Can Merit Pay Work in Education?

Hide Document

..

CAN MERIT PAY
WORK IN EDUCATION?
Research shows it's difficult-and risky

By DAPHNE SIEV WHITE
IKE MOTHERHOOD, apple pie, and ice cream, who can be again~t the principle of merit pay? Xerox executive Nathan B. Winstanley, for one. Winstanley posed the question in a recent issue of Personnel Administrator, adding that "even though a merit pay system is difficult (or all but impossible) to administer, it is a nice thing in which to believe." Noted industrial psychology researcher Herbert H. Meyer is another. "Despite the apparent soundness of the simple principle on which merit pay is based, experience tells us that it does not always work out with such elegant simplicity," Meyer wrote in "The Payfor-Performance Dilemma," an article in Organizational Dynamics. "In our society, it is difficult to argue with the proposition that higher performance deserves higher pay," A. Mikalachki argued in The Business Quarterly. "The paradox of this system is that most white-collar workers want merit pay until they get it, and then they do not want it at all." The business world, like education, has had a -long experience with pay-for-performance plans, dating back to the turn of the century. But while there is very little empirical research on such programs in school settings, there is a broad research base in business and industry. Early research was glowing about the theoretical possibilities of merit pay, but more recent studies have shown that a great many plans do not work out in practice the way they do on paper. Yet some plans are more effective than others at motivating employees and increasing productivity. One of the keys that is most effective in unlocking resistance to merit pay is. employee participation, research has shown. The same plan can succeed or fail

L

-i I
.\

I
I

I

l
-I

l
A j

depending on the degree of staff participation in planning, writes Edward Lawler in "Participation and Pay," which appeared in Compensation Review. ∑'When workers participate in the design and administration of a system, they are more likely to trust it for two reasons: They have more information about it, and they perceive they have control over what happens," Lawler says. In addition, partiCipation leads to feelings of ownership, and this leads people to view the plan as fair, he said. Other researchers have found that trust is the one element most often missing when management decides to implement a merit pay system. Yet Lawler, one of the most respected authorities in the field, suggests that the best way to solve that problem is to let employees participate in the design of the plan. "What better indication of the seriousness of an organizational development effort and of the trustworthiness of management is there than for management to turn over pay administration to employees?" Lawler asks. "It can provide dramatic proof to employees that management is 'for real' when it talks about participation and that it trusts them to handle a very important facet of organization existence." EVELOPING A pay-for-performance plan is deceptively difficult, because pay is one dimensional while human behavior is multidimensional. In designing a plan, it is necessary to define precisely what aspects of performance are to be evaluated and what levels are acceptable-and this is much easier said than done. If employees are to accept such a plan, they must feel that the standards are fair, that the evaluation is carried out impartially, and that stan∑ dards will not be continuously raised beyond theil grasp-another tall order. "Success or failure of a system depends very large!)
WINTER

D

Daphne Siev White is an assistant editor in the AFT editorial department.
S I AMERICAN EDUCATOR

198:

on acceptance of the criteria established, and that's more difficult than we think," Meyer says. "Even in sales, where objective measurements are possible, it is hard to agree on what is expected." Management tends to set very high standards, while employees want to

lower the standards as much as possible so they can achieve a merit raise, he says. If it is difficult to agree on standards in sales and manufacturing, it should not be surprising that school boards and teachers have trouble reaching a consensus on evaluation methods. "Research on effective teaching tells us that there is not one best system of instruction, " says Rand Corporation researcher Linda Darling-Hammond. "No one has discovered the set of teacher attributes that promotes student learning." Teaching is an art as well as a science, and "a behavior that is effective when used in moderation can produce significant and negative results when used too much or-as others have found-when applied in the wrong circumstances," Darling-Hammond et al. said in their recent review of the literatur'e on teacher evaluation. "This kind of finding makes it difficult to develop rules. for teaching behaviors that can be generally applied." Because of the problems associated with defining what makes an effective or outstanding teacher, it is even more difficult to design an evaluation tool that can rate teachers and compare one to another. As Paul Thompson and Gene Dalton put it in the Harvard Business Review, .. An organization or even a group requires a carefully coordinated set of widely different technical and social skills and abilities to operate effectively. Each individual contributes a rich combination not necessarily correlated with performance. For the most part, the really significant rewards come through promotions, and the really outstanding people rise up the ladder rather than move up in salary in the same grade level." HE "SIMPLE elegance" of merit pay fades rapidly when exposed to the workaday world precisely because people and their motivations are so very complex. For example, a merit pay plan designed to reward the better workers and send a message to less productive workers often leaves everyone feeling that they didn't get what they deserve, Meyer said. In a study that has become a classic in its field, Meyer showed that 90 percent of workers in a General Electric plant felt their performance level was in the top 50 percent, and further studies with accountants showed that 37 percent of them felt they were in the top 10 percentile. "The fact that almost everyone thinks he is an
10 I
AMERICAN EDUCATOR

