"Given the drawbacks of pay-for-credentials schemes, a natural alternative is to consider pay-for-performance, in which the best teachers are identified by their principals, school inspectors, or through some objective measure such as student test score gains."

Reading between the lines here, it seems that pay levels are the only thing considered when attempting to understand the declining quality in teachers. I think it's the imbalance between the quality of work expected of teachers, where the paper trail to determine little johnny's social studies grade is held to the same standard as the engineering analysis of a bridge design, and yet along with the pay, the professional respect, support & social standing are just nowhere near the same level. They are treated with suspicion by the parents, contempt by the students and are frequently ignored or targeted by the administration. They are given little authority over what they teach, how they teach it or how they manage their class and yet society seems determined to lump them with all of the responsibility.

Teaching is unlikely to ever become a profession that people chose for the money, so using small pay variations to draw quality people is unlikely to work. The more schools become overloaded with systems of bureaucracy the less rewarding it is for those who have a passion for teaching.

"But don’t take my word for it. If we want to know the best way to identify the best teachers, let’s run a series of randomised trials: pitting the current system against various alternatives to see which one comes out on top. In the same way that we test new drugs before putting them on pharmacy shelves, we ought to be sure that strategies to attract and keep talented teachers actually work before rolling them out nationwide."

Teachers are just like pharmaceuticals? Interesting analogy. Presumably there will be opportunities to sue when the thalidomide teacher happens? The nonsense belief in so "test scores" while convenient for managers and quantitatively inclined researchers is a serious problem. Just what is being measured? How do we know? Does it correlate with anything other than other test scores? Look at test scores (where you can find them) of folk we regard as really smart or who have made huge contributions to society. You mean a B+ in history does not make me a good citizen? If we can get past the nonsense accountabilism (Weinberger: http://harvardbusinessonline.hbsp.harvard.edu/hbrsa/en/issue/0702/article/R0702A.jhtml?type=F#section20)
that plagues any sensible debate in education we might begin to ask more important questions about the role of various forms of schooling in this century. And heaven forbid, we might even begin to develop some imaginative educational approaches, pace the Rudd revolution.

"the evidence from places as diverse as Israel, India and the US suggests that the advantages outweigh the disadvantages"

Wwwwwwwwhere, exactly?

Why provide a wealth of footnotes for your earlier points, yet remain coy about supporting the (in my view, questionable) thrust of your argument?

Unfortunately although this is a well written piece it bears no relation to the facts involved.

Student performance factors are based on more things than just teacher performance, and while I am the first to admit that if a teacher performs poorly the student will also it is not that simple. If I may give an example to illustrate my point:

Lets say little Jimmy's parents have just had a nasty divorce where there was a bitter custody battle. At the beginning Jimmy though that it was something that he did to cause it (as kids do). Then there came the fights in which Jimmy could not sleep for the shouting and the fear. Then little Jimmy worried about the prospect of having the move away from his friends at one school because one of his parents moved away from the area. Finally, after the divorce his remaining parent turned to drink due to the stress of the divorce and began beating him. All of this had the effect of causing Jimmy stress and causing his in class learning to be reduced and his test scores to be lowered.

This is unfortunately a story that is played out many times over Australia each day. Who is it that will determine if the teacher has performed to the appropriate level? Someone just looking at test scores would see that little Jimmy has been slipping and under the proposed system would rate them as a ineffectual teacher when in fact the root cause is out of the teachers control.

A second question may be do you wish your childs history revealed to an “objective measure” consultant. The problem with comparing teaching with other professions is that teaching involves variables that are often either legally or physically unmeasurable. In the case of a doctor or a lawyer performance indicators are easier, how many case taken and how many reached a successful conclusion.

Teaching on the other hand often involves factors like student home life and school yard relationships. Performance based pay will see teachers leave the profession, then where will we be?

C'mon guys. It seems quite reasonable to pay teachers for the grades their students get on tests. it is measurable and quantifiable and should be used in lots of professions. let's see:

1. Pay dentists according to the lowering rates of decay among their patients.
2. Pay doctors according to the health of their paitents - not the time they take to treat them.
3. Pay journalists based on how well informed their readers are based on an independent random survey of readers.
4. Pay Real Estate agents a commission based on the difference between what they tell you you'll get for your house and the final sale price (they could owe you money!)
5. Pay economists (like the author0 according to how wealthy the rest of us are.

Andrew Leigh asserts that teaching standards were higher in the 1960's. Really? he wouldn't say that if he had been there. Teaching standards were supposedly higher because talented women were limited to teaching, nursing. Yes well those were the only options spruiked to convent girls i.e. about a quarter of the school girl population.

Talented women are now able to establish careers in the highest paid professions like medicine and law so these women are lost to teaching. In the 1960s, before married women had the right to work, women were lost to the teaching profession when they got married usually by age 23.

Andrew then dismisses increasing pay with senority - and presumably the school leadership position, he questions whether teachers with a Masters degree teach any better, then mentions the Business Council of Australia's idea of an accreditation scheme like the American National Board for Professional Teaching Standards which has accreditted 2% of teachers. Newsflash to Andrew and the BCA - Australian teachers must complete a Graduate Diploma in Education or a Bachelor of Education before they can teach and sometimes they also need a Cert IV in Workplace Training as well. In Victoria teachers are accredited by the Victorian Institute of Teaching which demands that teachers duplicate the project work they completed for their university degree for the VIT. The VIT is not well regarded amongst teachers.

Andrew doesn't like pay-for-credentials so naturally that leaves pay-for-performance as measured by student test performances. That sounds reasonable except as Andrew and every other educator in Australia knows there is great variation in the standards of work demanded of students depending on whether the school is for academic, middle class or factory fodder.

I believe the education standards in academic and middle class schools are far higher now than they were in the 1960s but we don't have a manufacturing sector any more. Students from industrial areas are still unlikely to get places in medicine and law.