Research has shown time and again that experience matters in good teaching. What it hasn’t shown is that every experience matters equally. In fact, a teacher’s first few years on the job are by far the most important, and it has been demonstrated repeatedly that the vast majority of teaching improvement comes in the first few years on the job.
Unfortunately, districts have yet to utilize this research in any way. Instead, they set arbitrary teacher salary schedules that are based purely on a teacher’s years of experience and education credentials. They mostly do not reflect actual teacher performance year-to-year, but they don’t even take into consideration the career paths of the typical teacher.
To show what this looks like, I’ve graphed teacher salaries in four DC-area suburban districts (Montgomery and Prince George’s Counties in Maryland and Fairfax and Prince William Counties in Virginia) with average teacher effectiveness scores in mathematics. The effectiveness scores are value-added measures that compute a teacher’s ability to maintain or increase student scores on standardized tests of achievement. They are graphed according to differences from the average teacher, and they come from a recent paper which found effectiveness scores in line with prior research.
The maroon bars represent salaries for fully certified teachers with bachelor’s degrees only, and the blue bars are average salaries for fully certified teachers with Master’s degrees only. As the chart shows, salaries in these four districts increases almost linearly, although they grow much faster after teachers have 14 or 17 years of experience than when they have only a few.
On the other hand, teacher effectiveness makes nearly all of its gains in a teacher’s first two years on the job. They are barely more effective at 25 years in as they were after two or three, and no more effective at 22 than they were at four.
Teacher salaries are 2007-8 base salaries for teachers with full credentials in Montgomery County, MD; Prince George's County, MD; Fairfax County, VA; and Prince William County, VA.
Remember that these are for the average teacher. We could go even further and talk about the effectiveness of individual teachers, but that’s fraught with quite a bit more controversy. If common sense and all statistical research shows a teacher learns 90 percent of what they’re ever going to learn in the first few years on the job, shouldn’t districts start paying them that way too? See Chart 6 here for how this might work on a district-wide basis.
Another important lesson from all this research is that tenure after a few years on the job really should mean something special. Since we know that teachers are not likely to improve later in their careers as quickly as they did in their first few years, we would want to invest a lot of resources into making sure we awarded tenure to teachers who really deserved it.
Right now we only value credentials; policies following from this body of research would value excellence in actual teaching performance.
Lowering Student Loan Default Rates: What One Consortium of Historically Black Institutions Did to Succeed 
College and Career-Ready: Using Outcomes Data to Hold High Schools Accountable for Student Success
03/11/10 | Kevin Carey is quoted on the rising costs of education in The Economist.
03/10/10 | Chad Aldeman Aldeman's False Fronts is quoted on Accuracy in Academia.
03/09/10 | Ben Miller is quoted on loan origination on the Firedog Lake blog.
03/07/10 | Kevin Carey calls for a Race to the Top in higher education in a column in The Chronicle of Higher Education.
03/08/10 | Andrew Rotherham's Eduwonk post on RTT finalists is quoted on Class Struggle.
03/07/10 | Andrew Rotherham is quoted on the success of MATCH in The Times-Picayune.
03/06/10 | Eduwonk's reaction to Ohio's status as a RTT finalist is cited in a Wall Street Journal editorial.
03/05/10 | Andrew Rotherham's reaction to New York's status as a Race to the Top finalist is cited on National Journal.
That is a good point Teaching Serendipity.
Also as I looked at the graph the first thought that came to my mind was that the first year teachers might really be woefully under-prepared to teach. They might actually do most of their learning on the job. After that in general it looks like most teachers are increasing in effectiveness at about the same rate as they increasing in pay.
YES, BUT — Do some veteran teachers contribute value IN ADDITION TO student test scores? An effective teacher in her 15th year might contribute significantly more than an effective teacher in his 4th year in terms of mentoring/ developing other teachers, writing curriculum, helping shape school policies, collecting school resources, etc. All of these factors are key in a well-run system, and all contribute to the in-class effectiveness of that 4th year teacher.
As a 3rd year teacher who is relatively effective in terms of student test scores, I know anecdotally that my more experienced colleagues contribute much more than I do to the school as a whole, and to the development of my own effectiveness.
To be sure, not all veteran teachers contribute in these ways, but many do, and maybe all should. Perhaps we need additional measures of teacher value?
Chad,
The first thing you need to do as your point sinks in is to grin, and appreciate the capriciousness with which human beings make decisions, and make choices against our self interest. There is plenty of cognitive science on why we retain outmoded systems. It does no good to curse the darkness and complain that people aren’t irrational.
Unions leaders frequently try to explain the value of reducing pay steps and frontloading pay increases. The difficulty of making that case, I’d argue, is comparable to explaining to people who still hold on to rational models to explain human systems why they continue to fail.
So, we humans need to first laugh at ourselves, forgive ourselves for our foibles, and then talk through methods that will be smarter.
As I’ve said before, I’d put more effort into issues that research says are more likely to raise student performance, such as rebuilding and improving teacher quality and autonomy. But if younger teachers want performance pay, so be it.
We’re a lot better off focusing on incentives. If a rational model fails, and its just used for bonuses, people being people will get upset. But that’s not a disaster. Try to use those models to fire teachers and you risk a disaster.
And you implicitly suggest a compromise. It would be much easier to give pay raises and protect seniority status for teachers who are increasing performance. But like you say, this data becomes much less valid on single teachers over single years. So, we’d still need to negotiate tough-minded contracts. And you know what I’m going to say next. Give those imperfect measures to imperfect peer review committees, and I see a lot of good things that can happen. If that imperfect data remains in the hands of administrators who are imperfect at best – and often completely irrational as many featherless bipeds are – and I don’t see a solution.
[...] Effectiveness as a Function of Time By juscohen Over at The Quick and the Ed, Chad Aldeman maps teacher effectiveness measures against a representative salary scale. The result is as you would [...]
[...] the Quick and the Ed, Chad Alderman makes a few more points about teachers compensation and the effectiveness of teachers through a chart. [...]