Administrative records are increasingly being used by Federal agencies in lieu of, to augment, or as a quality control tool for survey data. As agencies’ use and reliance on administrative records grow, it becomes increasingly important to understand how the data are assembled, organized, managed, and reported. This paper will focus on answering the question: “How does the mode in which administrative records are reported or obtained from record-keeping entities affect the quality of those data?”
We analyze the data from the Energy Supplier Survey (ESS) of the Residential Energy Consumption Survey (RECS). During the ESS, administrative records are reported by companies who supply energy to the households surveyed for the RECS. Respondents self-select mode from a fixed set of mode options, the most common being online teleform, electronic file submission, and paper/pen. Reported data were edited by analysts using tools such as bills collected during the household interview. In this paper we explore whether the reporting mode had an effect on the number of edit failures and errors in the data, by controlling for factors such as number of administrative records reported. We find that errors in the data did differ by the mode selected by the respondent.