This article reports on the findings obtained from an exploratory study on the effectiveness of feedback in a computer-based learning environment. The main aim of the study was to gain insight into learning outcomes and student perceptions in relation to different types of feedback. This aim was operationalised in terms of an investigation amongst Dutch-speaking Belgian second-year students of translation studies. While translating grammatically difficult sentences from Dutch into English in a computer-based environment, the students were confronted with five feedback types, which differed in their degree of explicitness concerning the grammatical problem hidden and the mistake made, and the degree of learner involvement required towards finding the correct solution. Questionnaires and grammar test and delayed post-test data yield insights into the learners' reactions and learning outcomes in relation to the different feedback types. We found that more explicit feedback, combined with adequate depth of processing, led to better learning outcomes and more positive student perceptions. (Contains 3 figures, 1 table, and 3 notes.)