There is still another possible intervening factor which is of importance and of great interest to teachers. It is suspected that the effects of type of learning may be moderated by proficiency. In a simple mean comparison, the result is only about a main effect, assuming that a particular type of learning is good for all the learners. However, it may be possible that a certain type of learning works well for students at a low proficiency level, while students at a higher proficiency level may benefit more from
another type of learning. This is suggested by Robb and Ercanbrack (1999), who conducted an experiment of the effects of direct test preparation on TOEIC score gains by Japanese college students. They did the experiment on two samples: English majors and non-English majors. Results showed that non-English majors, whose proficiency level was lower than English majors’, improved their scores at a statistically significant level, but English majors did not appear to have benefitted from the instruction. These results from the two different samples suggest, as Robb and Ercanbrack (1999) say, that test-preparation may be beneficial for students of a low proficiency level, not students of a high proficiency level. In other words, it is likely that the effects of test-preparation are moderated by students’ proficiency. If this is the case, it is possible that the effects of other types of learning are also moderated by students’ proficiency.