David Bacon Stories & Photographs
The Money in Testing
by David Bacon

LOS ANGELES (5/28/00) - This has been the year that California schools went test-crazy.

In every district, students have taken the new state-mandated STAR (Standardized Testing and Reporting) test by the thousands. Based on their scores, every school in the state has been rated and placed on a scale from the lowest to the highest. Dozens of LA district schools found themselves at the bottom of the list.

And this year Governor Davis' latest pet education reform, the high school exit exam, will be implemented for many students, guinea pigs who will test it for all California students in years to come.

It all may appear to be the result of Pete Wilson's, and then Grey Davis' belief that a highly-publicized commitment to the latest fad in education reform is the key to the hearts of California voters. But if that's the case, they're just two among a horde of politicians around the country who've arrived at the same conclusion.

By the turn of the millennium in January, every state in the U.S. but one had adopted standards for what public school students were expected to learn in at least one subject. Forty-one of those states had gone on to adopt tests to measure student performance, presumably on the yardstick those standards provided.

Like California, Ohio's gone test-crazy this year too. In mid-March, every fourth grader in the state took the Ohio Proficiency Test, Ohio's bid to do what everybody seems to think will improve education.

In the fourth, sixth, ninth and twelfth grades, students take standardized tests, which supposedly measure what they know. Increasingly, the curriculum in the classroom is being reoriented to produce good test results in five areas - reading, writing, math, citizenship and science. Ever greater amounts of classroom time are now devoted to teaching strategies for taking tests and scoring well, beginning in the fall, when students in some schools are given pre-tests, and continuing through January and February, when they go through practice drills.

But what do the tests actually measure? And even more important, do standardized tests really do much to improve the quality of education?

Two Ohio mothers say the tests are hurting students, not helping them. "We used to have a wonderfully rich program in our schools," says Mary O'Brien, whose five kids are students in the Upper Arlington City Schools outside of Columbus. "Now it's all oriented to test-taking. We think these tests are racist and sexist. They just rank and sort students - they don't actually teach them much at all. It's really just politics."

O'Brien and her fellow parent, Teri Ziegler, started a movement three years ago to junk the Ohio Proficiency Test. "We purposely didn't give our group a name," she says. "We're just two moms against tests. We're accidental activists."

But O'Brien and Ziegler's conclusions aren't just the subjective reactions of a couple of parents. Last week they gained support from an exhaustive study by Youngstown State University Professor Randy Hoover, who examined the correlation between scores on the proficiency test and the income level of students. He found that the poorer the family of the student, the lower the test score was likely to be.

The problem isn't that poor students are less intelligent, Hoover says. It's that the tests don't measure what they know, and are geared to the experiences and knowledge of wealthier students.

Youngstown and Cleveland schools, for instance, ranked far down on the state's report card, based on 27 standards - mostly proficiency test scores. As Hoover discovered, schools in affluent neighborhoods do predictably well, and schools in poor, minority neighborhoods don't.

Is the ranking of schools going to lead to diverting resources from wealthy communities to poorer ones? Not necessarily.

Next year, promotion to fifth grade will depend on passing the reading portion in the fourth grade. Since students who don't pass are likely to be concentrated in schools with the least resources, those schools will have increasing problems marshalling the additional teachers, classrooms and other materials needed to help those students make up for lost time.

Furthermore, school districts which rank low on the state's report card could lose funding, and have the state allow charter schools in their area, diverting even more students and resources. Proposals are now being considered to tie pay increases for school personnel to test ranking as well.

"Wealthier schools will forge ahead, and poorer ones fall further behind," O'Brien predicts. "Our district is pretty well-off, but we're not just concerned for ourselves. We're worried about all of Ohio's children."

Ohio isn't alone in confronting this dilemma. This year every state in the U.S. but one had adopted standards for what public school students were expected to learn in at least one subject. Forty-one of those states had gone on to adopt tests to measure student performance, presumably on the yardstick those standards provided.

Such a widespread and rapid change has rarely swept through the nation's schools. And it hasn't done so without resistance. Parents in many states have protested the extensive use of standardized tests, especially since so many decisions involving students' lives are now determined by test performance. Graduation from one grade to another, and from high school itself, is now often test-determined. Test scores increasingly determine the ranking of schools, the resources available to them, and even the ability of parents and teachers to control the local curriculum.

Meanwhile, politicians vie with each other to appear more concerned about education, and position themselves as would-be "education governors," and now, in this year's national election, "education president."

Driving this almost obsessive interest in testing are factors ranging from political ambition to genuine frustration by parents and teachers with the ability of the public school system to teach its students. But testing is getting a big push from another important source, which gets much less media coverage - the testing companies themselves.

Districts and states are spending huge sums on testing and standards. That money is going to a few large companies, who are also the publishers of the texts that schools use for instruction. Dominating the field are three big publishers. McGraw-Hill and its subsidiary, CTBS, publish the Terranova test series. Harcourt Inc.'s Education Group publishes the Stanford-9 test, and Houghton-Mifflin's Riverside division publishes the Iowa Test of Basic Skills and the Metropolitan Achievement Test.

