Data on farm household income and characteristics come from USDA’s annual Agricultural Resource Management Survey (ARMS). Changes in survey methodology, sampling, or response rates can influence estimates derived from survey data. In 2012, more sampled farms responded to the survey than in recent years, as is usual during years when the Census of Agriculture is conducted. This increased the number of family farm records used to calculate farm household statistics--from 19,441 in 2011 to 20,561 in 2012. The number of records enumerated through the mail also increased, from 14.2 percent in 2011 to 42.6 percent in 2012. These and other survey design changes make comparisons between farm household statistics calculated using 2012 survey data with those from earlier surveys difficult to interpret. The changes also affect the precision of estimates. The coefficient of variation (which measure the size of the standard error relative to the mean) for total household income and off-farm income was larger in 2012 relative to 2010 and 2011. See more statistics on the ARMS from recent years.