To study the effects of inaccuracies in the parameter probabilities of a Bayesian network, often a sensitivity analysis is performed. In such an analysis, one or more parameter probabilities are varied systematically, by means of which their functional relationship with an output probability of interest is established. For reasons of computational complexity and difficulty of interpretation, sensitivity analysis of a Bayesian network is restricted to a single parameter, or to two parameter probabilities at most. From the results of such restricted analyses however, it is not easily predicted how inaccuracies in multiple parameter probabilities will interact and jointly affect the output probability of interest. Another general technique for investigating the effects of parameter inaccuracy is to perform an uncertainty analysis. Taking a sampling approach, this technique is less prone to computational problems than sensitivity analysis is. There being little experience as yet with uncertainty analysis of Bayesian networks, we re-consider this technique for studying the effects of inaccuracies in a network’s parameter probabilities and provide some insights for the interpretation of the results obtained.