Applying non-parametric statistical methods to the classical measurements of inclusion complex binding constants.

A study on using non-parametric statistical methods was carried out to calculate the binding constant of an inclusion complex and to estimate its associated uncertainty. First, a correct evaluation of the stoichiometry was carried out in order to ensure an accurate determination of the binding constant. For this purpose, the modified Benesi-Hildelbrand method had been previously applied. Then, four statistical methods (three non-parametric methods: two bootstrap approaches, the jackknife method and a parametric one: Fieller's theorem) were employed in order to compute the binding constant. The results obtained from applying these methods and the combination of the methods: jackknife after bootstrap and bootstrap after jackknife were compared. The best results in terms of accuracy were obtained from the application of a bootstrap method: the resampling residuals approach. These procedures were applied to the inclusion complex 2-hydroxil-propyl-beta-cyclodextrin-2,4-dichloro-phenoxyacetic, which shows photochemically-induced fluorescence.