Академический Документы
Профессиональный Документы
Культура Документы
Abstract. This article provides a commentary on issues regarding the use of response to intervention (RTI) with English language
learners (ELLs). The commentary draws on current literature on
reading instruction, ELLs, RTI, and students with learning disabilities and highlights key points from the articles in this special
issue. The discussion includes future directions for research.
Using a response-to-intervention (RTI) approach to and procedural concerns (see, for review, Kavale &
identifying students with learning disabilities has been Forness, 2000; Keogh, 1987).
Several critical issues have been discussed at length
a welcome shift in policy and practices that have
otherwise largely relied on arbitrary discrepancy crite- in proceedings of the President's Commission on
ria in determining eligibility for special education. Excellence in Special Education (2002) and a summit of
Several potential benefits of RTI hold promise for learning disabilities researchers (Bradley, Danielson, &
improving educational outcomes for children. First, Hallahan, 2002), both of which contributed greatly to
early screening assessment and intervention for strug- recent changes in federal law. Subjectivity in teachers'
gling readers raise the possibility of identifying learn- referral of students, inaccuracy in assessment practices,
ing disabilities early and avoiding a prolonged wait for and lack of consistency in the nature and quality of
students to qualify for special education services. general education instruction prior to placement often
Second, early identification may lessen the impact of result in a "wait-to-fail" phenomenon, whereby stua learning disability by keeping the achievement gap as dents who qualify for special education are placed long
narrow as possible. Third, the use of RTI may con- after entry to school (Gresham, 2002; Vaughn &
tribute to improving literacy rates by providing inter- Klingner, 2007).
Problems with a discrepancy approach are well
vention for all students who experience difficulty with
reading, not just those whose difficulties are severe documented, and the concept of RTI holds promise
for addressing these problems (e.g., Gresham, 2002;
enough to warrant special education.
How we define learning disabilities and operational- Vaughn & Fuchs, 2003). The use of discrepancy criteria
ize the criteria for eligibility has been a controversial is dependent on accurate and valid measurement, an
topic for decades. Consider, for example, how many issue of particular concern when applying discrepancy
times a multidisciplinary panel such as the National criteria to English language learners (ELLs). LinanJoint Committee on Learning Disabilities has raised crit- Thompson and colleagues (this issue) suggest that other
ical issues in connection with defining learning disabil- factors, such as language of instruction and the school
ities (e.g.. National Joint Committee on Learning context in general, need to be considered as well. Under
Disabilities, 1987, 1991, 1998). Prominent authors and the reauthorized IDEA legislation, P.L. 108-446, and
other panels have frequently raised similar definitional accompanying federal regulations, a state may not
213
require the use of discrepancy criteria to identify stu- instruction in the early grades. This report makes a
dents with learning disabilities and must permit the strong recommendation for research and development
use of a RTI approach. A multidisciplinary team that to "carry promising practices and validated practices
includes the child's parents must examine evidence through to classroom applicability" (p. 382), including
regarding the child's progress toward grade-level stan- research "on educational improvement, particularly in
dards and the child's responsiveness to scientifically schools with large numbers of children from lowbased instructional practices. The team should examine income families" (p. 383).
evidence to identify patterns of strength and weakness
Considering that the RTI process begins with highin the child's academic or behavioral competence and quality reading instruction in the general education setuse multiple measures and a comprehensive evaluation ting, important questions arise regarding ELLs. What do
to make the final determination.
we really know about effective reading instruction for
When the students under consideration for eligibility ELLs? Are the curricula, practices, and assessments used
are also classified as ELLs, there may be additional with non-ELLs as effective with ELLs?
factors to consider in determining eligibility, but these
The research base on effective reading instructional
are not specified in the federal guidelines. The articles practices for ELLs is fragmented and fraught with
in this issue examine some of these factors, including controversy. In the most comprehensive effort to date
the nature and quality of the general education reading to examine the research on ELL literacy development.
program, the assessments used to determine students' Snow (2006) summarized the work of the National
responsiveness to instruction, and the content and Literacy Panel on Language-Minority Children and
procedures of follow-up intervention. In this commen- Youth: "The literature we reviewed reveals remarkably
tary, I first provide an overview of critical issues in little about the effectiveness of different aspects
reading instruction and special education for ELLs of instruction, and provides only limited guidance
and then attempt to draw from these articles some key about how good instruction for second-language speakprinciples regarding the use of RTI models with ELL ers might differ from that for first-language speakers
(p. 638) ... Most discouraging, the research we reviewed
students.
