students have high rates of migration. As a result, a fraction of the sample is missing on-track data for each grade level.

Research Design

Students who reclassify and students who stay ELs may differ systematically in terms of English proficiency, prior schooling, and other factors that contribute to academic outcomes. Thus, directly comparing the mean outcomes of EL and Reclassified Fluent English Proficient (RFEP) students may lead to biased estimates for the impact of EL status. To address this concern, I use an RD design with the eighth grade binding test score as the running variable. The RD design mimics a randomized experiment by restricting the analysis to students with scores within a narrow range of the reclassification threshold, relying on the assumption that the potential outcomes are a continuous function of the test scores (Lee & Lemieux, 2010). Test scores have two components: a student’s true proficiency and random noise created by idiosyncratic elements such as the weather or the student’s health and energy level. Students who score just above and just below the classification threshold may have the same true English proficiency yet be classified as EL or RFEP due to some random condition. An experimental research design would have randomly assigned EL or RFEP status to students, and their high school outcomes would be compared to determine the effect of EL status. The RD design leverages the test-based status assignment mechanism such that students are as good as randomly assigned just around the preset cut score that separates ELs from RFEPs.

I use each student’s lowest centered score to form a binding RD (Papay, Willett, & Murnane, 2011; Porter, Reardon, Unlu, Bloom, & Cimpian, 2017; Reardon & Robinson, 2012). Table 2 reports the binding score itself and the domain score that became the binding score for student subgroups by home language and grade of initial EL classification. The binding RD approach is preferable to the frontier approach in this context because using the binding score allows for estimation without discarding all observations below the threshold in one or more domains. As the estimates obtained with this approach may be sensitive to scaling, I standardize the scores to minimize potential bias.

Model

I use the following general model to estimate the effect of reclassification on high school outcomes:

Y_{i} = α + β_{0} {Above}_{i} + f (S_{i}) + X_{i} + T_{y} + ε_{i},

(1)

where ${Above}_{i}$ is an indicator for scoring above the cut score for student $i$ , $β_{0}$ is the coefficient of interest which estimates the impact of reclassification, and $f (S_{i})$ is a function used to flexibly model $S_{i},$ the centered score. $Y_{i}$ represents each outcome. I also include test year fixed effects ( $T_{y}$ ) and student demographics $(X_{i})$ in the model. Heteroskedasticity robust Huber–White standard errors are reported.

There is imperfect compliance to EL status assigned based on test scores because other factors, such as teacher and parent recommendations, were considered in the reclassification decision. Some students who scored below at least one score threshold were reclassified, and some students who scored above all the thresholds stayed as ELs. The first-stage estimates show that scoring below the cut score in eighth grade did significantly induced maintaining EL status (see Supplemental Table A1 and Figure A2 in the online version of the journal for estimates and graphs). In the rest of this article, I report the reduced-form effects of reclassification at the end of eighth grade, which are analogous to intent-to-treat effects in an experiment.

Effect Heterogeneity

To see if the effects of eighth grade EL reclassification differed by the time spent as an EL, I divide my sample into two EL subgroups: kindergarten to third grade entry (long-term ELs) and fourth to seventh grade entry (more recent immigrants). I then test for effect heterogeneity by interacting reclassification with K–3 entry in the following model:

\begin{array}{l} Y_{i} = α + β_{0} {Above}_{i} + β_{1} {Above}_{i} \times K 3_{i} \\ + β_{2} K 3_{i} + f (S_{i}) + X_{i} + T_{y} + ε_{i}, \end{array}

(2)

where $β_{0}$ is the main effect of reclassification on the omitted category, or Grade 4–7 entry ELs. $β_{1}$ is the coefficient of interest and the differential effect on K–3 entry ELs. Similarly, I interact reclassification with students’ home language to see if effects differ for users of Spanish, Chinese, and other languages. Using Spanish as the omitted category, I estimate the differential effect of reclassification on users of Chinese and users of other languages.

The validity of the RD design depends on the assumption that students just above and just below the score threshold do not differ systematically in unobservable characteristics that could influence both assignment and outcomes. I test this assumption by checking for score density balance and observed covariate balance around the cut score (McCrary, 2008). Test score densities and observed student covariates are balanced (see Supplemental Table A2 and Figure A1 in the online version of the journal). This suggests that test scores have not been precisely manipulated and that students on either side of the cut score are likely to be similar in unobserved characteristics.

Comparing estimates from several model specifications, I choose as my preferred specification a model with linear splines, which fits the data better than higher order polynomials according to likelihood ratio tests. I use the Calonico, Cattaneo, and Titiunik (2014) bandwidth selection procedure to calculate the optimal bandwidth. The procedure yielded 0.50 SDs around the cut scores. I report estimates obtained using the optimal bandwidth, as well as a few other larger and smaller bandwidths, including the full sample.

