Example 2.6.2 For the data set in Table 2.3.2, both classical and inverse methods give an estimated concentration of the unknown sample outside the range of standards. It is recommended that another experiment be performed to provide a better estimate of the unknown sample. Based on the results from Table 2.3.2, the unknown concentration is approximately 17. Hence we may choose 17 as the middle design point. As a result, the low and high design points may be chosen as 12 and 22, respectively. Also, from Example 2.3.1, the estimated slope and mean-squared error are 5.43 and 15.85, respectively. Suppose that the expected slope is 5. Based on this information, the sample size required to yield 90% power for hypothesis (2.3.11) at the 2.5% level of significance can be determined as follows.