Batya Elbaum, NCSEAM Pam Roush, West Virginia Part C OSEP National Early Childhood Meeting
description
Transcript of Batya Elbaum, NCSEAM Pam Roush, West Virginia Part C OSEP National Early Childhood Meeting
How to Explain the Numbers:Helping Staff, Parents, and Other
Stakeholders Understand the Results of the NCSEAM Surveys for
Part C and 619
Batya Elbaum, NCSEAMPam Roush, West Virginia Part COSEP National Early Childhood
MeetingArlington, VA, December 2007
Purpose of the session To provide participants with
strategies for explaining the measures and percentages that come from the NCSEAM rating scales addressing Indicators C4 and B8.
Part C Indicator #4
“Percent of families participating in Part C who report that early intervention services have helped the family a) know their rights, b) effectively communicate their children’s needs, and c) help their children develop and learn.”
Part B Indicator #8
“Percent of parents with a child receiving special education services who report that schools facilitated parent involvement as a means of improving services and results for children with disabilities.”
WV Part C Indicator 4 4A – Know rights
2006 76.5% 2007 73.8%
4B – Communicate 2006 72.8% 2007 71.0%
4C – Help child 2006 86.3% 2007 78.7%
Change in mean measure by region
Region 2006
N 2006 Mean
2007 N
2007 Mean
Change in Mean
1000 Part C Impact on Family Measures 62 716.41 44 705.32 -11.09
1100 Part C Impact on Family Measures 64 702.91 44 715.36 +12.45
1200 Part C Impact on Family Measures 50 665.87 47 670.60 +4.73
1300 Part C Impact on Family Measures 93 717.19 58 688.85 -28.33
1400 Part C Impact on Family Measures 28 755.98 24 637.00 -118.98
1500 Part C Impact on Family Measures 25 615.88 25 627.71 +11.84
1600 Part C Impact on Family Measures 48 695.17 33 651.53 -43.63
1700 Part C Impact on Family Measures 90 706.44 53 681.74 -24.69
Percent on 4A by Region in 2006 and 2007
2006 2007
Region N 4A % N 4A %
1000 62 72.58 44 70.45
1100 64 71.87 44 86.36
1200 50 72.00 47 74.47
1300 93 .78.49 58 77.59
1400 28 92.86 24 58.33
1500 25 60.00 25 64.00
1600 48 79.17 33 63.64
1700 90 81.11 53 79.25
Why use a measurement analysis?
Why use a measurement analysis?
Use of a measurement framework ensures that measures will mean the same thing, regardless of how many items, or which specific items, are administered.
Why use a measurement analysis?
We can’t assume that all survey items are equally agreeable.
A measurement analysis gives us a measure of each item’s overall agreeability.
Order of Impact on Family items from least to most agreeable
Item Calibration
Item Stem: Over the past year, Early Intervention services have helped me and/or my family:
678 Participate in typical activities for children and families in my community.
656 Know about services in the community.
640 Know where to go for support to meet my family's needs.
625 Keep up friendships for my child and family.
609 Know where to go for support to meet my child's needs.
584 Be more effective in managing my child's behavior.
576 Make changes in family routines that will benefit my child with special needs.
576 Do activities that are good for my child even in times of stress
570 Improve my family's quality of life.
565 Feel that I can get the services and supports that my child and family need.
563 Get the services that my child and family need.
562 Feel that my family will be accepted and welcomed in the community
559 Feel more confident in my skills as a parent.
559 Feel that my child will be accepted and welcomed in the community.
556 Communicate more effectively with the people who work with my child and family.
553 Understand how the Early Intervention system works.
546 Understand the roles of the people who work with my child and family.
539 Know about my child's and family's rights concerning Early Intervention services.
534 Be able to evaluate how much progress my child is making.
516 Understand my child's special needs.
498 Feel that my efforts are helping my child.
498 Do things with and for my child that are good for my child's development.
Why use a measurement analysis?
Use of a measurement framework allows us to test whether all the items are measuring the same thing.
Measuring the 3 subindicators
NCSEAM’s measurement analysis of pilot data from thousands of families showed that items that are related to the three OSEP outcomes all fit into a single scale of family outcomes.
Measuring the 3 subindicators
Families appear to achieve these outcomes in a very consistent order. Families who report that EI helped them
know their rights also report that EI also helped them help their child develop and learn.
Families who report that EI helped them effectively communicate their children’s needs also report that EI also helped them know their rights and help their child develop and learn.
Location of key items related to Indicator C4
Indicator 4c: Help their children develop and learn. IFS items: “Understand my child's special needs.” [516] “Do things with and for my child that are good for my child's development.” [498]
Indicator 4b: Effectively communicate their children’s needs. IFS Item: “Communicate more effectively with the people who work with my child and family.” [556]
Indicator 4a: Know their rights.IFS Item: “Know about my child's and family's rights concerning Early Intervention services.” [539]
556
539
516
Why use a measurement analysis?
We can’t assume that the response choices (very strongly disagree, strongly disagree, disagree, agree, strongly agree, very strongly agree) indicated by equidistant circles or numbers on a page are really equidistant.
