By Margaret Renkl

After our youngest child was born two years ago, the nurses whisked him to a warming table to give him his first test. He scored a 9 on his first APGAR evaluation, a 10 on the second. “He’s a very organized baby,” a nurse called to my husband and me about our 20-minute-old son, and we beamed at each other. We had another wonderful little boy, already so smart and organized, a boy who would no doubt go far in the world.

Though as veteran parents we understood perfectly clearly that APGAR scoring of newborns is no prediction of either future intelligence or future success, we both felt a kind of primitive, utterly irrational pride when our baby’s initial score of 9 magically rose to a perfect 10. Our child: the gold medalist, the Rhodes scholar, the Nobel Prize winner in the great book of life.

Among the dozens of middle-class parents I know, there’s not a single couple who didn’t at least start off entirely persuaded of their own child’s brilliance. “He walked at 8 months,” they brag at the park, in the grocery store, in the preschool parking lot, at family reunions. “She spoke in complete sentences by her first birthday and taught herself to read by watching Sesame Street.” But none of these qualities, child-development experts tell us, necessarily indicates a particularly smart child—not walking early, not talking early, not reading early. Children develop according to a broad human outline, but the details inside those lines are infinitely various.

Still, we Americans, obsessed with success—especially early signs of success—persist in believing that achievement can be measured, that faster is better than slower, and that more is better than less. Even worse, we often make the mistake of assuming that something which cannot be quantified is somehow not there at all. Consequently, test anxiety for a lot of middle-class parents starts very early. For well-off Nashville children whose parents choose private schools, there are entrance exams to sweat—starting at the ripe old age of 4. And while by definition public schools are not selective, they force kids and parents to contend every year with their own brand of test anxiety, the Tennessee Comprehensive Assessment Program, otherwise known as TCAP.

There’s been a lot of talk about TCAP lately, partly because individual score reports went home Monday in students’ report cards, and partly because the testing company from which Tennessee buys the tests, CTB/McGraw Hill, was several months late in reporting test results, for the second year in a row. We’ve been hearing repeatedly in local news reports about how damaging it is to a school system when scores arrive too late to be used in determining which kids should go to summer school, or how teachers should plan their lessons. We’re hearing state officials demanding to know why the scores are late and what the testing company intends to do to make sure they aren’t tardy again this year. But we’re not hearing any of them ask whether our kids ought to be taking these tests in the first place.

Because Metro administrators tend to be middle-class people who have to answer to the state politicians—themselves largely middle-class people—who fund our schools, it’s not all that surprising that they operate according to the typical middle-class faith in external measures of intellectual aptitude and achievement. The two bedrock principles underlying a test like TCAP are the belief that what kids have learned can be accurately measured, and that those measurements can accurately determine the academic success of a child, a teacher, a school, and a city’s school system.

So with increasing intensity, Metro teachers emphasize and prepare for this annual set of fill-in-the-bubble tests that are supposed to tell both schools and parents how well our children in grades K through eight are learning. But if we look beyond the political rhetoric of “raising standards” and “school accountability,” there’s little reason to believe that the TCAP does a very good job of telling us what’s really happening in our classrooms. As a former student, as a former teacher with 12 years’ classroom experience, as a journalist who researches child-development and education issues, and as the parent of a child in a Metro school, I don’t believe the TCAP—nor any standardized, multiple-choice test—truly can give an accurate measure of what a particular child has learned, or of how well that child’s teacher has taught him.

But when both major-party candidates for the U.S. presidency claim that they will be “the education president”—and bolster the claim with promises to institute mandatory testing for every child, every year, in every school—it’s hard to blame Tennessee’s state Department of Education for spending $13 million over three years to buy these tests for kids in grades three through eight. It may even be hard, in this political climate, to blame Metro officials for kicking in another $148,000 to buy the tests for grades one and two as well, despite the fact that they aren’t required to do so, and the fact that even testing experts agree that any standardized diagnostic tool for very young children is extremely unreliable.

And I certainly don’t blame the teachers, who in some circumstances must devote an obscene amount of teaching time to test preparation. Their feet to the legislative fire of “high standards” and “accountability,” educators and education officials here are only doing what education officials are doing all over the country now: responding to political pressure to raise standards in the public schools, standards that are measured by standardized tests.

