This paper presents a Bayesian modeling approach to quantify the level of disinfection achieved from raw infectivity data. Drinking water disinfection experiments often use animal tests and a logistic regression data analysis to evaluate the performance of a disinfectant. The proposed Bayesian method addresses statistical problems in the commonly used logistic regression approach and uses more realistic probabilistic assumptions about the data to predict disinfectant dose and pathogen response relationships. The new methods are compared to the commonly used methods using one published data set for UV light disinfection of Cryptosporidium. The Bayesian approach yielded comparable log inactivation estimates to those from the conventional methods. However, the Bayesian method estimated UV dose required to achieve 3 log inactivation significantly different (smaller) from the same using conventional regression modeling approach based on the estimated log-inactivation as the known response variable values. Includes 6 references, table, figures.