Class size or teaching?

I've found that one of the most controversial debates when it comes to education, especially in developed countries, is what is the best way to spend the scarce money we have on education. The debate often boils down to a choice between two things: smaller classes, or better teachers.

Personally, I've always felt it's obvious that we should pay teachers better, especially in Malaysia. There is no reason to hold teachers to a rigid pay scale based on seniority, let alone the pay scale of the civil service. Teachers perform a much more important job than civil servants, and I daresay deal with a lot more stress. It only makes sense to pay them more, and especially to pay the outstanding teachers more.

The argument for smaller class sizes seems rather vague to me, and it's predicated essentially on the notion that it's hard for good teachers to pay attention to more than 20 pupils at a time. While I think there is probably room to rethink the traditional classroom dynamics of one teacher lecturing/supervising a classroom of pupils, I think that diminishing marginal returns kick in pretty quickly past the point of 30 pupils. While I don't have much data, from my personal experiences, the rowdiest classrooms have been those of 35 or more, with not much difference below that, especially not below 30 students.

Recently my attention was drawn to a University of Auckland study, which claims to have brought together 50,000 different studies and looks at a total of 83 million pupils around the world. The study's conclusion is that class size is not very important, and that the quality of teaching easily outstrips class size in terms of importance. The professor who authored the study is John Hattie, who I have not heard of before. He seems to have published quite a bit on this subject, but I was unable to turn up much about this latest study of his, so it is worth taking with a grain of salt.

Nevertheless, I think it's still a question worth asking: what matters more, teaching quality or class size? Obviously Malaysians have some other things to worry about, such as our horrible exam system, but there already seems to be a general consensus that we need to raise the standards of our exams, and try to orient them away from general memorisation of facts. There is, however, nothing close to a consensus on the question I'm raising.

What do you think? Should we spend more money on hiring good teachers to teach to classrooms of 30 or 40, or should we spend more money to hire the same quality of teachers but with smaller classrooms of 20 to 30?