Title: Exploring issues of identification and treatment of children with lateemerging reading disabilities
1Exploring issues of identification and treatment
of children with late-emerging reading
disabilities from an RTI perspective.
- Donald Compton
- Vanderbilt University
- This research was supported in part by Grant
H324U010004 from the U.S. Department of
Education (OSEP), Grant R324G060036 from the
U.S. Department of Education (IES), and Core
Grant HD15052 from the National Institute of
Child Health and Human Development to Vanderbilt
University. Statements do not reflect the
position or policy of these agencies, and no
official endorsement by them should be inferred.
2(No Transcript)
3Overview
- Very short introduction on response-to-interventio
n (RTI) - Study 1 Results of a 1st grade universal
screening study - Study 2 Exploring serious challenges to early
screening procedures Children with late-emerging
reading disabilities (LERD) - Study 3 Results from a study investigating the
use of a multicomponent reading intervention
program for struggling readers in late elementary
school.
4Typical RTI Procedures
- All children in a class, school, district are
tested once in the fall to identify students at
risk for long-term difficulties. - The response of at-risk students to GE (Tier1) is
monitored to determine whose needs are not met
and therefore require more intensive tutoring
(Tier 2). - For at-risk students, research-validated Tier 2
tutoring is implemented. Student progress is
monitored throughout intervention. - Those who do not respond to the validated
tutoring are identified and further evaluated by
a multi-disciplinary team evaluation for possible
LD designation and special education placement.
5RTI Three Tiers
- Tier 1 General education
- Research-based program
- Faithfully implemented
- Works for vast majority of students
- Screening for at-risk pupils, with monitoring of
at-risk response to general education - Tier 2 Small-group preventative tutoring
- Monitoring of at-risk response to tier 2
intervention - Tier 3 Special education
6Tertiary Prevention Specialized
Individualized Systems for Students with
Intensive Needs
CONTINUUM OF SCHOOL-WIDE SUPPORT
5
Secondary Prevention Specialized Group Systems
for Students with At-Risk Behavior
15
Primary Prevention School-/Classroom- Wide
Systems for All Students, Staff, Settings
80 of Students
7Advantages of RTI
- A stronger focus on prevention
- Earlier identification of children with
disabilities - An assessment process with clearer implications
for academic programming - The end of a wait-to-fail model of identification
8Questions Remaining Regarding Implementation of
RTI
- Who should enter the RTI process?
- What should effective tier-2 instruction look
like? - How should one determine whether effective
secondary prevention has been conducted? - What is a valid and practical method of
monitoring responsiveness to secondary
prevention? - What is a defensible definition of
nonresponsiveness (i.e., RD) in terms of key
reading skills? - How should tertiary prevention be conceptualized?
9Study 1 Using Universal Screening to Identify
1st Grade Children at Risk for RD
10The First Step in RTI Screening
- RTI success depends on accurate specification of
this risk pool. - Perfect screening would result in 100 accurate
identification of True Positives (those who
will develop RD) who will go into Tier 2
interventions and True Negatives (those who
will not develop RD) who will be excluded from
Tier 2 intervention.
11Two Types of Screening Errors
- Two types of errors challenge the accuracy of
procedures for determining risk. - False positives
- Children who eventually become good readers score
below the screening cut score and are falsely
identified as at risk. - Undermine RTIs prevention purpose by increasing
the number of children identified at risk and
thereby stressing school resources to provide
intervention to an inflated percentage of the
population. - False negatives
- Children who later exhibit reading problems score
above the cut score and are falsely identified as
not at risk. - Diminish the utility of RTI prevention by failing
to provide intervention to children who will
eventually develop RD.
12Previous Attempts
- The vast majority of early identification studies
have reported classification accuracies outside
the acceptable range for RTI to work - False positives ranging from 20 to 60
- False negative rates ranging from 10 to 50
13Identifying Children Who Are Responders (not at
risk for RD) and Nonresponders (at risk for RD)
to Tier 1 InstructionHypothetical Case Studies
14Sarahs Progress on Words Read Correctly
Sarah Smith
Reading 2
Words Read Correctly
Sep
Oct
Nov
Dec
Jan
Feb
Mar
Apr
May
15Jessicas Progress on Words Read Correctly
Jessica Jones
Reading 2
Words Read Correctly
Sep
Oct
Nov
Dec
Jan
Feb
Mar
Apr
May
16Do we have evidence that we can accurately
identify children in the fall of first grade who
are at high risk for becoming RD (Jessica) while
limiting false-positives (Sarah) ?
