Evaluating phrasal verb exercises: An investigation into the effectiveness of error-free and trial-and-error learning
Although phrasal verbs are perhaps the most challenging type of verb phrase for L2 students to learn, only a handful of studies have looked into the effects of methods to enhance their acquisition, but these studies focused exclusively on the use of exemplary study input materials. The present thesis investigates this topic by examining and comparing learning procedures that consist of a study trial followed by a test trial (retrieval conditions) with learning procedures that include a test trial followed by a study trial (generation conditions). In the field of cognitive psychology, these two procedures have received considerable attention for the learning of single words but less so in the field of applied linguistics for the learning of phrases. In essence, the use of retrieval and generation conditions for the learning of phrasal verbs by non-native speakers of English was examined in three separate studies in this thesis. Additionally, it investigated the extent to which versions of these conditions occur in general ESL/EFL course textbooks. The studies were carried out in L2 classrooms and data were collected electronically from a group of L2 learners whose L1 system lacks phrasal verbs. In the first study, 199 students from five parallel classes were assigned to either a retrieval condition or a generation condition. In the retrieval condition, the study trial presented a phrasal verb and a paraphrase of its meaning, and then the test trial displayed the initial letter of the verb followed by the paraphrase of the phrasal verb’s meaning. The generation condition was comprised of the same two trials; however, the order of their presentation was reversed, so the test trial preceded the study trial. In the second study, 153 students from four parallel L2 classes were assigned to one of four conditions. All the conditions were comprised of the same study trial and test trial. The study trial presented a phrasal verb along with a paraphrase of its meaning. The test trial displayed the verb of the phrasal verb followed by the paraphrase of its definition. In the study-test condition, the study trial and the test trial occurred consecutively, while in the study-delay-test condition, they were separated by approximately 6.5 minutes. In the test-study condition, the study trial occurred immediately after the test trial, while in the test-delay-study condition, a 6.5-minute interval separated the two trials. The last study examined contextualized versions of the retrieval condition and the generation condition on 172 L2 students from six parallel L2 classes. This thesis produced the following main findings. First, a vast majority (72% to be exact) of phrase learning conditions in existing course textbooks are generation-oriented. Second, the experimental studies showed that retrieval learning conditions offer significantly better short-term learning of phrasal verbs than generation learning, although no such advantage was found in long-term learning retention. Third, contextualized learning (i.e., learning phrasal verbs with the contextual support of exemplars) yields more effective learning than decontextualized learning does. Fourth, errors produced in generation conditions are difficult to unlearn. Overall, these findings provide us with some new insights about the learning and teaching of phrasal verbs. These results also have direct, meaningful pedagogical implications for the teaching of phrasal verbs as they show specifically which teaching procedures are more effective and which ones are less effective or ineffective for the learning of these difficult English verb phrases.