Ohio's tests are scored by National Computer Systems, a company with a close relationship to a Harcourt, one of the big three. The NCS contract is estimated to be worth about $10 million a year.

Ohio's ninth and twelfth grade students have their essays graded by Measurement, Inc. For $1.4 million a year, the company employs temporary workers at close to minimum wage in a North Carolina strip mall. These workers, who have no teaching experience or education credentials, spend about two minutes looking over each paper. Many states other than Ohio contract for the services of this same company.

Together, test publishers divide a testing market that was estimated at $218.7 million for 1999 by the Association of American Publishers. The publication of standardized tests is considered part of the market for instructional materials, which, at $3.4 billion, is over 15 times as large. But the market for tests has been growing at an average of 7% a year for over a decade -- much faster than the market for textbooks, whose annual growth rate was 3% over the last four years, according to the executive director of the association's school division, Steven Dreifler. That growth rate alone makes it an attractive market for publishers, and one which promises to become much larger and more important in relation to their traditional business of publishing books.

Publishers are very secretive about the money they make on testing, and hide it within the income figures they report for educational publishing generally. But Houghton-Mifflin's Riverside testing operation, which sells the Iowa test, grew at a phenomenal 17% last year, while its overall textbook division grew at 9.2%. In 1997, McGraw-Hill's testing division had gross income of $95 million, and its overall educational publishing group grew 5.7% to $832 million.

The rising profits of test publishers have had a profound effect on the process in which the tests themselves are developed and adopted.

In November, 1997, the California Board of Education overruled both state Superintendent of Public Instruction Delaine Eastin, and the superintendents of numerous districts, adopting the Stanford-9 test to administer to 4 million of the state's students annually. State educators had spend years developing a set of standards for core curriculum, and their intention was to move on to develop tests which would reflect what was actually being taught.

Governor Pete Wilson, and the state school board he dominated, cut the process short. To force the legislature to immediately adopt an off-the-shelf test, he withheld $200 million in school spending until lawmakers agreed. "We think it's a big waste of time, energy and money," said Lloyd Porter at the time, a Yorba Linda teacher monitoring board actions for the California Teachers Association.

Wilson insisted on adopting the Harcourt test specifically, and school districts around California were forced to sign contracts with the company, worth $12 million a year, for a guaranteed period of 5 years. The state even insisted that all children take the test in English, including those who spoke only Spanish. Obviously, the test didn't assess the real knowledge and skills of those children. But the kids fulfilled a more important function. They consumed the product.

The year following adoption, Harcourt's revenues from its Education Group division shot up $85 million (18%), and its profits jumped $34 million (58%).

It wasn't all roses. Harcourt was later penalized $1.1 million in August, 1999, for late reporting of test results, and for 100,000 mistaken reports of results which were sent to parents, and had to be recalled.

Testing and scoring errors are not that uncommon. Last year 8000 New York City children were mistakenly sent to summer school when McGraw-Hill printed their math scores on their reading test forms.

Harcourt added to its remarkable revenue increases in recent years by participating in one of the most highly-touted examples of this new test-driven trend in education - what Republican Governor, now presidential candidate, George Bush Jr. claims as Texas' "education miracle." Beginning as early as 1985, the company's subsidiary, Harcourt Brace Educational Measurement, was involved in developing the now-famous Texas Academic Assessment Skills test.

Texas currently contracts for test development with National Computer Systems, which scores the Ohio Proficiency Test as well. In Texas, NCS subcontracts test development to Harcourt, and gets about $20 million a year; Harcourt's cut is not public. In addition Harcourt gets about $2.8 million a year for developing TAAS study guides.

Being the test developer can be very advantageous. The company's textbooks were marketed to local districts around the state with a flier stating "Why choose Harcourt Brace for your math program? . . . (It is the) only program to have tests written by the same company that helps to write the TAAS tests and actually wrote the Parents' Study Guide for TAAS: Harcourt Brace Educational Measurement."

Stating the connection so baldly is considered bad form, and Harcourt later said it would discontinue the promotion. But the fact is that, according to the Texas Education Agency, the company sold $25 million of elementary school math textbooks in the state last year alone.

Other contracts are similarly lucrative. McGraw-Hill won a $30 million contract for tests in Kentucky this year. In Mississippi, with one of the lowest per-pupil spending levels in the country, the state is about to sign a 10-year testing contract with McGraw-Hill for a total of $29.4 million. Despite releasing test scores late this year, Wisconsin not only renewed its McGraw-Hill contract, but also increased the annual payment from $1.25 million to $1.5 million.

A rebellion by Wisconsin parents, however, finally forced the legislature to kill a proposed $10.1 million test high school students would have had to pass to graduate. In California, on the other hand, the state's new Democratic governor, Grey Davis, last year rammed a similar test through the state legislature, along with a measure similar to Ohio's, ranking the state's schools based on test results.

