The Super Safety Net of Safe Harbor

1
The Super Safety Net of Safe Harbor

• Peter Cincotta
• Office of Accountability, Research, and Testing
• Baltimore County Public Schools

2
Objectives

• Overview of No Child Left Behind (NCLB)
  accountability process in Maryland
• Explanation of Safe Harbor process and how it
  fits into the determination of Adequate Yearly
  Progress (AYP).
• Share some interesting findings on the
  proficiency rates needed to make AYP through Safe
  Harbor.

3
Adequate Yearly Progress

• NCLB provides flexibility to states in many areas
  of the law.
• The Maryland Department of Education (MSDE) has
  taken advantage of these areas.

4
(No Transcript)
5
Safe Harbor

• The principle behind Safe Harbor is to recognize
  subgroups (and/or schools) that demonstrate a
  significant increase in achievement even if the
  achievement does not reach the level of the
  Annual Measurable Objective (AMO).

6
Safe Harbor
AMO 53.6
40.3
15
7
Safe Harbor

• The rule is
• The percentage of students scoring in the basic
  category (the basic rate) must decrease by 10
  from the previous year.

8
Safe Harbor

• Example
• In 2004 the Special Education subgroup for
  Mathematics at Ann M. Wells Elementary had a
  proficiency rate of 8.0.
• Therefore, this subgroup had a basic rate of
  92.0. (100.0 8.0 92.0)
• This basic rate must decrease by 10.

9
Safe Harbor

• 10 of 92.0 is 9.2
• That is 10 X 92.0 9.2
• Therefore, the basic rate must decrease by 9.2
  percentage points.
• 92.0 - 9.2 82.8 .
• A basic rate of 82.8 would be a proficiency rate
  of 17.2.
• 100.0 - 82.8 17.2

10
Safe Harbor
40.3
AMO 53.6
15
11
Safe Harbor
Ann M. Wells Elementary School 2005 AYP
Mathematics
AMO 53.6
40.3
17.2 - Safe Harbor Proficiency Rate target
15
8 - Proficiency rate last year
12
But wait. Theres more!
13
Safe Harbor

• MSDE asked for and received permission to use a
  confidence interval around the Safe Harbor
  Proficiency Rate target

14
Safe Harbor

• The length of a confidence interval is dependent
  on the size of the subgroup.
• Large subgroups produce small confidence
  intervals
• Small subgroups produce large confidence intervals

15
Confidence Intervals
Small groups have larger confidence intervals
Large groups have smaller confidence intervals
AMO is 43.8
16
Safe Harbor

• Lets say there are 100 Special Education
  students among grades 3, 4, and 5 at the Ann M.
  Wells Elementary school.

17
Safe Harbor
Ann M. Wells Elementary School 2005 AYP
Mathematics
AMO 53.6
11.0 is the lower end of the Safe Harbor CI with
a subgroup size of 100 students and a proficiency
rate of 8 last year.
40.3
17.2 - Safe Harbor Proficiency Rate target
15
8 - Proficiency rate last year
18
Safe Harbor

• Therefore, this subgroup made AYP through the
  Safe Harbor provision
• To recap, this proficiency rate was
• Lower than the AMO,
• Lower than the lower end of the (AMO) CI,
• Higher than last years proficiency rate, and
• Higher than the lower end of the Safe Harbor CI.

19
Safe Harbor

• Two strings attached
• All Students must make AYP in the same subject
  of the Safe Harbor subgroup
• Other Academic Area must improve over last year
  for the Safe Harbor subgroup.

20
Safe Harbor
Ann M. Wells Elementary School 2005 AYP
Mathematics
AMO 53.6
Notice how close the lower end of the Safe Harbor
CI is to last years proficiency rate.
40.3
15
8 - Proficiency rate last year
21
Safe Harborwith a Proficiency rate of 8 last
year
22
Safe Harbor

• Under some circumstances, a very small increase
  in the proficiency rate would be enough to make
  AYP through Safe Harbor.
• Under some circumstances, even flat achievement
  or a decrease in the proficiency rate would be
  enough to make AYP through Safe Harbor.

23
Minimum Proficiency Rates Needed to make AYP with
Flat Achievement
24
Lower End of Safe Harbor CI for a subgroup of 100
students
25
Even large groups will require a relatively small
increase to make AYP
26
Safe Harbor

• Note that Safe Harbor is determined based on two
  data points
• Proficiency rate last year, and
• Subgroup size this year.

The AMO is not part of the calculation.
27
Contact Information

• Peter Cincotta
• Office of Accountability, Research, and Testing
• 1940 Greenspring Drive
• Timonium , Maryland 21093
• (410 ) 887 7755 , ext. 209
• Fax (410 ) 561 5769
• E-mail pcincotta_at_bcps.org