AYP Workbook Membership and Ncount Revisited

1
AYP WorkbookMembership and N-count Revisited

• CCSSO-ASR
• Orlando, FL
• 13 JUNE 2008

Research in Action, Inc.
2
So Just Tell Me with the Data

• Further investments in the technology and human
  infrastructure underpinning education
  accountability and improvement efforts will be
  necessary. However, most elected officials will
  not have the expertise to grasp the technical and
  operational trade-offs associated with these
  investments.
• (Robert Palaich, Dixie Good, and Arie van der
  Ploeg, June 2004,pg. 1)

3
Going Retro with Robin RolfMinimum N-counts
4
Going Retro with Robin RolfMinimum N-counts
5
Going Retro with Robin RolfMinimum N-counts
6
Going Retro with Robin RolfMinimum N-counts
Rich Hill (Center of Assessment), 2002
Simulation When a minimum N of 10 was used, the
correlation increased to .50, and to .56 when the
minimum N was 20. A consistent decision was
made for about 70 percent of the schools (254 out
of 814 either passed one year and failed the next
or vice versa). When the minimum N was 10,
this decreased to 197 out of 814 for a minimum
N of 20, the result was 167 out of 814.
7
FlashbackRationalizing the Decision
Sampled over 100,000 records for a single
year Removed any record not meeting FAY
requirement (n3,227) Examined n-count and
multiple membership for 2007 Final sample had
97,729 students and 606 schools
8
Minimum N-counts
9
Minimum N-countsEconomically Disadvantaged
10
Minimum N-countsStudents with Disabilities
11
Minimum N-countsEnglish Language Learners
12
Going Retro with Robin RolfMultiple
Membership
13
Going Retro with Robin Rolf2002
14
Going Retro with Robin Rolf2006
15
Going Retro with Robin Rolf2006
16
Multiple Membership 3-4-5 Groups