There are weaknesses in all types of research, but we can still learn a fair amount from it.
With this in mind, Norris and Ortega published a meta-analysis of L2 instruction of form (ie, grammar) in Vol 50 No 2 of Language Learning (2002).
Norris and Ortega felt they could substantiate trends more accurately if they did a cumulative analysis of short-term instruction in which the research methodology lent itself to being combined into a larger analysis. (Their article is about 90 pages long, so I have had to do a bit of condensing here.)
Note in the quotes below, when they say "L2 instruction" they refer to studies that looked at teaching form (ie, grammar). Their conclusions, based on 49 different research studies conducted between 1980 and 1998:
"In general, focused L2 instruction results in large gains over the course of an intervention. Specifically, L2 instruction of particular language forms induces substantial target-oriented change."
"The effects of L2 instruction seem durable. This can be concluded from the cumulative empirical observation that, although such effects tend to marginally decrease over time...it is the case that average effect sizes for delayed post-tests remain relatively large."
"On average, instruction that incorporates explicit (including deductive and inductive) techniques leads to more substantial effects than implicit instruction."
Again, the research is not perfect, but there is enough here to give solid indications that effects of grammar instruction are not just minimal.
And let me repeat: Instruction of grammar alone is *never* enough. However, it can help improve the speed of acquisition and help students achieve an ultimately higher level of accuracy in production.
Karen
http://karen.stanley.people.cpcc.eduCharlotte, North Carolina, USA
