Error-prone and error-free exercises on verb-noun collocations

Onderzoeksoutput: Unpublished abstract


Many contemporary L2 study materials include exercises which require learners to match the constituent parts of collocations, for example by supplying the appropriate verb (e.g., meet) to precede a given noun phrase (e.g., the deadline) in gapped sentences. However, a study by Boers, Demecheleer, Coxhead and Webb (2014) found that wrong choices made in this type of exercise can leave undesirable memory traces, despite corrective feedback. This finding suggests that, in the case of collocation learning, at least, it may be advisable to design procedures where wrong exercise responses are avoided. The quasi-experimental study reported in this presentation investigates this possibility. Two groups of high-intermediate EFL learners (n = 19) were given exercises consisting of gapped sentences where the verb of a collocation was missing. In one treatment condition (the error-prone procedure), the students tried to supply the verbs and were subsequently given (corrective) feedback on their responses. The students crossed out any wrong responses and wrote down the correct ones on their work sheets. In another condition (the error-free procedure), the students were given a hand-out with examples of the target collocations to help them complete the blanks in the exercise. Both groups of students were given four exercises each targeting 14 verb-noun collocations, spread over four lessons. They took a pre-test two weeks prior to doing the first exercise and a post-test two weeks after completing the fourth exercise. Like the exercises, the tests consisted of gapped sentences with the verb of the verb-noun collocations missing. Comparisons of pre-test to post-test performance on all the item responses which were wrong in the pre-test (totalling 505) show a significant advantage of the error-free over the error-prone exercise procedure, in par- ticular where wrong exercise responses were given under the latter procedure. Only 16.5% of the incorrect exercise responses in the error-prone condition were followed by correct post-test responses, indicating that the corrective feedback often failed to entrench the correct collocation in the learner's memory. In addition, almost 4% of these students' correct pre-test responses were replaced by wrong ones in the post-test. We ar- gue that the lack of semantic distinctiveness of the verb (e.g., make, do, have), in many collocations as well as competition from semantically related verbs (e.g., say, tell, speak) makes it particularly hard for learners to block unhelpful cross-associations during the deliberate learning of sets of verb-noun collocations. While the exemplar-modelled, error-free procedure yielded generally better outcomes – 38.5% of incorrect pre-test responses were replaced by correct post-test responses – it evidently did not work wonders either. The direct provision of exemplars probably made the task too easy and shallow. We conclude by propos- ing ways of designing and implementing collocation exercises that can promote more engagement and yet minimize the risk of error. Boers, F., Demecheleer, M., Coxhead, A, & Webb, S. (2014) Gauging the effects of exercises on verb-noun collocations. Language Teaching Research, 18, 54-74.
Originele taal-2English
Aantal pagina's1
StatusPublished - 28 aug 2015
Evenement25th Conference of the European Second Language Association - Aix-en-Provence, France
Duur: 26 aug 201529 aug 2015


Conference25th Conference of the European Second Language Association


Duik in de onderzoeksthema's van 'Error-prone and error-free exercises on verb-noun collocations'. Samen vormen ze een unieke vingerafdruk.

Citeer dit