This article first appeared in the St. Louis Beacon, Feb. 6, 2009 - Tying bonuses to job performance has long been a practice in the corporate world. But as the news out of Wall Street shows, the dangled-carrot approach can sometimes backfire. One potential reason is highlighted in a recent study co-authored by Judi McLean Parks, professor of organizational behavior at Washington University's Olin School of Business.
Performance-based compensation can motivate employees to do better work, but if they are unable to perform you've also given them an incentive to cheat on results, the study shows. The findings have implications for CEO compensation plans and the financial difficulties many companies are experiencing as a result of misreported financial information, McLean Parks said in an interview.
But they also might provide new grist for critics of pay schemes in which the performance bonus is more modest.
Take elementary and secondary education. Paying teachers based on their performance -- a concept embraced by Secretary of Education Arne Duncan -- promises to be at the forefront of a policy debate under the Obama administration.
Missouri law currently requires school districts to adhere to salary schedules "applicable to all teachers," a provision limiting merit pay. State courts have not responded favorably to challenging the status quo.
State Sen. Jane Cunningham, R-West County, has introduced a bill to allow a salary schedule to include other qualifications in addition to experience and credentials, such as measurable classroom performance. It also would allow districts to provide hiring incentives to attract and retain teachers in shortage areas, such as math and science. The bill has been referred to the Senate education committee, and Cunningham hopes to get a hearing in the next few weeks.
Cunningham first introduced this idea when she was a state representative. She said with a new administration that is warm to the idea of merit pay, the timing is right.
"Across the whole country, merit pay is a huge issue," Cunningham said. "My goal is to give this tool to our school boards and superintendents so they can reward teachers who teach in difficult assignments, who are in shortage areas such as math and science, and who have been recognized for high performance." (Missouri courts have ruled that schools aren't allowed to give incentive money for teachers in high-demand areas like math and science.)
Public school teachers chronically complain about being underpaid, and supporters of merit pay believe rewarding teachers for good performance will motivate people to enter the field and excel in their work, hopefully improving student outcomes.
"To allow school boards to develop some sort of differential pay schedule is an idea worth considering," said Brent Ghan, a spokesman for the Missouri School Boards' Association, which supports merit pay. "One of the big issues is retaining excellent teachers, and we'd like to see mechanisms in place to reward teachers who are doing an outstanding job."
Opponents cite the perceived subjectivity and favoritism in the process of assessing merit, and worry that merit pay may discourage teachers from working in schools with many lower-achieving students, particularly when student outcomes are measured by standardized test scores.
The Missouri State Teachers Association passed a resolution last year opposing legislation to promote any use of merit pay, including the use of standardized test scores or other "subjective" criteria as a measurement of teacher performance or to determine future salary increases. Todd Fuller, an association spokesman, said getting all teachers to "a decent level of salary" is the top priority (noting the state's near-bottom average teacher pay ranking). "Then we'll have a discussion about merit pay."
A 'Cure Worse Than the Disease'
It might seem implausible that a teacher would cheat to earn extra pay, given the small payoff compared to that for business and financial executives. Still, McLean Parks said the study showed that "any time someone implements a pay scheme, even on a mundane level, when outcomes are tied to a performance metric that external conditions may prevent them from attaining, you run that risk" of cheating.
Concerns about teacher cheating on student performance measures are not new. The Washington Post reported 25 years ago that when a school board in Dallas deliberated whether to adopt a merit pay plan that awarded teacher bonuses on the basis of student test scores, teachers bluntly warned they'd cheat if this were done.
The New York Times has chronicled what it termed "scandal-ridden" merit schemes in Arkansas as early as 1969, as well as Kentucky's 1997 effort to award teachers cash bonuses for higher test scores as part of an incentive program that was plagued by inflated test scores. More recently, the newspaper noted how the specter of evidence of cheating at four schools in Houston, leading to the firing of six teachers, hung over a 2006 debate about a plan to distribute increased incentive payments of up to $3,000 a year per teacher and $25,000 per administrator.
Michael J. Podgursky is an economics professor at the University of Missouri-Columbia who studies incentive schemes that tie teacher pay to achievement gains by students. He found that while some performance plans resulted in improved student performance, they also may result in more cheating -- teachers focusing excessively on a single test, altering test scores, assisting students with test questions, disciplining low-performing students to ensure their absence on test day and misreporting data.
