do we use uncertainty of calibration curve to determine the title value or by using standard error derived from repeat injection of the same sample?