Research Support for Evidence-Based Education Policy: The North Carolina Education Research Data Center

1
Research Support for Evidence-Based Education
PolicyThe North Carolina Education Research
Data Center

• Clara G. Muschkin, Director
• November, 2009

Bridging the gap between research and public
policy to improve the lives of children and
families
2
NC Education Research Data Center- A Partnership
for Research

• A partnership of Duke University and the NC
  Department of Public Instruction
• Established in January 2001
• Renewed in January 2006 and 2009
• Funded by the Spencer Foundation

3
Why is the Data Center Necessary?

• Administrative data are a valuable resource for
  education research
• Research and evaluations require merging data
  from varied sources.
• The state does not have resources to support
  outside research.
• The state does not have resources to create new
  files by linking data from various sources.

4
A Fruitful Partnership

• Supported 120 original research projects
• Over a third are doctoral dissertations
• Over a third from institutions outside of NC
• Wide range of study topics relevant to education
  policy and reform in NC and beyond
• Unlimited data access for NCDPI research and
  evaluation staff

5
Ongoing Data Center Activities

• Acquire data collected annually by the NCDPI on
  all school districts, schools, students, and
  teachers
• Encrypt confidential data, use unique identifier
  to link information across sources and over time
• Develop standards for researcher access, review
  proposals, advise researchers on data use, and
  ensure commitment to data security.

6
Encryption

• Randomly assign student and teacher IDs
• Not derived from birth date, name, social
  security number, school
• Each year, when we receive new data, we match to
  the previous years
• Use name, student id, birth date, school and
  grade to match
• Assign new IDs to unmatched records

7
Consistency checking

• Check measures across files to ensure consistency
  (charter school codes)
• Check measures across years to note any changes
  (new values)

8
Codebooks

• Number of records
• Unit of analysis
• Variable names and values
• Changes or additions by ERDC

9

• END OF GRADE TEST 2008. GRADES 3, 4, 5, 6, 7, 8
•  
• Computer file name and date
• EOG3PUB08.sas7bdat. (End of Grade Test. Public
  Data. Grade 3. Year 2008.) June 30, 2009
• EOG4PUB08.sas7bdat. (End of Grade Test. Public
  Data. Grade 4. Year 2008.) June 30, 2009
• EOG5PUB08.sas7bdat. (End of Grade Test. Public
  Data. Grade 5. Year 2008.) June 30, 2009
• EOG6PUB08.sas7bdat. (End of Grade Test. Public
  Data. Grade 6. Year 2008.) June 30, 2009
• EOG7PUB08.sas7bdat. (End of Grade Test. Public
  Data. Grade 7. Year 2008.) June 30, 2009
• EOG8PUB08.sas7bdat. (End of Grade Test. Public
  Data. Grade 8. Year 2008.) June 30, 2009
•  
• Codebook date (most recent update)
• July 7, 2009
•  
• Summary
• Each student in grades 3 through 8 is tested in
  reading and math each year. The North Carolina
  Education Research Data Center has a separate
  file for each of these grades, and each file has
  one record for each student who was a member of
  that school at the time of the test. Data
  include records for students who were absent or
  exempt from the test for various reasons. If a
  student takes a retest, this information is
  included in a separate student record within the
  file.
•  
• The data include the following information the
  individual students test records the use of any
  testing modifications the students classroom
  activities, such as using calculators in math
  class or having written assignments in reading
  class and the students demographic data,
  including birth date, learning disabled status,
  and exceptionality status.
•  
• Origin

10

• MASTID Encrypted Student Identifier
• This variable is an encrypted master student id
  that permits identifying students with multiple
  test records in one year and tracking the
  students progress from one year to subsequent
  years, even if that student changed schools or
  districts. The NCERDC created this variable.
  For more information on how it was created,
  please see Technical Report 2 Creating a
  Longitudinal Student Database.
•  
• MATHABSA Student Absent from Math Calculator
  Active Test
• 1 Yes
• 0 No
•  MATHABSI Student Absent from Math Calculator
  Inactive Test
• 1 Yes
• 0 No
•  MATHACH Math Achievement Level
• Grade 3
• 1 I (311-328) Insufficient Mastery
• 2 II (329-338) Inconsistent Mastery
• 3 III (339-351) Consistent Mastery
• 4 IV (352-370) Superior Performance
•  
• Grade 4
• 1 I (319-335) Insufficient Mastery

11
Merging data

• Aggregate files to facilitate research
• EoG, EoC go from one file per test and school gt
  one file per test
• Licensure/Salary go from database with all
  transactions gt yearly snapshots

12
Merging data

• Merge existing files to new datasets requires
  written consent of participants
• Link academic records to participation in Talent
  Identification Program
• Link academic records to student surveys of their
  friends and behavior
• Link UNC system data to teacher records

13
NCERDC Holdings

• Full listing of data files provided on the Data
  Centers website
• nceddatacenter.com
• Temporarily
• childandfamilypolicy.duke.edu/project_detail.php?i
  d35

14
District Information

• Per pupil expenditure
• Salary Supplements
• District report cards
• LEA Universe

15
School Information

• Accountability score
• Public School Universe
• School Report Card
• Membership

16
Student Information

• End of Grade and End of Course Tests
• Dropouts
• Suspensions/ Expulsions
• Course Membership
• Exceptionality

17
Teacher Data

• Pay history
• Education
• Certification
• Tests (Praxis)
• Activity Assignment
• Working Conditions Survey

18
Linked Databases

• With files from the mid 1990s to the present, the
  Data Center can
• Track teachers careers over time.
• Track students progress over time.
• Aggregate student or teacher data to a school or
  district level of analysis.
• Link school, teacher, and student data to files
  from other sources.

19
Procedures for Obtaining Data

• Write a proposal stating the research questions,
  specifying the data elements that you will need,
  and how they will be used to address the
  questions.
• Prepare a data security plan.
• Sign a confidentiality agreement.
• Obtain your institutions IRB approval (must be
  reviewed).
• Agree to send DPI and the Data Center copies of
  any published paper.
• Ensure faculty sponsor for graduate student
  proposals.