The skeptical take of The American Prospect‘s Dana Goldstein on the Obama adminstration’s promotion of individual merit pay for educators has led to a widely read exchange on a number of different blogs. Take a look at Matt Yglesias’ two posts, here and here, as well as the comments by The Quick and Ed’s Kevin Carey, the Core Knowledge blog’s Robert Pondiscio, and the conservative Atlantic blogger Megan McArdle.
What is interesting about the exchange is the commonplace conceptual slippage between the idea of individual merit pay and the broader concept of differentiated pay. The ever reliable voice of “Know Nothingism” on all matters educational, the professional anti-teacher union Mike Antonucci, attempts to use this slippage to suggest that the opposition of teacher unions to individual merit pay means that we are opposed to all differentiation of pay and committed to nothing but a single salary scale, based only on educational credits and years of experience and service. In fact, a simple search of the American Federation of Teachers web site demonstrates that this is clearly not the case: for many years we have been on the record as supporting differentiated pay for what teachers know and what teachers do, provided that it is jointly developed by management and labor and based on principles such as universal availability, credible and agreed upon standards of practice and adequate base compensation. Read the 2002 AFT Convention Resolution and the AFT Fact Sheet on the subject, as well as the piece on “Rethinking Teacher Compensation” by the late AFT President Sandy Feldman. Note that the AFT is on record supporting the use of differentiated pay as one component of a broad-based effort to address the needs of high poverty, hard to staff schools and as part of the development of a teacher professional career continuum that includes roles such as a master teacher. Here in New York City, the UFT negotiated a school wide bonus plan for high needs schools.
In short, the issue is not whether teacher pay should be differentiated, but how it should be differentiated. Individual merit pay is only one form of differentiation, the most problematic form. It is problematic not simply because it misreads the motivation and psychology of educators, as Goldstein correctly points out, but because it is based on a set of untested and incorrect assumptions and first principles about the use of such individual performance incentives. For all of the talk of what “the evidence” demonstrates, there is precious little attention paid to the most current research. In an Economic Policy Institute study published last May, Teachers, Performance Pay and Accountability, Richard Rothstein, Scott Adams and John Heywood demonstrate that the experience with individual performance pay in the private sector and in other economic settings is actually quite different from what is assumed in much of these conversations. In reality, individual merit pay plays a minor role in the private sector [only 6% of private sector workers have their pay determined in this fashion], and for very good reasons, since the available quantitative and objective measures commonly capture only a small portion of what employees are doing. When employees are held accountable to such partial and radically incomplete measures, serious distortions are introduced into their work: widespread “gaming” of the system and measurement inflation occurs. Richard Rothstein’s contribution to the volume shows how these effects can be seen in a variety of different settings, such as health care. Of particular interest to many readers will be Rothstein’s examination of how quantitative performance goals in the Soviet command economy introduced the most dysfunctional economic distortions — when given a quota of shoes to produce, Soviet shoe factories would produce large numbers of small size shoes and small numbers of large size shoes, since the smaller shoes consumed smaller amounts of resources such as leather. But, needless to say, the economic need was for a whole range of different sized shoes.
The relevance here to the most common form of individual merit pay for teachers, one based on their students’ scores on standardized tests, is clear. With the exception of NAEP, the New York City Department of Education school progress reports are certainly among the most developed set of measurements of educational performance based on standardized test scores in the United States, one which has had hundred of millions of dollars invested in it and is now in its third year, with sufficient time to work out the design problems of any new system. One would think, therefore, that the recent results of those progress reports for elementary and middle schools, where standardized test scores make up 85% of the grade, would give pause to those, such as the Ed Sector’s Kevin Carey, who have been such outspoken cheerleaders for the use of such scores as a basis for teacher evaluation and teacher pay. Who really believes that 97% of the elementary and middle schools in New York City are performing at ‘A’ and ‘B’ levels, or that schools careen from ‘F’ to ‘A’ and ‘A’ to ‘F’ in a single year, or that there is a negative correlation in schools’ performance, one year to the next, such that their grades would appear to be completely random? Is it any wonder that teachers would not want their base salaries determined by such measures?
But it is an entirely different matter for teacher pay to be differentiated along lines of different roles, such as that of a master teacher or a mentor to novice teachers; along lines of exceptional knowledge and skill, such as that demonstrated by National Board for Professional Teaching Standards certification; along lines of undergoing special professional development to improve one’s skills; or along lines of taking on difficult assignments, such as work in a high poverty, hard to staff school. All of these forms of differentiation, and many more, conform to the principles laid down by the AFT, and are completely unobjectionable to us.
One last point, regarding the complaint made by Yglesias that Masters’ Degree and advanced education do not translate into increased teacher effectiveness and thus should be eliminated as a basis for calculating teacher salaries. There is a real problem here, even given the very limited quantitative measures of teacher effectiveness we have, but the shallowness of the analysis — its acceptance of aggregate statistics when a drill down into the details of the matter is required — leads to a misspecification of the nature of that problem. It is interesting that those who argue so strongly for a differentiation of teacher salaries fail to engage in the most basic differentiation of this data. There is no attempt made to see if different teacher education programs have different value added to a teacher’s effectiveness, so we don’t know if the problem is truly universal or the artifact of particular teacher education programs that are used as ‘cash cows’ by their institutions of higher education. There is no attempt to see if different types of Masters Degrees have different effects, so we don’t know if a degree in Mathematics, as opposed to a degree in Mathematics Education, makes a difference at the secondary level, where a solid grasp of subject matter is essential. Studies conducted in states where Masters Degrees are optional for educators are assumed to hold true to states where they are a requirement for holding a permanent teaching license.
The real problem is that in too many instances, teacher education programs are not of the quality we need. The solution is not to produce less educated teachers, but to fix teacher education. We would accept nothing less if it were medical education or legal education programs that were not working as they should, and we should accept nothing less here.