Inside states’ efforts to expand and strengthen longitudinal data systems

State and local governments have long relied on longitudinal data systems to identify and address public issues, advocate for funding, measure impact and assist other critical operations. But it’s time that government leaders consider how to expand such systems to further enhance their quality and impact, experts said in a recent webinar. 

Longitudinal data systems generally track and monitor statewide education and workforce data over long periods of time, and common use cases for them include creating public dashboards that “display and clarify” outcomes of a particular program or policy and provide research and evaluation, which analyzes “the factors that are impacting outcomes,” said Kathy Booth, senior director for economic mobility and educational data systems at WestEd, during a webinar hosted by Actionable Intelligence for Social Policy.  

But governments are recognizing that a myriad of factors can influence residents’ education and workforce outcomes, like their health, family wellbeing or access to public benefits, Booth said. That’s why states are increasingly exploring how to integrate more diversified data sources into their longitudinal systems in an effort to foster more holistic and comprehensive policies and programs aimed at improving a resident’s outcome in life. 

A major challenge to doing that is the hurdle of building trust between data sharing “partners who are operating in different contexts, facing different pressures and are accountable to different missions,” said Della Jenkins, executive director of Actionable Intelligence for Social Policy. 

Iowa is one state that has addressed such concerns to support the creation and participation of its longitudinal data system, said Sharon Zanti, research associate at Iowa State University. 

“Since about 2015 we've been partnering with our state's early childhood programs to run an integrated data system that is focused a lot on early childhood and family well being, but more and more, we are expanding partnerships and thinking about how to integrate data from education, workforce and other health and human service agencies,” she said. 

The Iowa Integrated Data System for Decision-Making, or I2D2, began as a university-state partnership between Iowa State University and Early Childhood Iowa, a statewide initiative aimed at coordinating public and private organizations’ efforts to improve childhood outcomes, Zanti said. The system helped early childhood programs improve the quality and standardization of data, and that work lead state and university leaders to develop the Iowa data drive, a site where users can access dashboards tracking data that informs, for example, needs assessments and community planning.  

The data drive has shifted the mission from “helping our partners collect the data [to] go out and do … a lot of work on the ground themselves,” Zanti said. 

Iowa and its university partners are continuing to explore how to expand the data system’s scope, she said. Under Early Childhood Iowa, researchers are exploring partnerships with other state agencies like the Iowa Department of Health and Human Services to build a more integrated data system across the state. 

Behind the state’s growing data sharing capabilities is the participating partners’ commitment to using shared data responsibly. For Iowa’s integrated data system, state and university partners sign a memorandum of agreement that establishes broad legal and governance frameworks. Additionally, stakeholders also establish data sharing agreements and data use licenses for individual projects to ensure initiatives are vetted extensively to align with partners’ specific data needs, Zanti explained. 

Connecticut has shared a similar experience, said Scott Gaul, the state’s chief data officer at the Office of Policy and Management. The office has managed the state’s longitudinal data system since 2021, which originally centered around education and workforce issues but has evolved to also incorporate health and human services data and criminal justice data. 

Connecticut’s DataLinkCT system aims to establish a consistent approach to how data is collected and shared across agencies, but the partnership model remains “very decentralized,” Gaul said. This method ensures that “agencies control their data and decisions” while Gaul’s office’s role is to facilitate the collaboration and review of data sharing and requests, he explained. 

Additionally, a decentralized approach allows agencies to tailor data policies and governance to their individual needs, helping them determine if they can or want to participate in a data sharing project, said Susan Smith, director of business intelligence and analytics for the Connecticut Department of Social Services. 

For instance, Smith said her agency must comply with a state law that limits their ability to share data if its use directly advances or relates to the administration of social service programs. Under Connecticut’s longitudinal data system model, DSS is still able to adhere to the state law without having to forfeit participation in data initiatives totally. 

Indeed, when building or growing longitudinal data systems, state leaders should consider how to “crosswalk” different agencies’ guardrails around the use of their data to ensure privacy, equity and ethical use and “build a bridge between … governance frameworks that already exist” to streamline system creation and participation, Jenkins said.

In North Carolina, building agencies’ trust in expanding longitudinal data systems can start with a “carrot” approach, said Trip Stallings, executive director of the state’s longitudinal data system. Leaders, for example, should consider how to demonstrate to potential partners that such systems offer the services and technical support that they may not have otherwise, as agencies grapple with budget and resource shortages. 

Among North Carolina state agencies, “some of them are starting to see that there is value in [participating], and that we really are here to help and make their lives easier,” Stallings said. For instance, he said officials hope to add geospatial data, health data and criminal justice data to the data system by this time next year. 

He also underscored the value of offering participants training or assistance under a longitudinal data system. Officials in North Carolina, for example, developed “a path for onboarding new data owners and new data sources” to help ease agencies’ understanding of how to leverage the system, he explained. 

Ultimately, when building longitudinal systems or expanding them, if agencies’ data security, privacy and trust “aren't at the beginning of your thinking about it, you really shouldn't go forward, because we want to make sure that people can trust that when their data are linked, and it's going to be put to good purposes,” Booth said.