I also perform two additional tests for the robustness of my results. First, I rerun my models using a false threshold of −0.25 SDs below the true threshold. Second, I assign off-track status to 10th and 11th grade for students who had 9th grade on-track data but were missing 10th or 11th grade on-track data. Results are reported in Supplemental Tables A9 and A10 in the online version of the journal.

Main Findings

Test Participation and Attrition

Table 3 reports the balance of missing test scores or on-track status. Outcome missingness, especially in later grades, is a proxy for early departure from high school. At the reclassification threshold, outcome missingness was balanced with one exception. For 9th grade ELA, the estimate on missingness is negative and significant, which means that reclassification in eighth grade had a positive reduced-form effect of 4.8 percentage points on taking the test. Because reclassified students were more likely to take the test, which might affect the test score distribution, I interpret estimates on 9th grade ELA score as suggestive evidence. For 10th and 11th grade ELA and SAT reading test-taking, as well as having 10th and 11th grade on-track status, the small positive effects were imprecisely estimated. There is insufficient evidence for differences in test participation or persistence in these later grades.

Table 3

Missingness of Test Scores and On-Track Status at Reclassification Threshold, Linear Model (Optimal Bandwidth = 0.50 SDs)

Outcomes	(1)	(2)	(3)	(4)
Outcomes	9th grade ELA	10th grade ELA	11th grade ELA	SAT reading
Binding score above threshold	0.048**	0.005	0.043	0.065
	(0.019)	(0.025)	(0.033)	(0.044)
Adj. R²	.547	.556	.274	.031
Observations	1,975	1,975	1,975	1,975
Outcomes	(5)	(6)	(7)
Outcomes	9th grade on-track	10th grade on-track	11th grade on-track
Binding score above threshold	−0.003	0.020	0.039
	(0.023)	(0.026)	(0.030)
Adj. R²	−.003	−.001	.001
Observations	1,975	1,975	1,975

Note. Robust standard errors are given in parentheses. Models do not include student covariates. The dependent variable is the probability of missing each test score or on-track outcome. ELA = English language arts; SAT = Scholastic Aptitude Test.

p < .1. **p < .05. ***p < .01.

Test Scores

The estimated effects of EL reclassification based on eighth grade test scores are shown in Table 4. Discontinuity graphs for the linear specification are shown in Figures 1 and 2, and graphs for the quadratic specification are shown in Supplemental Figures A3 and A4 in the online version of the journal. The reduced-form estimates on ELA test scores, obtained using the linear model and the optimal bandwidth, were small and not significant (Table 4, Column 1). Reclassified ELs were 4.8 percentage points more likely to be missing 9th grade ELA test scores, so I take the estimated reclassification effect of −0.013 SDs to be suggestive. For 10th and 11th grades, the estimated effects were 0.013 SDs and −0.041 SDs, which were not significantly different from zero. For SAT reading, the estimate was a moderate −0.112 SDs but imprecisely estimated. The quadratic model yielded small and statistically nonsignificant effects (Table 4, Column 2). As shown in Table 5, estimates from bandwidths between 0.25 SDs and 1.50 SDs were qualitatively similar. For the full sample, which includes students who scored far away from the reclassification threshold, effects on all ELA test scores were positive and significant.

Table 4

Effect of Eighth Grade EL Reclassification on Test Scores and On-Track to Graduate Status (Optimal Bandwidth = 0.50 SDs)

	(1)	(2)
	Linear	Quadratic
RQ 1: test scores
9th grade ELA
Reclassification	−0.013	0.037
	(0.050)	(0.073)
Adj. R²	.222	.223
Observations	1,748	1,748
EL mean	−0.492	−0.492
10th grade ELA
Reclassification	0.013	−0.031
	(0.063)	(0.090)
Adj. R²	.200	.200
Observations	1,461	1,461
EL mean	−0.418	−0.418
11th grade ELA
Reclassification	−0.041	0.018
	(0.065)	(0.092)
Adj. R²	.215	.214
Observations	1,411	1,411
EL mean	−0.382	−0.382
SAT reading
Reclassification	−0.112	−0.038
	(0.080)	(0.119)
Adj. R²	.199	.198
Observations	831	831
EL mean	−1.208	−1.208
RQ 2: on-track to graduate
9th grade
Reclassification	−0.041	−0.080
	(0.044)	(0.066)
Adj. R²	.100	.100
Observations	1,831	1,831
EL mean	0.540	0.540
10th grade
Reclassification	0.015	0.010
	(0.043)	(0.064)
Adj. R²	.166	.168
Observations	1,786	1,786
EL mean	0.448	0.448
11th grade
Reclassification	0.047	0.050
	(0.044)	(0.065)
Adj. R²	.184	.184
Observations	1,736	1,736
EL mean	0.478	0.478

Note. Robust standard errors are given in parentheses. Model includes test year fixed effects and student demographics. EL refers to students scoring below the cut score. EL = English learner; RQ: Research Question; ELA = English language arts; SAT = Scholastic Aptitude Test.

p < .1. **p < .05. ***p < .01.