above-average performer probably lies at the root of most of our problems with merit pay plans," Meyer wrote. "Since the salary increases most people get do not reflect superior performance (as determined by interpersonal comparisons or as defined in the guidelines for the pay plan), the effects of the actual pay increases on motivation are likely to be more negative than positive. The majority feel discriminated against because it appears that management does not recognize their true worth." The negative effects of some merit pay plans include employee stress, subversion of long-term company goals for short-term personal gains, loss of significance for any reward except merjt pay, competition among colleagues who are supposed to be working together, . and serious morale problems. Because of this dilemma, researchers in both business and education are urging adoption of plans such as differentiated staffing and career ladders rather than one-dimensional, win-lose systems like merit pay. For in reality, most companies that claim to have merit pay plans wind up giving the majority of employees very similar pay increases, "and these increases are not necessarily correlated with performance," Meyer says. He agrees that in most companies the really significant rewards come through promotions and not merit increases. "Every organization, certainly every large organization, should have well-defined hierarchies of positions through which individuals can advance. Such advancement opportunities should be available to as high a percentage of members of the organization as possible, " Meyer has written. "Most promotions would probably involve moving up to positions within one's function that are more complex or that involve higher levels of responsibility and more important decisions. The outstanding accountant, salesman, technician, engineer, and so on is often much more valuable to the organization for his technical contributions than for his supervisory or managerial skills." NOTHER ARGUMENT for advancement through promotion is the fact that promotional decisions are made much more carefully than merit pay decisions, Meyer said. Such findings are supported by researchers in the education community. "I have never seen any evidence that shows merit pay promotes excellence in teaching," said Gary Sykes, former head of the teaching policies team at the National Institute of Education. The best way to reward outstanding teachers is to offer them expanded responsibilities and opportunities to develop professionally, he says. Darling-Hammond agrees. "One of the greatest intrinsic rewards for teachers is to give them the autonomy to do what they think is best for their students," she says. But the accountability movement (concerned with the removal of incompetent teachers) has tied the hands of the competent majority and is preventing them from having the independence of judgment enjoyed by most professionals. Teacherproof curricula, teaching to tests, and other accountability measures are taking much of the joy out of teaching, she says.
WINTER

A

T

1983

Yet it was the anticipation of helping students and making an individual contribution that brought many teachers into the profession. In his classic study, Schoolteacher, Dan Lortie found that the "psychic," intrinsic rewards of teaching were more important to teachers than payor convenient scheduling. The recent Carnegie Foundation study High School, by Ernest Boyer, also recommends improving the conditions of teaching, rewarding and recognizing outstanding teachers publicly, and creating a career path for teachers in addition to increasing teacher pay. The best reward for teachers is to let them become as competent as they can be and to remove those impediments-discipline problems, frequent interruptions, excessive paperwork-that make good teaching difficult, Sykes says. Setting up an evaluation plan that would help teachers become more effective-rather than one designed to weed out incompetent teachers-could have beneficial side effects. "The process of developing evaluation systems is an occasion for many things in an organization, such as the interaction of constituencies, celebration of important values, and joint recognition of problems," wrote M.S. Knapp in his book Toward the Study of Teacher Evaluation as an Organizational Process. "Whether or not performance objectives are met by a specified proportion of a school district's teachers, the indirect results of such efforts may have considerable impact on staff enthusiasm, beliefs or behavior, with ultimate benefits for students." However, any evaluation plan must be set up carefully and with a particular goal in mind-one plan should not be used for multiple goals. A fascinating study by Michael Gallagher showed that the results of a performance evaluation were very much influenced by the stated goal of the review. When three groups evaluated the same videotaped performance of a faculty member teaching a class, the group told that the professor was to be punished (put on-probation, laid off, or immediately dismissed) rated him higher than the group that was told he was to be rewarded (promoted, given tenure, or a salary in-

formance evaluation should not be used for different purposes, since the stated purpose of the evaluation can affect the actual performance rating," Gallagher says. For example, an appraisal made to determine the need for additional training may emphasize a person's weakness in a particular area to justify the training, but that same appraisal should not be used in salary or layoff decisions, he stresses. "The only solution to this problem .. .is to conduct different kinds of performance evaluations for different purposes," Gallagher concludes. Yet a recent NIE study of thirty-two school districts in twenty-four states shows that a majority of the districts use one evaluation tool for four goals-personnel decisions, staff development, school improvement, and accountability, although different districts attach different priorities to each of the goals. "How realistic is that expectation?" asks Milbrey Wallin Mclaughlin, who conducted the study. "Can a single evaluation system address all four purposes equally well?" VALUATIONS THAT tend to work the best are those that focus on future performance, with the staff member and supervisor setting goals together, rather than focusing on past actions. The much-quoted General Electric study showed that criticism of past performance had a negative effect on goal achievement, while praise of past performance had little effect on achieving future goals. "The more criticism a man received in the performance appraisal discussion, the more defensively he reacted," the researchers found. And the more defensive an employee felt, the more his performance deteriorated. On the other hand, "one of the most significant findings in our experiment was the fact that far superior results were observed when the manager and the man together set specific goals to be achieved, rather than merely discussed needed improvement," Meyer et aI., concluded. In some cases, peer review has worked better than appraisal by a superior and has "often proved to be more effective in relating pay to performance than the typical one-on-one superior-subordinate pay administration decision process," Lawler reports. Peers have better information and, when they are properly trained, can make better judgments than the supervisor can alone, Lawler adds. Meyer agrees, saying that "research shows evaluations done by peers often have more validity than evaluations done by superiors." Those doing the evaluations-whether they are peers or supervisors-must be trained to do ratings based on objective criteria and to be as bias free as possible. "The evaluator is the most important part of the evaluation instrument," Gallagher says. The evaluator is just as much a factor in the rating process as the person being evaluated because, just as in the Olympics, the same performance can garner different scores from different judges. For that reason, it is of crucial importance to carefully select and train evaluators, Gallagher says. One way to get away from peer competition is to im(Continued on page 42)
AMERICAN EDUCATOR I

