Over the last several years, I've had the opportunity to see some really good elementary science lessons...and some truly awful ones. As I think about what made the difference, some of it is due to general instructional expertise, some is due to the orientation of the classroom (teacher vs. student centered), but a lot of it seems to be tied to content knowledge.
In elementary schools, science comes in a box. I feel like that is a dangerous symbol to implant in young minds, but that's another battle for another time. Teachers have a manual with step-by-step instructions (including what to say) and prepackaged materials. Science by convenience. Teacher pull out the items, read from the manual, kids fill in worksheets...and Voila! they've taught science.
This would all be well and good except for one thing: kids have questions. Kids want to know why and how. They have their own hypotheses (often misguided, but at least they're thinking) and ideas. Some teachers are very good about allowing kids to ask and predict. Others are terrified to leave the scripted lesson and have some real exploration.
The other issue I've seen that gets in the way of real learning is the quality of the "output" provided. By this, I mean the worksheets that come with the curriculum (specifically those which come with the FOSS kits---STC is a far superior program). There is little or no critical thinking required by students. The lessons are wasted opportunities to have kids capture the process that is science. Again, this is not necessarily the teacher's fault. S/He is using what is there---and has been assured that the materials have quality. But so many were developed before a standards-based era where more rigorous thinking was required. The experiments themselves are still strong---but the lesson structure is not.
It is overwhelming to think about the kind of out of the box thinking that would be necessary to better support student learning in science. Teachers don't have time (or knowledge/expertise) to revamp things...and I'm only one person. I think that if we can get teachers to ask better questions during and after science lessons---even if they themselves don't know the answers---we'll be a lot closer to student achievement.