1

P20 SLDS – How to Get Started2012 MIS Conference

Jeff SellersSLDS State Support Team

Carol JennerWashington Education Research & Data Center

Charles McGrewKentucky P-20 Data Collaborative

Photo credit: iStockPhoto.com

2

Carol JennerWashington Education & Research Data Center

3

The Washington Setting• Education Research & Data Center (ERDC) is the focus of P

20/W work in Washington State. • The Office of Superintendent of Public Instruction (OSPI) is

developing a K-12 SLDS, which will serve as the source of K-12 data for the P-20/W system.

• ERDC is a state education authority by statute, based in the research division of the Office of Financial Management (also home to the state’s unit-record public baccalaureate data system).

• ERDC has been funded by the Legislature since 2007 and has had 2009 ARRA SLDS grant funding since 2010.

• We anticipate that we will soon begin to build our data warehouse.

4

Think P-20!

• P-20/W: Longitudinal, cross-sector work– Sectors: early learning, K-12, higher education (2/4-year), workforce

• Focus on cohorts – a group of individuals sharing a common characteristic and observed over time (longitudinally) as a group. Two examples:– Graduates from a particular year– Entering 9th graders in a particular year

• Two approaches– Outcomes for a cohort (looking forward)– Previous experiences of a cohort (looking backward)

• Focus on transitions across sectors: entering Kindergarten students, transition out of high school, transition from 2-year to 4-year higher ed, bachelor’s degree recipients, early leavers

5

P-20 = Longitudinal, Cross-Sector

Year-to-year change in 4th grade assessment scores Longitudinal Cross-sector Trend P-20

High school assessment scores of 2009 graduates taking pre college mathematics in Fall 2009

Longitudinal Cross-sector Trend P-20

Year-to-year postsecondary enrollment rate of previous spring high school graduates

Longitudinal Cross-sector Trend P-20

6

The Foundation: Identity-Matching

Personal information• Name, birthdate, ID numbers, demographic characteristics

Contextual information• Location, entry and exit dates, program participation, graduation status,

GPA

Deterministic and probabilistic approaches• Currently using Link Plus software and home-grown routines

Use of reference data• Confirmation of birthdate, last 4 of SSN, name changes

An on-going process

7

Washington P-20 Concepts• Person-Role-Organization (PRO) conceptual model

– Roles include student, teacher, employee, client– Organizations include schools, colleges, classes, firms– A person may have several roles simultaneously, including several roles within

an organization– Program participation, assessments, achievements, characteristics, and

identifiers are all related to Person-Role-Organization

• Each person is associated with one P-20 ID– the person entity is used only to assign a unique P-20 ID since all

characteristics and events are captured at the PRO level

• The P-20 ID is the product of identity matching of data both within an organization and across organizations

• Linking – using the P-20 ID to relate data for each individual across education sectors and employment

8

CohortsCohorts may be• Fixed – the members do not change• Updatable – members may change (based on identity-matching updates,

receipt of, for example)

Cohorts may be• Standard – well-defined and pulled on a routine basis for reporting (high

school graduation class, for example)• Ad hoc – the result of a special query or of a submitted list of individuals

(bachelor’s degree recipients who started at a community college or participants in a particular after-school program, for example)

Individual characteristics for cohort members• Retained at the Person-Role-Organization level (usually no attempt to

determine a “best” value for a characteristic across roles and organizations)

9

Benefits of Centralized P-20

• Dedicated to true P-20/W work – research that addresses critical questions. No within-sector transactional activity conducted using the P-20/W data warehouse.

• Lean and mean – only elements required for P-20 research and policy analysis are included in data warehouse.

• Investment in identity-matching is retained and improved over time.

• Prior studies can be replicated – “from” and “to” dates stored for all elements.

• Persistent research IDs can be maintained, allowing for updating of research data sets.

10

What’s Working for ERDC• Engaged and supportive leadership• Communication

– ERDC Newsflash, press releases– Well-defined critical research and policy questions– Persistence is critical

• Data governance– Organized consensus-based approach to data governance– Master data-sharing agreement template covering release of de-identified

data (data going out)– Standard request process that applies to all

• Data familiarity attained through use of data• Use of data to inform data warehouse development – we are

“data-ready” for mapping

Data Governance Committee StructureOffice of Financial

Management

Education Research& Data Center (ERDC)

Data StewardsCommittee

Experts directly familiar with data from their agency used in research.

Data CustodiansCommittee

Technical experts responsible for the technical delivery of data to and from the warehouse.