provides little basis for deciding whether or what
English Language Learners, Reading Instruction,
kinds of accommodations or adaptations are most
and Learning Disabilities
helpful to second-language learners" (p. 639). To some
The number of ELL students in schools around the extent, the lack of answers from research may be
nation has increased significantly in recent years, espe- related to constraints imposed by the politics of educacially in urban centers (Donovan & Cross, 2002). tional policy, most notably arguments over bilingual
Studies and databases continually demonstrate the per- versus second-language instruction (Gersten, 2006;
vasive academic difficulties of ELLs. Many students who Gersten & Baker, 2000).
enter school with a primary language other than
There is some evidence that word-level instructional
English score below competency markers on various
components
validated with native English-speaking
aspects of academic achievement; over 50% score in the
children
are
also
effective with ELLs, such as explicitly
bottom third in reading or mathematics (National
teaching
phonological
awareness, letter-sound relationCenter for Education Statistics, 2005) with a continued
ships, and decoding, especially when done along with
gap between ELL and non-ELL reading achievement.
meaningful experiences in engaging text. We know less
The Hispanic population in U.S. schools has grown
about how to effectively teach ELLs to gather meaning
from 2% to 15% from 1950 to 2000, and an estimated
from text in their second language (Chiappe & Siegel,
31% of students have difficulty speaking English, the
2006; Chiappe, Siegel, & Wade-Woolley, 2002; Snow,
language of instruction in most classrooms containing 2006). Not only do we need to know what practices are
ELLs (Donovan & Cross, 2002).
effective, but as Klingner and colleagues noted, we need
The disproportionate representation of minority stu- to know what works "with whom, in what contexts,
dents, many of whom are classified as ELLs, in special and under what circumstances" (Klingner, Sorrells, &
education has been a nagging concern in the field. The Barrera, 2006, p. 223).
2002 National Academy of Sciences panel (Donovan
The beginning of the RTI process. Tier 1, involves
& Cross, 2002) charged with examining disproportion- high-quality reading instruction in the general educaate representation issues concluded that the most tion classroom. A series of observational studies in
effective means of reducing the disproportionate repre- first-grade classrooms with large proportions of ELLs
sentation of low-income, ethnic-minority students in investigated specific instructional practices that correboth special education and gifted education programs lated significantly with ELL students' reading gains
would be to improve the core elements of classroom (Baker, Gersten, Haager, & Dingle, 2006; Gersten, Baker,
214
Haager, & Graves, 2005; Graves, Gersten, & Haager, Assessing Responsiveness of ELLs to Intervention
Any form of RTI must include assessment procedures
2004; Haager, Gersten, Baker, & Graves, 2003).
Six clusters of observed teaching practices, or sub- to identify potential candidates for tiers of intervention
scales, were found to relate significantly to beginning and to monitor students' progress toward established
reading instruction for ELLs. One cluster. Explicit goals. Curriculum-based measures (CBMs) are used
Teaching/Art of Teaching, included such practices as extensively for both screening and progress monitoring
modeling, making instruction explicit, and prompting (see Fuchs & Fuchs, 2006, for review). All of the studies
students, and correlated with reading gains (r = .62, in this issue used the Dynamic Indicators of Basic Early
Haager et aL, 2003; r = .75, Baker et al., 2006). Only one Literacy Skills (DIBELS) in English as a form of CBM to
item of this subscale truly seems unique to ELL measure the progress of ELLs who received intervention,
students - the extent to which the teacher adjusts his but not all used the DIBELS assessments as a screening
or her use of English during a lesson to make it under- tool to identify potential candidates for intervention.
standable to the students. The other items represent Screening is generally viewed as a means of catching
what would typically be considered simply good students who may be at risk for reading difficulties so
instruction (e.g., prompting, modeling). Instruction that no child is allowed to slip through the cracks only
Geared Toward Low Performers (r = .65/.60), Phonemic to find much later that intervention would have helped.
Given that these assessments are conducted in
Awareness and Decoding (r = .63/.49), Interactive
English
and in the early grades when ELLs are at the
Teaching (r= .57/.62), Vocabulary Development (r= .51
beginning
stages of acquiring English competency, one
.64), and Sheltered English Techniques (r = .49/.67) also
might wonder about the validity of screening using
correlated significantly with reading gains for ELLs
tools that have not included ELLs adequately in norm(Baker et al., 2006; Haager et al., 2003). Other signifiing samples and development procedures. False posicant findings came from contrasting high-gain classtives are a concern in general when using universal
rooms with less effective classrooms (Gersten et al.,
screening tools (Speece & Walker, 2007), and this con2005; Graves et al., 2004). That is, students in the highcern might be greater when using these tools with ELLs.
