For the determination of the moisture contents of honey, the Brazilian legislation adopts the refractometric method proposed by AOAC. However, in other work, the authors observed that the crystallization interferes in the refractive index measurement when the honey sample if encountered crystallized and, therefore, moisture contents overestimated were obtained. The European Honey Commission (EHC), adopt other refractometric method that use a pre-treatment of sample when this one was crystallized. So, the objective of this work was to compare the precision of these refractometric methods by different statistical techniques and it establish thus the more adequate procedure for moisture analysis in honey. The results of t-tests analysis in the 95% confidence level for the average moisture contents of the honey samples suggested that there were significant differences between the two-refractometric methods (AOAC e EHC) only for the crystallized samples. The analysis of standard deviations by application of F-test and building of confidence intervals shown that the EHC method was more precise than the AOAC methodology for the crystallized honey samples. In this way, it may be suggested the adoption by Brazilian legislation of refractometric method of EHC, as official method, by this one not present systematic errors.
honey; refractometry; precision; physical state; statistical analysis