Table 5

Effect of Eighth Grade EL Reclassification, Other Bandwidths

	(1)	(2)	(3)	(4)	(5)	(6)
	bw = 0.25 SDs	bw = 0.75 SDs	bw = 1.00 SDs	bw = 1.50 SDs	bw = 2.00 SDs	Full sample
RQ 1: test scores
9th grade ELA
Reclassification	0.040	−0.012	−0.018	0.020	0.054*	0.201***
	(0.069)	(0.042)	(0.038)	(0.033)	(0.032)	(0.031)
Adj. R²	.169	.305	.355	.415	.435	.410
Observations	926	2,497	3,067	3,825	4,171	4,493
EL mean	−0.403	−0.596	−0.669	−0.772	−0.817	−0.855
10th grade ELA
Reclassification	−0.017	−0.010	−0.035	−0.022	0.015	0.155***
	(0.088)	(0.052)	(0.048)	(0.041)	(0.039)	(0.038)
Adj. R²	.172	.265	.309	.378	.399	.370
Observations	773	2,095	2,575	3,179	3,456	3,707
EL mean	−0.349	−0.517	−0.584	−0.680	−0.718	−0.746
11th grade ELA
Reclassification	−0.008	0.017	0.040	0.035	0.067	0.161***
	(0.091)	(0.054)	(0.049)	(0.043)	(0.042)	(0.040)
Adj. R²	.148	.251	.276	.331	.340	.331
Observations	756	1,986	2,427	2,972	3,226	3,481
EL mean	−0.283	−0.469	−0.529	−0.604	−0.633	−0.673
SAT reading
Reclassification	−0.018	−0.117*	−0.086	−0.078	−0.053	0.057
	(0.111)	(0.069)	(0.063)	(0.055)	(0.052)	(0.051)
Adj. R²	.123	.257	.290	.351	.363	.336
Observations	444	1,111	1,312	1,550	1,658	1,749
EL mean	−1.116	−1.290	−1.323	−1.395	−1.416	−1.423
RQ 2: on-track to graduate
9th grade
Reclassification	−0.029	−0.013	−0.009	0.005	0.014	0.021
	(0.060)	(0.037)	(0.033)	(0.029)	(0.028)	(0.027)
Adj. R²	.100	.103	.115	.120	.115	.100
Observations	973	2,608	3,229	4,054	4,444	4,844
EL mean	0.579	0.514	0.506	0.493	0.489	0.477
10th grade
Reclassification	0.019	0.039	0.051	0.072**	0.069**	0.077***
	(0.059)	(0.037)	(0.033)	(0.029)	(0.028)	(0.027)
Adj. R²	.163	.168	.178	.180	.174	.156
Observations	944	2,549	3,155	3,971	4,351	4,738
EL mean	0.488	0.422	0.412	0.401	0.398	0.391
11th grade
Reclassification	0.057	0.037	0.038	0.064**	0.074***	0.104***
	(0.059)	(0.037)	(0.034)	(0.030)	(0.028)	(0.027)
Adj. R²	.178	.193	.210	.217	.219	.207
Observations	918	2,481	3,063	3,845	4,200	4,577
EL mean	0.514	0.440	0.418	0.397	0.388	0.380

Note. Robust standard errors are given in parentheses. Model has linear splines and includes test year fixed effects and student demographics. EL refers to students scoring below the cut score. EL = English learner; bw = bandwidth; RQ: Research Question; ELA = English language arts; SAT = Scholastic Aptitude Test.

p < .1. **p < .05. ***p < .01.

On-Track to Graduate Status

As shown in Table 4, Column 1, the estimated reclassification effects on on-track to graduate status were negative for 9th grade (–4.1 percentage points) and positive for 10th and 11th grades (1.5 and 4.7 percentage points, respectively). Quadratic estimates are in the same direction as linear estimates (Table 4, Column 2). None of the estimates are statistically significant at bandwidths of 1.00 SD or smaller. At larger bandwidths, the effects on 10th and 11th grade on-track status were positive and significant (Table 5, Columns 4–6). Effects estimated using a wide range of larger and smaller bandwidths are within the confidence intervals of the optimal bandwidth estimates (see Supplemental Figures A5 and A6 in the online version of the journal).

