Tradeoffs between Using Administrative Data and Survey Data to Construct Outcomes for Impact Analysis
Thursday, November 8, 2018: 9:30 AM-10:15 AM
Atrium - Exhibit Level (Marriott Wardman Park)
*Names in bold indicate Presenter
Participants: Laura Peck1, Shawn R Moulton1 and Nicole Constance2, (1)Abt Associates, Inc.(2)U.S. Department of Health and Human Services
Description of Research: In any evaluation, there are tradeoffs when choosing between administrative and survey data sources to define and analyze outcomes. Survey data faces potential recall error and non-response bias, whereas administrative data sources may be limited by the type and scope of data collected. Due to these factors, outcomes constructed from survey data may differ from outcomes constructed from administrative data, even when designed to capture the same information (e.g., annual earnings). These differences may occur in both the level of the outcome, and in the estimated impact of the program (i.e., the difference between the treatment and control group means). Because impacts on a survey-based outcome and an administrative-based outcome may differ, the choice of whether to rely on administrative or survey data to construct outcomes may influence the findings of an evaluation and, hence, policy decisions informed by these findings.