17Compton et al. (2006) Study
- Explored issues affecting development of decision
rules for selecting 1st graders for tier-2
intervention within an RTI model of LD
identification. - What is the added predictive utility of including
initial word identification fluency (WIF) or 5
weeks of WIF PM to a multivariate screening
battery (that already includes phonemic
awareness, rapid naming skill, and oral
vocabulary)? - Can we develop screening procedures which are
accurate enough to work within an RTI framework?
18Example of a WIF Probe
19Overview of Study Methods
- In 42 classes in 16 middle-TN schools, identified
low study entry 1st graders. - The 6 lowest students per class on one or both
measures, also judged as such by the teacher,
were designated low study entry (n 252). - At end of grade 2, a total of 206 were
found/tested (attrition rate of 18 over 2
years). - In October, administered a multivariate screening
battery initial WIF, phonemic awareness, rapid
naming, oral vocabulary. - Monitored progress with WIF, each week for 5
weeks calculated 5-week slope and level. - At end of grade 2 classified children as
RD/non-RD based on a composite score of reading. - Applied logistic regression to classify RD/non-RD
at end of grade 2, using 1st-grade prediction
battery and short-term PM as predictors. - Evaluated differences in classification accuracy
using sensitivity, specificity, and area under
ROC curve.
20Results
TN/FN/TP/FP HR Sens Spec AUC Initial
Screen 145/5/15/41 77.7 75.0 80.0 .863 Sound
Matching Rapid Digits Vocabulary Add Initial
WIF 150/3/17/36 81.1 85.0 80.6 .904 Add 5-Week
PM 154/2/18/32 83.4 90.0 82.7 .912 Classification
Tree Analysis 174/0/20/12 96.8 100.0 93.5 .982
21Conclusions Study 1
- Using the 6 screening variables and logistic
regression we are able to accurately identify 90
true positives and 85 of true negatives
correctly! - These results have been replicated on a more
representative sample. - Seems to me we should stop here, pat ourselves on
the back, and proclaim mission accomplished! - Hmmm, maybe there are children who develop RD
after the end of second or third grade and we
miss them via early screening?
22 Study 2 Exploring Subtypes, Prevalence Rates,
and Early Identification of Children Exhibiting
Late-Emerging Reading Disability
23 Who are we missing using early identification
procedures focusing on word ID skills?
- Are there student who do not display reading
deficits until later in elementary school (i.e.,
late-emerging RD LERD)? - Can we identify separate subtypes of LERD and
what is the prevalence of each subtype? - Can we identify variables that distinguish LERD
from typically developing (TD) readers in
kindergarten or 1st grade? - What is the effect of early intervention on LERD
identification?
24Literature on the Existence of Late-Emerging RD
- Children with LERD undermine the intent of early
prevention by depriving at-risk children of the
early intervention they require (i.e., false
negatives in early grades). - Estimates suggest that the prevalence of children
with late-emerging RD range from 20 to 46 of
those identified with RD during elementary school
- Several studies have identified students who were
not identified until fourth grade (Catts Hogan,
2002 Compton et al., 2008 Leach et al., 2003
Lipka et al., 2006) - Previously these high prevalence rates had been
attributed to flaws in the identification process
(due to high intelligence, good behavior, or
compensatory strategies).
25Conceptualizing Subtypes of LERD using the Simple
View of Reading
- Goughs simple view of reading postulates that
two general types of skill are required for good
reading comprehension - accurate words identification
- general language comprehension ability
- Evidence suggesting partial independence of word
ID and comprehension skills allowing 3 major
subtypes - Specific deficits in word-reading (LERD-W)
- Specific deficit in comprehension (LERD-C)
- Deficits in word-reading and comprehension
(LERD-CW).
26Prevalence of LERD
- Catts Compton (2008) employed mixture latent
transition modeling (i.e., mover-stayer model),
to examine the stability of classes associated
with RD across time (grades 2-10) and to identify
the prevalence of the three LERD subtypes. - 493 children were followed longitudinally as part
of an epidemiologic study (Tomblin et al., 1997).
- 5 standardized reading measures (2 word
reading/decoding and 3 requiring passage reading
and question answering) were used to estimate
latent class membership in grades 2, 4, 8, 10.
27What is Latent Transition Analysis?
- LTA, an extension of latent class analysis,
allows modeling of the movement of individuals
across latent classes as a function of time - Because LTA relies on latent variable analyses,
classes are not directly observable, but rather
are inferred from a combination of manifest
indicators - LTA produces parameter estimates corresponding to
the proportion of individuals in each latent
class initially (RD and TD), as well as the
probability of individuals changing classes with
time.
