Escherichia coli (E. coli) and intestinal enterococci (ENT) are the indicator bacteria used to test the microbiological quality of bathing water. In Croatia, the quality of bathing water is assessed by parametric method i.e. on the basis of the 90th and 95th percentile values which are calculated using the corresponding arithmetic mean and standard deviation of the logarithmic bacterial concentrations. There are also non-parametric methods for determining the value of a particular percentile, and one of them is the so-called Hazen method. In this work, we study how the different methods of percentile calculation affect the assessment of coastal bathing water quality. We have chosen a "problematic" location with frequently elevated bacterial concentrations as a test site.The results show that in the case of E. coli, the 95th percentile values calculated using the Hazen method are significantly lower than those calculated using the parametric method (Wilcoxon rank test, p < 0.05), while for ENT there was no statistically significant difference. For case of annual assessment, the difference in coastal bathing water quality was statistically significant for both ENT and E. coli (McNemar test, p=0.004 and p=0.02 respectively). In the final classification, the difference in coastal bathing waters quality was statistically significant only in the case of E. coli (McNemar test, p=0.025).