We report an experiment investigating how concurrent verbalisation during a task can affect performance (a so-called "reactivity" effect). Participants studied three-variable line graphs while (a) concurrently thinking aloud or (b) silently studied the graphs and provided an interpretation once they felt they had understood it. Results showed that verbalisation hindered performance significantly compared to the silent condition. To support the claim that the act of verbalising was hindering performance, competing explanations were also tested, which confirmed thinking aloud as the most likely cause. This contradicts claims by Ericsson and Simon (1993) that thinking aloud reflects but does not affect performance and provides further evidence that verbalising thought processes can hinder performance.