Statistical analysis of the extended Hansen method using the bootstrap technique.

In this study, simple bootstrap techniques are combined with the extended Hansen solubility approach to calculate biases, standard errors, and confidence limits of the partial solubility parameters and to obtain bias-corrected values for these solubility parameters. The bootstrap method is rather new in its application to problems in the pharmaceutical sciences and, therefore, is described here in some detail. This method provides measures of the statistical variation of ratios of regression coefficients without making unwarranted assumptions about data variability. The bootstrap can be used in many statistical packages such as MINITAB, SPSS, SAS, BMDP, or GLIM, all of which are widely available, and could be useful in other areas of the pharmaceutical sciences where regression analysis is employed.