AbstractThis chapter compares two national data sets commonly used in the USA for empirical research on farm workers: the Current Population Survey (CPS); and the National Agricultural Workers Survey (NAWS). As the major national survey of employment in the USA, the CPS covers farm workers as well as employees in other industries in the USA. These data are compared with the NAWS, a major survey of crop workers employed on US farms. The samples for the CPS and NAWS are shown to be very different, even if crop workers (only) from the CPS Annual Earnings File are compared with the NAWS data. Unlike the CPS sample, the NAWS sample includes the (large and significant) segment of the farm worker population that shows a marked degree of mobility and is very likely to be unauthorized. Farm workers in the NAWS are more likely to be Hispanic, have very low levels of education and are younger than in the CPS. The NAWS also specifically shows that, compared with the past, US farm workers are more likely to be solo males working on farms in the USA but spending at least part of the year 'at home' in Mexico. They are also not likely to access public support systems in the USA, despite their low incomes. Finally, their earnings and incomes are difficult to compare with those of other low-income workers in the USA because only US earnings are included in the NAWS data. Inter-industry comparisons with the CPS are possible but, again, a portion of the farm worker population is missed.