IDS Basics

Administrative data are collected by government agencies in the normal course of delivering social services and programs. Administrative data are often Personally Identifiable Information (PPI), meaning that, without the proper protections, they could potentially identify a specific individual in a data set. Therefore, agencies must comply with stringent federal regulations and privacy laws when they endeavor to share or link data across domains.

When administrative data sharing and data linkage is done safely and legally it becomes a powerful tool to inform policy and practice.

An IDS is designed to do just that; it provides the governance process, legal framework, technology, and human capacity to safely link administrative data across multiple agencies to monitor and track how services are being used and to what effect.



An IDS does not link and store all administrative data from across domains permanently. Instead, it links specific extracts of administrative data across multiple domains for approved projects and an approved timeframe. These domains may include but are not limited to: juvenile justice, homelessness, health & vital statistics, adult justice, behavioral health, education, assisted housing, workforce development, employment & earnings, and child welfare.

Mature, fully-functioning IDS sites need an infrastructure that promotes:

  • Shared governance processes
  • Legal compliance and data security
  • Data management and analytics
  • Political and economic sustainability

IDS that are fully-functioning:

  • Integrate data from three or more agencies on an ongoing basis
  • Are governed by a master Memorandum of Understanding (MOU) between data-holding agencies that establishes a commitment to data sharing in compliance with federal mandates, such as the Family Educational Rights and Privacy Act (FERPA), the Health Insurance Portability and Accountability Act (HIPAA) and the Privacy Act, and sets policies and procedures for decision-making on everything from project approval to review and dissemination;
  • Utilize separate Data Use Licenses (DUL) to authorize and set the terms of individual research projects;
  • Create outlined procedures for data retrieval, record linkage, cleaning, sharing, and protecting data;
  • Employ strategies for securing executive buy-in to sustain their efforts across different administrations and budget cycles
Executive-based Model

Entities with strong state or county executive support or leadership that have existed over several administrations. These systems often have reliable revenue streams for operating expenses and routinely reside in secure, non-partisan government offices. This arrangement safeguards these systems against sudden political or economic changes. Example: Center for Innovation Through Data Intelligence (CIDI), NYC

Agency-based Model 

Typically originate and reside in departments of health and human services (HHS). Most agency-based models were originally built to help caseworkers manage their workload by providing a clearer picture of clients that utilize services from multiple government systems and programs. Over time, these sites initiate further data sharing agreements with other agencies, like school districts or state education departments, streamline legal processes, and promote the use of integrated data for policy analysis more broadly.  Example: Department of Human Services (DHS), Allegheny County, PA

University-based Model
Do not typically have formal governing boards. Instead, they develop a research agenda that is based on their funding revenues, research interests, and partner-agency needs.  Example: Chapin Hall at the University of Chicago

For more examples and information on the benefits of each model, see our Expert Panel report.