Quality Issues in Survey Data Collection

Jaki S. McCarthy

Chief Cognitive Research Methodologist

USDA's National Agricultural Statistics Service
Research and Development Division

 
A statistician working with a survey dataset must assume that data are accurate and of high quality.  Survey organizations such as the US Department of Agriculture's National Agricultural Statistics Service do many things to help ensure that this assumption is justified.  Data quality can be dramatically impacted in the data collection phase of a survey, and this is the area that will be the subject of this talk.  Because survey data are usually collected from people, understanding what respondents must do to report those data is important to ensure quality.  This talk will begin by answering the question:" What do respondents have to do to provide the data you want?"  Following this, we will examine how data quality can be impacted in this response  process and what NASS does to develop procedures for collecting good data.  This talk will give the audience an appreciation for the work that goes into ensuring quality in survey data collection and examples from NASS surveys will be used for illustration.