I used to think that standardized tests were not suitable to check on the improvement made by reading. Although my students are required to read over 2000 pages for one year, the average used to be only 1500 pages (a little over 300,000 words). I was not sure whether such a small amount of reading done by my students would make any significant change on standardized tests. Furthermore, students usually take all sorts of different English classes. It is hard to conclude what caused the gain even when the score goes up. Still I was fortunate to have several students who believed in the power of reading and read during the time when they were not taking classes or during their vacation. They gained significantly on the TOEFL from reading alone. The students read 100 to 150 pages per week and gained about 3.2 points on the TOEFL per week. That was about the same rate as that of international students at a TOEFL preparation course in the US (Mason, 2006, International Journal of Foreign Language Teaching).

There are many research reports all over the world about the effects of reading. There is no doubt that reading is good for improving language competence. However, I feel that some people have regrettably used these results to justify employing ER in skill-building contexts without appreciating the FVR/comprehension-based contexts in which the studies were done.

Returning to the issue of pre- and post-testing, another big reason to do it is that students like to see their improvement. For example, when they see that their fluency and accuracy in writing improves from only listening to stories and reading books, they are amazed. They start having more confidence in the method that they are using.

You've done a great deal of research with ER. What's driven you in that area?

I get joy out of finding an answer to my question from data analyses. Also, I think that it is my job. I believe that unless we try to find out whether Krashen's theory is correct, using studies, there will be no advancement in the SLA field.

Some teachers who teach SLA theory say that the Input Theory is old. I wonder what makes them say such a thing? Have they come up with a new theory? Can they offer something better than the Input Theory? I think that there are only two theories in the SLA field right now, the skill-based approach and comprehension-based approach. Other approaches are only derivations of the skill-based approach. No one has proven that the Input theory is wrong.

Some say that the Input Theory cannot be proven to be false therefore it is not a theory. I wonder why they give up on the theory so quickly without thorough investigation. Some say that what Krashen says is so obvious that it is not interesting. Then they go off and investigate the effects of trained self-feedback on revision (?!) and forget about comprehensible input and do not use reading or listening in their course and have students do skill-building activities! If it is so obvious, why don't they use it?

A popular argument against the Input Theory is that Comprehensible Input (CI) works but it is inefficient (Long and Robinson, 1998). People agreed with this and other criticisms, did not investigate the issues, repeated them to others, and told others that the input theory is old. However, my studies show that CI is a lot more efficient than the skill-based approach. I have found that input alone is the most efficient. When I was reading Krashen's book that was published 20 years ago, I saw that he had already said this. This kind of finding uplifts my spirit.

7