*From an interesting piece on study habits come a few simple lessons on what helps people actually learn things:
1.) studying the same thing in different places (so that you learn thing in different contexts)
2.) studying more than one thing at once (so that you make connections between the different subjects)
3.) studying the same thing at different times (so that you forget and re-learn things -- "forgetting is the friend of learning")
The piece would seem to support the idea that we should "spiral" when we teach students instead of "scaffold" -- things should be briefly introduced and re-introduced in different contexts instead of teaching one subject one day and a different one the next. I'd also add that this seems like a pretty good argument for more interdisciplinary teaching/learning -- something that's difficult to do at the K-12 level and for which there's little incentive for college professors to try (indeed, there's probably a discentive since more specialization often equals a greater chance of earning tenure).
*Aaron Pallas proposes a good rule on his new blog: "the weight accorded to any one element of a teacher evaluation system should be proportional to the uncertainty about the inference which is drawn from that element". In other words, if we're not sure about how meaningful something is then we shouldn't place the whole weight on that item (that goes for both principal evaluations and value-added scores)
* Are middle schools hurting students? My experience teaching in a middle school has convinced me that those findings are at least plausible (as do my experiences when I was 12-14). This isn't to suggest that K-8 schools are a silver bullet, just that, in theory, they have more potential than middle schools to advance the learning of 6th-8th grade students (especially if those students can take a leadership role in the school by tutoring younger kids, serving as crossing guards, etc.).