This really is one of my biggest pet peeves about ELA teaching - over-correcting student writing mistakes. What do I mean by "over-correcting"? Let me back up a bit and describe how I avoid over-correcting.
When dealing with usage problems (and this includes punctuation, grammar, mechanics, etc), it is important to deal with one problem at a time and spend a lot of time on it. I'll spend a month or more helping students with one problem (not every lesson or for a whole period, mind you). A big one for my students - one that usually takes more than a month - is subject verb agreement (SVA). At the beginning of the year, I'll cover basic SVA. Students will practice in groups, those who have the down helping those who don't. After spending a few lessons doing group work, I'll move students into individual work. Now, on writing assignments, I only correct SVA errors as that is the only problem we have dealt with as a class. I don't correct run-ons and fragments even though my students have problems with those. When I conference with students about their formal writing assignments and they have SVA errors, I will spend some time in the conference focusing on those errors. If in a final draft of a paper the student has those errors, I will mark the error with a reiteration of why I'm marking it. This process may take a month or more in order to really focus on SVA.
The next step, after I feel that we've spent enough class time on SVA (again, usually a month or more - not all lessons during that month - just here and there as needed, once or twice a week mini-lessons), I'll move on to run-ons and fragments. The process is repeated with group work then individual work and then in conferences. The difference is that now I'm helping students in conferences and in my written comments with two skills: SVA and run-ons/fragments. Even if I see loads of comma errors, I don't mention them or mark them. I put them in the back of my mind for later areas of study.
This represents teaching and learning about mechanics that is focused and effective. Why? When students reach the secondary level, we can assume two things. First, they've probably received grammar instruction before (no, you aren't the first teacher to notice they have grammar problems). Second, many of their errors are so ingrained in how they use language that one mini-lesson or mark on a paper is not going to do the trick in reprogramming their brain and how it understands the use of language. At this point, they've been making the mistake so much that it looks and sounds correct to them. This will take time and focus to switch.
In addition, I don't correct errors that we haven't discussed in class. Yes, I can assume that students have had grammar instruction before, but I don't always know the topics that were covered or whether this grammar instruction was good. In the best of scenarios, I would be able to ask the teachers who had the students before me what they covered and how, but we in education are used to not having the best of scenarios.
Actually, let me highlight something ... I don't even correct errors on papers. I mark them. There's a difference. Correcting is when you put the 'correct' answer on the student's paper; marking is when you mark that they made a mistake but do not give them the 'correct' answer.
There is this distinction (and this gets me back to my original point) because correcting student mistakes is not instruction. Correcting their mistakes is editing, and if that is what you enjoy you should have gotten into book publishing. We are teachers and our job is to teach students how to use the English language. Correcting mistakes on their paper - and worse, over-correcting every single mistake we can find whether or not we have covered it in depth in class or not - is not teaching. If you correct errors on their papers on a first draft, they will go back and fix them on their computer mindlessly. They then turn in a paper that is 'perfect,' and you feel good about yourself. But, did they learn? Some would argue, yes they did. So, how? Osmosis? They learned just because they fixed the mistake? If that is learning, then they would never make that mistake again. How many times have you corrected an error on a student paper and that student makes the same exact mistake over and over? (Now, there are situations in which the student made a silly error and you 'caught' it. The student really knows what to do but was just a bit careless. There, your problem is not grammar knowledge it is proofreading skills. That's different.)
I heard a teacher at the end of last year say, "I'm sick of fixing all their mistakes. After twenty, I'm going to stop." I laughed to myself because everyone is so impressed with how well her students write, when in fact her students write so well because they have an incredible editor - her! We are not editors; we are teachers. We need to be focused with our instruction on mechanics. One topic at a time for a significant period (when I say significant period, I mean that we take 20 minutes once or twice a week for a month or more if needed - not every day all day for weeks on end) and work with students on conferences to explore their errors. Marks on a paper are not instruction. Teacher-student interaction - either whole group, small group, or individually - is instruction.
Focus on grammar instruction, not correcting mistakes on paper.