Jim Lloyd_2007 Educational Value Added Assessment System (EVAAS) Phase 2 Presentation.

Jim Lloyd_2007

Educational Value Added Assessment System (EVAAS)

Phase 2 Presentation

Learning OutcomesUnderstand the difference between

achievement and growthUnderstand how teachers can have an effect

on growth varianceUnderstand that growth data is displayed in

various methodsAnalyze student value added dataBegin to consider how you could apply/use

value added data in order to leverage student growth

Value Added 101It is very complicated statistical procedure

to measure student growth, but it is highly reliable and valid.

It’s not meant to be a hammer or stick to prove that people are not doing their jobs. It’s another data tool to help us improve.

Why is measuring achievement & progress important?Gives a snapshot of achievement at a

particular point in time.

A balanced way of measuring student learning.

Achievement tests don’t tell growth.

A measure of progress begins to fill in missing pieces to the puzzle.

It is the law.

Suppose this box represents a group of students for a given

grade level and subject area…

Low

High

High

Low

… and we fast forward to the end of the given school year

An Achievement Status Consideration

Considerations for Gauging Effectiveness:

1) External Standard (proficiency level)

2) Student’s Ending Point of Achievement

? (Yes/No)

… sets a proficiency level students are expected to meet.

Teacher Effects

Factors related to student learning - District, School, and Teacher Influence

on Student ProgressFollowing inferences were shared at the

Governors Education Symposium (2004)

Based on 22 years of Value-Added Study, Dr. Sanders draws the following conclusions:

Variation in student academic progress can be attributed this way:

5% attributed to District quality

30% attributed to School quality

65% attributed to Teachers quality

Difference in student achievement levels after Experiencing qualitatively different teacher

sequences of effectiveness, Dallas

Source: Heather Jordan, Robert Mendro, & Dash Weerasinghe, “Teacher Effects On Longitudinal Student Achievement” 1997.

76

27

15

90

Beginning 3rd Grade

Score (Percentile)

Ave

rag

e M

ath

Sco

re 3

yrs

late

r in

Per

cen

tile

sDallas StudentsAssigned To 3Highly EffectiveTeachers In ARow

Dallas StudentsAssigned To 3IneffectiveTeachers In ARow

Rivers (1999)Can varying sequences of teacher quality

vary student chances of passing a high stakes test? o Rivers (1999) followed students from 4th through 9th grade

from two large urban districts.

o Students were grouped into quartiles on the basis of achievement on the 4th grade standardized test.

o Math teachers were identified for grades 5-8 and were linked to effectiveness levels (Low, Avg, High), based on their success in facilitating value-added achievement with a prior group of students.

o Do these quality distinctions apply to another cohort of students and offer reliable information about how a student will do on a high stakes in the future?

Probability of Passing High School Exam based on Sequence of Teacher Effectiveness Experienced

Jim Lloyd_2006

How is value added different than what we’re currently doing?

Current system measures only achievement (what percentage of kids pass the test).

New system measures progress towards the standard, so can more accurately reflect the growth of ALL students, especially those who may be hard-pressed to ever reach the passing rate.

Jim Lloyd_2006

How will this help me as an educator?Measures growth from

where the student enters the grade level.

Measures progress at building/grade levels.

Shows if programs/instruction are impacting student growth.

Aligns classroom resources to help all.

Jim Lloyd_2006

EVAAS

How to view a summary report

Jim Lloyd_2006

Concepts to RememberThis isn’t about proving…it’s about improving.

The power of the scores is in the sample size. The scores become less reliable as the size of the

population decreases.

We will not be looking at the individual growth of individual students because it is not reliable.

Jim Lloyd_2006

Colors to RememberA score of “0” isn’t bad because we have a 0

growth model which means that if all students grew as expected the color would be yellow.

Red isn’t goodYellow is no detectable differenceGreen is good

Jim Lloyd_2006

Terms Mean NCE Gain—the average gain that the group

achieved compared to where they started. This is the growth measure for the entire group of kids who have at least 2 data points.

Standard Error—a statistical calculation used to adjust for measurement error. As score dispersion increases, the SEM increases to adjust for this.

In EVAAS, we us this to determine how confident we are that the score didn’t occur due to error.Impacted by sample size and dispersionOhio’s model uses 1 SEM for significance

Jim Lloyd_2006

Terms cont.State Base Year—the average amount of growth that

happened by the grade in Ohio (expressed in NCE units)

2005 Mean—the average amount of achievement that the students came in with.

2006 Mean—the average amount of achievement that the students left with.

How is the student GAIN determined?It is the difference between the 2005 and 2006.

Types of ReportsCompositeDiagnostic Subject Specific

ReadingMath

Performance Subject SpecificReadingMath

Subgroup specific (reading/math)

Jim Lloyd_2006

Composite Reports

Jim Lloyd_2006

Jim Lloyd_2006

Composite ReportShows the total amount of growth by

combining all achievement tests administered.

We’re going to look at our composite report now.

What are Building Summary Reports?

Jim Lloyd_2006

Jim Lloyd_2006

Building Summary ReportsThere are summary reports for each building

as well that are subject specific.

Here is reading building summary report...

Jim Lloyd_2006

Subject Specific Reports

Jim Lloyd_2006

Jim Lloyd_2006

Subject Specific ReportsShows the amount of growth obtained in a

particular subject area.We’re going to look at a Summary Reading

and Math Report now.

Jim Lloyd_2006

What is this? Growth standard of 0 is what we would expect if the student made a years growth

Jim Lloyd_2006

EVAAS

How to view a Building Diagnostic Report

Jim Lloyd_2006

Pos Gain

Neg Gain

This shows us the # of OFCS students in each quintile

Total # of students in the testing pool is divided into 5 groups called quintiles

A “0 gain model” means that if students made about a year’s worth of growth we would see a gain within +1 or -1 of the standard error

Disaggregated Building Report

You can look at value added data for different reported groups as well.

Jim Lloyd_2006

Note that without at least 5 in a group, no data is reported

Disaggregated Building Report

You can ask to show you who the students are too

Jim Lloyd_2006

Example—Scattered spread w/a low performing spike

Jim Lloyd_2006

EVAAS Math Reports

District Reports 2006Current 6th Grade Students

Jim Lloyd_2006

Building Summary ReportThere are summary reports for each building

as well that are subject specific.Here is mathematics.

Jim Lloyd_2006

Math Diagnostic Reports

Jim Lloyd_2006

What is the gain pattern?

Remember to look @ these things

Implications of EVAASQuestions

What sorts of organizational strategies will lead to increasing district, building and teacher effectiveness?

What are the professional development implications of instituting EVAAS in Olmsted Falls?

What is the short and long term plan to prep the administrative team and building staff for this additional accountability system?

Responses Teachers and principals make

the difference. Build capacity! Make a concerted effort to get

ahead of the curve and use this and other data to our advantage.

Foster and reinforce a data-driven culture across the district.

Build upon and expand our staffs’ capacity to implement and monitor a standards-based educational program for all students.

What’s Next?EVAAS Phase 3

Review of Spring 2007 value added data

Discussion of the implications of the data

This number is underestimated and will likely change.