66
independent variables and interaction terms were tested. The selection of the final equation for each crop was based upon the extent to which it was believed to describe observed conditions in the area, and upon statistical indicators of significance. The statistical criteria included the student's "t" values for individual variables, and R values for the function as a whole.
Generally the level of significance of the regression coefficients was a =.05 or lower. However, several variables were included in the final production functions for which the coefficients were significant at the a =.20 or a =.25 levels. These latter levels were considered acceptable given the nature of the data used. The data were obtained by interviews and were based on recall, not written records. The sizes of the samples were small and this may have contributed to the lower confidence levels for a number'of coefficients. Also, it is believed that these results would provide useful guidance in any future investigation and experimentation.
Cotton
The following production function was derived for cotton:
Y -3266.82 + 253.10 MP 25.00 MP2+ 23.92 L .038 L2 + 1.812 (NCI)
2
-.0011 (NCI) + .00000022 (Nc!)3
where:
Y = yield of seedcotton in kilograms per hectare
Mp = tractor hours per hectare in soil prepa ration (pre-planting)
L = hours of labor per hectare
NCI = value of fertilizer and insecticide per hectare, in pesos.
In Table 9 the values derived from this equation and the mean value of each input based upon sample data are summarized.