Anderson-Darling test

Finally, we may consider a parametric analysis of the relative differences (Figure 14-14 and Table 14-9). A goodness-of-fit test (Anderson-Darling test, P greater than 0.5) showed that the relative differences did not depart significantly from a normal distribution, which in this case supports the parametric approach (Figure 14-15). The parametric 2.5 and 97.5... 

Here we focus only on the Anderson-Darling test, which has been recommended by the IFCC. The procedure is as follows ... 

K-S Test. The Kolmogorov-Smirnov goodness of fit test assesses the ability of a probability distribution calculated from WEiBULL STATISTICS (q.v.) to fit the experimental data. It and the A-D (Anderson-Darling) test (which is more sensitive to discrepancies at low and high probabilities of failure) are used as part of the CARES (q.v.) computer program for failure prediction. For a discussion of... 

Three tests of shape differences that can be used to evaluate goodness-of-fit are the Kolmogorov-Smirnov test, the Cramer-von Mises test, and the Anderson-Darling Computer programs for aU three... 

---