This paper investigates how the level of self employment effects the measurement of official unemployment statistics. The components of the unemployment rate are identified and the rate is recalculated to reflect the labor market performance for wage and salary workers only. Data from the Survey of Income and Education, conducted in 1976, are used to demonstrate that sub-national unemployment rates contain a systematic bias which is highly correlated with the level of self employed agricultural workers. After adjustment for self employment, rates increased by over 20 percent in nonmetropolitan areas and only 10 percent in metropolitan areas. The implications for manpower funding allocations resulting from this consistent understatement of rural unemployment rates are discussed.