Monday, January 18, 2016

The Unique Insanity of Teacher Merit Pay

Governor Brownback and some of his ultra-conservative allies in the legislature have been in the news lately for promoting a merit pay system for Kansas teachers as part of the new funding formula (local districts currently have the option of implementing such a system for their own teachers, though I don't know of any that do). The knee-jerk response that many educators and public education advocates likely have is "You first."

They make it sound like it's a no-brainer. We all get paid based on how well we perform, right? This will attract the highest quality teachers, right? It may sound logical on the surface but a century's worth of history and extensive research in multiple disciplines doesn't back it up. Yet teacher merit pay tied to ill-defined student success measures will not go away. There's an infectious insanity to it that seems to defy attempts at intellectual inoculation. Nevertheless, here's another attempt at inoculation.

Brief History of Teacher Merit Pay

While somewhat trite, the often repeated saying that “those who forget the past are doomed to repeat it” certainly applies to the use of merit pay in American public education.

As pointed out by Goldstein (2014), Murnane and Cohen (1986) and Ravitch (2013), there have been repeated attempts and failures to implement merit pay since the early 20th century. A large number of studies on merit pay programs implemented through the 1970s found that most failed within six years of implementation. A few examples of these studies include Bacharach et al. (1984), Doremus (1982), U.S. Commission on Civil Rights (1977) and Young (1900). They all generally faced the following limitations: “excessive administrative paperwork, low funding, disagreements about how to judge good teaching, and strong opposition from teachers themselves” (Goldstein 2014).

For those who would claim that teacher unions were responsible for these failures, up through at least the 50’s teacher unions were not widespread, and where they existed they generally weren’t well organized. Teacher opposition through unions was not responsible for the early failures of merit pay systems.

Murnane and Cohen (1986), as reported by Ravitch (2013), divide these programs into two general categories: test score merit pay systems and principal evaluation merit pay systems. Test score systems are essentially “piece-rate compensation systems” better suited to manufacturing or similar jobs where it’s easier to measure an individual’s contribution to a company’s output. Examples include the number of “widgets” assembled per hour or the number of calls made per hour.

Research and case studies have demonstrated that the consequences of such merit pay systems in education include:
  • With bonuses and test scores taking center stage, there is a focus more on the students who easily respond and less on those who are more difficult to work with. 
  • Collaboration among teachers/staff is damaged or destroyed because they’re in essence competing to be the best at achieving a limited measure of student success and competing for money.
  • All other education goals are downplayed, such as critical thinking skills, positive behavior, citizenship, etc. 
Some similar as well as unique problems arise with the principal evaluation system, and they tend to revolve around problems with actual fairness as well as perceptions of fairness. These include:
  • Teacher tasks and goals can’t easily be evaluated relative to pay. The rate of “wiidget” production is very different from improving student performance. The former is much more in the control of the worker than the latter, which has many other variables outside of a teacher’s control. Not to mention there still isn’t agreement on how to measure student success.
  • Poor ratings, often associated with perceptions of unfairness (real or not) resulted in resentment and teachers actually working less hard.
  • Many principals were used to giving higher ratings to inspire teachers to live up to expectations (a not uncommon management practice in general). However the merit pay bonuses commonly decreased this use of higher ratings due to budget constraints.
  • It greatly weakens the trust between teachers and principals (and even among teachers); as a result, teachers are less likely to ask for help.
By the mid-1980’s, 99 percent of U.S. teachers were in districts using a uniform salary schedule based on education and experience. Yet despite the history and evidence against merit pay, it again saw a resurgence in the 1990’s that continues to this day. According to Ravitch (2013), the assumed underlying cause and effect is that offering bonuses to teachers will motivate them to do what it takes to obtain the money. As a result, they will increase their effort and teach more effectively, raising student test scores in the process.

