used as testing material and short-answer format was employed. Hansen (1994)
used both detail and global questions to assess students micro as well as macro
understanding of the lecture and the researchfindings validated the researcher’s
hypothesis that high proficiency students performed well on both while low pro-
ficiency students performed much worse on global questions. A quite popular
assumption is that low proficiency students rely upon bottom-up process, decoding
phonemes and words instead of synthesizing clauses. But global questions need
summarizing information across clauses. Confronted with summarization tasks,
high proficiency students tend to use top-down processing in listening. In the
current research, high proficiency students actively build semantic association
between propositional units and tend to synthesize them and integrate them into the
existent discourse structure in mind. Generally, when the participants’higher
cognitive processes are targeted, bottom-up information processing is far from
enough, top-down processing becomes indispensable, for their schema knowledge
is also fully engaged.
We already discussed that discourse construction is the key to lecture compre-
hension in teachers’eyes according to their questionnaire data. They emphasize a
holistic comprehension of an academic lecture through identifying key points and
concepts, linking those points and then building a macro structure of it. For stu-
dents, the discourse construction factor is the fourth main component while items
concerning summarization are packed in thefirst main component, note-taking.
Thus quantitative data and qualitative data are triangulated for the importance of
reaching test-takers’ discourse construction process by designing appropriate
summary gaps. Albeit difficult, summary gaps cannot be traded off for convenience
in writing test items.
8.4 Triangulation of Task Demands and Cognitive
Processes
All the four main factors representing task demands of academic listening on which
teachers and students reach their consensus can witness the triangulation with
cognitive processes. Besides decoding and discourse representation, note-taking
represented by skills such as using symbols, outlining notes, retrieving information
from notes and transferring notes to the gaps by itself is a complicated cognitive
process. On the whole, these skills correspond to different depth of cognitive
processes. Therefore, the quantitative data do present key components that help us
define the task demands, but we also need to probe into the qualitative data, i.e.,
TAP data with the purpose to observe whether the task demands are realized in the
test-takers’test-taking process.
8.3 Linking Task Targets, Task Demands... 135