This study investigated the efficacy of retrieving EFL vocabulary from memory as a long-term retention strategy. Three learning treatments, rereading, recognizing and recalling target words, were compared with the assistance of 74 university students who underwent the treatments to learn academic American English during a 15-week semester. In addition to investigating the efficacy of the learning treatments, the study explored the effects of recognition and recall testing in relation to the treatments for possible interactions between learning treatment and test format. The study found that while rereading was the preferred student study strategy, recalling words was a more efficacious learning practice. Recognition learning was also less effective than recalling words, which suggests that the use of recall tests will promote long-term retention more than recognition tests such as multiple-choice tests. Learning treatment and test format comparisons suggested that the retrieval of words as a learning strategy was likely to be the most effective study practice regardless of test type despite transfer-appropriate processing inconsistencies.