To paraphrase George W. Bush, American taxpayers just aren’t willing to pay for something that isn’t demonstrably working, or to reward someone who hasn’t demonstrably achieved something. So for politicians, tests like the TCAP are supposed to do two things: 1. identify students who are struggling, and 2. identify schools where there seems to be a whole lot of struggling going on.

This is a polite way of saying that these tests are supposed to identify failing kids and failing schools. If the kids are failing, say proponents of the standards movement (that would be virtually every candidate this election season), they need to go to summer school, or they need to repeat a grade. If the schools are failing, they need new leadership, or they need new teachers—or at the very least they need to be embarrassed publicly. Coming soon to your local daily newspaper: the Metro report card, a list of all area schools ranked by their aggregate TCAP scores.

The standards movement is a back-to-the-basics education agenda first espoused by political conservatives but now embraced by almost everyone. Many states that have adopted a standards-based curriculum—Gov. George W. Bush of Texas was one of the first governors to push for such reforms in his state—also use the results of standardized test scores alone, with no regard for grades, to determine summer-school attendance or promotion status. This use of diagnostic data is known in education circles as high-stakes testing, and it’s supposed to end the practice of social promotion, which by now everyone outside the actual world of education—liberal and conservative alike—agrees is a bad thing. Kids shouldn’t go on to the next grade if they haven’t learned what they need to know in the grade they’re already in.

It sounds like a “no-duh” kind of idea, as does the standards movement itself. But high-stakes testing is coming under increasing fire among educators and students of social policy, and for good reason: It turns out that holding back kids doesn’t necessarily mean they’ll learn the material any better the second time around, and it puts them at a shockingly high risk of dropping out later on.

Last month Education Week reported a new study of the country’s most notorious high-stakes testing philosophy at work in a large urban system, the Chicago public schools. The report found that students held back under the new high-stakes testing policy actually did no better on subsequent testing than kids who had been socially promoted. In some cases, the socially promoted kids did even better than those who’d been held back for another year to develop the skills they needed.

Believing that standardized tests truly can measure what students have learned and thus—more crucially—what they are likely to achieve in later grades seems, on the surface, no great leap of faith. It’s pretty simple: Enough wrong bubbles filled in with that No. 2 pencil, and it’s clear a student needs another chance to learn the material. Enough kids in a class fill in enough of the wrong bubbles, then there must be something wrong with the teacher. Enough kids fill in enough of the wrong bubbles year after year, from teacher to teacher, then there must be something wrong with the school. Surely a first-grader can understand this logic.

The standardized-test juggernaut is driven mainly by politicians and, to some extent, parents. It has not been overwhelmingly embraced by professional educators, who might acknowledge the political expedience of such testing, but who decry its insidious long-term influence on school curriculum. Politicians, they say, don’t know how to educate kids; elected officials, they say, should run the government and let experienced teachers run the schools. For their part, politicians argue that teachers oppose testing only because it identifies ineffective teaching and exposes them for the lazy incompetents they really are.

But veteran teachers, unlike politicians, understand that there are myriad reasons to distrust the easy answers that standardized testing appears to provide. Among them:

Scoring errors abound. Last year CTB/McGraw Hill, the textbook publisher/test provider from which Tennessee buys the TCAP, misscored the Tennessee tests. The error, caught before scores were distributed, nonetheless caused a delay of several months before the correct results were reported.

Last year CTB/McGraw Hill, it turns out, misscored not just the Tennessee tests, but those of eight other systems as well. And some school systems didn’t discover the error until some high-stakes decisions had already been made. New York City sent 8,600 children to summer school last year who didn’t need to go, and officials there only discovered the scoring error because a single education official, the city’s testing director, persistently (and unpopularly) argued that, in his opinion, the scores were lower than they ought to be.

David Taggert, the president of CTB/McGraw Hill, has admitted that scoring errors are normally detectable only when someone in the school district, reading the scores in the context of other information he or she has about the school, notices that something is amiss. Which of course begs the question: What happens if no one notices that something is amiss?

