Given the reporting and regulatory requirement included in public education, student and unit level data sets have long been a necessary resource on which educators, administrators, policy makers and researchers have drawn. The development of these data systems require significant legislative and public support as they’re both expensive to establish and maintain, and include sensitive data regarding the inputs and outcomes of educating our nation’s youth. Initially constructed of sufficient K-12 data to meet state and federal reporting requirements, federal-level education policy makers and legislators began providing sufficient financial support for expanding the utility of these data sets beginning with the National Center for Education Statistics (NCES) Statewide Longitudinal Data Systems (SLDS) Grant Program 4 in 2005. By 2017 the US Department of Education5 had allocated more than one billion dollars6, in total, to forty-seven (47) state public education systems7 to standardize data collection and access, and began linking student-level data across state agencies.
Such a coordinated system of data, a State Longitudinal Data System (SLDS), may reasonably be expected to include data on the student experience from the time a student enters the public education system, pre-kindergarten, through primary, secondary, and higher education, and into the labor market. When linked to private and out-of-state8 education data available through the National Student Clearinghouse9 the breadth of a student’s education experience becomes available. For educators and policy makers these data systems offer the opportunity to assess the effect of education policy decisions. For social science researchers they hold great potential in examining the effects of changes on a developing portion of our society. For students and households they offer promising developments in ever more efficient and effective delivery of costly education inputs.
Under the standard inferred by the Institute of Education Sciences10, an agency of the US Department of Education11, State Longitudinal Data Systems were conceptualized as data resources linking student and subsequent workforce data records between various state agencies and education institutions. These linkages form a P-20 longitudinal system12 rather than the more limited K-12 system, previously the de facto standard; a restriction still observed in the Missouri, Arizona, California and Idaho SLDSs. This linkage allows state agents, administrators, researchers, and other interested parties to evaluate student outcomes across the breadth of their education and workforce experiences. Subsequent to receiving federal grants and legislative support, several states, including Utah, Washington, Kentucky, and Minnesota formed SLDSs successfully linking data across these education and workforce agencies. However, many similarly supported data systems do not exhibit this capability.
The lack of robust linkages within many SLDSs is primarily the result of disassociated two factors: 1) the difficulty of creating data linkages in individual level records as students progress through primary, secondary and higher education systems, and 2) concern over those privacy risks associated with capturing and storing sensitive individual-level information in an era during which such data is increasingly valuable to all parties, including those with ill intents. Creating linkages between state agencies at the primary, secondary, and post-secondary education levels is an arduous task. It requires the joint effort of each data provider throughout the data system implementation process and the employment of information technology workers with experience designing and creating complex data systems architecture.
Absent significant ongoing state and federal level supports, the required allocation of time and budget to successfully complete this process often extend already resource stretched agencies past their output capacity and may result in poorly managed or inadequately maintained data systems. Evidence of such outcomes may be observed repeatedly in select states and state agencies building SLDSs in the decade since receiving initial financial and legislative support. While some state agencies successfully developed the initial stages of an SLDS, connecting education data throughout K-12 education systems in their respective states. Other states, including Alabama, Arkansas13, Connecticut, and Maine were unable to successfully compile sufficient data resources to complete critical linkages.
This development stage requires the cooperation of a state’s department of education and the public local education agencies (LEAs) that comprise most K-12 education providers in the state. These LEAs enjoy relationships formed with their respective department of education, making the linkage of data between these entities less onerous. Linking student data between other state agencies such as departments of higher education or economic development is a more difficulty process and often requires the joint consent and cooperation of each agency’s stakeholders and directors, and the formal approval from state legislation due to privacy concerns. In addition to those previously noted, Alaska, Illinois, Texas and Virginia SLDSs include fully operable (linked) data systems.
Most state data systems include a unique student identification number allowing linkages throughout the state’s public education system. This identifier follows students into the state’s public higher education system, where it is linked to the student’s social security number, often through a student’s Federal Student Aid (FAFSA)14 application15, and makes possible linkages to labor force data through the state’s department of workforce services or economic development arm, and private and out-of-state public education data through the National Student Clearinghouse. This student identifier to social security number linkage is tenuous and results in at least two substantial issues: 1) for those students who do not progress into public higher education the linkage is not made and labor force and private and public out-of-state education data is made increasingly difficult to capture, and 2) the inclusion of a student’s social security number in an aggregated data repository heightens security concerns and privacy risks.
Though US federal income tax reporting has required the inclusion of dependent’s social security numbers since 198716, many state public education departments and local education agencies do not require the student’s social security number be collected or reported. For those that do, or for those for which capturing the data point is optional, including the number in the state’s data system is inconsistent, if present. Were this not the case, data linkage would become a far simpler and less expensive matter.
The privacy risk associated with these data collection and linkage processes present difficulties to state agencies in their SLDS development efforts. Though student level data, including identifiable and de-identified student and household information, is covered by federal protections through the Family Educational Rights and Privacy Act (FERPA)17, state legislatures and public education policy makers have grown increasingly sensitive to student and household privacy issues. Heightened negative public sentiment towards privacy concerns have impacted how local, state, and federal governments may act. Government agencies concerned over the risk of leaking personally identifiable student data must take extreme measures to ensure student data is collected, analyzed, and distributed appropriately. These measures add to the expense and difficulty of linking data across multiple agencies and education institutions in an SLDS, creating disincentives for these entities to attempt such an effort.
Select states have overcome these hurdles and developed SLDSs successfully linking data across multiple state agencies. In these cases, state legislatures have opted to create sustainable funding solutions in support of state education agencies, departments or employees charged with developing and maintaining the data systems. These funding solutions ensure the data system receives sufficient attention to continuously maintain and improve the system to meet the need of all participating state agencies.
7 Alabama, New Mexico and Wyoming have not received NCES SLDS funding grants in the 2005 – 2015 funding rounds. All other states received SLDS funding in one or more of the six funding rounds ; https://nces.ed.gov/programs/slds/stateinfo.asp
8 Out-of-state public education refers to student participation in public higher education outside of the state in which the student completed public secondary education
11 The US Department of Education Statewide Longitudinal Data Systems grant program outlines requirements for receiving federal grants in support of developing SLDSs; https://www2.ed.gov/programs/slds/factsheet.html
12 P-20 longitudinal systems include individual and unit level data ranging from pre-school participation and into the labor force and have become the de facto standard for State Longitudinal Data Systems
13 Arkansas’s SLDS exhibited linkages across agencies until 2015, when the Arkansas legislature and new management in the Arkansas Department of Education deemed that these linkages created major privacy concerns. The Arkansas SLDS no longer has linkages for new data between agencies, however, it still contains past data records from multiple agencies.
15 In excess of 90% of public and private higher education students file a FAFSA application, some as the result of institutional requirements, and others resulting from the student’s preference for federal student financial aid; https://nces.ed.gov/fastfacts/display.asp?id=31
16 The Tax Reform Act of 1986 (PL 99-514-OCT. 22, 1986) requires social security numbers for all persons represented on individual federal income tax returns; http://slds.rhaskell.org/documents/2017/07/tax-reform-act-of-1986.pdf
17 Family Educational Rights and Privacy Act (FERPA); https://www2.ed.gov/policy/gen/guid/fpco/ferpa/index.html