Discussion

This study reports two notable findings. First, 8th grade reclassification had a small and statistically nonsignificant effect on 9th, 10th, and 11th grade ELA test scores, and the negative effect on SAT reading scores was imprecisely estimated. Second, there is suggestive evidence that reclassification improved the probability of being on track to graduate in 10th and 11th grades. The effects of reclassification did not differ significantly by the time spent as ELs before high school or home language group, but positive estimates suggest that long-term ELs benefit differentially from reclassification.

Some of my coefficients were imprecisely estimated, but the magnitude of estimates and standard errors were on par with results from previous RD reclassification studies (e.g., Carlson & Knowles, 2016; Reyes & Hwang, 2019; Robinson-Cimpian & Thompson, 2016). For instance, Robinson-Cimpian and Thompson (2016) estimated the post-2007 intent-to-treat effect of reclassification in third to eighth grades on year-after ELA test scores to be 0.025 SDs, with a standard error of 0.054. My estimate for eighth grade reclassification on ninth grade (year-after) ELA test score is −0.013 SDs, with a comparable standard error of 0.050. For ELA test scores, my confidence intervals allow me to rule out effect sizes that are less than −0.17 SDs and more than 0.09 SDs. As these confidence intervals are fairly wide, I would caution against concluding from this study that reclassification has no effect. Instead, I emphasize that my findings ought to be interpreted in tandem with recent causal research on EL reclassification, for this study extends quasiexperimental research on EL college and career preparation in several ways.

First, this study complements and contrasts with earlier RD papers. Using a sample from Wisconsin, Carlson and Knowles (2016) found that reclassification in 10th grade led to higher ACT scores and provided suggestive evidence for increases in college enrollment. This study adds to the literature by showing that reclassification at the eighth grade juncture had little effect on ELA test scores. My estimates for 11th grade ELA test scores and SAT reading are in the opposite direction as estimates on ACT scores from Carlson and Knowles (2016). In fact, my negative estimate was fairly large, with a confidence interval between −0.27 SDs and 0.05 SDs. I can thus rule out positive effects larger than 0.05 SDs, compared with their estimate of positive 0.96 points on ACT reading, equivalent to about 0.18 SDs. This contrast is not surprising because my sample is drawn from a district in California with an extensive history of serving diverse EL populations. Its EL policies and services likely differ from the State of Wisconsin, a newer destination for immigrants. This goes to support the conclusion from Cimpian et al. (2017), which states that reclassification should not be expected to have universally positive or negative effects. Instead, reclassification policies and implementations should be interrogated with consideration of the local context and specific practices.

This study uses a sample from the same district as Umansky (2016b), which found that initial EL classification upon kindergarten entry has a significant negative impact on both short- and long-run (2nd through 10th grade) ELA test scores. This study complements Umansky (2016b) by showing that for students previously classified as an EL, 8th grade reclassification had very small and nonsignificant effects on short-run (9th and 10th grade) ELA state test scores. In other words, for students initially classified as ELs and remained ELs until 8th grade, whether or not they exit EL status makes little difference to ELA test scores in 9th or 10th grade. When interpreted alongside Umansky’s (2016b) results, it is possible that ELs, as a group, fall behind never-ELs during elementary and middle school, but 8th grade reclassification makes little difference within the subgroup who maintain EL status to that juncture.

The group of ELs in this study is very different from those who are initially classified in kindergarten or 1st grade and reclassify before middle school. The composition of my sample makes the results from this article comparable to those of Reyes and Hwang (2019). Both studies found no reclassification effect on 9th grade ELA test scores. Reyes and Hwang (2019) found the effect of reclassification on the ELA section of the high school exit exam (taken in 10th grade) to be negligible; I found the same to be true for 10th grade CST-ELA. In addition, their estimates on math course placement and mine on on-track to graduate status during 9th to 11th grades were not statistically significant.

The provision of EL services offers a potential explanation for the small and nonsignificant estimates for the effects of EL reclassification at the end of eighth grade. First, like the district in the study by Reyes and Hwang (2019), the district in this study has a long tradition of serving ELs. From conversations with district administrators, I learned that middle and high school ELs are closely monitored both before and after reclassification. The district also offers various academic programs that target new immigrants or long-term ELs. For instance, newcomers who have lived in the United States for 3 years or less are eligible to attend intensive summer courses that integrate ELD with academic content. When support services are well matched to students’ needs, we would expect reclassification to have little effect.