28LTA Model
Four Classes Time 1 ? Time 2 TD ? TD RD ?
RD TD ? RD RD ? TD
29Catts Compton
30LERD Subtype Prevalence
31Summary of LTA Modeling
- Overall, 82 of the sample remained in the same
class (i.e., stable) across grades 2-10 (69
typically developing 13 RD) - Approximately, 13 of the population transitioned
from the typically developing class to one of the
LERD classes - Comprehension deficits (LERD-C LERD-CW) made up
the largest transitioning group (8.5 of the
population) - The vast majority of transitions took place
between grades 2-4 - Once children transitioned into a class they
tended to stay in that class
32Can We Identify Late-Emerging RD Before the
Problem Emerges?
- We attempted to identify late-emerging RD
subtypes using kindergarten variables - Oral Vocabulary (CELF)
- Receptive Vocabulary (PPVT)
- Grammatic Understanding (CELF)
- Grammatic Completion (CELF)
- Sentence Imitation (CELF)
- Narrative Comprehension (CELF)
- Narrative Recall (CELF)
- Phoneme Deletion
- Letter ID
- IQ
33Kindergarten Predictors of LERD
34Conclusions Study 2
- RD and TD latent classes are fairly stable across
time - Our findings indicate the existence of LERD
children who fall into 3 major subtypes (LERD-W,
LERD-C, LERD-CW) - Clear differences exist between TD and EIRD
across all 10 kindergarten measures - LERD subtypes with comprehension deficits (LERD-C
and LERD-CW) resembled the EIRD subtype scoring
poorly on early language measures but exhibited
elevated performance on letter ID measures - Compared to LERD-C, the LERD-CW subgroup tended
to score lower on all measures except letter ID - LERD-W subtype mirrored the EIRD subtype profile
across kindergarten with deficits being almost
half as severe as EIRD
35Study 3 What Can We Do Instructionally For Poor
Comprehenders?
36What Skills Are Needed for Skilled Reading
Comprehension?
- Comprehending written material requires the
coordination of a complex set of skills - Accurate and fluent identify the words
- General language comprehension ability
- Background/domain knowledge
- Application of specific reading strategies
- Inference making
- Monitor comprehension processes
- Deficits in any one of these areas can contribute
to comprehension failure
37What Should We Do?
- Design instruction to address the diverse set of
literacy skills - Effective reading programs designed to improve
struggling readers comprehension skills must
provide integrated work across (Pressley, 2000) - Decoding and word recognition
- Reading fluency
- Vocabulary
- Knowledge development
- Reading comprehension strategy
38Can We Do This and Does it Work?
- Consensus reports suggest that teachers have
difficulties integrating multiple instructional
components in an effective manner to teach
reading to struggling readers (NICHD, 2000 RAND,
2002 Snow, Burns, Griffin, 1998). - Few studies have systematically studied the
effects of multicomponent reading programs on the
reading skills of struggling readers - Our long-term research objective is to develop
multicomponent reading programs to address the
diverse needs of late elementary school students
who are struggling readers
39What Have We Done Thus Far?
- We have been examining the effects of in-text
reading activities, embedded within
multicomponent reading programs, to promote
vocabulary and knowledge learning from text - Our primary objective in this study was to
develop instructional dialogues, strategies, and
materials that increased the probability that
struggling readers would derive and retain
vocabulary and declarative knowledge while
reading text - I present efficacy data from two studies
evaluating the efficacy of a multicomponent
instructional program for students in grades 3-5
who are struggling readers - Study 1 We isolated the effects of metacognitive
strategy instruction on the retention of
knowledge - Study 2 We examined the effects of text-level
instructional program designed to increase
vocabulary and declarative knowledge acquisition
during reading in expository text
40Study 1 Research Questions
- What are the benefits of including a
metacognitive strategy component to comprehension
instruction on the retention of declarative
knowledge in children who are poor readers - Where do the effects come from?
41Study 1 - Procedure
- Groups of struggling 3rd and 4th grade readers
consisting of 2 3 children were randomly
assigned to one of three interventions - Instruction was provided by trained RAs
- Instruction was provided outside the general
education classroom - Groups received either 30 minutes or 60 minutes
of instruction per lesson depending on the
treatment group assigned. - Groups met 3 4 times per week for approximately
10 weeks. - Groups received a total of 25 lessons
42Interventions
- Decoding Only Decoding Fluency (30
minutes/lesson) - Decoding Traditional Comprehension Decoding
Fluency (30 minutes/lesson) Comprehension (30
minutes/lesson) - Decoding Reciprocal Teaching Decoding
Fluency (30 minutes/lesson) Comprehension
Metacognition (60 minutes/lesson)
43Decoding Only
- Word Identification Strategies
- (WIST Gaskin et al., 1986
- Lovett et al., 2000).
