This study investigates non-parametrically the optimizing behavior of a sample of 289 Kansas farms under profit-maximization and cost-minimization hypotheses. The study uses both deterministic and stochastic non-parametric tests. The deterministic results do not support strict adherence to either optimization hypothesis. The stochastic tests suggest that all 289 farms fail the profit-maximization hypothesis, whereas 171 farms failed the cost-minimization hypothesis. Allowing for non-regressive technical change does not alter the basic results; 276 farms violate the profit-maximization hypothesis and 138 violate the cost-minimization hypothesis. The evidence against cost-minimization behavior seems to be far less substantial than that against profit-maximization behavior.