The well-known Allais paradox is reformulated as a daily farm decision problem. Only 26% of the farmers exhibit violations of the expected utility hypothesis. Moreover, the tendency for violation decreases with the farm operator's education, experience and family size. No effects of the farm main crop or its scale were detected. Finally, when taking into account the possibility of choice errors, we find that the violation rate is statistically insignificant.