In 2013, the Research and Development Division of the National Agricultural Statistics Service conducted a study to determine if targeted data collection for likely nonrespondents could positively impact response rates for the Agricultural Resource Management Survey (ARMS) Phase III. The methodology for the study was the same as a study done in 2012, but was implemented with only four participating regions, encompassing 12 states. This smaller sample was used in 2013 due to the unforeseen complications with carrying out the 2012 study across the United States. Likely nonrespondents were flagged using classification trees and their relative importance to the calibration target was taken into consideration with the use of impact scores. Flagged records were then assigned to targeted data collection methods including the use of enumerator incentives and having state or regional office staff make in person contacts to attempt to increase response rates for those records. Given our small sample size, this is an exploratory study.


Downloads Statistics

Download Full History