Maintaining the Power of One- on-One in a Group of Three: Next Steps Triads (available on:

32
Maintaining the Power of One-on-One in a Group of Three: Next Steps Triads (available on: www.uurc.edu/Educators/Research.php )

Transcript of Maintaining the Power of One- on-One in a Group of Three: Next Steps Triads (available on:

Maintaining the Power of One-on-One in a Group of Three:

Next Steps Triads

(available on: www.uurc.edu/Educators/Research.php)

Authors

Kathleen J. Brown

Matthew K. Fields

Grace T. Craig

University of Utah Reading Clinic

Darrell Morris

Appalachian State University

Theoretical Frame: Readers

University of Virginia Intervention Assisted reading on instructional level Word study: systematic, isolated Fluency work: repeated readings

2-3x per week; 45 minutes

(Brown, Morris, & Fields, 2005; Invernizzi, Juel, & Rosemary, 2001; Morris, Shaw, & Perney, 1991; Santa & Hoien, 1995; Morris, Tyner, & Perney, 2003)

Theoretical Frame: Educators

University of Virginia Prof. Development

Clinical practicum in schools Modeling, Observation, Coaching (36

hours) Tutoring (45 hours – minimum)

(Brown, Morris, & Fields, 2005; Morris, Shaw, & Perney, 1991; Morris, Tyner, & Perney, 2003)

Theoretical Frame: Group Size

University of Virginia Model 1:1 tutorial

Elbaum, Vaughn et al., meta-analysis (2002) no advantage for 1:1 over small group 2 unpublished doctoral dissertations Fountas & Pinnell (1996) secondary

finding

Theoretical Frame: Group Size

Vaughn et al., (2003) Assisted reading, phonics Group size: 1:1 vs. 1:3 vs. 1:10 No differences between 1:1 and 1:3;

both more effective than 1:10

Research Question: Readers

Is 1:3 grouping as effective as 1:1 for improving the performance of struggling readers who receive Next Steps?

Research Question: Educators

Can non-certified paraprofessionals deliver Next Steps in a 1:3 format effectively--

--when supervised by an intervention specialist?

Methods: Readers

N = 129 14 Title 1 and non-Title 1 schools Public & parochial; rural & urban Grades 2-8 Diverse SES, ethnicity, ELP At baseline, range = primer to early

2nd Triads matched on instructional level

Methods: Educators

N = 34 Classroom teachers, literacy coaches,

paraprofessionals, UURC staff Each already certified in Next Steps 1:1 71% tutored 1:1 and 1:3 Full lessons observed 7 times over year

Methods: Intervention

45 minute lessons 45 lessons over 1 year

Assisted reading Word study Fluency

Triad: rotating “target student” & partnership

Methods: Pre-Post Measures

Criterion-referenced Word recognition automaticity (Flash) Passage reading level Spelling

Norm-referenced Woodcock Word Attack (WRMT-WA) Woodcock Passage Comp. (WRMT-PC)

Methods: Passage Reading Criteria

Acc. (% )

Rate (wpm)

Mid G1 90 30 End G1 90 40 Mid G2 93 65 End G2 93 90 Mid G3 93 90 End G3 93 110 Mid G4 95 110 End G4 95 120 Mid G5 95 120 End G5 95 130

Methods: Analyses 3-Level HLM

Student, tutor, school 1:1 vs. 1:3 – Level 1 Variable Certified vs. Non – Level-2 Variable

Regression analysis Maximum likelihood (not OLS)

Model reduction method Run full model w/ all covariates Remove non-significant covariates Retain variables of interest

Results: Reduced Model HLM-3 Coefficients for Post Passage Reading

Variable SE p

Intercept -.387 .484 .438

Grade .612 .173 .001

Grade * Certified/Non -.525 .183 .005

Number of Sessions .024 .007 .001

Pretest Score .252 .200 .211

Pretest Score * Certified/Non .823 .332 .015

Certified/Non -.211 .604 .729

Triad/Single .290 .193 .135

Triad/Single * Certified/Non -.165 .324 .612

p-value for Level-2 R (Tutor Effect) = .001

2 p-value for Level-3 U (School Effect) > .500

Results: 1:1 vs. 1:3 on Passage

Reading

Single Triad Baseline

M

(SD) 1.66 (.38)

1.80

(.49)

Exit

M (SD)

2.68

(.85)

2.85

(.91)

