What is the best class size for EFL classes at universities in Japan?
The intuitive answer is most likely, the smaller the better. However, a more nuanced answer might have to consider issues that are not normally considered in such a discussion. For example, one issue for EFL that can be usefully related to class size is placement, provided that such placement is based on valid assessment of the students abilities, experiences with various task types, and motivation (which links to needs). This is especially important at the upper beginning to upper intermediate range of levels because at these levels students tend to become quite heterogeneous in their ability to learn in class, work together, and participate cooperatively in activities. Even their language abilities, although labeled under one term, such as 'high beginner', can be quite different and strengths and weakenesses can seem idiosyncratic.
In my experience, it always seems that having placement is better than having no placement (the latter is often the case in EFL in Japan at universities). Especially if you are asked to teach a large EFL class. With valid and reliable placement, expansion of class size is possible because you can make predictions about how the students will behave as learners in the classes you are planning to run.
However, placement requires assessing all skills, plus factoring in needs, wants, motivation, etc., if at all possible. That is to say, groups of upper beginners through intermediate levels seem to individualize in many ways, such that it does help to make classes more manageable for communicative methods and materials if you can keep it to 15 or below but above 5 (and in the case of odd numbers, as a teacher I often become a pair partner, but I also use groups of three for a lot of discussion activities).
Two, it seems that most people in education will cite the need for smaller class size, while decision-makers typically push class sizes up. That has always been the case for me both as a student and as a teacher. In the case of EFL in Japan, with the dominant pattern being teacher-led, teacher-dominated, heavy use of Japanese, etc., large classes are manageable within that pattern.
The most common cause of negative critiques of a Japanese teacher of English would be 'the students didn't understand the material' and/or 'the students didn't do well on the department's exam (in the case of a senior high school, for example). Different expectations are projected onto 'English native speakers' (NS was fresh, energetic, made me want to learn English, gave me insights about his/her culture, etc.).
The profession in Japan is dominated by Japanese teachers of English, so programs, departments, institutions evolve their internal culture towards the dominant group.
Still, to sum up the discussion so far: most teachers want smaller classes. My basic rule at university EFL in Japan: once you give me more than 15 students, it doesn't much matter if it is 20, 25, 30, 35, 40--up to 60. I'm basically going to be demonstrating and running the same activities. The larger classes do take more energy and tend towards more off-activity behaviour, but after 20 years of looking at the same, I would be dead by now if I let that worry me too much.
This also brings to mind the ongoing debate about class size relative to academic performance in the Anglophone global media. You often hear it discussed that certain E. Asian countries (e.g., Japan, S. Korea, Taiwan) have large class sizes compared to the 'west' and yet continue to outperform on achievement tests (that have been standardized enough to allow for international comparisons).
However, it's a constant here in Japan when the larger society debates the education system: teachers want smaller class sizes because of the management and teacher-student relation issues (overcrowded, understaffed schools have more bullying problems). And most people who support educational reform consider classes that are too large as undesirable (with a class size of 30 often being the 'magic' number). Moreover, Japan has been overall been moving down in international 'league tables' of academic performance. Even as class sizes have become smaller--because of official changes to make classes smaller, but also demographic decline). More often than anything to do with class size, less rigorous curriculum and teaching methods are usually criticized as the cause.
Another aspect of the issue, though, is how is teaching and learning a second/foreign/additional language is the same as or different from the regular 'content' of the curriculum.EFL in Japan is largely handled as simply another curriculum subject that is on high stakes tests. In western academia, where most of the meta- and pseudo-theories about LT, LL, SLA etc. are created and disseminated, however, the issues under discussion have more to do with how second language learning is different from native language acquisition. Also largely ignored in the schools of theory and experimental research are issues such as the relationships (1) between general learning and language acquisition and (2) literacy and overall language development (including second languages).