gain classrooms had higher rates of accurate passage
Yet, some preliminary evidence suggests that screenreading, or oral reading fluency. Effect sizes were large ing assessments have some utility with ELLs for identifor all subscales except Sheltered English Techniques, fying students who may benefit from additional
where no difference was found. This is interesting, reading intervention, when interpreted and used with
given the emphasis in teacher education and pro- caution. For example, assessments in English of rapid
fessional development on sheltered instructional letter naming and phoneme segmentation skills, using
techniques as a means of making instruction compre- the DIBELS assessments at the middle of kindergarten,
hensible for ELLs. Thus, sheltered English instruction were highly predictive of end-of-kindergarten and
may not be sufficient for truly starting ELLs off in end-of-first-grade reading outcomes (Oh, Haager, &
beginning reading. To make significant reading gains, Windmueller, in press). In first grade, nonsense word
ELL students need teachers to be using effective instruc- reading was a strong predictor at the beginning of the
tional techniques in general, adjusting their instruction year, but by the middle of the year, oral reading fluency
for individuals having difficulty, engaging their stu- (passage reading) was the most predictive (Oh et aL, in
dents in interactive and engaging vocabulary and press). Similarly, Windmueller (2004) found that from
comprehension development, and providing high- first to third grade, nonsense word reading and oral
quality explicit instruction in phonemic awareness and reading fluency were strong predictors of reading outdecoding.
comes.
We know even less about effective instruction for
Although the predictive value in using these tools is
ELLs with learning disabilities (ELL/LD), particularly in encouraging, it is important to be cautious about overreading (McCardle, Mele-McCarthy, & Leos, 2005). The reliance on such measures for making high-stakes deciresearch base on effective reading instruction for stu- sions (i.e., labeling students as "at risk" or "disabled," or
dents with LD in general must be extrapolated in trapping them into a remedial track that is difficult to
designing instruction for ELLs with LD (Klingner et al., escape). Nevertheless, the use of data to guide supple2007; Vaughn, Mathes, Linan-Thompson, & Francis, mental instruction for an ELL student who may need
2005). What we know more about is preventing reading more practice with reading in English is valid. If it
failure that might lead to later, unnecessary identifica- turned out to be a case of false positive identification,
tion of learning disabilities in ELLs, as demonstrated by that is, a student who appears in the data to be below
the articles in this special issue of the Learning Disability grade-level benchmarks but is actually achieving as
expected, the use of effective progress monitoring, the
Quarterly.
215
cantly greater gains in reading than the control classrooms, providing some evidence that this flexible teaching approach may be useful with ELLs in Tier-1
instruction in RTI models. In this study, the ELLs most
benefited in word- or letter-level processing. Reading
fluency gains lagged somewhat, but this may be consistent with other studies showing that ELLs may be
slower to develop skills beyond word-level processing
(e.g., Chiappe et al., 2002).
Kamps and colleagues (this issue) followed 318 students receiving Tier-2 supplemental instruction. Most
important, these authors demonstrated the benefits of
providing Tier-2 intervention to ELLs by examining
students' reading gains in both decoding and reading
fluency. The scope of the study allowed for an examination of specific reading programs and approaches that
led to the significant gains. Programs that followed the
principles of direct instruction were found to be highly
effective with ELLs of varying linguistic and cultural
backgrounds.
Looking within just four classrooms. Graves and colleagues (Mclntosh, Graves, & Gersten, this issue) provide a close-up look at how two of the teachers not only
provided high-quality tier-one instruction, but also supplemental Tier-2 instruction for struggling ELL readers.
The two teachers received high-quality ratings on their
overall reading instruction. The extra support and
instruction that they provided for ELLs experiencing
difficulty greatly contributed to these students' growth.
One important instructional issue that emerges from
the studies in this issue is fidelity of implementation.
Both Calhoon et al. and Kamps et al. included careful
and systematic examination of fidelity in their studies,
which greatly enhanced their ability to attribute outcomes to the treatment. Fidelity of implementation
continues to be a research-to-practice issue, particularly
when bringing practices developed in empirical investigations to scale (Denton, Vaughn & Fletcher, 2003;
Sloane, 2005). RTI will only be as effective as the implementation of the intervention part of the equation.
Therefore, in implementing any tier of the RTI model,
school personnel must carefully consider the importance of monitoring and supporting teachers' use of
the prescribed instructional procedures. Additionally,
schools should plan for ongoing examination and support of their model to refine and adjust it.
Another issue that emerges is the nuanced and individualized nature of ongoing support that teachers will
provide for struggling ELLs in RTI. Mclntosh et al. provide vignettes of teachers skilled in not only knowing
when a student needs support but also in knowing how
to provide that support in that classroom and for that
student. Contextual factors are critically important and
include school-, teacher- and student-level variation.
216
217
218