Regression analysis using auxiliary information from time studies of cable yarding operations
Time studies are commonly employed in the analysis of cable yarding productivity. For a given site and system, productivity is a function of the yarding distance and the load size. Field conditions make it more difficult to sample load volume than yarding distance, which leads to substantial differences in sample size for the two variables. When multiple linear regression is used to fit equations for predicting yarding productivity to these data, many of the observations must be omitted. Two alternatives to conventional multiple linear regression that make use of the auxiliary information in the additional observations on yarding distance were compared on the basis of statistical efficiency. A mixed estimator was found to be most efficient for all yarding systems studied.