469 – Official Statistics: Policy and Practice
Developing Response Metrics for the Economic Census
Justin Ward
U.S. Census Bureau
Diane K. Willimack
U.S. Census Bureau
In the past, the U.S. Census Bureau used the check-in rate to measure the performance of data collection operations for the Economic Census. The check-in rate is the proportion of mailout that has returned a form and serves as a proxy indicator of response, but does not necessarily indicate receipt of usable data. We are developing unit and item response metrics to provide more meaningful information about data quality. Since we calculate the unit response rate after completion of all data collection and processing operations, we developed a proxy metric for the unit response rate that can be calculated in real time and, along with the check-in rate, used to monitor and manage collection activities. We use 2012 Economic Census data to examine the behavior of these new response metrics. The unit response rate and its proxy will always be lower than the check-in rate, which includes units returning forms without usable data. During collection, the proxy performs similarly to the check-in rate. After collection, the proxy metric and the unit response rate are similar with their differences mainly explained by limitations of information available during collection.