On-Track Status

Although not statistically different from zero at the optimal bandwidth, my coefficients for 10th and 11th grade on-track status were in favor of RFEPs. This points to the need for more research on ELs’ pathways to high school graduation and access to college entrance requirement courses. The use of on-track to graduate indicators can be leveraged to support future studies. The district in this study incorporates the total number of credits and the number of credits in each academic subject, which takes into consideration the types of courses students are taking and if they are earning passing grades. In other words, on-track status captures access to and performance in core courses. This is an example of a measure that can be calculated and reported easily by the school and informs stakeholders of students’ academic access and progress in a timely manner. Future studies can interrogate equity in academic access by examining to what extent subgroups of ELs are on track, in terms of the quantity and the quality of courses they take and complete. The State of California has recently prohibited schools and districts from excluding ELs from engaging with the standard curriculum (California Legislative Information, 2018). On-track indicators that capture both the quantity and quality of core classes reflect how well schools are providing access in compliance with this new provision. As of 2018, 11 states (AK, AR, CT, DE, IL, MD, MA, NV, OR, WA, and WV) include on-track indicators in their high school accountability system (Achieve, 2018). By adopting on-track indicators calculated using both course-taking and achievement, states can begin to identify and reduce education access gaps among subgroups.

EL Subgroups

As mentioned earlier, eighth grade ELs are composed of two main subgroups, recent immigrants and long-term ELs, both of whom require intensive monitoring and support. Until recently, policy and instruction tended to pool the two subgroups. But findings from this study highlight the need to disaggregate data to focus on each subgroup. Unlike former ELs who reclassified during earlier grades, long-term ELs continue to struggle with English language and literacy in middle and high school. They also take fewer rigorous courses and perform worse in other academic subjects compared with their EL and non-EL peers (Johnson, 2019b; Olsen, 2010, 2014). The instruction and services ELs receive in earlier grades are critical to shaping downstream outcomes, and inadequate support is a major factor contributing to students’ remaining in EL status for 5 or more years (Olsen, 2010). There is an urgent need to identify students’ linguistic and learning needs early and provide appropriate accommodations as they progress through each grade level.

This study is the first to explore the differential impact of reclassification on long-term and recent immigrant ELs. Unfortunately, the results from my interaction analyses were not precise enough to form the basis for strong conclusions (see Supplemental Tables A3–A8 in the online version of the journal). But they do offer suggestive evidence that reclassification in eighth grade leads to better high school outcomes for long-term ELs, especially in terms of on-track to graduate status. Long-term ELs with relatively high English proficiency might benefit from policies that allow them to reclassify sooner (e.g., modifying reclassification criteria) or changes in curriculum that increase engagement with rigorous academic materials (e.g., placing long-term ELs in classes alongside native and fluent English users and providing support). Kim and Garcia (2014) found that long-term ELs who identify as proficient English speakers feel demotivated by courses that fail to challenge them and disengage from school. Improving the quality of curriculum and services to which long-term ELs are exposed may address this problem. In light of the course-taking gaps found by recent research between long-term ELs and their peers (Johnson, 2019b), the potential positive impact of eighth grade reclassification on long-term ELs’ academic attitude and achievement is worth further investigation. Future studies should leverage larger sample sizes to identify differential treatment effects on subgroups within the EL student population.

ESSA (2015) offers an opportunity for states to overhaul their accountability systems in support of college and career readiness for all students. As a subgroup that comprises nearly 10% of all K–12 students, ELs’ access to and success in rigorous high school classes is of primary importance. More rigorous studies are needed to build a body of precise estimates on the various effects of EL classification and reclassification on student subgroups across educational contexts. To ensure that all language minority students receive the support they require to prepare for college and career, further research is also necessary to uncover best practices in identifying and serving students with different prior academic experiences. Qualitative evidence would be especially valuable to this line of inquiry. Future research might examine the extent to which EL services affect students’ access to college and career counseling, college entrance tests and test preparation, and completion of college applications.

Concluding Remarks

A few study limitations should be considered. First, the data for this study come from one district in California with a unique demographic composition. Spanish-speaking ELs only comprised 40% of the sample, making the results unlikely to be generalizable to statewide or nationwide policy contexts. Second, indicators of students’ socioeconomic status were unavailable. Therefore, I was not able to use this important background characteristic to increase estimation precision or test for effect heterogeneity. Third, the identification strategy relies on the assumption that students with test scores in a close range around the cut score are as good as randomly assigned to EL and non-EL status. This implies that the results would only be generalizable to a subpopulation, that is, students with relatively high levels of English proficiency. Finally, the analytic sample was small relative to some previous studies (e.g., Robinson-Cimpian & Thompson, 2016) and as a result, some of the analyses may have been underpowered.