- WIST teaches children to recognize larger subword
units by applying four different decoding
strategies including Compare and Contrast (rime
unit), Peeling-Off affixes, Vowel Variation, and
Seek the Part You Know. Approximately 25 minutes
per lesson. - QuickReads (Hiebert, 2003)
- QuickReads feature short, high-interest,
nonfiction texts designed to be read quickly
(multiple times) and meaningfully. Approximately
5 minutes per lesson
44Decoding Traditional Comprehension
- WIST QuickReads (30 minutes)
- Reading Expository Text Orally (30 minutes)
- Each group read 3 books (50 pages/book)
- Book Titles Chasing Tornadoes, Young Pioneers,
Rain Forests - Children took turns reading text aloud
- Comprehension Instruction
- After each paragraph/passage the teacher directed
children to - Answer or generate a question about the
paragraph/passage - Summarize the paragraph/passage
- Make predictions
- Provide clarification if needed
45Decoding Reciprocal Teaching
- WIST QuickReads (30 minutes)
- Reading Expository Text Orally (30 minutes)
- Each group read 3 books (50 pages/book)
- Book Titles Chasing Tornadoes, Young Pioneers,
Rain Forests - Children took turns reading text aloud
- Comprehension Instruction
- Teacher modeled the metacognitive dialogue for
question generating, summarizing, predicting, and
clarifying - After each paragraph/passage the children would
- Generate a question about the paragraph/passage
- Summarize the paragraph/passage
- Make predictions
- Seek clarification
- Before proceeding to the next paragraph the group
had to come to agreement that the question,
summary, and prediction were of high quality - Children gradually assumed responsibility for
comprehension instruction
46Effects of Text Instruction
47Why is RT Superior?
- Increased Opportunities to Respond and Listen to
Others Respond?
48Increased Opportunities to Respond
- RT Dialogue
- Lesson 6
- Taylor (reads text in book) There were a number
of trails the pioneers could take. Most people
used the Oregon Trail, which was the best-known
route. The trail ran fro 2,400 miles over prairie
grass, mountain rocks, and desert sand. After
passing through the Rocky Mountains, the pioneers
came to the halfway point. Here the trail
divided. The northern part continued into Oregon.
The southern part went into California. - Teacher Thats good. So Taylor, can you give us
a summary? - Taylor Hmmmm.
- Teacher Brucker and Allie, I want both of you
guys to look to answer that. - Taylor A summary?
- Teacher From the paragraph. (waits) What trail
did they take? - Taylor The Oregon Trail.
- Teacher Very good. What happened to that trail?
- Taylor They came to the halfway point and they
divided the northern part - Teacher Where did the other part go?
- Taylor Southern.
49Increased Opportunities to Respond
- RT Dialogue
- Lesson 24
- Allie (reads text in book) Weather satellites
travel high above the earth. They take pictures
of the atmosphere below. They show where the
thickest clouds are. This is where storms are
most likely to develop. - Allie This is aboutThis paragraph is about a
satellite, a satellite in space taking pictures
of the earth. - Brucker That is what I was going to say.
- Taylor I like it.
- Brucker I like it too, but this is what I was
going to say. This paragraph is about a
satellite from space that is taking pictures of
the earths atmosphere and these where the
thickest clouds are. - Taylor Thats good.
- Allie Im not saying it is bad Brucker, but it
could be a little too long. I dont know how to
shorten it but - Taylor You could put both of yours together
somehow but make it short. Like you and Brucker
both said taking pictures. - Allie Are you saying maybe we should do like.
..This is about satellite in space taking
pictures of earth... - Taylor Say where the storm is going to hit
next. - Brucker Yea. What do you think?
- Taylor Yea, thats good.
- Allie Why did Who took What takes pictures?
- Taylor You could just say satellite.
- Brucker I got one. How do we know where a storm
is?
50Increased Opportunities to Respond
- Traditional Dialogue
- Lesson 24
- Trevor (reads text in book) Weather satellites
travel high above the earth. They take pictures
of the atmosphere below. They show where the
thickest clouds are. This is where storms are
most likely to develop. - Teacher You are really doing a great job. All
right, summary? - Trevor I read about what weather satellites
that they travel high above the earth. - Teacher They do. Weather satellites travel high
above the earth. But what do they do high above
the earth? They take... - Trevor Pictures.
