Why Should I Use My Administrative Data for Analysis and Evaluation?

August 31, 2017 China Layne

In the last decade, federal agencies have increasingly used administrative data to conduct analysis and evaluation. Administrative data is collected for non-research purposes such as: applications for services, records of service receipt, and outcomes after service. Under continued guidance to “make do with less”, these agencies have been encouraged to use administrative data as an alternative to collecting primary data for analysis. Seeing the direction of these changes, your organization might ask, “what are the advantages of using administrative data for analysis”.APDU blog 1 graphic (1).png

Administrative Data Advantages

Administrative data have several important advantages for use in analysis and evaluation.

  • Because your organization already collects this data as part of managing your programs, using the data for analysis and evaluation is the most cost-effective option.
  • As your organization collects a wide range of information directly about the programs (including participants, services provided, and outcomes), your data serves as the most relevant source of information on your programs.
  • Unlike a one-time survey, your organization’s administrative data is both the most up-to-date source of information on your programs and can provide longitudinal data on the programs.
  • As your organization’s administrative data includes information on both the program participants and providers, this data likely comprises the largest sample of the relevant population than any other single public or proprietary data source.
  • With proper preparation and forethought, administrative data can provide causal results in evaluation similar to using primary data collection.

Dr. Rebecca Bilbro, Lead Data Scientist at ByteCubed, shares her perspective on the advantages of using administrative data for analysis: “I sometimes hear organizations express concerns about the completeness and consistency of their administrative data. However, in my experience such datasets can often inform the development of very robust data products. From applications that use historic case management data to prioritize new incoming cases, to others that find patterns in data to help predict and even pre-empt shortages, surges, and other organizational challenges. Even if it encodes a fraction of the substantial domain expertise that goes into making everyday business decisions at an organization, administrative data can be an immensely powerful tool."

When to Employ Administrative Data

Administrative data can be employed at almost any time to support analysis and evaluation. The data can be used as the initial, primary source for analysis or as a later addition to primary data collection. Either way, administrative data will be most effective if considered first, and fully, as the data source while planning analysis and evaluation. Likewise, administrative data can be used for analysis years after the collection instruments are designed and the data collected. However, the data will be most effective if analysis and evaluation concerns are taken into consideration during design of the data collection instruments and processes. Doing this will ensure that your administrative data includes the most useful information, in the correct formats, to support a wide range of robust analyses.

This month, Summit Principal Anthony Curcio will discuss using administrative data for analysis and evaluation at this year’s Association for Public Data User’s (APDU) conference in Arlington, Virginia. Curcio will be moderating a panel, Data Integration for Program Effectiveness, where he will speak with panelists such as Dr. Rebecca Bilbro who are at the vanguard of using administrative data for analysis.

As outlined above, there are many advantages to using your organization’s administrative data for analysis and evaluation and this data can be employed at almost any stage of analysis. In the next post, we will discuss some of the many types of analyses that can be conducted using administrative data.

Share This: