Paying Teachers for Performance, Options and Rationale
In any given year, a teacher’s pay with a school district is primarily set on the basis of a “schedule” that determines the annual salary as a function of the teacher’s level of education and years of experience. A teacher with a bachelor’s degree and three years of experience, for example, would be paid $41,278 in East Lansing Public Schools, while one with a master’s degree and nine years experience would receive $58,522. (See table on page 36 for data from the 2010–11 ELPS Salary Schedule.) Pay might be supplemented for an additional special assignment such as coaching basketball, but for the most part the salary would be determined solely by level of education and years of experience.
This approach to setting teacher pay began in the 1920s, as a way to set compensation on an impartial, transparent basis. Before the adoption of a single schedule, differences in pay might have arisen due to favoritism, cronyism or discrimination by race or gender. By 1970, 97 percent of U.S. school districts used a single schedule to determine pay.1 The schedule is typically adjusted annually, with some increase in pay at all points in the schedule. In Michigan, as in many other states, the details of the schedule vary across districts, so that a teacher with three years of experience and a master’s degree might make $40,000 in one district, but $55,000 in a neighboring district.
In addition to its virtues of even-handedness and transparency, the typical single schedule can be seen as rewarding some desirable characteristics. It provides an incentive for teachers to complete additional coursework and adds to teachers’ pay as they gain job experience.
As policymakers have looked more closely at teacher pay, however, they have begun to wonder how much education and experience are worth, and whether other characteristics that demonstrate skill and effectiveness, especially teacher performance, should be given greater weight in determining compensation.
Some policymakers are now focusing on connecting teacher rewards to student scores on standardized tests. Using that criterion, a teacher earning a master’s degree in the subject taught (especially a master’s in mathematics or mathematics education for high school mathematics teachers) may have some effect on student achievement worth additional compensation, but getting just any master’s would not. Many studies have found a connection between student scores and teacher experience, at least in the ?rst years of teaching. But that bene?t levels off after three to ?ve years of experience. So the current bases of the single salary schedule have some empirically established connections to student test scores, but research has not established a tie to simply gaining master’s degrees, or to teaching experience beyond the initial years.
The Push for Performance-Based Pay
Recently, policymakers have insisted that the system of setting teacher compensation must be changed. Substantial changes are likely to occur. To some extent, the changes will be driven by political considerations. As the discussion unfolds, however, it would be prudent to keep the goals for a compensation plan in mind, and to look at the range of options that could be built into a new system.
Goals often mentioned for changing compensation are:
- Providing incentives for all teachers to be more effective
- Encouraging the “best” teachers to stay
- Encouraging the “worst” teachers to leave
- Attracting effective teachers to the occupation
- Drawing effective teachers to high-need areas, including high-poverty schools and subjects like mathematics, science, special education and bilingual education
Common-sense arguments suggest that moving away from a single salary schedule could help to achieve all of these goals. If pay were determined at least in part by some measure of job performance, teachers would have a ?nancial incentive to do well on that measure. Those teachers most rewarded for their performance would have a greater reason to keep teaching; those least rewarded might be more likely to consider other options. People considering whether or not to pursue a teaching career might ?nd it appealing to have the chance to improve their pay with top performance. If the rate of pay varied by teaching specialty and assignment, districts might attract teachers in areas of need.
Performance Pay Options
With these goals in mind, a range of options might be used to provide ?nancial incentives. The mix of options chosen would affect both the structure of the ?nancial incentives and the overall expenditure on teacher compensation. Three features of a compensation system are particularly important:
- Whether a reward is made as a one-time bonus, as a supplement tied to a teaching assignment or as an increase in base salary
- Whether rewards are given to individuals or to groups of teachers
- Which indicators are used to determine the award—ratings from a principal, structured classroom observation or a measure of student learning
Bonus, Supplement or Increase in Base Pay?
One characteristic of an incentive system is whether an increase is permanent, or tied to a single year or particular assignment. A pay increase might be given as a one-time bonus, as a supplement for particular classes or as an increase to the base salary. In the ?rst case, for example, teachers could be given a $5,000 bonus if their students met an achievement target. Getting a bonus the next year would depend on meeting a target again. In the second case, pay would be supplemented if a teacher took on an assignment in a high-poverty school or was assigned to teach mathematics or special education. The supplement would continue as long as the teacher had such an assignment, an incentive which might bring new teachers to a district or persuade teachers to gain the additional expertise needed to teach a new content area or to succeed with students in a challenging environment. In the third case, teachers could be given a $1,000 boost in base pay, which would increase their salary throughout their employment. The initial amount for this type of award would be smaller, but the long-term amount would be larger if the teacher continued to teach in the district. Under this option, there would be an incentive to perform well each year, plus an incentive for the top teachers to stay in the district as their annual rate of pay increases.
Rewards to Individuals or Groups?
Another characteristic of the incentive system is whether the ?nancial award is made at the individual level or at a group level, such as to all teachers in a school. Awards to the individual ?t the idea that every teacher should be rewarded for his or her own contribution to student learning. Those favoring individual rewards might also point to the possibility that some ineffective teachers would bene?t from the high performance of their colleagues under a group award model—the so-called “free rider” problem. Group level awards have several advantages, however. If rewards are based on the performance of all students in a school, teachers would bene?t from collaboration and be more likely to help each other succeed. Group awards also avoid the dif?culties in sorting out the contributions of individual teachers in schools where students are taught by several different teachers. In a middle school, for example, a student might improve in science both because of what happened in science class and because of gains in reading ability from an English class.
Measuring Quality: By Principals, Structured Observations or VAMs?
Everyone agrees that teacher quality is critically important, but we are far from consensus about how quality should be measured. To some extent, the single salary schedule itself can be seen as linked to teacher quality. To the extent that level of education and years of job experience are indicators of quality, the current schedule ties pay to quality. As noted earlier, however, evidence linking these to student outcomes is weak. Moreover, degree level and experience are only indicators of teacher characteristics, with uncertain connection to a teacher’s actual performance in the classroom.
Three alternative ways to measure teacher quality, all linked to performance, are evaluation by the school principal, structured classroom observations and analysis of student test scores. Each of these has some appeal; each also has weaknesses.
Evaluations by the school principal are the traditional way school districts evaluate the work of teachers. In most school districts, evaluations by the principal are weighted heavily in the crucial decision about whether a new teacher should be granted tenure. Principals are also expected to do annual evaluations of teachers. Principal evaluations have intuitive appeal, based on the thought that principals have regular contact with teachers. But research on principals’ ability to judge teacher effectiveness is mixed. Even studies generally supporting the validity of principal judgments acknowledge the dif?culty of making consistent evaluations of teachers in the middle of the distribution. Principal evaluations may also be subject to abuse, with principals punishing teachers for union activity or for expressing views contrary to their own.
The use of structured classroom observations can increase the sense that performance evaluations are objective, particularly if the observations are carried out by someone other than the principal, perhaps a team from the district central of?ce. A weakness of this approach is that little empirical evidence exists to demonstrate that any particular observation system is closely tied to student learning. The Gates Foundation is currently undertaking a massive study to examine the links between performance observation systems and student learning, but results from that study are still years away.2 In addition, most existing observation systems were developed for use with young children, with work yet to be done to modify them for use in high schools, where subject matter speci?city will pose further challenges.
The measures of teacher performance that have been recently highlighted by the news media are value-added measures (VAMs), which are increasingly being included as at least one component of teacher evaluation. Value-added measures start with gains in student test scores, then use statistical procedures to take account of the fact that classrooms vary enormously in the entering characteristics of their students. The adjustments are needed because a teacher with a class of low-performing students from low-income families will have a harder time producing learning gains than a teacher with students from high-income families who begin the year well above grade level. The idea of rewarding teachers for the value they add to student learning is tremendously attractive. Unfortunately, close examination of these value-added measures has shown that they appear to have a large amount of “noise,” with variability due to factors other than teacher performance. In addition, they depend on having student test data from prior years, which means that they can seldom be used for teachers in subjects other than mathematics and literacy, or for teachers below grade 4 or above grade 8. (States test literacy and mathematics in grades 3–8, but testing is much less frequent in other grades and subjects.)
The problem for policymakers is that each of these measures of teacher performance is imperfect. Using a variety of approaches may reduce the impact of any one measure’s weakness but it does not eliminate the dif?culties.
The Bottom Line
The basis for teacher pay is already changing in many districts. Greater change will occur in the next few years. Many of the changes will shift the system from being based solely on degree and experience to including links to teaching performance and assignment. Still to be determined is exactly how those links will be made.
Will strong performance lead to increases in base pay, or only to temporary bonuses or supplements? Will rewards be given on the basis of individual teacher performance, or on the basis of what an entire school does? Will performance be measured by principal ratings, systematic observations by central of?ce or student test results? Some of the consequences of these choices can be predicted, but research is still scant about the effects of the various approaches. For long-term improvements in education outcomes, it will be important to think carefully as these choices are made, then to do systematic studies of their effects.
Education Policy Center: education.msu.edu/epc
- Koppich, J. E. (2010). Teacher unions and new forms of teacher compensation. Phi Delta Kappan, 91(8), 22–26.
- For more information, see the Bill & Melinda Gates Foundation, Measures of Effective Teaching (MET) project. http://www.gatesfoundation.org/highschools/Documents/met-framing-paper.pdf.
A university investigation has found that three previously published New Educator articles written by Professor Sharif Shakrani contain instances of plagiarism. Those articles were found to contain material from other sources without proper attribution. They include “Teacher Turnover” (Fall/Winter 2008), “A Big Idea: Smaller High Schools” (Spring 2008) and “Education Stimulus Funds: An Opportunity to Plan Ahead” (Spring/Summer 2009).
The College of Education and the Education Policy Center are strongly committed to research integrity. Therefore, we have retracted those articles and extend a sincere apology to our readers.