This study used ELA and SAT reading test scores as measures of reading ability. It is possible that reclassified and continuing ELs are developing different skills but these measures were not able to show the differences. For instance, ninth grade EL curriculum might be focusing only on reading comprehension, whereas the non-EL curriculum moved beyond comprehension to critique and evaluation. If the test items only measured comprehension and not evaluation, the two groups might perform similarly, even though non-ELs were developing an additional set of higher order skills. Most of my sample had already moved onto higher grades before the Smarter Balanced assessments that aim to measure these skills were adopted. It would be worthwhile to examine Smarter Balanced scores for later cohorts to see if there are differences in the development of higher order skills. Unfortunately, reading achievement is not assessed annually in high school. As a result, it is hard to observe students’ reading development over time. Low SAT participation rates also limited the data in this study. Although missingness of the test scores was balanced across the cut score, the analyses did lose power. This points to a larger, more prevalent issue of missing EL data.

Data on EL academic outcomes, including attendance and graduation, are often reported with substantial missingness because of reclassification and attrition due to migration. An individual student’s EL status changes as he or she exits EL services. In state and federal data reporting, details such as reclassification date or years spent as an EL are seldom provided. Thus, it is not possible to make some of the distinctions (e.g., ever-EL vs. never-EL; newcomer vs. long-term) that would provide key insights into EL curriculum and policy. In addition, ELs tend to migrate more than non-ELs. Schools and districts face the challenge of tracking ELs as they move across districts and even states. Without proper support from the state and federal governments, schools and districts lack resources to monitor ELs’ academic trajectories and, as a result, many ELs simply disappear from data systems. This happens especially frequently during secondary school, which makes tracking the college and career outcomes of ELs even more difficult. Additional systematic EL data collection and curation are desperately needed to enable more robust research and better-informed policy discussions on college and career readiness for all. I hope this study can serve as a catalyst for rich discussions that center on programs and policies that foster ELs’ academic progress.

Supplemental Material

Angela_Johnson__Online_Appendix – Supplemental material for The Impact of English Learner Reclassification on High School Reading and Academic Progress

Supplemental material, Angela_Johnson__Online_Appendix for The Impact of English Learner Reclassification on High School Reading and Academic Progress by Angela Johnson in Educational Evaluation and Policy Analysis

Footnotes

Acknowledgements

The author is grateful for continued support from her research partners in the anonymous district. She thanks two anonymous reviewers for their helpful comments.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The author thanks the Stanford E. K Potter Fellowship for supporting this research. All views expressed in this article are her own and do not represent Stanford University or NWEA.

ORCID iD

Angela Johnson

Author

ANGELA JOHNSON is a research scientist at NWEA. Her research interests include college access, bilingual education, and education policy.

References

Abedi

(2008). Classification system for English language learners: Issues and recommendations. Educational Measurement: Issues and Practice, 27(3), 17–31.

Achieve. (2018). On track or falling behind? How states include measures of 9th grade performance in their ESSA plans. Retrieved from https://www.achieve.org/files/On-Track-Brief.pdf

Allensworth

A. M.

Easton

J. Q.

(2007). What matters for staying on-track and graduating in Chicago Public Schools. The University of Chicago Consortium on School Research. Retrieved from https://consortium.uchicago.edu/publications/what-matters-staying-track-and-graduating-chicago-public-schools

Balfanz

Fox

J. H.

(2011). On track for success: The use of early warning indicator and intervention systems to build a grad nation. The Everyone Graduates Center at Johns Hopkins University. Retrieved from http://new.every1graduates.org/wp-content/uploads/2012/03/on_track_for_success.pdf

Brooks

M. D.

(2017). “She doesn’t have the basic understanding of a language”: Using spelling research to challenge deficit conceptualizations of adolescent bilinguals. Journal of Literacy Research, 49, 342–370.

California Legislative Information. (2018). AB-2735 English learners: Participation in standard instructional program. Retrieved from https://leginfo.legislature.ca.gov/faces/billNavClient.xhtml?bill_id=201720180AB2735

Callahan

R. M.

(2005). Tracking and high school English learners: Limiting opportunity to learn. American Educational Research Journal, 42, 305–328.

Callahan

R. M.

Humphries

M. H.

(2016). Undermatched? School-based linguistic status, college going, and the immigrant advantage. American Educational Research Journal, 53, 263–295.

Callahan

R. M.

Shifrer

(2016). Equitable access for secondary English learner students: Course taking as evidence of EL program effectiveness. Educational Administration Quarterly, 52, 463–496.

10.

Callahan

R. M.

Wilkinson

Muller

(2010). Academic achievement and course taking among language minority youth in U.S. schools: Effects of ESL placement. Educational Evaluation and Policy Analysis, 32, 84–117.

11.

