### Dr. David Treder – Data Analysis Utilizing the New

```Michigan’s Accountability System
What Does It Take To
Be A Green School?
There were 98 “Green” Schools
schools:
-- 9th Grade Transition & Alternative Learning Center, Inkster
So, We’ll go with 96 Green Schools
Couple Quick Factoids
• Only 7 had a Top-to-Bottom Ranking
• Only 13 had AMO targets
96 Green Schools
28 - Had ZERO Possible Points (brand new in 201213, so no Proficiency data, no SIP Requirement, No
REP data Requirement
96 – 28 = 68
96 Green Schools
24 - Opened or reconfigured in Fall 2011, so their
AMO was zero
(AMOs were based on 2011-12 FAY percent proficient…
these schools had none, so actually their AMO was set
at the percent proficient in 2012-13)
68 – 24 = 44
Opened in 2011 (points awarded for zero proficient)
96 Green Schools
29 – Less then 30 FAY students (small schools)
- no bottom 30 % subgroups
44 – 29 = 15
AMO Targets for Small Schools
• A 95% confidence interval is used
• Compares school proficiency against a statewide target, based on based on the number
of FAY students:
> sqrt((p * (1 - p)) /n)
p = percent proficient, n = number of FAY students
• Confidence interval put around theTARGET, not the “sample” (Pct Prof of FAY kids).
• So, with 23 FAY Kids:
> 1.96* sqrt((.14*(1-.14)) / 23) = 14.2%
> Bottom end of the Math Target “Confidence Interval” = -0.2
> So, Zero Percent Proficient means the school made the target
96 Green Schools
6 – No AMO, with > 30 in 2012
- “Target” becomes small school target
- Again, bottom 30 subgroup can be
awarded 2 pts. with no students
proficient
15 – 6 = 9
(Not sure why Bottom 30% gets a green – a Sample Size of 12 is the
point where Zero kids will get you a green).
96 Green Schools
2 Closed in 2012, no data
available
9–2=7
Closed 2012
(same for all subjects)
96 Green Schools
6 – AMO targets for Math & Reading only
(No Bottom 30% in Science or Soc. Studies)
7–6=1
And the School Left Standing:
Webster Elementary, Livonia School District
Proficiency Rates
100% - Math, Soc. Studies, and Writing
94% - Science (bottom 30%, 82% proficient)
“To have ANY students proficient in the Bottom 30% subgroup, a
school needs to have 70% proficient in the ALL Students group”
This is not exactly accurate…
Bottom 30 Math
3
3
Math Std Error
46
50
Math Z Score
-1.609
-1.694
Math Proficient
Provisional
Provisional
But Pretty Close …
Genesee ISD-Wide Data
Math
- 32,337 FAY/Valid Math Scores
- Bottom 30%, proficient
2.8% growth
1.0% provisional
And REALLY Close in Science
- 13,399 Valid/FAY Science Scores
- 0.1% (20 students) Bottom 30%, Provisionally Proficient
- all -- but 1, from a gifted/magnet program -- had a SEM
of > 44
(13 had the maximum SEM of 50)
So, you want to be a
Green School next year?
It’s best, this year, you look like this
Arvon Township School
Skanee, MI 49962
Or This….
Accountability for Title III/Limited English
Proficient (LEP) Students
-- AMAOs (Annual Measurable Achievement Objectives)
AMAO 1: Progress
The district must demonstrate that the percentage of
its students making “progress” on the English Language
Proficiency Assessment (ELPA) meets or exceeds the
current year’s target.
Progress Calculation
The AMAO 1 progress calculation is done in the following manner for each
district:
1. Calculate the number of EL students in the district.
2. Calculate the number of those students gaining at least 4 points* on the ELPA
scale from the previous year (or a prior year).
3. Divide the result of (2) by the result of (1) to obtain the percentage of
students in each district making sufficient progress over the past year.
4. A district is identified as making sufficient progress if the target percentage of
its EL students gained a minimum of 4 points from the previous year’s (or a
prior year’s) ELPA.
*The rationale for choosing a cutoff of 4 points is that it is larger than the overall
standard error of measurement along the ELPA scale, indicating that if a student made
at least that much progress, it is attributable to student gains rather than measurement
error.
Statewide Mean Scale Scores, and Mean Across-Year “Growth” from Year to Year
State-Wide Mean Scale Score
K
531
2010
525
2011
525
2112
526
2013
529
2009 -2010
2010 -2011
2011 -2012
2012 -2013
1
559
556
556
559
560
25
31
34
34
2
586
586
591
593
590
27
35
37
31
3
607
606
608
607
609
20
22
16
17
4
620
621
622
622
622
14
16
14
15
5
629
630
632
632
632
10
11
10
10
6
614
614
615
615
617
-15
-15
-17
-16
7
621
629
623
624
623
15
9
9
8
8
628
629
632
631
630
8
3
8
6
9
631
628
632
632
633
0
3
0
3
10
637
637
640
639
639
6
12
7
7
11
643
643
646
645
643
6
9
5
4
12
643
644
648
647
643
1
5
1
-2
2006 Technical Manual, with Standard Setting
2006 Technical Manual, with Standard Setting, and 1st Years’ Scores
2010 Technical Manual: Mean Scale Scores
2013 Technical Manual: Mean Scale Scores
Purpose and Recommended Use
2007 Tech Manual
……………..
Because test results provide students, teachers, and parents with an objective report of
each student’s strengths and weaknesses in the English language skills of listening,
speaking, reading, and writing, the MI-ELPA helps determine whether these students are
making adequate progress toward English language proficiency. Year-to-year progress in
language proficiency can also be measured and documented after the MI-ELPA vertical
scale is successfully established.
2110 & 2013 Tech Manual
………………
This vertical development of the language tested allows the test to differentiate more
finely among students at different stages of language acquisition. Because test results
provide students, teachers, and parents with an objective report of each student’s
strengths and weaknesses in the English language skills of speaking, listening, reading,
writing, and comprehension, the ELPA helps determine whether these students are
making adequate progress towards English language proficiency.
A number of extensive, well documented Technical Manual have been done …
But don’t see anything on how the vertical scale was “successfully established.”
DESCRIPTIVE STATISTICS AT ITEM AND TEST LEVELS
4.1 Item-Level Descriptive Statistics
4.2 Higher-Level Descriptive Statistics
RELIABILITY
5.1 Classical True Score Theory
Internal Consistency Reliability
Standard Error of Measurement (SEM) within CTT Framework
5.2 Conditional SEM within the IRT Framework
5.3 Inter-Rater Reliability
5.4 Quality (or Reliability) of Classification Decisions at Proficient Cuts
CALIBRATION, EQUATING, AND SCALING
6.1 The Unidimensional Rasch and Partial Credit Models
6.2 Calibration of the Spring 2013 ELPA
Stability of Anchor Items
6.3 Scale Scores for the ELPA
6.4 Test Characteristic Curves for the ELPA by Assessment Level
IRT STATISTICS
7.1 Rasch Statistics
7.2 Evidence of Model Fit
7.3 Item Information
8.1 Validity Evidence of the ELPA Test Content
Relation between ELPA and Michigan English Language Proficiency (ELP) Standards
Validity of the Presentation of the Listening Stimulus
8.2 Validity Evidence of the Internal Structure of the ELPA
8.3 Validity Evidence of the External Structure of the ELPA
Relation between the ELPA and the MEAP Subject Tests
Relation between the ELPA and the MME Subject Tests
Distribution of Student Classification across Performance Levels
```