Author : Putu Sudira

Summary : The second and fourth degree multiple regression were used on the data to obtain the effect of climatic variables independently on soybean yield at different stages of crop development. Likewise, the selected crop stages method, using the stepwise regression procedure was utilized to establish a general prediction model for soybean yield. An attempt was made to compare the results obtained from the stepwise regression method with the results of the second and fourth degree multiple regression technique. Results obtained in the second and fourth degree multiple regression showed that only maximum temperature and sunshine duration equation model for the first growing season were significant at 5 percent level of significance. On the other hand, the stepwise regression model identified only the most significant climatic variables that affected the yield of soybean. Results showed that for the soybean growing season, out of twelve climatic variables entered, variables were only considered which explains 99.79 percent of the variation of the soybean yield, and it was significant at 1 percent level. The second and fourth degree regression model appear to be sufficiently reliable for practical purpose. However, the selected crops stages method using the stepwise regression procedure gave a relatively better result than that of the second and fourth degree multiple regression in predicting the yield of soybean.