The present studies directly test the usefulness of two English-language working memory capacity (WMC) assessments with two samples of students whose native language was not English. Participants completed two widely used complex span tasks, Reading Span (RSpan) and Operation Span (OSpan), in English. To determine whether the well-established relationship between WMC and Raven's Advanced Progressive Matrices (RAPM) would be observed when span tasks were not given in the native language of the participant, span scores were regressed against performance on the RAPM. Results indicated that while OSpan was a reliable and valid predictor of RAPM in non-native-English speakers, RSpan administered in English was not. (Contains 3 tables.)