another eye opener regarding standards/curriculum is that there's a lot of opportunity for triage. I've spent some time looking at the Arizona standards, which line up with the AIMS (Arizona's FCAT or w/e), along with data showing how much each area and standard show up on the AIMS. There's stuff that, as a teacher shooting for best overall scores and working within time constraints, would be prudent to actually skip teaching entirely. Teaching your 3rd graders about vertexes is pointless (something like 2% of the math portion) compared to spending an extra week on multiplication facts and word problems (something like 15% of the test).
The improvement standard for pay is better than overall performance, but you still run into the same issues of teaching for the test, potential dishonesty, and still having problems getting through to these keeeeedz.
Take ESL learners for example. Say you get a group of 5 or so fresh-from-Mexico kids with virtually no English ability. You work your ass off and by the end of the year, they can read and write a little bit, and maybe even participate in class every now and then. All that work will not show up on a standardized test because despite their huge improvement, at that point they are still probably lacking the lexical knowledge necessary to take the test, even considering the accommodations afforded to ESL students (teacher reading directions, reading math problems, etc).
And of course there's also the problem of students who plain don't give a shit and having no support in their home to back you up.
That's why it's so hard to say "oh he/she is a bad teacher, look at those test scores" when in reality, you need the best teachers with the most needy kids and even then you can't expect miracles.
XI Wiki


