Dale blogged about this very topic recently, after reading John Hattie's "Visible Learning for Teachers". The research seems to indicate that when students know where they're going, they are more likely to get there. It kind of makes sense: The video games our children so often play always begin with a goal or a "mission", so why not our math lessons?
On the other hand, the goals of a lesson can be manifold: A recent lesson on probability I taught to my grade 3 students focused on the vocabulary of the new concepts we were learning: certain, possible, impossible, likely, unlikely, equal chance and so on... So, one goal was to learn and correctly use these new words. But another goal was communication. I wanted students to communicate their solutions effectively, using a sentence to share their "answer" to a problem I had presented, and to include some sort of explanation (diagram or illustration, math term, or more words) to tell how they arrived at said answer. A third goal was reasoning. We've been working on asking ourselves, before we present solutions to a problem, "is this a reasonable answer? How do I know?" Finally, since students were working with a partner to discuss this particular problem, we were focusing on collaboration for learning skills; I was observing how well they listened and shared with one another during the work period, so that I could collect some data for report cards.
Am I to share ALL of these goals with my students? And if so, when?
After fiddling around with some different models, I have found it effective to highlight the mathematical goals of a lesson or unit early in the lesson, right after the "minds on", and before presenting the actual problem that students will work on. An addition, I will often draw students' attention to the criteria which we co-constructed earlier in the year, about how to best communicate thinking, which are permanently posted in the room for handy reference. I do this as they are working, but before they share their solutions.