Wim Gombert

92 CHAPTER 6 almost exclusively, and the learners have a great deal of L2 exposure and use. e AIM method has enough materials for the rst three years, and the rst author designed a follow-up program with similar L2 exposure principles for the nal three years. Key elements of this 6-year, DUB program were L2 exposure and active use. is integrated DUB program aimed at developing all skills, but activities focused on oral skills while writing, listening and reading skills were expected to develop implicitly. WRITING INSTRUCTION Even though the overall SB and DUB programs were quite di erent in terms of L2 exposure, both groups received exactly the same 30-hour intervention program with the same amount of L2 exposure to prepare for their nal written exams (the data used for the present study). is included receptive and productive activities on a variety of topics, including abortion, immigration, living for a thousand years, tattoos and stay-athome fathers (see Appendix B for an example of the materials used to prepare students for the respective exams). Both groups were presented with audio-visual exposure based on these topics during class, with follow-up activities, such as constructing arguments in the target language and guided writing tasks, in which students were prompted with ideas and the necessary vocabulary. e teacher selected the four favorite topics, based on students’ votes, to be used during the nal exams, facilitating test preparation by students. As the groups came from two di erent cohorts, they had di erent favorite topics. During the test, students were o ered two topics which were randomly chosen out of the four they had been able to prepare. In their nal year, there were two testing points for writing: one in December (a mock test) and one in April (the nal test). We used the December test as it was assumed to have generated more spontaneous language than the April test. During the exam, students were presented with two of the four topics discussed, and self-selected one for their essay in which they were required to write a minimum of 200 words within 50 minutes in the computer lab at the school. (See Appendix C for an example of such an exam). SB students had the choice between ‘living for 1000 years’ and ‘migrants crossing the Mediterranean’; DUB students had the choice between “legalizing abortion in Chili” and “stay at home fathers”. e researcher monitored both tests and could ascertain that the students made little to no use of dictionaries. MEASURING CHUNKS Even though chunks may be easy to de ne, they are di cult to operationalize as both relative frequency and association strength play a role, and they are di cult to distinguish. To be able to compare studies, we will follow Piggott (2019) and Hou (2017) (see Table 12 for an overview), who based their studies on Verspoor et al. (2012), which