THE GROWTH OF VALUE-ADDED MODELS IN PARK HILLJanuary 2011

Phillip Adam – Coordinator for AssessmentPark Hill School District

Presentation Outline

What is Growth Data? Park Hill’s growth history Benefits of growth/VA models Challenges and considerations

Park

Hill

’s G

row

th H

isto

ry

What is Growth Data?

Questions that Growth Data can answer

Purposes for calculating & reporting growth data at the following levels:

Student Teacher Classroom Building District State

Park

Hill

’s G

row

th H

isto

ry

How does my child’s growth compare to the typical child’s growth?

Is an intervention/ program working? Are the students in this classroom/building/district showing adequate growth?

Park Hill’s Philosophy

Use better data than we did last year to make better decisions than we did last year

Better decisions means better programs and processes for studentsPa

rk H

ill’s

Gro

wth

His

tory

Making Inferences We have been making data-based

judgments and decisions for decades. Is this program working? Which school has the best learning? Which grade level creates the most student

achievement? Which schools or classrooms need an

instructional makeover? What are our opportunities for improvement?

Our questions haven’t changed much, but the quality of the answers has.

Park

Hill

’s G

row

th H

isto

ry

Making Inferences

The method of analysis must match the intended inference.

Park

Hill

’s G

row

th H

isto

ry

Status & Improvement

Starting in the early 1990s ITBS SAT-9 MMAT MAPPa

rk H

ill’s

Gro

wth

His

tory

Status & ImprovementPa

rk H

ill’s

Gro

wth

His

tory

1998 1999 2000 2001 2002Grade 2 45.8

Grade 3 51.8

Grade 4 53.3

Grade 5 57.1

Grade 6 57.4

Grade 7 54.4

Grade 8 NA

Grade 9 57.0

SAT-9 NCEs6th grade did the best2nd grade needs to improve

What does this really tell us?

Status & ImprovementPa

rk H

ill’s

Gro

wth

His

tory

1998 1999 2000 2001 2002Grade 2 44.7 43.6 46.7 45.8 45.8

Grade 3 49.1 51.1 51.9 52.6 51.8

Grade 4 50.0 50.8 52.9 52.2 53.3

Grade 5 57.2 55.9 55.9 56.6 57.1

Grade 6 58.0 57.4 55.8 56.7 57.4

Grade 7 54.8 57.1 57.2 53.4 54.4

Grade 8 NA NA NA NA NA

Grade 9 55.2 56.2 54.9 57.1 57.0

SAT-9 NCEs4th grade did the best3rd grade needs to improve

-0.8+1.1

+0.7

+0.5

+1.0-0.1

+0.0

What does this really tell us?

Cohorts are different…Pa

rk H

ill’s

Gro

wth

His

tory

1998 1999 2000 2001 2002Grade 2 44.7 43.6 46.7 45.8 45.8

Grade 3 49.1 51.1 51.9 52.6 51.8

Grade 4 50.0 50.8 52.9 52.2 53.3

Grade 5 57.2 55.9 55.9 56.6 57.1

Grade 6 58.0 57.4 55.8 56.7 57.4

Grade 7 54.8 57.1 57.2 53.4 54.4

Grade 8 NA NA NA NA NA

Grade 9 55.2 56.2 54.9 57.1 57.0

SAT-9 NCEs

We need to control for students’ histories!

Cohort Improvement

Started in 2002 Stanford-9 NCEs EXPLORE to PLAN PLAN to ACT

The case of AYPPark

Hill

’s G

row

th H

isto

ry

Cohort ImprovementPa

rk H

ill’s

Gro

wth

His

tory

1998 1999 2000 2001 2002Grade 2 44.7 43.6 46.7 45.8 45.8

Grade 3 49.1 51.1 51.9 52.6 51.8

Grade 4 50.0 50.8 52.9 52.2 53.3

Grade 5 57.2 55.9 55.9 56.6 57.1

Grade 6 58.0 57.4 55.8 56.7 57.4

Grade 7 54.8 57.1 57.2 53.4 54.4

Grade 8 NA NA NA NA NA

Grade 9 55.2 56.2 54.9 57.1 57.0

SAT-9 NCEs3th grade did the best7th grade needs to improve

+6.0+0.7

+0.8

+4.9

-2.3-0.2

Cohort Improvement

Expanded in 2006 MAP Terra Nova Percentiles SAT-10 NCEs

Combined across multiple measures

Expanded further in 2009 Terra Nova Percentiles Grades 1 – 5 STAR – Reading and Math Percentiles

Park

Hill

’s G

row

th H

isto

ry

Pre K 1 2 3 4 5 6 7 8 9 10 115055606570758085

2008-2009

Growth by Grade SpanPrimary Intermedi

ateMiddle High

School

Growth Percentiles

Started in 2008 Now using regression-based model with matched data

Produced by OSEDA Growth per student Pa

rk H

ill’s

Gro

wth

His

tory

Growth PercentilesPa

rk H

ill’s

Gro

wth

His

tory

Growth Percentiles

Used for: Program evaluation Subgroup analysis

Park

Hill

’s G

row

th H

isto

ry

EVAAS (Educational Value-Added Assessment System)

Started in 2009

Park

Hill

’s G

row

th H

isto

ry

What’s the Difference?Growth vs. Value Added

0

1

2

3

4

5

6

Park

Hill

’s G

row

th H

isto

ry

+3.6

What’s the Difference?Growth vs. Value Added

0

0.5

1

1.5

2

2.5

3

3.5

4

Park

Hill

’s G

row

th H

isto

ry

75th Percentile

What is Value-Added?

K 1 2 3 4 5 6 7 8 9 10 11 120123456789

101112

Average

Above Average

Below AveragePa

rk H

ill’s

Gro

wth

His

tory

What is Value-Added?

K 1 2 3 4 5 6 7 8 9 10 11 12404550556065707580

Average Above Average

Perc

ent P

rofic

ient

What’s the Difference?Growth vs. Value Added

0

0.5

1

1.5

2

2.5

3

3.5

4

Park

Hill

’s G

row

th H

isto

ry

v

EVAAS

Used for: Evaluating school effectiveness

Evaluating teacher effectiveness

Evaluating program effectiveness

Park

Hill

’s G

row

th H

isto

ry

Benefits of Growth/VA Models More accurate data More accurate inferences Better decisions Cultural shift toward growth/value

added Challenges traditional beliefs about

effectiveness Good school/bad school Achievement gap

Bene

fits

of G

row

th/V

A

Challenges and Considerations Missing data Teacher-student association accuracy Complexity of the regression

methodology Traditions/upsetting longstanding

beliefs

It is a more complete picture of what is happening inside the Black Box.

Bene

fits

of G

row

th/V

A

THE GROWTH OF VALUE-ADDED MODELS IN PARK HILLPhillip Adam – Coordinator for AssessmentPark Hill School District(816) 359-5508adamp@parkhill.k12.mo.us