This paper presents a method for assessing causality in power quality under stationary conditions, resorting to an index tipically employed in graph theory, the centrality index. Subsequently, the method is illustrated by means of two real systems, allowing a comparison among the results obtained from other existing methods. A discussion regarding the results and its interpretation is presented as well.In comparison to existing methods, the proposed method does not discard not responsible agents, it directly assigns responsibility. Two application cases are presented, so as a discussion on the obtained results.