Research & ReportingCoordination CommitteePolicy experts who interact with agency decision-makers, stakeholders, and researchers.

ERDCGuidance Committee

Agency directors or deputies from agencies contributing data

12

Sustainability

• ERDC was established as a result of Washington Learns, a governor-led 18-month review of the state’s entire education system, including workforce (2005-2006).

• ERDC with its focus on P-20/W has been funded since its creation in 2007. The legislature has already “bought in.”

• 2009 ARRA SLDS grant funding has accelerated the ERDC agenda.• Increased demand for P-20/W information and efficiencies of a

centralized data system will eliminate agencies’ needs for contracting out for similar work.

• Productivity will increase once data warehouse is in production with topic-specific data marts.

Teachers Who Leave Teaching

Teachers who left teaching jobs matched against administrative records from Unemployment Insurance Program and Department of Retirement Systems.

Source: “Who Leaves Teaching and Where Do They Go?,” Washington Education Research & Data Center, January 2011.<http://www.erdc.wa.gov/briefs/>

Retired30.0%

UI benefits0.4%

Not located21.9%

Outside public edu-cation14.1%

Public education28.4%

Both public & non public education

5.2%2006n=4,291

Retired25.5%

UI benefits1.8%

Not located23.2%

Outside public education10.0%

Public education33.9%

Both public & non public education

5.6%2008n=3,411

“Feedback” Reports

1. High School Feedback Reports

2. Reports for community & technical colleges- Incoming students- Transfers to baccalaureate institutions

3. Reports for baccalaureate institutions- Incoming students- Workforce outcomes

erdcdata.wa.gov

Adding Student Characteristics

See www.erdc.wa.gov/briefs/pdf/201005.pdf for full report.

Low-Income Not Low-Income Total0%

20%

40%

60%

80%

100%

31% 31% 31%

11%22% 19%2%

3%3%

5%

12%11%

50%

31% 36%

All high school graduatesBy income status and type of institution.

(Universe: 2008-09 public high school graduates)

Low Income Total0%

20%

40%

60%

80%

100%

22%29%

63% 49%

4%5%

11% 17%

Distribution of enrollmentBy income status and type of institution.

(Universe: 2008-09 public high school graduates enrolledin any postsecondary institution)

Not enrolled Out of state Washington private Washington public 4-year Washington CTC

Adding Community Characteristics

Community-Level Median Household Income(for households with children)

Census 2000 Educational Attainment and Median Household income at school district level obtained from NCES School District Demographics System. nces.ed.gov/surveys/sdds/index.aspx

Lowest Highest0%

20%

40%

60%

80%

100%

1,429 2,007 2,176 2,6633,722

3,7324,220 4,238

4,0453,594

601919

10681108

21426,058

5,557 4,742 4,0413,155

Lowest Highest0%

20%

40%

60%

80%

100%

1,871 2,027 2,009 2,4303,655

3,768 3,978 4,2294,014

3,834

739970 1010

10462072

5,691 5,206 5,158 4,1523,327

Educational Attainment Household Income

(Universe: 2008-09 public high school graduates in districts where educational attainment data are reported)

(Universe: 2008-09 public high school graduates in districts where median household income is reported)

Not enrolled Out of state Washington private Washington public 4-yearWashington CTC

Percent of Population Age 25 and OverWith Bachelor’s or Higher Degree

Student Employment

GPA <2.00

GPA 2.00 - 2.49

GPA 2.50 - 2.99

GPA 3.00 - 3.49

GPA 3.50 - 4.00

Total

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

56%

58%

61%

59%

52%

57%

5%

5%

5%

6%

7%

6%

39%

37%

34%

35%

41%

37%

Employed during school year Employed summer only Not employed

School-Year Employment (in high school) by Student GPA at H.S. GraduationUniverse: 2009 High School Graduates for whom employment status could be determined

Pre-College Mathematics

Lower-incomeschool

Medium-incomeschool

Higher-incomeschool

0%

10%

20%

30%

40%

50%

60%

42%35%

27%

StudentsLow-income

StudentsNot low-income

0%

10%

20%

30%

40%

50%

60%

45%

31%

Series10%

10%

20%

30%

40%

50%

60%

48%43%

38% 38%33%

26%

Lower-in-comeschool

Lower-incomeschool

Middle-incomeschool

Middle-incomeschool

Higher-incomeschool

Higher-incomeschool

Students: Not low-incomeStudents: Low-income

By school income level [based on percentage of graduates eligible for free or reduced-price lunch (FRPL)]

By student FRPL eligibility

By combination of school income level and student low-income status

Pre-College Math Course-Taking Rates,2008-09 Public High School Graduates enrolled in public postsecondary institutions (both CTCs and 4-year institutions) in 2009-10.Student low-income status is tied to FRPL eligibility.Schools are classified into three categories based on the percentage of graduates FRPL-eligible. Approximately 31,000 high school graduates are included in this analysis.