E

I

.

I !

..

crease). The control group, which was not told the purpose of the evaluation, rated him midway between the other two groups . Those evaluators who were told that the faculty member was to be punished rated him highly in order to keep him from being penalized, while those who were told he would be rewarded gave him lower grades to keep him from getting the increase, Gallagher concluded in Personnel magazine. "The results of this study indicate that a single perWINTER

1983

11

CAN MERIT PAY WORK?

(Continued from page 11)
plement a system of group rewards. Group incentive plans have been more effective than individually based incentive payments in many instances, reports Bruce Ellig in Compensation Review. And "peer expectations can have a powerful impact on performance," David Belcher notes in the same publication. "One expert, in fact, found that peer expectations acted as a better predictor of grades in accounting than did the expectancy that high effort would result in good grades. In an employment relationship, the social contact could be expected to be even more powerfuL" HATEVER EVALUATION system is selected by a school district, teachers must understand what tasks they are responsible for, the criteria by which they will be evaluated, what data will be collected, and how the rating system works, says education researcher Gary Natriello. Finally, teachers must get feedback from the evaluation process (something that is not now done by many districts), and they should be given the opportunity to discuss with an administrator a plan to strengthen their weaknesses and

W

improve their teaching. The more specific the suggestions on this score, the better, Natriello says: A district should be able to provide teachers with. inservice education or point them to specific college courses. Natriello has been studying the only school district in the country that has had a merit pay plan in effect for thirty years: Ladue, Missouri. Some of the reasons the system has lasted this long, he says, are that teachers help develop the criteria by which administrators rate them; administrators are given extensive training in evaluation and this task is an integral part of their jobs; principals are held accountable by their supervisors for the degree to which their reports are based on the stated criteria; and teachers do get feedback and suggestions for improvement. The researchers say their study is not yet far enough along to tell whether teachers in Ladue are satisfied with this plan-but they have found that most teachers consider the plan objective. Ladue is a wealthy, predominantly white suburban school district, however, and the lessons learned there may not apply to other districts, Sykes warns. "The moral of the story is that performance-based pay is not for everyone or every situation," Lawler says. Meyer predicted that if a merit pay plan were to be mandated on a statewide level, "it could be disastrous." When plans are imposed from above, they almost invariably fail and are abandoned after a few years, he says. A merit pay plan should involve as many people as possible in its development, the procedure should be pilot tested, and only then should the plan be implemented "slowly and carefully," Meyer says. When dealing with mice, one researcher said, all that is necessary to get them to run differently is to move the cheese around. But when trying to motivate human beings, merely moving the cheese around may not be sufficient. 0

l

"

RESOURCE LIST
The fonowing references were cited in this article: Belcher, David, "Pay and Performance," Compensation Review, Third Quarter, 1980 Darling-Hammond, Linda, Wise, Arthur and Pease, Sara, "Teacher Evalution in the Organizational Context: A Review of the Literature," The Review of Educationa.l Research, Fall, 1983 Ellig, Bruce, "Pay Inequities: How Many Exist Within Your Organization?" Compensation Review, Third Quarter, 1980 Gallagher, Michael, "More Bias in Performance Evaluation?" Personnel, July/August, 1978 Knapp, M.S. Toward tbe Study of Teacber Evaluation as an Organizational Process: A Review of Current Researcb and Practice, SRI International, 1982 Lawler, Edward, "Merit Pay: An Obsolete Policy?" unpublished paper, 1980 Lawler, Edward, "Participation and Pay," Compensation Review, Third Quarter, 1975 Meyer, Herbert, "the Pay-for-Performance Dilemma," Organizational DynamiCS, Winter, 1975 Meyer, Herbert, Kay, Emanuel and French, John, "Split Roles in Performance Appraisal," reprinted in Performance Appraisal, Harvard BUSiness Review Mikalachki, A., "There is No Merit in Merit Pay!" Business Quarterly, Spring, 1976 Thompson, Paul and Dalton, Gene, "Performance Appraisal: Managers Beware," reprinted in Performance Appraisal, Harvard Business Review Winstanley, Nathan, "Are Merit Increases Really Effective?" Personnel Administrator, April, 1982 WINTER

1983

Hide Document