child outcomes: what we are learning from national, state, and local data

Presented at Division for Early Childhood National Harbor, Maryland

November, 20111

Child Outcomes: What We Are Learning from National, State, and

Local DataKathy Hebbeler

ECO at SRI International

Lynne Kahn ECO at FPG Child Development Institute

Topics

• Why outcomes data are collected?

• State approaches• National data• Use of the data at state

and local levels

2Early Childhood Outcomes Center

Let’s get acquainted


Early Childhood Outcomes Center

Why does the federal government want data on child outcomes?

• Government Performance and Results

Act (GPRA)

• Program Assessment Rating Tool

(PART)

• Individuals with Disabilities Education

Act (IDEA)4

Requires goals and indicators be established for

federal programs, including IDEA

Indicators and data collection for school age

population included data on outcomes

Previously, for early childhood data had been

reported on:

• Number of children served (Part C)

• Settings (both Part C and 619

Government Performance and Results Act (GPRA) passed in 1993

Early Childhood Outcomes Center

OSEP: PART evaluation results (2002)

130 programs examined in 2002; 50% programs had

no performance data

Programs looking at inputs, not results

Part C and Section 619

No child outcome data - “results not demonstrated”

Department of Education needs to develop a

strategy to collect annual performance data in a

timely manner

6

9

SEC. 616. <<NOTE: 20 USC 1416.>> MONITORING, TECHNICAL ASSISTANCE, AND ENFORCEMENT. ``(a) Federal and State Monitoring.…..

(2) Focused monitoring.--The primary focus of Federal and State monitoring activities described in paragraph (1) shall be on-- (A) improving educational results and functional outcomes for all children with disabilities;

IDEA 2004

The word “results” appears 65 times in the legislation.

Early Childhood Outcomes Center 8

$373,351,000

$438,548,000

State and Local Uses

• Accountability– Justifying the investment in EI and ECSE

• Program Improvement– Using data to identify program strengths and

share them– Using data to identify program weaknesses

and address them


Reporting Requirement for

Early Intervention and

Preschool Special Education


11

OSEP Reporting Requirements: Child Outcomes

– Positive social emotional skills (including positive social relationships)

– Acquisition and use of knowledge and skills (including early language/ communication and early literacy)

– Use of appropriate behaviors to meet their needs

What States Report:OSEP Reporting Categories

Percentage of children who:

a. Did not improve functioningb. Improved functioning, but not sufficient to move

nearer to functioning comparable to same-aged peers

c. Improved functioning to a level nearer to same-aged peers but did not reach it

d. Improved functioning to reach a level comparable to same-aged peers

e. Maintained functioning at a level comparable to same-aged peers

3 outcomes x 5 “measures” = 15 numbers12


0 6 12 18 24 30 36 42 48 54 60 660

10

20

30

40

50

60

70

Developmental Trajectories

Functioning like same aged peersImproved functioning to that of same aged peersMoved closer to function-ing like same aged peersImproved functioning, no change in trajectoryDid not improve function-ing

Age in Months

Gro

wth

in

Ou

tco

me

Reporting details

• Progress for all children who exited between July 1, 2009 and June 30, 2010

• Stayed in the program at least 6 months• Data reported to OSEP in February of

each year.• Progress data first submitted in 2008


The Summary Statements

1. Of those children who entered the program below

age expectations in each outcome, the percent who

substantially increased their rate of growth by the

time they turned 3 [6] years of age or exited the

program.

2. The percent of children who were functioning within

age expectations in each outcome by the time they

turned 3 [6] years of age or exited the program.


Formula for SS 1

(c+d)/(a+b+c+d)

1616

16

17

Formula for SS 2

(d+e)/(a+b+c+d+e)

17

Timeline


When Critical Event

January 2004 – January 2005 Stakeholder input gathered on 3 child outcomes

July 2005 (revised September 2006) OSEP releases reporting requirements for state programs

February 2008 States submit first data on 5 progress categories: Children who exited between July 1, 2006 and June 30, 2007

February 2010 States establish baseline and set targets on the Summary Statements for first time.

February 2011 States submit data on 5 categories for the 4th time.

State Approaches to Measuring Child Outcomes

Approach Part C

(56 states/jur)Preschool

(59 states/jur)

One tool statewide 7/56 (13%) 9/59 (15%)

Publishers’ online analysis

3/56 (5%) 6/59 (10%)

COS 7 pt. scale 41/56 (73%) 38/59 (64%)

Other 5/56 (9%) 7/59 (10%)

…and now there are national data


…one day we get a request

• From the U.S. Department of Education• Include the child outcomes data as a

GPRA indicator?• Also, in President’s budget justification for

Part C and Preschool 619 funding?• Initial request received in 2010, repeated

in 2011.


What would be the impact of state variation in data quality on the national

numbers?

The Dilemma

• Variations in quality of state data– Some states started earlier– Some states had devoted more attention to

improving quality


Our Response: Compute the analyses several ways

Identify the states with the highest quality data and use only their data. Stratify by number of children served and weight data to produce national estimate.


Use data from all states. Weight data to represent the nation.• Weighting necessary

because a few states are sampling. Also, many states not reporting data on all children.

