Do Teachers Need Extrinsic Rewards like Performance Pay?

Richard Rothstein is among the most responsible skeptics when it comes to educator performance pay.  In a taping for a forthcoming PBS television show on performance pay, he posed an interesting question.  He asked, in effect, “why do performance pay advocates assume teachers need added motivation?  Is there not evidence already that they are motivated?”

As support for his position that teachers, generally, are motivated, Richard referred to gains in the fourth grade mathematics scores on NAEP for African American youngsters. He points out, quite correctly, that these scores have increased over time.  Even though the achievement gap has not narrowed—black youngsters now score on NAEP where white fourth graders did several decades ago— are not the black student achievement gains evidence that, when teachers know what to do, they do it?

In thinking about Richard’s postulate, I consulted with a number of the nation’s experts regarding reading and mathematics instruction (e.g., Ted Hasselbring, Dan Reschly, and David Chard).

The general consensus (and I will provide some of the details below) is that the rise in NAEP math scores for African American students provides virtually no support for the contention that teachers “do the right thing when they know what is right, and do not need extrinsic incentives to do it.”

If anything, it may be the other way around.  Teachers may badly need extrinsic incentives, in addition to information, to do the right thing, even when the right thing is well known.

Let me first address some possible explanations for the math gains demonstrated by African-American students.

The big point here is that presently, no one knows why these gains have occurred. We are all pleased that they have occurred, but no one has studied the effect sufficiently to offer an empirical analysis.

There are several prominent hypotheses.

Given the long-standing and wide gap between white and black student mathematics scores, the NAEP gain may represent nothing more than regression to the mean.

Recent improvements in mathematics standards may have contributed significantly to the ability of black students to score higher.  The mathematics curriculum, as expressed in state guidelines, textbooks, student workbooks, and teacher guides, has been modified to make more sense, and this alteration may place math instruction closer to what NAEP tests are measuring, irrespective of what teachers are doing.

If this latter hypothesis should ever be proven correct, it would testify to the efficacy of a standards-based strategy for raising student achievement.

Mathematics is less culturally dependent than is reading.  Both reading and mathematics require our brain to create new neural patterns. (We are not born with these pathways.) In reading, the 26 human-invented squiggles we label letters, and the 44 sounds associated with them, must be learned.  However, once the letters and sounds are learned, there is still an enormous world of meaning—culturally determined meaning—that a child must acquire in order to read.  Possibly one percent of children can acquire this comprehension on their own, but most need instruction to get there.  The less rich a child’s cultural milieu, the longer it may take to acquire reading comprehension skills, because there is such a great deal of culture to be learned also.

Mathematics, conversely, has a far greater internal logic that permits an individual to advance in skill without being completely stymied by deficits in cultural awareness.

Because low-income African-American youngsters often have less of an opportunity than middle-class children to acquire cultural awareness, when motivated they may learn math faster than they learn reading, absent any consideration of teacher effectiveness.

Here is a second point.  I am responsibly informed that there is a great deal more known about how to teach reading, to all kinds of students, than is known about how to teach mathematics.  If teachers will simply “do the right thing” in the absence of extrinsic inducement, why are NAEP reading scores stagnant?

Perhaps the answer to the foregoing question is that teachers do not have sufficient professional development activities on what is known about effective reading instruction. Possibly, but this path also seems fraught with complexity.  The most significant IES piece of research on this suggests that professional development presently has little or no effect on reading instruction and achievement.  (See The Impact of Two Professional Development Interventions on Early Reading Instruction and Achievement.)

This is a controversial finding, and experts question the degree of implementation fidelity in the studies involved.  However, it just may be that professional development is more effective when operating in tandem with extrinsic incentives.  Here is a quote from one of the emails from experts I received when exploring Richard’s question:

Massive professional development without incentives (and these may or may not be financial) doesn’t seem to be sufficient. This question puts me in mind of the DESSI work done by Huberman and Miles in the 70s and 80s.  Particularly in the area of innovation.  I recall they talked about how we might have to take the “low road” to convince people to do things differently (this might involve sufficient motivation positive or negative) that, in turn, will lead to positive outcomes and will help teachers see why they were persuaded to do it in the first place.  Conversely, incentives without knowledge would also seem shortsighted.

Of course, I do not have evidence one way or another regarding the issue of extrinsic incentives, but that is one reason why the upcoming release this spring of experimental results from the National Center on Performance Incentives (NCPI) is so important.  The NCPI is conducting a randomized field trial in which teachers in the treatment group (who are all middle school math teachers in Nashville) will receive bonuses based on student achievement gains. The extrinsic incentives involved in this experiment are big ones.  We may know more then than now.

Meanwhile, I want to thank Richard for elevating the issue because I never otherwise would have thought about it so much or learned so much from others.

Last Updated


Notify Me When Education Next

Posts a Big Story

Business + Editorial Office

Program on Education Policy and Governance
Harvard Kennedy School
79 JFK Street, Cambridge, MA 02138
Phone (617) 496-5488
Fax (617) 496-4428

For subscription service to the printed journal
Phone (617) 496-5488

Copyright © 2024 President & Fellows of Harvard College