Changes in sexual signals have the potential to promote rapid divergence and reproductive isolation among populations of animals. Thus, identifying processes contributing to variation in signals is key to understanding the drivers of speciation. However, it is difficult to identify the processes initiating changes in signals in empirical systems because (1) the demographic history of populations under study is usually unclear, and (2) there is no unified hypothesis-testing framework for evaluating the simultaneous contribution of multiple processes. A unique system for study in the Hawaiian Islands, the planthopper species Nesosydne chambersi, offers a clear demographic context to disentangle these factors. By measuring variation in male vibratory sexual signals across different genetic populations on the island of Hawaii, we found that that multiple signal traits varied significantly between populations. We developed a mixed modelling framework to simultaneously test competing hypotheses about which processes contribute to changes in signal traits: genetic drift, sensory drive or reproductive character displacement. Our findings suggest that signal divergence proceeds along different axes for different signal traits under the influence of both neutral and selective processes. They are the first, to our knowledge, to document the relative importance of multiple processes on divergence in sexual signals.