Ratio and product estimators that made use of information on auxiliary variable and variable of interest not yet drawn were compared with the conventional ones that made use of information on auxiliary variable and variable of interest already drawn when sampling without replacement from a finite population with known mean. From the two numerical examples used (when the population correlation coefficient is positive and negative), it was observed that the ratio and product estimators that made use of information on auxiliary variable and variable of interest not yet drawn always performed better once all of their conditions were met.