Re: (teach) how helpful is grammar instruction? (Norris & Ortega)

Karen described the meta-analysis that Norris and Ortega (2000) did on the effects of grammar instruction, concluding that it found instruction very helpful. I'd like to offer a different view on what N&O found.

The key phrase in all this is the vague 'effects of instruction'. To understand N&O's findings (and form-focus research in general), you have to ask 'the effects of instruction on what?' For most of the studies that N&O used, the 'what' was performance on formal grammar tests. To no one's surprise, these studies found instruction very helpful. And this is why their overall results look so impressive: studies that relied on formal grammar tests made up the great bulk of their sample. When you focus on the six other studies, which asked how instruction affects more realistic, communicative use of language, the results are not impressive at all. And I've argued elsewhere that these unimpressive results probably overstated the effectiveness of instruction, substantially.

This is not about whether the research had weaknesses, as all research does. It's about what questions were and weren't addressed in that research. If the question you're interested in is whether instruction improves learners' ability to use 'explicit declarative knowledge under controlled conditions' (N&O, p. 486), then N&O provide a clear positive answer. If you're interested in how instruction affects practical ability to use the language, the findings point to a very different answer.

John Truscott