Calonico

Cattaneo

M. D.

Titiunik

(2014). Robust data-driven inference in the regression-discontinuity design. The Stata Journal, 14, 909–946.

12.

Carlson

Knowles

J. E.

(2016). The effect of English language learner reclassification on student ACT scores, high school graduation, and postsecondary enrollment: Regression discontinuity evidence from Wisconsin. Journal of Policy Analysis and Management, 35, 559–586.

13.

Cimpian

J. R.

Thompson

K. D.

Makowski

M. B.

(2017). Evaluating English learner reclassification policy effects across districts. American Educational Research Journal, 54(Suppl. 1), 255S–278S.

14.

Conger

Long

M. C.

Iatarola

(2009). Explaining race, poverty, and gender disparities in advanced course-taking. Journal of Policy Analysis and Management, 28, 555–576.

15.

Dabach

D. B.

(2014). “I am not a shelter!” Stigma and social boundaries in teachers’ accounts of students’ experience in separate “sheltered” English learner classrooms. Journal of Education for Students Placed at Risk, 19, 98–124.

16.

DeCapua

Marshall

H. W.

(2010). Limited formally schooled English language learners in U.S. classrooms. The Urban Review, 42, 159–173.

17.

ED Data Express. (2019). State graduation rates. Retrieved from https://eddataexpress.ed.gov/state-tables-main.cfm

18.

ESSA. (2015). Every Student Succeeds Act of 2015, Pub. L. No. 114-95 § 114 Stat. 1177 (2015–2016).

19.

Estrada

Wang

(2018). Making English learner reclassification to fluent English proficient attainable or elusive: When meeting criteria is and is not enough. American Educational Research Journal, 55, 207–242.

20.

Flores

Kleyn

Menken

(2015). Looking holistically in a climate of partiality: Identities of students labeled long-term English language learners. Journal of Language, Identity & Education, 14, 113–132.

21.

Flores

S. M.

Drake

T. A.

(2014). Does English language learner (ELL) identification predict college remediation designation? A comparison by race and ethnicity, and ELL waiver status. The Review of Higher Education, 38, 1–36.

22.

Freeman

Y. S.

Freeman

D. E.

Mercuri

(2002). Closing the achievement gap: How to reach limited-formal-schooling and long-term English learners. Portsmouth, NH: Heinemann.

23.

Gamoran

(2009, August). Tracking and inequality: New directions for research and practice (WCER Working Paper No. 2009-6). Madison: Wisconsin Center for Education Research, University of Wisconsin–Madison. Retrieved from https://files.eric.ed.gov/fulltext/ED506617.pdf

24.

Hartman

Wilkins

Gregory

Gould

L. F.

D’Souza

(2011). Applying an on-track indicator for high school graduation: Adapting the Consortium on Chicago School Research indicator for five Texas districts (Issues & Answers Report, REL 2011-No. 100). Washington, DC: Regional Educational Laboratory Southwest, National Center for Education Evaluation and Regional Assistance, Institute of Education Sciences, U.S. Department of Education. Retrieved from https://ies.ed.gov/ncee/edlabs/projects/project.asp?ProjectID=264

25.

Johnson

(2019a). The effects of English learner classification on high school graduation and college attendance. AERA Open, 5(2), 1–23. doi:10.1177/2332858419850801

26.

Johnson

(2019b). A matter of time: Variations in high school course-taking by years-as-EL subgroup. Educational Evaluation and Policy Analysis. Advance online publication. doi:10.3102/0162373719867087

27.

Kanno

Kangas

S. E.

(2014). “I’m not going to be, like, for the AP”: English language learners’ limited access to advanced college-preparatory courses in high school. American Educational Research Journal, 51, 848–878.

28.

Kim

W. G.

Garcia

(2014). Long-term English language learners’ perceptions of their language and academic learning experiences. Remedial and Special Education, 35, 300–312.

29.

Lee

D. S.

Lemieux

(2010). Regression discontinuity designs in economics. Journal of Economic Literature, 48, 281–355.

30.

Lillie

K. E.

Markos

Arias

M. B.

Wiley

T. G.

(2012). Separate and not equal: The implementation of Structured English Immersion in Arizona’s classrooms. Teachers College Record, 114(9), 1–33.

31.

Linquanti

Cook

H. G.

Bailey

A. L.

MacDonald

(2016). Moving toward a more common definition of English learner: Collected guidance for states and multi-state assessment consortia. Washington, DC: Council of Chief State School Officers.

32.

Martin-Beltran

Peercy

M. M.

(2014). Collaboration to teach English language learners: Opportunities for shared teacher learning. Teachers and Teaching: Theory and Practice, 20, 721–737. doi:10.1080/13540602.2014.885704

33.