Gain M (SD)

1.02 (.57)

1.12 (.81)

Results: Reduced Model HLM-3 Coefficients for Post

Word Rec Automaticity

Variable SE p Intercept .539 .368 .167

Number of Sessions .027 .007 .000 Pretest Score .769 .074 .000

Certified/Non -.544 .317 .095 Triad/Single -.247 .207 .235

Triad/Single * Certified/Non .620 .347 .076

p-value for Level-2 R (Tutor Effect) = .066

2 p-value for Level-3 U (School Effect) > .500

Results: 1:1 vs. 1:3 on Word Recognition Automaticity

Single Triad Baseline

M

(SD) 1.93 (.89)

2.03

(.88)

Exit

M (SD)

3.03

(.92) 3.03

(1.09)

Gain 1.10 1.00

Results: Reduced Model HLM-3 Coefficients for Post Spelling

Variable SE p Intercept 1.45

8 .319 .000

Number of Sessions .024 .006 .000 Pretest Spelling Score .475 .049 .000

Certified/Non -.263 .275 .346 Triad/Single -.199 .177 .263

Triad/Single * Certified/Non .444 .298 .138

p-value for Level-2 R (Tutor Effect) = .114

2 p-value for Level-3 U (School Effect) = .142

Results: 1:1 vs. 1:3 on Spelling

Single Triad Baseline

M

(SD) 2.09

(1.31) 1.82

(1.13)

Exit

M (SD)

3.40

(.93) 3.21

(.88)

Gain 1.31 1.39

Results: Reduced Model HLM-3 Coefficients for Post

WRMT Word Attack

Variable SE p Intercept 12.923 1.563 .000

Pretest Word Attack Score

.642 .057 .000

Certified/Non -1.295 2.152 .551 Triad/Single -.289 1.402 .837

Triad/Single*Certified/Non .050 2.361 .983

p-value for Level-2 R (Tutor Effect) = .052

2 p-value for Level-3 U (School Effect) > .500

Results: 1:1 vs. 1:3 on WRMT Word

Attack

Single Triad Baseline

M

(SD) [GE]

16.75 (7.58) [2.8]

16.35

(7.73) [2.8]

Exit M

(SD) [GE]

23.25

(6.57) [4.0]

22.58

(7.16) [3.8]

Gain 1.2 1.0

Reduced Model HLM-3 Coefficients for

Post WRMT Passage Comprehension

Variable β SE p Intercept 10.088 2.605 .002

Number of Sessions .157 .041 .000 Pretest Passage Comp.

Score .574 .061 .000

Certified/Non .983 1.926 .613 Triad/Single -.817 1.169 .486

Triad/Single * Certified/Non -2.011 2.002 .317

p-value for Level-2 R (Tutor Effect) = .001

2 p-value for Level-3 U (School Effect) = .137

Results: 1:1 vs. 1:3 on WRMT Passage Comprehension

Single Triad Baseline

M

(SD) [GE]

22.71 (7.21) [2.2]

22.76

(6.55) [2.2]

Exit

M

(SD) [GE]

30.29

(5.53) [3.3]

28.54

(6.17) [2.9]

Gain 1.1 0.7

Results: Passage Reading Gain

% Gain Single Triad < .5

0% 2%

≥ 1.0

62.5% 60.6%

≥ 2.0 4.2% 18.2%

Discussion: Readers

Replicates Vaughn et al., 2003

No advantage for 1:1 over 1:3

Discussion: Educators

Replicated Brown, Morris, & Fields (2005)

Paraprofessionals were able to deliver triad reading intervention effectively

…when supervised by an intervention specialist

Implications for Ed Practice

Growing evidence that 1:3 is an effective grouping format for intervention

more efficient use of resources allows more students to receive intervention

Implications for Ed Practice

Paraprofessionals can effectively extend the reach of certified educators in helping struggling readers improve…

…with training and supervision.

Implications for Ed Practice

>1 group size requires educator management skill & reduces individual attention

Odd-number grouping allows educator to retain some luxury of 1:1 tutorial Address individual student needs Progress monitor

Implications for Ed Practice

Benefits of 1:1 tutorial

Professional development opportunity to focus solely on reading development—not on management issues.

Students who “don’t fit” a group

Future Research

Economies of Scale - 1:3 vs. 1:5 advantage?

Intervention that targets earlier phases of development pre-alphabetic readers? partial alphabetic readers?