Srivastava proposed an estimator &ymacr;α for estimating the population mean [&Tmacr;y] of a study variable ‘y’. This estimator uses auxiliary information and depends on the values of a parameter α. He also showed that &ymacr;α attains its minimum mean square error value when α is suitably chosen. However, in practice that value is unknown. In this work is proposed an estimator of the optimum value of α, which is evaluated from the selected sample. A study of the performance of Srivastava's estimator when it uses the estimation of the optimum value of α is included.
