Federated vs. Centralized SLDS Governance

While state public education agencies have formed State Longitudinal Data Systems (SDLS) through a variety of methods, time frames, and partnerships, differentiated decisions with respect to system architecture have resulted in the production of distinct SLDSs, often unique in their capabilities and functions. However, virtually all state public education managers have opted to form the governance structure of their SLDSs through one of two models: 1) a federated or 2) centralized governance structure4. While these governance structures serve the same basic purpose, to manage an SLDS effectively and safely, they diverge significantly in terms of how they attempt to produce these outcomes.

A centralized governance structure refers to an SLDS that is contained within a single, physical data warehouse. This data warehouse, maintained by a primary manager, is most often a specialized division within the state agency overseeing public K-12 or post-secondary education.  In rare cases, a separate center operating under the purview of a state agency, but not legally a part of that agency, is granted SLDS management responsibilities. The SLDS manager is responsible for ensuring the collection and quality of required data inputs from public local education agencies (LEAs) forming the respective state’s K-12 education system and any other education or workforce agencies and institutions having agreed to provide data to the SLDS. Alaska, Kentucky, Indiana and Utah each enjoy SLDSs with successfully managed, centralized governance structures.  There are currently 42 public education data systems operating under a centralized governance system.

A federated governance structure refers to an SLDS not contained within a physical data warehouse. Instead, this type of structure depends on the cooperation of participating state agencies and other data providers. These separate data providers temporarily link their data to create combined data sets as necessary in support of required agency, state, and federal reports and research questions important to the participating data providers.   Upon completion, reports are saved but the underlying data used to generate the reports will be destroyed. This ensures that no partners are at risk of leaking sensitive data records to the public. Examples of states that utilize federated governance structures include North Carolina, Virginia, and Idaho.  There are currently 11 state public education data systems operating under a federated governance system.

The primary difference between centralized and federated governance structures lies in the level of control placed in the hands of one agency or agencies to manage the SLDS. In the centralized model, one member of the SLDS partnership enjoys significant control over the day-to-day operations of the SLDS, the projects the SLDS supports, and those initiatives undertaken to improve the SLDS itself. Among the benefits of this model are a reduced level of bureaucracy to which the SLDS management team must adhere and increased attention paid to the SLDS from a direct overseer. However, this model can suffer when the managing partner of the SLDS is a separate entity from the predominant education agency in the state. This partner may have control over the management of the data system, but will be unable to perform any significant functions without the support of its state education agencies.

The Arkansas Research Center (ARC)5, the manager of Arkansas’s centralized SLDS, is a prime example of this scenario. From 2009 to 2015, the ARC pushed to the forefront of SLDS development, establishing best practices for system architecture and secure data linking. Many other states modeled their own SLDS configurations and governance structures off the success of the ARC during these years. In 2014 and 2015, key changes at the state legislative and administrative levels produced a much less favorable regulatory environment towards personal privacy and data management. The Arkansas Department of Education (ADE)6 ceased sharing data records with the ARC in 2015 and has placed the center in a difficult position moving forward. While it still has active partnerships with other state education and workforce partners, it no longer has access to critically important K-12 student data records that form the foundation of longitudinal research and analysis. The ARC will continue to conduct research with the prior data records it has received from the ADE and the new data records it receives from other partners but going forward, its future as an independent SLDS manager is now in question.

The federated governance structure is often adopted in states in which the state legislature has engineered constraints against the permanent establishment of linkage practices regarding sensitive student data. In an effort to circumvent around these prohibitions, SLDS partners, including state education and workforce agencies, postsecondary education institutions and other related agencies agree to link data records on an ad hoc basis and often file a formal memo of understanding (MOU). SLDS partners may refuse to participate in a linkage process at any time should they believe the purpose of the linkage does not serve the best interest of their stakeholders. The benefits of this governance structure are that SLDS partnerships can be established more quickly and with more participating members. However, this governance structure often suffers from situations in which one or more critical SLDS partners chooses not to participate in a research or report effort due to concern of regulatory examination or stakeholder disapproval.

Both centralized and federated SLDS governance models produce advantages and disadvantages depending on the details of the governing agreements, however, states have shown both models may be used to establish effective and successful SLDSs. Each governance model has unique needs that SLDS partners must account for when developing their SLDS. These include matters such as legal data sharing agreements between partners, active data governance committees, and annual partner meetings to discuss matters relevant to the SLDS. However, both governance models share at least one common need to create a successful SLDS: direct oversight and accountability from a primary manager or a group of SLDS partners to establish responsible handling of the SLDS and its developments.

 

State Longitudinal Data Systems – by governance structure

Centralized Governance
Alabama Louisiana Oklahoma
Alaska Maine Oregon
America Samoa Maryland Pennsylvania
Arizona Massachusetts Puerto Rico
Arkansas Michigan Rhode Island
California Minnesota South Carolina
Colorado Mississippi South Dakota
Delaware Missouri Tennessee
District of Columbia Montana Texas
Florida New Hampshire Utah
Georgia New Jersey Washington
Hawaii New Mexico West Virginia
Indiana New York Wisconsin
Kentucky  Ohio Wyoming
 .
Federated Governance
Connecticut Kansas North Dakota
Idaho Nebraska Vermont
Illinois Nevada Virginia
Iowa North Carolina

 

1 State Longitudinal Data Systems Research, Bill & Vieve Gore School of Business, Westminster College, Salt Lake City, Utah; www.slds.rhaskell.org

2 Richard Haskell, PhD, Associate Professor of Finance, Gore School of Business, Westminster College, Salt Lake City, Utah (2017);; http://www.rhaskell.org/haskell

3 Peter Seppi, BS Finance & Economics, Westminster College, Salt Lake City, Utah (2017): https://www.linkedin.com/in/peter-seppi-11000091

4 Centralized vs. Federated: State approaches to P-20W Data Systems, Institute of Educational Sciences, National Center for Education Statistics, October 2012; http://slds.rhaskell.org/documents/2017/07/federated-centralized.pdf

5 Arkansas Research Center (ARC); https://arc.arkansas.gov/

6 Arkansas K12 Education; http://www.arkansas.gov/education