New blog: Test pattern

There is a relatively new blog around called the Test Pattern, that gets a few individuals together to debate specific public policy issues in written essays.

The current debate on the blog is around performance pay for teachers, one such discussion around its merit (or lack of) is given by Robbie Allan here.

What do I think?  I have no problem with performance pay in of itself, but it relies on low cost measures of performance.  In an area like primary and secondary school education, how observable are outcomes?  If we can’t observe actual outcomes well, performance related pay may in fact make outcomes worse – not just by leading teachers to fake results, but by convincing teachers to invest in the “wrong parts” of education.

Hell, when it comes to education do we objectively even know what areas provide the most value?  Until we really face the issue of what performance really is, the entire concept of “performance” related pay is relatively redundant.

  • MarkS

    Hi Matt,

    You know, most private employers have individual contracts, and they are negotiated and reviewed annually.  Good employees get pay rises, bad don’t.   Your description of performance pay supposes that pay is directly linked to particular KPIs that are set at some kind of national level.  Any good principal would be able to go through an annual performance review process and give different teachers different pay rises depending on performance within the school (as opposed to how long the bum has warmed the seat as is the case now).  

    Cheers,

    Marl 

    • This all relies on the observability of performance – which is really the key of what I am after here.

      Also, if its the principals choice, we have possible incentive issues with him.  In fact, it is pretty important that the principals incentives are aligned with what is in the “social need” – and so their is probably a strong case for them to face some type of incentive based pay.  That or there needs to be a clearer separation between them and teaching staff methinks.

  • This paper that was presented at the GEN conference may be of interest.

    • Interesting.  When it comes to implementation I have two questions though:

      1)  How do test scores relate to value add from education?

      2)  How do we ensure that teachers cannot manipulate results in ways that work against any said value add.

      If those are answered well, then the justification for performance based pay exists, and I’d say it should go forward.

      Given I know nothing about the field all I can do is frame, and I’m happy to be educated 😀

      • Best is to watch the Youtube at that page of the man himself explaining it. He really is an extremely good presenter, although I don’t know how it will come across on grainy video.

  • Test scores don’t relate to value add. You can figure out value add, (basically the change in learning outcomes) but there are statistical questions about whether it can be measured with any precision. Moreover, a range of factors such as attitude and inspiration are particularly hard to measure.
    A supposition exists that the private sector gets this right. I don’t think this is necessarily true.
    My point is this: it might not be bad, but it’s not clearly good and so we should focus on other things that are more effective. No one has managed to prove conclusively that merit pay has a significant impact on performance in teaching. There are known harms with poor implementation. Teachers are against it (which is a surprisingly important part of any successful merit pay system). Despite all this, we may well be able to design a system that does work. Just that it’s probably not the best use of our time compared to, say, providing better professional development.

    • Have you seen Chetty’s paper that I linked to above, which explicitly deals with Rothstein’s concerns about measurement problems and shows a clear link between test scores and VA?

  • Kimble

    Hey Matt, welcome back!

    Look, none of us are experts on what everyone should be paid. So why not just allow unconstrained pricing of teachers, and leave it up to the parties involved to create their own, efficient method of determining teacher quality? Emergent order ftw.

    This works for the following professions; lawyers, programmers, actors, prostitutes, sportsmen, sales executives, shearers…

    Oops, I accidentally listed one profession twice.

  • Just keep the teacher’s pay the way it has. Looks like they want to turn teaching into inside sales. However, in Europe teachers are paid more I think.

  • Richard29

    The good news is that we are doing well currently:
    http://en.wikipedia.org/wiki/Programme_for_International_Student_Assessment
    In fact we are consistently rated higher than countries like the US and UK who’s policies we sem set on copying.

    It seems to me that there are two ways to improve performance:
    – Improve Overperformance (Excellence)
    – Eliminate Underperformance

    The latter seems to be a better return on investment. Generally speaking it’s easier to lift a D to a C student than it is to lift a B to an A.

    Lifting underperformance involves a different prescription. Too often the underperforming kids we hear about in the media are turning up to school hungry and cold, sick or with a lack of sleep from poor housing. They often have family problems and may suffer from a lack of educational resources and little focus on the value of education. In short the problems of educational underperformance are mostly problems of poverty not teaching.

    The example of Finland is instructive – they didn’t set out to achieve excellence, they wanted to achieve equality of opportunity but by ensuring there were no barriers to every kid achieving to their potential excellence resulted:
    http://www.theatlantic.com/national/archive/2011/12/what-americans-keep-ignoring-about-finlands-school-success/250564/

  • Richard29

    The same underperformance/overperformance argument can also relate directly to teacher pay. If we want to remove teacher underperformance will merit pay make a difference?

    I suspect probably not unless you were to implement punitive pay cuts for teachers whose students perform poorly in tests. We don’t know enough about the long term relevance of test scores and the extent the teacher is an influence to make that an even slightly viable option.

    The big problem of poor teacher performance is not that principals don’t know who the better and worse teachers are, but that they can’t replace the poor performers. The trick here is that to replace a poor teacher you need the opportunity to replace, that;s easy enough, just have the ability to hire on renewable 1-5 year contracts by school year. The second part is who to replace with – if a school gets 7 applicants for 7 jobs then the opportunities for improvement are limited. If they get 50 applicants for 7 jobs they can pick the wheat and loose the chaff.

    So that comes back to the issue of raising the prestige and public perception of the job and lifting pay at all levels so that the best people want to become teachers.