- Teacher Good, weather satellites take pictures
of the atmosphere to show where a storm might
start. Courtney, what do weather satellites do? - Courtney The weather people probably have this
little TV and monitor and they can look from
the space ...I forgot the name of it... - Teacher Satellites
- Courtney Satellites and if they take a picture
of it making a funnel they can probably see it
and when it is coming back on they can say there
is going to be a tornado and can tell people to
get covered.
51Study 2 - Research Questions
- What are the effects of text-level instruction
aimed at increasing vocabulary and declarative
knowledge acquisition during reading in
expository text? - Are there trade-offs between vocabulary learning
and declarative knowledge acquisition based on
the type of text-level instruction provided to
children during reading in expository text?
52Procedure
- Groups consisting of 2 4 struggling readers
grades 2 - 5 were assigned to one of three
interventions. - Instruction is provided by trained RAs.
- Instruction is provided at Vanderbilt University
after school hours. - Groups receive 90 minutes of instruction per
lesson. - Groups met 2 times per week.
- Groups receive a total of 24 lessons for a total
of 36 hours of instruction.
53Outcome Measures
- Experimenter created vocabulary measure
- 40 multiple choice items made up the vocabulary
assessment. - Questions were generated directly from the books.
- To control for reading ability all items were
read to the children at pre- and posttest. - Experimenter created declarative knowledge
measure - 30 multiple choice items made up the assessment.
- Questions were generated directly from the books.
- To control for reading ability all items were
read to the children at pre- and posttest.
54Interventions
- Word Identification Strategies (30 min)
- WIST (Lovett et al., 2000) teaches children to
recognize larger subword units by applying four
different decoding strategies (Approximately 20
minutes per lesson). - QuickReads (10 min)
- QuickReads (Hiebert, 2003) feature short,
high-interest, nonfiction texts designed to be
read quickly (multiple times) and meaningfully
(approximately 10 minutes per lesson) - Reading Expository Text Orally (50 min)
- Each group reads 3 books (50 pages/book,
Chasing Tornadoes, Young Pioneers, Rain Forests)
with children taking turns reading text aloud. - Text-level Instruction
- During text reading children were instructed in
one of 3 text-level instructional strategies
(Vocabulary, Reciprocal Teaching, or Traditional
Instruction).
55Traditional Instruction(Control Condition)
- After reading all of the days passages students
completed a work sheet that required answering
questions regarding key concepts read during the
day. - Students are allowed to look back at the text to
answer questions and the teacher helps guide
students in finding the relevant information. - Care was taking in constructing the worksheets
that there was no overlap between vocabulary and
declarative knowledge covered on the worksheet
and that targeted on during VOC and RT training.
56Traditional Instruction Worksheet
57Reciprocal Teaching
- Strategies Taught to the RT Condition (Same as
Study 1) - Questioning
- Prediction
- Clarifying
- Summarizing
- Metacognitive Dialogue
58Example of RT Instruction
59Vocabulary Instruction
- Strategies Taught to the Vocabulary Condition
- Word Analysis
- Word Parts
- Parts of Speech
- Context Clues
- Inference Making
- Finding Signal Words to Identify Definitions,
Synonyms, Antonyms, or Examples - Look-up
- Metacognitive Dialogue
60Vocabulary Instruction
- Instructional Procedure for Vocabulary Condition
- Teacher modeled the metacognitive dialogue for
word analysis, contextual clues, and look-up - After each paragraph/passage the children use the
vocabulary strategies to ascertain the meaning of
a predetermined word in text. - Children gradually assumed responsibility for
vocabulary instruction.
61Vocabulary Strategy Chart
62Contextual Signal Words
63Contextual Signal Words
64Example of a Look-up
65Example of Vocabulary Instruction
66(No Transcript)
67(No Transcript)
68Conclusions Study 3
- Across groups reading instruction had a positive
effect on vocabulary learning, declarative
knowledge acquisition. - Results tend to suggest devoting significant time
to vocabulary learning while reading text does
not negatively affect declarative knowledge
acquisition while allowing more vocabulary
learning. - There is more to worksheets than meets the eye.
69Conclusion Across the Three Studies
- We can identifying accurately a group of children
who will experience early onset on RD - However we should be prepared to continue
screening children to identify children with
late-emerging RD. - We need to continue to develop effective
interventions to address the diverse needs of
children with poor comprehension skills due to
language and metacognitive deficits.
70Thank you!