Purpose: To investigate whether working memory capacity (WMC) modulates the effects of to-be-ignored speech on the memory of materials conveyed by to-be-attended speech. Method: Two tasks (reading span, Daneman & Carpenter, 1980; Ronnberg et al., 2008; and size-comparison span, Sorqvist, Ljungberg, & Ljung, 2010) were used to measure individual differences in WMC. Episodic long-term memory of spoken discourse was measured by requesting participants to listen to stories masked either by normal speech or by a rotated version of that speech and to subsequently answer questions on the content of the stories. Results: Normal speech impaired performance on the episodic long-term memory test, and both WMC tasks were negatively related to this effect, indicating that individuals with high WMC are less susceptible to disruption. Moreover, further analyses revealed that size-comparison span (a task that requires resolution of semantic confusion by inhibition processes) is a stronger predictor of the effect than is reading span. Conclusions: Cognitive control processes support listening in adverse conditions. In particular, inhibition processes acting to resolve semantic confusion seem to underlie the relationship between WMC and susceptibility to distraction from masking speech.