The natural calamities and conflicts faced by Chinese farmers were overwhelming during 1929 to 1933. In this thesis I compile village level data from John Lossing Buck's nationwide survey, using techniques of regression with robust standard error and analysis of covariance to compute changes in village yields to normal or best yields as a function of agricultural area, province, time, and catastrophes. The analysis shows the geo-political governance of a province had greater impact on agricultural productivity than agricultural area did. Also time of the investigation was taken matters, which is in line with the dynamic events that arose over the 1929-1933 timeframe. Surprisingly the key impacts on agricultural productivity were natural calamities despite the wide spread warfare of the time. This is however due to a selectivity problem in that a survey team could not be sent into an area with high risk of warfare.