Spacing of response choices
◘ ◘ ◘ ◘ ◘ ◘ 1 2 3 4 5 6
Spacing of response choices
◘ ◘ ◘ ◘ ◘ ◘ 1 2 3 4 5 6 1K 5K 10K ½ Mar. Marathon Ultra
The logic of distributions
Example: Distribution of height
Distribution of Measures of Reading Achievement - 1
X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X >5’ 5’0 5’2 5’4 5’6 5’8 5’10 6’0 6’2 6’4 6’6 6’8 6’10 7’ >7’
Distribution of Measures of Reading Achievement - 2
X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X >5’ 5’0 5’2 5’4 5’6 5’8 5’10 6’0 6’2 6’4 6’6 6’8 6’10 7’ >7’
West Virginia results 2006
West Virginia results 2007
WV without extreme cases 2006
WV without extreme cases 2007
Instrumentation and approach to data analysis make a difference
Choice of instruments: Part B NCSEAM K-12 and 619 9 NCSEAM K-12 21 Customized NCSEAM survey 11 State-developed or adapted 18 ECO Family Outcomes Survey 1
States’ reported baseline data onIndicator 8B
Choice of instruments: Part C NCSEAM Survey 25 ECO Family Outcomes Survey 18 State-developed or adapted 10 Combination 1
Analysis of states’ baseline data:Indicator C4
Sub-Indicator a. Know their
rights b. Communicate children’s needs
c. Help children develop and learn
Mean 79% of families 78% of families 85% of families
Range 45% - 99% 42% - 99% 53% - 99%
Analysis of states’ baseline data:Indicator C4
Figure 2: Baseline by Survey Tool and Criteria for Positive Response
0
10
20
30
40
50
60
70
80
90
100
ECO 3 & above
N=2
ECO5 & above
N=16
NCSEAM alternate analysis
N=6
NCSEAM standard analysis
N=17
State Survey
N=11
Tool and Analysis
% a
% b
% c
APR Family Outcome Results – Part C States Using NCSEAM Survey and RASCH analysisAPRs submitted February 2007
4A 4B 4C Response Rate
#
California 48% 42% 71% * 5,413 Connecticut 73% 71% 84% 27% 712 Florida 56% 52% 57% 25% 402 Idaho 56% 52% 68% 14.74% 355 Kentucky 82% 73% 89% 22% 786 Mass. 80% 77% 86% 30% 290 Michigan 56% 51% 73% 40.8% 2,439 Nebraska 76% 73% 86% 48% 921 New Mexico 78% 72% 92% 23.8% 183 New York 71% 65% 82% 25.6% 580 Pennsylvania 67% 63% 79% 23.81% 3,018 Utah 74% 71% 79% 25.4% 605 Wyoming 78% 85% 86% 30% 230 South Carolina 74% 70% 86% ? ?
Reference to a standard
Percent of measures above the adopted standards
standards
What is the relationship between % at or above the NCSEAM standard and % agreement on the “threshold item?”
Sample descriptive interpretation related to Indicator 4A Approximately 90-95% of families agreed,
with approximately two-thirds of families expressing strong or very strong agreement, that early intervention helped them:
Understand their child’s special needs. Do things with and for their child that
are good for their child’s development. Feel that they can handle the
challenges of parenting a child with special needs.
Part C Example: WV data
Q44 - know about my child's and family's rights
AKnow Response Category Frequency Percent
% Agree, SA, VSA
1 5 6.0 2 1 1.2 3 1 1.2 4 65 78.3 5 9 10.8
6 2 2.4 91.5 Total 83 100.0
Below cut-score on this standard
Missing 3
Total 86 4 15 6.4 5 57 24.5
6 161 69.1 100.0 Total 233 100.0 Missing 9
At or above cut-score on this standard Total 242
Q42 - communicate more effectively
BComm Response Category Frequency Percent % Agree, SA,
VSA 1 4 4.8 2 3 3.6 3 6 7.1 4 67 79.8
5 4 4.8 84.6 Total 84 100.0 Missing 11
Below cut-score on this standard Total 95
3 1 .5 4 15 7.0 5 62 29.0
6 136 63.6 99.5 Total 214 100.0 Missing 19
At or above cut-score on this standard
Total 233
Q46 - understand my child's special needs
CHelp Response Category Frequency Percent
% Agree, SA, VSA
1 6 9.2 2 1 1.5 3 4 6.2 4 50 76.9 5 4 6.2 83.1 Total 65 100.0 Missing 5
Below cut-score on this standard
Total 70
3 1 .4 4 24 10.4 5 52 22.5
6 154 66.7 99.6 Total 231 100.0 Missing 27
At or above cut-score on this standard Total 258
Part B ExampleItem: “The school explains what options parents have it they disagree with a decision of the school.”
SEPPS Measure
% VSD
% SD
% D
% A
% SA
% VSA
Total % A/SA/VSA
>=350<400 21.9 8.8 47.4 20.2 0.9 0.9 22.0 >=400<450 7.6 6.7 51.4 33.3 0.0 1.0 34.3 >=450<500 4.1 2.0 27.6 57.1 7.1 2.0 66.2 >=500<550 3.3 3.3 18.0 62.0 8.7 4.7 75.4 >=550<600 0.0 0.0 16.0 54.0 16.0 14.0 84.0 >=600<650 0.0 3.1 4.6 26.2 44.6 21.5 92.3 >=650<700 0.0 0.0 3.3 30.0 20.0 46.7 96.7 <600 (n=704) 704/903 = 78%
22.2 6.3 28.4 35.9 4.4 2.8 43.1
>=600 (n=199) 199/903 = 22%
0.5 1.0 4.5 15.6 24.1 54.3 94.0
ALL 17.4 5.1 23.1 31.5 8.7 14.2 54.4
The NCSEAM standard is a stringent standard Cut score for B at 600 Cut scores for C at 539, 516, 556 The standard is set so as to ensure
that approximately 95% or more of the responses will be in one of the agree categories on the item designated as the “threshold item”
Simple agreement with the “threshold item” is not enough
How precise are the percentages that states are reporting?
Error in estimation Whenever data are based on a
sample (whether owing to sampling or to a less than 100% return rate), there is some amount of error in generalizing to the population.
“Error,” or imprecision, can be represented by a confidence interval.
As sample size decreases, error of estimate - and its representation as a confidence interval - increases.
The decrease is not linear!