1 2

3

Uses: CCER Demonstration Project

The Community Center for Education Results (CCER) is a community partnership focusing on increasing student achievement in South Seattle and South King County, Washington. For complete CCER report see www.erdc.wa.gov/briefs/pdf/R201101_CCER_Demonstration_Project.pdf.

20

Memorandum of UnderstandingResponsibilities and Principles for Sharing and Using P-20 and

Workforce Data

Principle 1: Education Research and Data Center provides cross-sector, linked data to all data consumers in a consistent, transparent way.

Principle 2: Education Research and Data Center maintains the P-20 and workforce data warehouse.

Principle 3: Protecting the privacy of individuals is a priority. Principle 4: Partner agency data contributors (at the state and local

levels) are experts at understanding and explaining the data.

Principle 5: Common understanding and use of data increases its value.

Complete document: www.erdc.wa.gov/datasharing/pdf/workgroup/mou_final_201109.pdf

21

For More Information about P-20 in Washington

Carol JennerEducation Research & Data Center

Carol.Jenner@ofm.wa.govwww.erdc.wa.gov

22

Charles McGrewKentucky P-20 Data Collaborative

23

Kentucky’s Data Sharing History

• ACT High School Feedback Reports (1990+)

• The MAX Project (2003)

• KEN Project and P-20 (2005/07)

• K-12’s KIDS Project – 1st Round SLDS Grants (2006)

• Kentucky’s High School Feedback Reports (2007)

• Postsecondary Migration Study (2007)

• P-20 Data Collaborative -- SLDS Grants (2009)

• Many other items – financial aid ad hoc studies, etc.

24

Kentucky P-20 – Funding and SustainabilityCurrent and Previous Funding

• K-12’s KIDS Project – 1st Round SLDS Grants (2006)

• P-20 Data Collaborative – SLDS Grants (2009)

Creating a Sustainable Business Model

• Governor’s proposed budget includes $600,000 per year general funds recurring

dollars to sustain the P-20 Collaborative’s work.

• Early Childhood funds support part of the infrastructure and shared component

costs.

• Evaluating other options including charging fees, seeking additional grant

opportunities, and partnering on grants.

25

P-20 Data Collaborative Governance

Joe Meyer (chair)Education and Workforce

Development Cabinet Secretary

Dr. Terry HollidayKentucky Department of Education, Commissioner

Dr. Phil RogersKentucky Education Professional

Standards BoardExecutive Director

Robert KingCouncil on Postsecondary

Education,President

Kentucky’s P-20 Data Collaborative is governed by a committee of four with the three founding agency heads and is chaired by the Education and Workforce Development Secretary. All project and fiscal decisions are made by this group and must be unanimous.

26

Legal Authority• Collaborative established through an MOA in 2009 for

agencies participating in the grant.• P-20 Shared Repository established in 2010 by an

Executive Order with authority to match data across collaborative and other agencies.

Data Ownership• Each participating agency is a member. Each member

owns and maintains control of its own data.

Data Governance

27

What is the Role of P-20?Provide better, more timely information to inform policy makers than has ever been available before by:

1. Providing a secure way to link data across agencies to show a more complete picture than any agency can do;

2. Identifying the state’s critical policy questions and developing processes to answer them either by the agencies or by bridging the gaps between agencies; and

3. Working with policy makers and stakeholders to provide the information they need in the format they need it to improve programs and services. This requires tools and expertise that each agency may not possess.

P-20 is a complement and not a replacement for agency data systems unless the agencies want it to be otherwise.

28

What is the Role of P-20?P-20 can centralize a number of functions that would be much more work if each agency did it themselves such as:

1. Focusing more time and resources on processes related to data matching and de-identification than agencies can currently do.

2. Providing state-level, big-picture perspective of the state’s needs without being too entrenched in a single agency perspective.

3. Maintaining resources for analysis and research centrally that can be used by all the agencies and other groups.

29

• Providing objective, comparable, actionable data back to schools, districts, colleges, universities, and communities that they can use to improve education.

• Maintaining linked data that allows for cross-agency statistics like linking high school performance to college success or linking postsecondary training to employment and earnings.