Early Childhood Outcomes Center 24Note: Based on 29 States with highest quality data

Social relationships Knowledge and skills Action to meet needs

a 0.0165481765021159 0.0148974177446063 0.0145225115393538

b 0.179791044501508 0.199542744006336 0.174681665026758

c 0.185932320590048 0.247993957138401 0.214074779812331

d 0.294138593838746 0.369311305766878 0.370708975497373

e 0.323587775084597 0.168236271568793 0.22601676946091

3%

8%

13%

18%

23%

28%

33%

38%

Estimated Data for Part C, 2009-10


Note: Based on 29 States with highest quality data


SS1 0.70973347275746 0.742180533377054 0.755546347980436

SS2 0.61772636892334 0.537547577335669 0.596725744958282

5%

15%

25%

35%

45%

55%

65%

75%

Estimated Summary Statement Data for Part C, 2009-10



a 0.0173423475321513 0.0182367034048671 0.0164717096861928

b 0.114520004403404 0.133727833654954 0.108153036923834

c 0.281561785450637 0.32470410889358 0.208844242808853

d 0.343437919521324 0.34397760817479 0.356458315050575

e 0.243152978902516 0.179335862613809 0.310081127789363

3%

8%

13%

18%

23%

28%

33%

38%

Estimated National Data for Early Childhood Special Education, 2009-2010

Note: Based on 33 States with highest quality data

Early Childhood Outcomes Center 27Note: Based on 33 States with highest quality data


SS1 0.825777563121165 0.81482333429858 0.819365394291995

SS2 0.586590898423841 0.523313470788598 0.666539442839937

10%

30%

50%

70%

90%

Estimated National Summary Statements for Early Childhood Special Education, 2009-2010

Can we trust these data?


Pattern checking for validity

• Checking across years– How do the 2009-10 data compare to the

data for 2008-09?• Checking across methods

– How do the data for all states compare to states with highest quality data?


All states 19 best All states 29 bestFFY 08-09 FFY 09-10

0

20

40

60

80

64.570.2

64

71

SS1: % who Increased Growth Rates

Part C: Social Relationships

Part C: Social Relationships


0

20

40

60

80

6761.3

64.761.8

SS2: % who Exited at Age Expectations

Part B Preschool: Knowledge and Skills


0

20

40

60

80

100

76.282.7

78.381.5

SS1: % who Increased Growth Rates

Part B Preschool: Knowledge and Skills


0

20

40

60

80

100

55.851.2

55.5 52.3

SS2: % who Exited at Age Expectations

Possible interpretation of the data

• Nationally, a high proportion of children who receive Part C and ECSE services are showing greater than expected progress

• Nationally, many (over half) are exiting the program functioning like same age peers in at least one of the outcomes.


Would you agree?


Should each state’s data look like the national data?

• Probably not• More important that each state continue to

focus on the quality of its own data– Getting outcomes data on all children who exit– Working with programs whose data look

unusual to address possible data quality issues


For more information

• For updates to the framework and the self-assessment and resources to support the quality indicators:

www.the-eco-center.org


http://www.the-ec-center.org/

Additional Information


Criteria for States with Quality Data

1. Low percentage of missing data

2. No odd patterns in “a” or “e” categories

3. Did not use questionable data collection methods


Calculating Missing Data for Part C

Proxy for missing data =

Number with data for C3/

Exiting Data (618)


• Do not expect this number to be 100%

• ..but we don’t expect it to be 10% either

Part C: Percent of Exiters included in Outcomes Data

08-09<10% = 10*

10- 20% = 4

20- 30% = 8

30- 40% = 11

40- 50% = 8

50- 60% = 8

60- 70% = 4

70- 80% = 2

>80% = 1

09-10<10% = 5*

10- 20% = 4

20- 30% = 6

30- 40% = 8

40- 50% = 5

50- 60% = 11

60- 70% = 9

70- 80% = 1

>80% = 0

*3 states are sampling for Part C. Cut off was > 27%.

Calculating Missing Data for 619

Proxy for missing data =

Number with data for B7/

Child count


• Do not expect this number to be 100%

• ..but we don’t expect it to be 10% either

Odd Patterns in a or e

• a = % of children who show no new skills– Except this to be very small.

• e = % of children who maintained functioning comparable to age expectations.– Don’t expect this to be large.

• Quality defined as <10% in a and <65% in e.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49

0%

2%

4%

6%

8%

10%

12%

14%

16%

Percent Reported in "a“ for Knowledge and Skills for ECSE by State

Percent of Child Count included in Outcomes Data for ECSE

08-09<10= 11*

10- 20%= 15

20- 30%= 12

30- 40%= 12

40-50% =1

>50% = 2

09-10<10= 6*

10- 20%= 11

20- 30%= 12

30- 40%= 16

40-50% =4

>50%= 0

*4 States are sampling for 619Cutoff was > 11%.

child outcomes: what we are learning from national, state, and local data

Documents

early childhood data

topicswhy outcomes data

child outcome data results

annual performance data

family outcomes

functional outcomes

data collection

themusing data