McLean Parks agreed that if merit pay is based largely on student test scores, teachers at disadvantaged schools could be put in a position where they can't perform because of a one-size-fits-all mentality in setting performance standards. Her study showed that punishment for not meeting standards has a much more pronounced effect on the incidence of cheating than does a bonus for good performance.
"You stop and think of when funds are withdrawn under No Child Left Behind, and that's punishment," McLean Parks said. "This could encourage schools to cook the books to meet requirements. You need to be careful when constructing a merit pay system that you have appropriate measures so that people who are willing to tackle more challenging students aren't penalized."
Prior to conducting the study, McLean Parks said she had a feeling that compensation packages had something to do with the increasing cases of corporate fraud. To test the hypothesis, she and a colleague assembled a random sample of students who were paid for solving an anagram according to one of three different compensation plans.
Participants receiving a "flat salary" for their work were the most honest about score reporting, while many who received a performance-based bonus cheated when self-reporting. Additionally, those who were penalized based on low performance not only cheated but also stole pricey pens that were supposed to be returned.
McLean Parks said the results were striking. "We thought for a long time that if you aligned employee incentives to that of the company, you didn't have to do much else, and people would work really hard," she said. "But we found that if you give them a set of conditions that they can't meet or are unrealistic, you put people in a position where they can cause harm."
Pay contingent on performance, McLean Parks added in the report, "may be short sighted in that such contingencies may encourage behaviors such that the cure itself may be worse than the disease."
In Missouri, a Different Take on Performance Pay
Conversations about performance pay in elementary and secondary education in Missouri generally haven't focused on the issue of unethical teacher behavior. Ghan, the Missouri School Boards' Association spokesman, said if compensation plans are structured correctly teachers have no reason to be dishonest.
"Unfortunately, though few and far between, there may be individuals who have incentive to go outside the boundaries," Ghan said. "Certainly test scores should be a factor [in performance pay models], but if it's not the only factor and others come into play, it lessens the incentive of teachers to cheat."
The Missouri National Education Association has historically opposed merit pay, though it doesn't cite teacher cheating as a concern. Chris Guinther, the group's president, said that if teacher pay is based at all on student performance, long-term measures of achievement such as graduation rates should be the metric, rather than test scores. The association supports locally determined pay incentives for employees working in low-performing or hard-to-staff schools, or for teachers who achieve certifications available to everyone. Ghan said he favors giving school boards flexibility to give extra pay to teachers in high-demand subjects such as math and science.
In 2000, St. Louis Public Schools adopted a policy authorizing financial awards "to provide incentives" to schools and staff to improve student achievement. These "accountability awards" were to be based on objective evaluation data, including student test performance, that show teachers' efforts resulted in improved achievement. Some money was given to schools (but not individual teachers) in the first few years, but the awards haven't been funded since the new board of directors took over, according to Bill Purdy, a longtime board member.
While performance pay advocates are pushing for across-the-board changes to the current law, one school district has been putting its own twist on the compensation model for more than 50 years without legal challenge. The Ladue School District's system of pay raises doesn't take into account student performances on standardized tests -- though if nearly all students in a given class were underperforming, it would likely be a red flag to administrators, said Kathy Reznikov, director of communications for the district.
Ladue administrators visit classrooms to evaluate teachers' interactions with students, the methodologies used and their contributions to the development of the school. Teachers also meet with administrators to discuss their job performance. Administrators assign a point total upon which pay raises have historically been based. (In many years, there's been an across-the-board raise plus money given per point earned.) Reznikov said a new system being devised will put teachers into defined categories that explain their job performance.
"The idea is that if you work to make yourself a better professional then you'll be recognized for that," she said. "That's why we believe this system has continued to work."
Reznikov said she isn't aware of any incidents where teachers have cheated to receive higher pay. Still, the prospect that any performance pay system will increase teacher dishonesty can't be ignored. In a 2003 study, researchers at Harvard University and the University of Chicago found that serious teacher cheating occurred in a minimum of 4-5 percent of classrooms and that incentives induced this behavior. Cheating became more frequent with even relatively minor incentive increases.
As McLean Parks said of incentive pay: "It's not the panacea that people have assumed it to be."
Elia Powers is a freelance writer in St. Louis.