Results (
Thai) 2:
[Copy]Copied!
8.2 Error Analysis The means of measuring accuracy in past literature has been the number of error free clauses
as a percentage of the total number of clauses in a passage (Skehan & Foster, 1999). In the
Discourse Analysis section of this paper, this very same formula will be implemented. For
the time being, the researcher has calculated the total number of errors, instead of a
percentage, as some clauses contain multiple errors that would be missed if the previously mentioned formula were used. When looking at errors, there were significant results over the five weeks of study. Week 1
had a relatively high number of errors. Errors increased significantly on Week 2, remained
high on Week 3, but gradually declined on Weeks 4 and 5. The following graph indicates average number of errors per student over the five weeks of
study:198
Naturally, the teacher did give some grammatical explanations and corrections in class, as
well as corrections and feedback on weekly email writing assignments, but the bulk of
classroom time was spent on speaking skills, alongside vocabulary development in context
of the lessons.
Perhaps the decrease in errors near the end of the study could in part be the result of greater
selfmonitoring of speech acts, in other words, greater metacognition (Oxford, 1990). Also,
there seems to be some shift that students are making, on a shortterm basis at least, from
declarative knowledge to procedural knowledge. If true, this would agree with Anderson’s
theory that the more students speak the better they get with practice (Anderson, 1983; 1995). Unfortunately, the timeframe of the study was too short to show such change. In addition,
the results of the present study, in regards to accuracy, contradict the results of Crookes’
study; Crookes’ study showed little variation in accuracy over time whereas the present
study showed considerable improvement in accuracy over time. A much more plausible answer is that, as the speech acts were carried out for homework on YackPack, students were doing more functional planning, including using dictionaries,
grammar books, and other provisions. Also, some students could have been reading off
scripts, or rehearsed their speech ahead of time.199
8.3 Discourse Analysis to Determine Use of CSs From now, the study will undertake a discourse analysis to determine whether CSs are being
used by students, and how effectively they are using them.
Included in CSs are compensatory strategies to compensate for lack of procedural
knowledge; one of the most obvious compensatory strategies, that even native speakers
frequently use, is discourse markers. Unfortunately, many Japanese EFL students code
switch, and use Japanese discourse markers in their English speaking. The instructor has
noted this phenomenon in his classes annually, at several universities. Therefore, he has
made a point of teaching discourse markers explicitly in all his general communication
classes.
Interestingly, four out of the 16 students being studied used Japanese discourse markers the
first week. This number dropped to two students on the second week, and virtually no
students were using Japanese discourse markers by the third week.
Additionally, other CSs began to diminish, as well. In fact, the first week’s speaking tasks
seemed the most natural, and easiest to analyze. Here is one example from the first week:
A: My house is surrounded by moun…green…nandatakke…and…
This simple example has five CSs in it. The student almost said the word “mountain”. However, “mountain” was an approximation for another word the student really wanted to
say. The student then attempted a repair by using the word “green”, indicating transferring;
“green” is a literal translation of the Japanese word “midori”. The student then tried code
switching by saying the Japanese discourse marker “Nandatakke?”, with the literal meaning
“Uh…What is it?”, which could, also, be interpreted as an appeal for assistance. Finally, when the student couldn’t find the correct word, the student resorted to abandonment. Still, we can observe definite use of interlanguage which would be expected in such unplanned
speech.
It was very difficult to analyze any of this student’s later speaking tasks as they all
appeared scripted. In fact, the more the study progressed the more scripted several students’
speaking tasks appeared. Therefore, it would be interesting to undertake a qualitative analysis of two students: one, whose speaking appears to be unscripted throughout the study, and another, whose speaking
appears to be scripted throughout the study. In addition to a general descriptive analysis, the
researcher has decided to investigate factors such as lexical density, lexical complexity,
Being translated, please wait..
