Abstract
Survey data are widely used for health research because they seek to cover entire populations of people, and they contain policy-relevant information on people, families, and households. Administrative data typically entail records collected or maintained by federal, state, tribal, or local government agencies, or commercial entities; not for the purpose of demographic statistics, public health surveillance, or policy analysis, but for administering programs or providing services. This chapter outlines a research agenda for assessing the quality of linked data files, using the sources of survey error as a guide. It focuses on the concerns about the quality of survey data, familiar to all researchers who use survey data. The chapter explains the following potential problems and some possible solutions associated with linked data files: coverage error in the sampling frame, sampling error, nonresponse error, measurement error, data processing issues, imputation procedures, editing rules, documentation, dissemination, and timeliness.
| Original language | English (US) |
|---|---|
| Title of host publication | Health Survey Methods |
| Publisher | Wiley & Sons |
| Pages | 695-716 |
| Number of pages | 22 |
| ISBN (Print) | 978-1-118-59462-9 |
| DOIs | |
| State | Published - 2014 |
Keywords
- administrative data
- data documentation initiative
- health research
- health surveillance
- survey data
- survey error