Great theory, if we’re talking 1925. As with the previous sets of implementations, contemporary merit pay programs haven’t been any more effective. Studies documenting this include:
  • The National Center on Performance Incentives at Vanderbilt University’s study of Nashville public schools (Springer et al. 2010).  A $15,000 bonus was offered to teachers for improving math test scores. Over three years teacher performance was compared between a control and experimental group, with no significant different in student test scores occurring.
  • A Rand Corporation study (Sparks 2011) documenting the failure of New York City’s implementation of merit pay (no impact on reading and math test scores; no improvement of teacher satisfaction). It was generally seen as a waste of public money.
  • A four-year study (Glazerman and Seifullah 2012) of Chicago’s Teacher Advancement Program found a somewhat higher teacher retention rate but no significant impact on student test scores in the schools where merit pay was implemented.
References to additional studies can be found in Goldstein (2014), Murnane and Cohen (1986) and Ravitch (2013). You can also review the referenced studies on the research page of the National Center on Performance Incentives.

Evidence from the Workplace

Looking at research from the business/management world, we find a lot of work demonstrating the difficulties associated with effectively constructing a monetary focused incentive based system. Kohn’s (1993) article in the Harvard Business Review provides a good overview of some of this research.  A lot of the research methods and interpretations are rooted in psychology as well as what we now term Organizational Development Theory. After reviewing a large amount of research documenting some of the what and why of incentive program ineffectiveness in the workplace (over 25 references are given), he then lays out the following six specific reasons underlying why reward-driven compensation systems end up failing:
  • Pay is not a motivator: The assumption that more money is a strong motivator for people to work better or more over the long term is somewhat controversial, with context playing a huge role in the degree and type of motivation it provides.
  • Rewards punish: “Rewards [can] have a punitive effect because they, like outright punishment, are manipulative.”
  • Rewards rupture relationships: An over emphasis on competition for rewards, recognition and ranking, exacerbated by being highly publicized within an organization, often results in employees seeing each other as “obstacles to their own success.” The end result is the destruction of cooperation and teamwork.
  • Rewards ignore reasons: Relying on incentives (particularly financial incentives) to gain desired results doesn’t address underlying problems. It also has been substituted for giving employees what they really need to be effective – “providing useful feedback, social support, and the room for self-determination.”
  • Rewards discourage risk-taking: “Whenever people are encouraged to think about what they will get for engaging in a task, they become less inclined to take risks or explore possibilities, to play hunches or to consider incidental stimuli. In a word, the number one casualty of rewards is creativity.” 
  • Rewards undermine interest: Employees also tend to place almost complete focus on the specific criteria being evaluated for the incentives, as opposed to the job overall and any intrinsic motivations. “[A]ny incentive or pay-for-performance system tends to make people less enthusiastic about their work and therefore less likely to approach it with a commitment to excellence.”
Some may question that I’m referencing a 1993 article. But contemporary studies and overviews, including Van Dyke (2011), while they have evolved and not all in agreement with Kohn (1993), continue to emphasize the difficulties associated with creating and implementing a fair and effective incentive based system. Many of those difficulties still hinge on a) being able to effectively articulate what the desired outcomes are, b) having agreement among the key stakeholders that the outcomes being focused on are in fact the correct ones, c) effectively measuring the outcomes and d) determining what it is that employees actually value and how that varies by context.

I should also be clear that none of this says money is without any motivational value. What the history of teacher merit pay and the workplace/management research cited above indicate is that a) it is only one piece of the puzzle and b) measured success must equal actual or desired success. If the incentive programs aren’t carefully constructed to effectively account for the multitude of contextual sources of motivation at play (monetary and non-monetary sources), the contextual goals and indicators of success at play, effective measures of this success, organizational operations, the adequacy of base pay levels, the level of trust and respect within the organization, social/cultural factors, etc., then they will have unintended consequences. And unintended consequences have been the norm throughout the history of teacher merit pay, up until the present.

