Although I wish I had the time to put my two cents in on a number of questions that have recently been raised, I don't. I do, however, want to comment on the question of quantitative research. There has been very little done recently, for a number of reasons:

1.) Cost.  Collecting and analyzing samples (oral or written) is time-consuming and expensive.

2) Legal issues -- getting permission to use (analyze) samples of writing from an entire class is not as easy as it was 30 years ago.

3.) Deciding what "grammar" is to be studied.  (Someone already raised this question, but it is complex, especially since, even in this group, there is little agreement on the definition of grammar. Are we talking usage, or syntax?)

4) Samples -- how will samples be chosen? How will we know that the students have, or have not, had instruction in precisely those constructions that are to be analyzed during the preceding 6 months?)

5) Are the samples (raw data) available for inspection. As I suggested in my essay on the definition of the T-unit, in the previous, famous research, the researchers all defined the T-unit differently. Unless we can see the raw data, the studies are highly suspect.  (See: http://www2.pct.edu/courses/evavra/ED498/Essay009_Def_TUnit.htm)

Note too that the recent discussion (and different opinions) on how language is "mastered" in the first place also affects any quantitative research. One of the reasons that O'Hare's study is flawed is that he used the previous work of Hunt, O'Donnell, and Loban which showed that subordinate clauses naturally blossom between seventh and ninth grade. Probably for that reason, he chose seventh graders to study. But, once one sees what is going on, his doing so created a "sling-shot" effect which invalidates any application of his research to anything other than seventh graders. (And it may not be valid for them either.)

   I suggest that any quantitative study of the effectiveness of any approach to teaching grammar will be flawed until we get a better understanding of the natural development of what Hunt called "syntactic maturity." And we are a long way from that because we can't even agree on how to define basic terms.
Ed V.