If Students Are Paid, Test Scores Rise, Study Finds

When it comes to explaining poor test scores, politicians, parents and educators often complain that schools need more funds to produce better results.

Now, a newly published study by a USC researcher shows that money does indeed make a difference--at least when it is paid to students for their correct answers.

The study by education professor Harold F. O’Neil Jr.--the first to measure the potential for financial incentives to boost test scores--found that students who were offered $1 for every correct answer on a national math exam racked up scores that were 13% higher than students who were simply urged to do their best.

Among the 749 Los Angeles County eighth-graders studied, those who were promised money also were more likely to try several approaches to solving difficult problems and to check their answers.

“Small wonder,” O’Neil said, “that lackluster scores have inspired calls for reform.”

But his study shows that poor test scores may be caused, in part, by a lack of motivation rather than a lack of knowledge.

And the key to raising standardized test results--which spur political debate, establish school reputations and influence housing prices--may not be sweeping reforms, but old-fashioned bribery.

O’Neil said he got the idea for the study--which will be published this month in an academic journal of testing called Educational Assessment--when he looked at international comparisons showing that students in countries such as China and Japan score far better in math than American students.

“I always had suspected that one reason Asian kids did better on tests in general was because they saw it as a thing to do, they were motivated and they worked harder,” he said.

*

To measure that, he and coauthors of the report at UCLA gave comparable groups of Los Angeles eighth-graders and 12th-graders different sets of instructions aimed at spurring greater effort and engagement in a sample math test drawn from the National Assessment of Educational Progress, a test known informally as the “nation’s report card.”

One test group was told that the test was challenging and, if they did well, they would “feel a sense of personal accomplishment and feel good.” Another group was told that their scores would be measured against those of students around the world and they could feel proud if they performed well. A third group was offered a certificate from UCLA for a high score.

The final group was told simply that they would be paid for each correct answer, as much as $41 for a perfect score.

The first three sets of instructions made very little impact on the test-takers; in fact, a third of the eighth-graders could not even remember afterward what they had been told.

But the offer of money did register--at least with the eighth-graders. “It was almost like magic,” O’Neil said of the higher test results produced by the offer of money.

Among the 12th graders, the offer of money had no impact on test scores, and O’Neil believes that is because the small amount of money offered was not enough to interest high school seniors accustomed to the high cost of such things as car insurance and dating in Southern California.

To many parents, it is common sense that money motivates learning--they have watched their children improve their grades in response to offers of cash rewards for good report cards.

But O’Neil cautions that those gains may be short-lived. “There’s evidence that, if you pay kids to learn, they do well but then, after a while, they won’t learn unless you pay them,” he explained.

“On testing . . . you really want to know what the kid knows so . . . you want to make sure the kids are trying hard. And for eighth graders, a dollar per item works best,” he said.

Officials with the National Assessment of Educational Progress acknowledge that the lack of motivation among students means that test results may understate how much students actually know, especially in the upper grades.

Students who have taken the NAEP exams are routinely asked whether they invested as much effort in it as in their normal school work. Typically, fourth-graders say they try just as hard, eighth-graders try less hard and 12th-graders care very little about the tests.

But NAEP officials say that does not invalidate test results, because they are used primarily to track long-term trends in academic performance and make comparisons among states. And the motivation problem, presumably, remains constant year to year, state to state.

David Berliner, an Arizona State University education professor who coauthored a controversial 1995 book called “The Manufactured Crisis,” which contends that American schools are doing far better than they are given credit for, said he was not surprised at O’Neil’s results.

“I’ve long argued that if we wanted our kids to do well against the Germans or the Japanese we ought to pay them $50 to take the test,” he said. “American kids are tested to death, they see there aren’t great consequences to the tests and they’re bored by them.”

Lawrence Feinberg, assistant director of the National Governing Board of the NAEP, agreed. “Without consequences, of course, it’s human nature that people aren’t going to work as hard as they can,” he said.

But, he said, money is not the only incentive that can be used.

*

Some educators have argued that schools ought to devise nationwide exams that students would have to pass to graduate and that could be used by employers to decide whom to hire. With a job or a diploma at stake, it is argued, students would work harder in class.

“Everywhere else in the world where you have high performance, the test is connected to something very major, like getting a job, getting into a university or getting into an apprenticeship program,” said Ruth Wattenberg, deputy director of educational issues for the American Federation of Teachers.

“We need to be worried about how to create incentives for students to work hard throughout their academic career,” she said, “so they can go on and have a productive life.”

Making tests count for students or their schools, though, has political implications.

Some students would have to fail if graduation exams were allowed to hold them back, for example. And if test scores were used to measure a school’s effectiveness, poor student performance could cost some campuses funding. A bill now before the state Legislature would allow students assigned to poor-performing schools to use state-funded vouchers to attend private schools or public schools in other districts, taking their state funding allocations with them.

With consequences such as those, teachers and administrators would have to find a way to motivate the students to do well.

Among education experts, it has long been suspected that the NAEP test results were lower because of a lack of motivation, said Bob Schaeffer, a spokesman for FairTest, a national group that is generally critical of standardized testing. The O’Neil study, he said, is the first to confirm that.

“It’s troubling that you can boost performance for a relatively low investment,” he said, especially when some educators are pushing for the creation of a national examination whose scores would be used to guide teaching and set academic standards.

“We need to back off from this hysteria and the focus on one-shot test score results . . . and move toward more comprehensive and meaningful forms of assessment based on actual performance over time . . . which are less susceptible to short-term manipulation,” Schaeffer said.