Ravitch (2013) does point out some rare instances where merit pay programs have had some degree of success relative to a) the longevity of the programs, b) teacher satisfaction and c) no negative impact on student success. They were successful because they addressed many of the issues already discussed, some of which follow:
  • Merit pay was tied specifically to the extra work that teachers did. This is much more contextual, objective and easily agreed upon compared to either test scores, which are a very limited and controversial measure of student success, or the evaluation of a principal, which is more subjective and less consistent. Not using principal evaluations also eliminated any damage that such a merit based pay system could do to the trust between principals and teachers.
  • Almost everyone received some type of monetary award, it was often relatively small compared to base pay, and the awards were kept secret to avoid causing hard feelings. This minimized the perception of manipulation sometimes associated with merit based pay as well as the potential for disrupting relationships and collaboration.
  • Teachers and staff had a hand in designing the system. This increased their ownership of the system and desire to see it operate successfully (including working with administration to monitor and tweak it over the long term). Including teachers and staff also helped them feel respected.
  • They also generally occurred in districts that paid well to begin with and provided good working conditions. In other words, merit pay wasn’t used as a supplement for fair and equitable base pay, also helping teachers and staff feel respected. As Jim McNiece, chairman of the Kansas State Board of Education, and education commissioner Randy Watson recently pointed out, raising the pay of some without addressing overall compensation levels in the profession, will likely have a negative impact on morale (Llopis-Jepsen 2016).
Conclusions

Falling back on another trite saying, “insanity is doing the same thing over and over again and expecting different results.” With such a long, unproductive and repeated history, it would seem that teacher merit pay is its own special brand of insanity. The evidence suggests that if you want to positively impact student success, across a wide range of measures, start by providing teachers and staff with a competitive base pay and benefits package, with starting salaries and raises based on education and experience.

If additional incentives are then added, they should contextually be applied per the success stories provided in Ravitch (2013). Tying merit pay to indicators of student success doesn’t work. There are too many contextual variables impacting the various measures of student success, that vary in importance by district. Nor do we all agree on what those success indicators should be or how they should be measured. If these hurdles could be addressed, perhaps some type of merit based system could be implemented based on such indicators. However, I’m skeptical because of the other negative impacts that will potentially remain, on everything from collaboration to the ability to meet all student needs.

The evidence is fairly clear on where to go from here. The question is whether or not Kansas's governor and ultra-conservative legislators will accept it, recognize the limitations of teacher merit pay tied to student success indicators, and drop it as part of the new funding formula. Or will they continue pushing forward with the idea because they're driven by an ideological view of small government and/or a view that privatization should play a much bigger role in preK-12 education? Whichever they choose, Kansans should demand they be honest about what their motivations are.

References


Bacharach, Samuel B., David B. Lipsky, and Joseph S. Shedd. Paying for Better Teaching (Ithaca, NY: Organizational Analysis and Practice, 1984), 28– 29, 37– 38.

Richard R. Doremus, Whatever Happened to Kalamazoo’s Merit Pay Plan? Phi Delta Kappan 63, no. 6 (February 1982).

Glazerman, Steven and Allison Seifullah. An Evaluation of the Chicago Teacher Advancement Program (Chicago TAP) After Four Years (Washington, D.C.: Mathematica Policy Research, March 7, 2012)




Murnane, Richard J. and David K. Cohen, Merit Pay and the Evaluation Problem: Understanding Why Most Merit Pay Plans Fail and a Few Survive. Harvard Educational Review (Spring 1986).


Sparks, Sarah D. Study Leads to End of New York City Merit-Pay Program. Education Week, July 20, 2011.

Springer, M. G., D. Ballou, L. Hamilton, V. Le, J. R. Lockwood, D. McCaffrey, M. Pepper, and B. Stecher, Teacher Pay for Performance: Experimental Evidence from the Project on Incentives in Teaching (Nashville, Tenn.: National Center on Performance Incentives at Vanderbilt University, 2010).


Young, Ella Flagg (1900) Isolation in the School. PhD Dissertation, University of Chicago.

United States Commission on Civil Rights. School Desegregation in Kalamazoo, Michigan (April 1977).