Mavrogordato

White

R. S.

(2017). Reclassification variation: How policy implementation guides the process of exiting students from English learner status. Educational Evaluation and Policy Analysis, 39, 281–310.

34.

McCrary

(2008). Manipulation of the running variable in the regression discontinuity design: A density test. Journal of Econometrics, 142, 698–714.

35.

Menken

Kleyn

(2009). The difficult road for long-term English learners. Supporting English Language Learners, 66(7). Retrieved from http://www.ascd.org/publications/educational_leadership/apr09/vol66/num07/The_Difficult_Road_for_Long-Term_English_Learners.aspx

36.

Norbury

Wong

Wan

Reese

Dhillon

Gerdeman

(2012). Using the freshman on-track indicator to predict graduation in two urban districts in the Midwest Region (Issues & Answers Report, REL 2012–No.134). Washington, DC: Regional Educational Laboratory Southwest, National Center for Education Evaluation and Regional Assistance, Institute of Education Sciences, U.S. Department of Education. Retrieved from https://ies.ed.gov/ncee/edlabs/projects/project.asp?projectID=277

37.

Olsen

(2010). Reparable harm: Fulfilling the unkept promise of educational opportunity for California’s long term English learners. Long Beach: Californians Together.

38.

Olsen

(2014). Meeting the unique needs of long term English language learners: A guide for educators. Washington, DC: National Education Association.

39.

Olsen

Jaramillo

(1999). Turning the tides of exclusion: A guide for educators and advocates for immigrant students. Oakland: California Tomorrow.

40.

Papay

J. P.

Willett

J. B.

Murnane

(2011). Extending the regression-discontinuity approach to multiple assignment variables. Journal of Econometrics, 161, 203–207.

41.

Porter

K. E.

Reardon

S. F.

Unlu

Bloom

H. S.

Cimpian

J. R.

(2017). Estimating causal effects of education interventions using a two-rating regression discontinuity design: Lessons from a simulation study and an application. Journal of Research on Educational Effectiveness, 10, 138–167.

42.

Reardon

S. F.

Robinson

J. P.

(2012). Regression discontinuity designs with multiple rating-score variables. Journal of Research on Educational Effectiveness, 5, 83–104.

43.

Reyes

Hwang

(2019). Middle school language classification effects on high school achievement and behavioral outcomes. Educational Policy. Advance online publication. doi:10.1177/0895904818823747

44.

Robinson

J. P.

(2011). Evaluating criteria for English learner reclassification: A causal-effects approach using a binding-score regression discontinuity design with instrumental variables. Educational Evaluation and Policy Analysis, 33, 267–292.

45.

Robinson-Cimpian

J. P.

Thompson

K. D.

(2016). The effects of changing test-based policies for reclassifying English learners. Journal of Policy Analysis and Management, 35, 279–305.

46.

Shin

(2018). The effects of the initial English language learner classification on students’ later academic outcomes. Educational Evaluation and Policy Analysis, 40, 175–195.

47.

Stuit

O’Cummings

Norbury

Heppen

Dhillon

Lindsay

Zhu

(2016). Identifying early warning indicators in three Ohio school districts (REL 2016–118). Washington, DC: Regional Educational Laboratory Southwest, National Center for Education Evaluation and Regional Assistance, Institute of Education Sciences, U.S. Department of Education. Retrieved from https://ies.ed.gov/ncee/edlabs/projects/project.asp?projectID=358

48.

Thompson

K. D.

(2015). Questioning the long-term English learner label: How categorization can blind us to students’ abilities. Teachers College Record, 117(12), 1–50.

49.

Thompson

K. D.

(2017). English learners’ time to reclassification: An analysis. Educational Policy, 31, 330–363.

50.

Umansky

I. M.

(2016a). Leveled and exclusionary tracking: English learners’ access to academic content in middle school. American Educational Research Journal, 53, 1792–1833.

51.

Umansky

I. M.

(2016b). To be or not to be EL: An examination of the impact of classifying students as English learners. Educational Evaluation and Policy Analysis, 38, 714–737.

52.

Umansky

I. M.

Hopkins

Dabach

D. B.

Porter

Thompson

Pompa

(2018). Understanding and supporting the educational needs of recently arrived immigrant English learner students: Lessons for state and local education agencies. Washington, DC: Council of Chief State School Officers.

53.

Umansky

I. M.

Reardon

S. F.

(2014). Reclassification patterns among Latino English learner students in bilingual, dual immersion, and English immersion classrooms. American Educational Research Journal, 51, 879–912.

54.

Umansky

I. M.

Thompson

K. D.

Diaz

(2017). Using an ever-English learner framework to examine disproportionality in special education. Exceptional Children, 84, 76–96.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.33 MB