Data with poor degree of precision of recording and with large rounding errors may violate the underlying distributional assumptions of an analysis. In this paper we examine a large set of agricultural data through distributions of final digits in the data which ought to be uniform if the degree of precision of the data is good.