This article analyzes the productivity growth in Korean agriculture withdata on multiple crops, over the five years, from 1998 to 2002.Measurements are obtained from the estimation of a stochastic multi-outputdistance function. This study corrects for sample selection bias in the contextof a translog functional form that appears when some farms produce only asubset of the potential outputs. The results find that technological change hascaused a significant productivity change. Larger farms experience thehighest rate of productivity growth by the greatest rate of scale effect. Morehuman capital also leads higher productivity growth rates.