• Answering state and federal mandated reports that cross sectors.

• Providing greater transparency both for development and improvement purposes as well as for policy makers.

P-20 Benefits

30

What are the Questions?

How well do AP, IB, and dual credit programs improve college going and college success and shorten “time to degree?”

What factors in high school are better predictors of college and career success?

How much do our college graduates earn and how long does it take for them to find full-time work?

Which early childhood programs have the greatest impact on preparing students for Kindergarten?

31

What are the Questions?

How many high school graduates leave Kentucky to go to college and return here to live and work?

How successful are programs like GearUp, TRIO, etc. in terms of encouraging more disadvantaged students to attend and succeed in college?

What proportion of our college graduates leave the state to work?

Which teacher preparation programs’ graduates have the greatest impact on student learning?

Kentucky High School Feedback Report

• In- and out-of-state college going rates

• Public and private colleges included• Rates broken out by ACT categories• College readiness rates by academic

subject (English, Math, Reading)• Rates broken by ethnicity, economic

groups, and other student categories• List of colleges attended

High School GPAs and College ReadinessHow well are High School GPA and ACT Composites Aligned?

2.25 2.45 2.65 2.85 3.05 3.25 3.4515.0

16.0

17.0

18.0

19.0

20.0

21.0

22.0

23.0

24.0

25.0

Do Students’ College Readiness Rates Vary Based on Where Their Teachers Went to College?

University 1 University 2 University 3 University 4 University 5 University 6 University 7 University 80

10

20

30

40

50

60

* Preliminary data from school districts where more than 50% of teachers completed their bachelor’s degree from a single institution34

Percent of Students Not Readyfor College Level Math by TeacherPreparation Program

Kentucky’s History of Using Data

Kentucky’s P-20 Architecture

CPEPostsecondary

CPEAdult Education

KDEK-12 Students

EPSBTeacher Cert. 24/7 Secure

Data Collection,Processing,

and Matching

De-IdentifiedFiles andReporting

System

Data Sources Data Users

Agencies

State

Researchers

PublicReports viaWeb Portal

EarlyChildhood

P-20Staff

Workforce

37

Why Kentucky Chose a Stand-Alone System Architecture• For agencies that don’t have their own system and reporting

tools, it is less expensive to bring their data into the P-20 structure and use our licenses than build their own “silos” and try to link them together. Think beyond P-20 to a true state data system.

• Stand-alone system is less prone to agency data system issues because it only depends on the agencies when new or updated data are needed.

• Matching processes become better over time and incomplete agency level data can be made more complete (holes can be filled in over time with other sources) when it is all together.

• Less upkeep and agency involvement in routine activities.• System runs regardless of any changes in agency data systems or

out-of-firewall network issues.• Centralize access makes it easier to monitor and audit third-party

use of the system for researchers and others.

38

Why Kentucky Chose a Stand-Alone System Architecture• In situations where some state agencies have had a poor

reputation for providing accurate data, it can help to insulate P-20 and gain more support from policy makers as an objective source of information.

• Centralized access makes it easier to monitor and audit third-party use of the system for researchers and others in one location than if each agency did it separately.

• Centralization provides analysts and researchers to support multiple agencies in their use of the data.

• Stand-alone systems allow for intensive analysis of the data in blocks of time to identify issues like changing definitions and data quality problems that are too time intensive to do “on the fly.”

39

Overview of the P-20 Core System

40

Matching Records Across Sources• Names (and all of their iterations)• SSNs• Agency Identification Numbers• Date of Birth• Gender• Ethnicity• School and District• County and Zip Code• Maintaining data that change over time to allow for additional

longitudinal matching of older records• Others as needed by the source

41

Sustainable P-20 Success• Data have to be available, understandable, and people have to be

aware of where they can find information.• Confidential data must be kept confidential either as files or in

prepared reports. Cell-size minimums are not enough.• There is a difference between reporting data responsibly and

interpreting it. What is P-20’s charge here?• The audience’s needs have to be identified. Use focus groups when

developing new reports. Allow for ample vetting time for reports and studies. Provide information to data owners and others affected when data requests are answered. Trust is an important factor.

• Presentations, publications, and having P-20 recognized as a source of vital information. Leadership has to be respected and actively pushing the value of the process and system. Agencies, policy makers, and others should publicly support the importance of the system and use it for decision making.

42

For More Information about P-20 in Kentucky

Charles McGrewExecutive Director

P-20 Data Collaborative

charles.mcgrew@ky.govkentuckyp20.ky.gov

43

Questions?