Drop standardized tests as the fundamental benchmark of teaching/teacher effectiveness - for many of the reasons commonly given, PLUS, and MORE IMPORTANTLY, because of the fact that they capture FAR TOO LITTLE of what learning actually does occur in and by students and, therefore, FAR TOO LITTLE of what must be known and considered.
Remember this symbol: "L-". It tells what we need to keep in mind and to follow and faithfully adhere to - if we are to reach right conclusions about teacher quality and reward. (And, by the way, remember that teachers are NOT motivated by money.)
"L-" means "longitudinal" and "-" means "subtract". If we're really serious, we must measure ALL the learning that occurs and then subtract Point A in time from a later Point B in time. This difference - considering ALL things taught and learned - should then be compared and analyzed to reveal the best vs. worst results year-over-year, year-after-year. This should be (and can be) done over the long haul and on a "ceteris paribus", i.e., "all-things-considered", basis. All that, plus, of course, teacher self-evaluations and competent on-the-scene managerial observation and reviews by Principals, are pertinent as well. (I close below with an observation - often unstated - that relates here, too.)
Use researchers and teachers' assistants - not the teachers to record and keep track of ALL the learnings (qualitative and quantitative) that occur. This is far more inclusive than "standards" now focused on too exclusively. Besides, doing this well will bring along performance against "standards" even more strongly. (NOTE: This is a call for FAR more clerical support. Teachers' time as teachers is way too valuable to waste on such data-intensive record-keeping.)
Simply stated, the focus of measurement must be on PROGRESS made per student, not on LEVEL. Class levels (say, 3rd grade) tests done regularly by standards-based systems are - by definition - measuring the status, i.e., the LEVEL, of DIFFERENT student bodies each year. That makes the comparisons and the conclusions commonly reached and publicly reported - as an indicator of teacher quality - deeply flawed - I mean seriously, inherently and incurably flawed. Data (qualitative AND quantitative) must be recorded, reported and tracked BY student to reach any valid conclusions on teacher effectiveness. That's "longitudinal" and it's indispensable - and it's broader based than merely relying on numerical test scores (standardized or otherwise).
A related but often unspoken observation: Principals are in the indisputably BEST position to evaluate teachers, so are they failing in that responsibility? By virtue of this TV program being on the air, the answer must be: "Apparently so." So, why is that? To what degree, then, and in what ways, are bad principals the underlying - if unspoken - problem? What to do about that?