Assessing the Effect of Calibration on Nonresponse Bias in the 2005 ARMS Phase III Sample Using 2002 Census of Agriculture Data

The United States Department of Agriculture’s National Agricultural Statistics Service conducts the annual Agricultural Resource Management Survey (ARMS) in three phases. The third phase of the ARMS collects detailed economic data which is highly sensitive, and thus this phase suffers from lower response rates. With the release of the 2006 Office of Management and Budget Standards and Guidelines for Statistical Surveys, response rates lower than 80 percent may not only result in nonresponse bias, but they can jeopardize the future of surveys carried out by federal agencies. Based on the assumption that the use of calibrated weights derived from appropriate targets addresses nonresponse bias, the effect of the 2005 Phase III ARMS calibrated weights on relative bias of the mean was tested for all cases sampled in the 2005 ARMS with Census 2002 expenditure data. The results showed that calibrated weights decreased bias levels so that they were no longer significantly different from zero for almost 90 percent of the “study variables” (p < .05).