Despite the money spent, uncertainty is rampant over what the test scores actually mean. Sandra Stotsky, a researcher at Harvard, says that the TAAS test, for instance, doesn't measure what politicians say it does, when they argue that rising scores in the state mean that children are learning more. "There may have been no real improvement in reading skills. There may even have been a decline," she notes, alleging that the test is made easier so more students pass.

Texas has been consumed by testing fever, in which districts and schools organize TAAS camps, hold TAAS Olympics, and bend the curriculum towards test-taking, in a high-stakes environment in which the penalties for low performance can be brutal. Texas has no collective bargaining for teachers, and promotion can depend on student scores. Urban Texas counties have even indicted a school board, fired teachers and a principal, and launched investigations over allegations of test tampering.

Underlying the hysteria are even more basic questions about whether standardized tests reflect a bias which favors white children over racial minorities, and English-speakers over immigrants.

These tests are not recent inventions. The Stanford-9 and Iowa tests go back over 60 years, and were originally developed in universities. Stanford psychologist Louis Terman, who wrote the first test in the Stanford series before World War One, was notorious for regarding racial minorities and Jews as "feeble-minded." Other early test developers were held similar racist views, and saw the tests as instruments to weed out the less intelligent.

Allegations of racial bias continue to dog the tests today, despite publishers' claims of an unprecedented level of objectivity. In Cleveland, the NAACP charged the Ohio Proficiency Test with being racially biased after no student in five schools in poor urban areas passed all sections.

Two Harcourt tests, the Otis-Lennon and Metropolitan Achievement, were recently charged with being Eurocentric and discriminatory to New Orleans's African-American students, when they were used as a basis for admission to a local high school. For admission during the 1997-98 school year, 763 students took both tests, of which 44 percent were black and 42 percent were white. Of those, 347 passed both tests. Among those who passed, 27 percent were black and 59 percent were white.

New Orleans was sued, a common experience - most states are sued over bias or problems with mistaken test results. The Mexican American Legal Defense and Education Fund, for instance, mounted a major legal challenge to the TAAS test, saying it has a discriminatory impact on Black and Hispanic children. According to Maureen DiMarco, formerly California's Secretary of Education under Wilson, and now vice-president for education and government at Houghton-Mifflin, "it's hard to have a test that doesn't get sued." But, she notes, it's the state or school district that has to mount a defense and bear the legal costs, not the publisher.

"Tests do measure social and economic conditions," she admits. "Children from poor communities go to schools which don't have resources, and use less effective methods of instruction. Lots of test scores can be explained by the lack of books. Poor children also move more often. The implications of what's being measured are very deep. Poor kids can learn just as well as higher income kids. They're just not getting the resources they need to learn."

Despite these questions, and growing opposition by parents and teachers, the test market looks good to publishers. Twenty states now work with publishers to come up with state-specific tests, called "criteria-referenced", rather than using off-the-shelf, "norm-referenced" tests, according to the AAP's Dreifler. Working to produce unique tests for a state can produce consulting fees for the publisher helping to develop the product, followed by sales of the test itself. "The trendline is that this type of testing will grow," he says.

Dreifler says that the states don't normally involve publishers in the writing of the standards themselves, but DiMarco says the connection is really tighter than that. "It's a wise state that seeks the advice of a publisher when formulating standards, to ensure they're rigorous, and not too vague," she explains. "Then they can issue a better request for bids." The company that helps develop the standards, presumably, has a better chance at getting the bid for the test, and has an advantage in selling textbooks which meet the curriculum requirements as well.

Are we heading for an era in which large companies increasingly tie all these elements together? Publishers already hold enormous economic power, and strive for a close relationship with the state authorities who choose instructional materials and tests. With personnel moving back and forth between the state and private sector, an education-industrial complex is beginning to emerge.

Backing the growth in testing is also what supporters refer to as the standards and accountability movement. But this is not just a grassroots wave sweeping the country from below. Pushing its agenda are organizations like the Pew Charitable Trust, the Heritage Foundation, and the Hudson Institute, and corporate CEOs like IBM's Lou Gerstner.

O'Brien points to the relationship between the National Governors' Conference and the Business Roundtable, who together have set up an education thinktank called Achieve, Inc. Former Ohio Governor Voinovich contracted with the organization to produce a study of education in the state, which was used as the basis for moving towards standardized testing.

"It's not a conspiracy," she says. "Big corporations, like IBM, Proctor and Gamble and Eastman Kodak are very up-front about their agenda. They want schools to educate student to their specifications. They want education centered on testing, and curriculum aligned to the tests. They've even moved to pressure university education schools to adopt this way of thinking, and to join their school reform movement."

But O'Brien believes that opposition is growing as well. "We've appeared before the state board of education and lobbied legislators. We've organized conferences in many districts. Our movement has really exploded in Ohio this year."

Nevertheless, "the standards and accountability movement is growing," DiMarco says. "Just look at how many education governors we have now. Even presidential candidates know they have to speak to these issues."

"Publishers definitely see tests as an opportunity," Dreifler concludes.

Top of Page


photographs and stories by David Bacon © 1990-

website by DigIt Designs © 1999-