Upload
lynette-campbell
View
215
Download
0
Tags:
Embed Size (px)
Citation preview
1
The Added Value of Measuring Growth within an RtI Framework
Vicki L. Collins & Janet K. Holt
Northern Illinois University
Imagine a World that Forgot to Consider Growth . . .
An Unsuspecting Pet Owner
Puppy Chap
Dog Chap
4
A Disgruntled Child?
5
A Crowded Shoe
6
Focus of this Presentation
Chronology of research leading to RtI
Key decisions within the RtI model
Why consider growth?
Illustrate early literacy growth patterns
Why does the metric matter?
Chronology of Research
1980 1990 2000
Deno, 1986
Fuchs & Fuchs, 1998
Fuchs et al., 1993
ISBE, 2009
NASDSE & CASE, 2006
IDEIA, 2004
PCESE, 2002
Good et al., 2001
NCLB, 2001Deno, 1986
Fuchs & Fuchs, 1998
Fuchs et al., 1993
8
Key RtI Decisions
Problem Identification
Diagnostic
Curriculum Effectiveness• Core• Supplemental
Current Practice
9
10
10
20
30
40
50
60
70
80
90Lynne's Beginning Year Screening Results
Nu
mb
er o
f W
ord
s R
ead
Co
rrec
tly
Current Practice
10
M T W H F M T W H F M T W H F M T W H F M T W H F M T W H F M T W H F M T W H F M T W H F M T W H F M T W H FBaseline 1 2 3 4 5 6 7 8 9 10
0
10
20
30
40
50
60
70
Lynne's Progress Monitoring GraphGoal: In 10 weeks, when given 1 minute to read aloud a randomly selected second grade
passage from AIMSweb, Lynne will read 60 words correctly.
Weeks
Nu
mb
er o
f W
ord
s R
ead
Co
rrec
tly
11
Why Consider Growth?
Improve decision-making accuracy
Minimize false positives and negatives
12
Our Methods for Studying Growth
Participants
Measures
Procedures
Data analysis
Early Reading Project
Supported by Project REAL
Initiated in 2004-2005
Designed to improve reading achievement and data based decision-making in grades k-2
Utilized DIBELS measurement system
14
Participants
Project YearGrade
Kindergarten First Second
1(2004-2005)
1 2
2(2005-2006)
3 1 2
3(2006-2007)
4 3 1
4(2007-2008)
4 3
15
DIBELS Measures
Phoneme Segmentation Fluency (PSF) Identify sounds in spoken language
Nonsense Word Fluency (NWF) Map sound to print and blend sounds in words
Oral Reading Fluency (ORF) Read connected text fluently
16
Procedures -- Data Collection
GradeTime of Year
Beginning Middle End
K PSF 1 PSF 2NWF 1
1 PSF 3NWF 2
PSF 4NWF 3ORF 1
PSF 5NWF 4ORF 2
2 NWF 5ORF 3 ORF 4 ORF 5
17
Procedures -- Scoring Accuracy
Training
Limited responsibility
Double scoring
Multilevel Growth Modeling
Times of measurement may vary across individuals (time unstructured)
All data are used (Do not need same # of observations/ person = unbalanced)
Can model change over time within-individuals and differences across individuals
Can readily incorporate time-varying covariates.
Growth Models
Previous studies assume linear growth Used polynomial growth models to determine if
curvilinear growth occurred in this sample on 3 fluency measures
Tested linear and quadratic models on PSF and NWF Identifies acceleration or deceleration patterns
Tested cubic pattern for ORF Identifies phases of growth
20
Early Literacy Growth Patterns
Phoneme segmentation fluency -- quadratic
Nonsense word fluency -- linear and quadratic
Oral reading fluency -- cubic
PSF BenchmarksPSF
0
5
10
15
20
25
30
35
40
Middle K End K Beginning 1st Middle 1st End 1st
PSF
PSF Growth
-2 -1 0 1 20
10
20
30
40
50
Time Centered at Beginning of 1st Grade
PS
F
PSF Group Growth Pattern (Cohort 3)
NWF BenchmarksNWF
0
10
20
30
40
50
60
End K Beginning 1st Middle 1st End 1st Beginning 2nd
NWF
NWF Growth
0
10
20
30
40
50
60
70
80
NW
F
Random Sample of Individual NWF Growth Trajectories
-2 -1 0 1 2
Time Centered at the Middle of 1st Grade
ORF Benchmarks
ORF
0
10
20
30
40
50
60
70
80
90
100
Middle 1st End 1st Beginning 2nd Middle 2nd End 2nd
ORF
ORF Growth
-2 -1 0 1 20
25
50
75
100
125
Time Centered at the Beginning of 2nd Grade
OR
F
ORF Group Growth Pattern (Cohort 3)
27
Why does the Metric Matter?
Captures the full continuum of possible student outcomes
Different conclusions can emerge from use of level, slope, or both metrics
Continuum of Possible Outcomes
*Dual discrepancy model
Study of Dual Classification System with ORF Data
Level of performance judged by traditional benchmarking system
Adequate growth judged by local norms Those in top 25% - Atypical high growth Those in bottom 25% - Atypical low growth Those in middle 50% - Typical growth
Growth by Level ClassificationPercent of Those Meeting Instructional Recommendations on ORF by Growth Designation
for the End of 2nd Grade
Growth Level
Atypical Low Typical
Atypical High IR Totals
IR Benchmark f 10 35 43 88
% 3.7 12.9 15.9 32.5
Strategic f 7 36 23 66
% 2.6 13.3 8.5 24.4
Intensive f 59 57 1 117
% 21.8 21.0 0.4 43.2
Growth Totals f 76 128 67 271
% 28.0 47.2 24.7 100.0
31
Intensive/Atypical High Growth
Atypical High Growth - Intensive Status
010
20304050
607080
90100
Beginning Middle End
2nd Grade
OR
F
32
Strategic/Atypical High Growth
Atypical High Growth & Strategic Status
0
10
20
30
40
50
60
70
80
Beginning Middle End
2nd Grade
OR
F
33
Benchmark/Atypical Low Growth
Atypical Low Growth & Benchmark Status
20
40
60
80
100
120
140
Beginning Middle End
2nd Grade
OR
F
Decisions about Intervention Based on Metric
Decisions about Core Curricula Based on Metric
Recommendations for Improving Evaluations of Core Curricula
Determine percent of students . . .• progressing as expected• in need of intervention
Use existing standards to interpret each outcome (NASDSE & CASE, 2006)
Look at both percentages within and across years of schooling
Added Value of Considering Growth
When considered along with level data, it minimizes false positive and negative errors
When considered alone, it Illuminates whether adequate growth is occurring in all students
38
Next Steps
Determine if growth patterns are universal
Make methods for measuring growth of all students accessible
Establish criterion-based growth standards
Decipher variability in fluency measures
Imagine a World that Remembers to Consider Growth . . .