Authors: Nicholas Ortiz, Ginger Elliott-Teague
This is a common question that data teams face as technology becomes more advanced and the need to answer complex questions increases.
In a previous blog post, we explored the differences between data sharing, data linking, and data integration, as well as how these terms are often used interchangeably despite having some key differences. It can be hard to tease apart which of these is already occurring with your data system and which may be the best options for enhancing the various pieces of your data system to connect data within and across programs/agencies. How do you make sense of what’s already happening with your data system and decide whether to level up?
Think of data sharing, linking, and integration as existing on a continuum of connection. The table below summarizes how data connections vary across many key features.
Continuum of Sophistication for Data Connections
|How sophisticated and complex is the data connection?
|Are paper or electronic files transmitted?
|Can record-level data be transmitted or only aggregate data?
|Are distinct records matched across databases?
|Yes, but usually manual
|Yes, but usually automated
|Is the data linking process facilitated by electronic automation?
|Is significant technical expertise required?
|Can updates to the source database(s) automatically refresh the linked data set?
|Are the original data sets preserved afterward?
|Are programmers required?
|Do the data sets need to be restructured?
|Do the data have a unique identifier?
Note: This table is not definitive for all data initiatives and data connection types. Rather, it’s intended to reflect the major distinguishing features.
Deciding Whether to Level Up
So, should you stay with your current data structure or go with a new plan? Well, as with so many things, it depends!
It’s too simplistic to ask whether you should level up from sharing to linking to integration without grounding that decision in your program’s needs. In reality, your questions of interest should drive your program’s demand for stronger data connections. The overarching question to ask yourself is: What do I need to know that cannot be answered with the data I already have? For example, you may want to know:
1. How many children who receive Part C early intervention or Part B 619 early childhood special education services are…
- in community childcare?
- in Head Start?
- receiving Supplemental Nutrition Assistance Program (SNAP) benefits?
- receiving other home visiting services?
And… Do services or outcomes differ across these groups?
2. What are the kindergarten through grade 3 outcomes for children previously enrolled in early intervention and/or early childhood special education?
Once you know your questions of interest and have fully fleshed them out (specific years, subgroups, etc.), then you can consider what may need to change with your data system. These may or may not require sharing, linking, or integration.
If you need to level up your system to answer your questions, next consider your team’s level of readiness and capacity for change, the current status of each of your data system’s components (governance, data entry, identity resolution, etc.), and which data connection features are appropriate and worthwhile to be enhanced. Then, you will be ready to align possible features to your data needs. Your data system will likely be at various points along the data connection continuum depending on the specific feature. For example, your system may do a good job with automating electronic updates but still have manual data entry in some places.
Also remember that data linking or integration may need to occur within your program, across programs within your agency, or even across agencies. Generally, data connections become more complex as you move outside your own program.
If you are thinking about leveling up your data connections, you will want to consider the tradeoff between the complexity of your solution and the benefit to your team’s ability to answer important questions. Ultimately, whether you should level up on the data connection continuum depends on several things, including your specific needs, specific features to enhance, available resources, governance processes, and net benefit. DaSy is available to help you figure out where you are on this continuum, what your specific needs may be, whether you are ready to advance, and how to get to the next level. Whether you are going big with an early childhood integrated data system or starting small with connecting a few data sets in house, don’t hesitate to reach out!
Part C and 619 Data Managers and Coordinators and other data team members invested in answering important programmatic questions.
- Data Linking Toolkit: Data Linking Partnership by Partnership Configuration
- DaSy Critical Questions About Early Intervention and Early Childhood Special Education
- Blog: What’s the Difference Between Data Sharing, Data Linking and Data Integration?
- Video: How DaSy’s Critical Questions Can Support Your Data Linking Efforts
Partner Centers and Other Resources
- Center for the Integration of IDEA Data (CIID)
- Statewide Longitudinal Data System (SLDS) Program
- Report: The Integration of Early Childhood Data
About the Authors
Nicholas Ortiz, MPA is an Education Researcher and DaSy Technical Assistance Specialist at SRI International. He is a former 619 data manager with experience in data system development, including data linking and integration. He championed several data integration projects working in state government.
Ginger Elliott-Teague, PhD is a Senior Education Researcher and DaSy Technical Assistance Specialist at SRI International. She is a former Part C and Part B Data Manager and has led state efforts to link and integrate data within and across EI and ECSE programs.