This paper reports on a quasi-experimental study of the effect of different vocabulary study materials and strategies used by upper-intermediate English as a foreign language students in higher education. Students were assigned a selection of 163 words from the Academic Word List and were provided with different types of study materials. They were informed that they would be tested on the vocabulary five weeks later. The vocabulary assignment was embedded in their coursework, and the tests--translation of passages of academic writing into L2--were similar to the end-of-year tests for this course. Students kept logbooks detailing their study efforts (type of activity, duration and timing) and were post-tested twice. Post-tests show better results for two types of materials--thematically organized lists with L1 glosses, and similar lists with example sentences instead of glosses--but the differences were significant only on the immediate post-test. Time spent studying did not significantly affect the results. A more fine-grained analysis of the qualitative data showed that students, given time, tailor the study materials to fit their ingrained study habits, and that a majority prefer shallow strategies. Overall, gains were modest at around seven words/hour on average, which is much less than reported elsewhere, but this is clearly a factor of the challenging nature of the productive task in this experiment. (Contains 9 tables and 3 notes.)