623 – Statistical Modeling: Benefits and Drawbacks
Utilizing Paradata to Examine and Improve the Web Data Collection Process in Agricultural Census and Survey Programs
Robyn Sirkis
National Agricultural Statistics Service
In today's world, statisticians must consider and gain insights from multiple, and sometimes complex, data sources. One such source is paradata which are metrics related to the process of collecting survey data. In web surveys, paradata can be classified into several types: contact (session date and time), device (e.g., device type, browser used), and questionnaire navigation (e.g., changing answers, error triggers, breakoffs). With an increasing focus on efficiencies in data collection, the USDA's National Agricultural Statistics Service (NASS) began collecting and analyzing paradata for its web data collections. This research was conducted to learn more about the web collection process at the respondent level and make informed improvements to the online questionnaire. The analysis includes a comprehensive evaluation across mobile and non-mobile device types. This paper focuses on the methodology and results using web paradata from the 2017 Census of Agriculture and 2018 June Crops Survey. Challenges encountered analyzing the paradata and recommendations based on our findings will also be discussed, illustrating the valuable role paradata can play in helping survey organizations.