Abstract:
|
The U.S. Census Bureau conducts a mandatory economic census every five years, collecting establishment level financial and product data. The 2017 Economic Census featured changes in data collection and statistical methodology, including updated calculations of several response metrics, enabling a more effective assessment of data quality. Implementation of these new metrics relied on documentation of sources used for correcting estimated or inconsistent data. Correction sources included respondents’ revisions, administrative data, other survey(s) or reliable secondary sources for the same reference period, or estimates by industry analysts based on prior year data, industry averages, or other alternative reliable sources. This paper presents preliminary evaluation of key updated response metrics, such as the Total Quantity Response Rate and the Imputation Rate. We compare analogous 2012 Economic Census metrics with those from 2017 using data and methodologies from each period, and provide initial assessment of the impact of various correction sources on data quality. We will discuss next steps for research and implications for future Economic Censuses.
|