Even accurately scored results aren’t necessarily accurate. According to David Rogasa, a Stanford University professor of educational statistics, for a child who reads exactly at grade level—whose “true” reading level, calculated as an average of many scores taken close together, is in the 50th percentile—there’s about a 60 percent chance that on any given test day that student will score either above the 55th percentile or below the 45th. Many variables affect how well a child tests: She may not have slept well the night before. She may have skipped breakfast. She may be daydreaming about the scooter her parents have promised her if she makes good grades. One Metro teacher says that for the last couple of years, Metro maintenance workers chose TCAP test week to mow the grass outside his classroom.

A difference of 6 or more points higher, or 6 or more points lower, than “true” may not seem like much to get worked up about, but educators do. These are carefully tested (for lack of a better word) tests, and unlike polls that often have an error margin of 5 points or more, a loss or a gain of 2 points is considered significant. So if a kid who last year had a reading-comprehension score in the 50th percentile comes up this year with a score in the 43rd percentile, that drop would be a matter of real concern to anyone—parent, teacher, or politician—who puts a lot of faith in these tests.

Perhaps more critically, though, standardized tests fail to give an accurate measurement of learning, at least partly because they cannot distinguish between the child who has filled in an outrageously wrong bubble and the child who has filed a wrong answer that’s a near miss. As Leo Bostein, the president of Bard College, pointed out in a New York Times op-ed piece this year, “A pupil who confuses World War I with World War II knows something more than one who mistakes World War I for the American Revolution.” A test like the TCAP, of course, misses this difference.

Nor can it accommodate the children who know all the correct answers but who work too slowly to answer every question in the time allotted. It’s a typical mistake of test-champions, says Bostein, to confuse speed with competence: “Knowledge and understanding are not about rapid reflexes; learning is not a sport. Quickness of recall does not indicate depth of understanding.”

Indeed, depth of understanding—which must surely be one clear hallmark of an educated person—virtually cannot come from rigorous test preparation, which emphasizes product over process and speed over deliberation. In their study of two similarly affluent suburban Birmingham elementary schools, University of Alabama researchers Constance Kamii and Barbara Ann Lewis found that there was actually an inverse relationship between high standardized test scores and higher-order mathematical thinking. The class of second-graders who demonstrated the best grasp of mathematical principles had been taught according to a less conventional math program that does not emphasize test preparation, and their tests reflected it: Their aggregate score was a 79 on the state achievement test. By contrast, the corresponding percentiles of children in the traditional test-prep classrooms were 85 or above, though in interviews with the research team the children clearly struggled with a variety of mathematical concepts and applications—including place value, story problems, and estimation—that the other children had mastered.

These tests can perpetuate negative social stereotypes. According to social scientists, some kids experience a kind of test anxiety that is virtually a guarantee of lower scores. Educators aren’t wild about bringing this up, but there’s a pronounced racial divide in standardized test scores: White and Asian kids consistently get markedly higher scores than black and Hispanic children do on these tests.

According to a U.S. Department of Education report, the single greatest predictor of low achievement-test scores at a school is a large student population that qualifies for free lunch. So, obviously, a great deal of the racial disparity can be traced to the suppressing effects of poverty.

But not all of it can. In a recent issue of The New Yorker, writer Malcolm Gladwell reported that two Stanford psychologists, Claude Steel and Joshua Aronson, have discovered a testing trend they call “stereotype threat.” Steel and Aronson conducted studies which suggest that when people are given a test that taps into a pervasive negative stereotype about any group of which they are a member, they will perform at a much lower level than when the same test is presented merely as a laboratory experiment.

For example, the pair gave a group of Stanford undergraduates a standardized test that would, they said, measure the students’ intelligence. As testing statistics suggested they might, the white students performed much better than the black students did. But when the same test was given as merely a questionnaire, with no claim of measuring intellect, both black and white students performed equally well; their scores were virtually identical. The difference? Black students are more likely to choke when confronted with a test that pressures them to prove their intelligence in the context of a persistent stereotype that says whites are smarter than they are.

A similar experiment showed that women, when given a standardized test that purported to measure mathematical ability, performed much worse than similarly qualified men, but their performance was virtually identical with that of the men when no mention of quantitative ability was made in reference to the test.

It’s impossible to say whether grade-school-age children are already subject to stereotype threat. But it certainly makes poor test results among blacks and Hispanics—or low math scores among girls—a more complicated matter than a tighter regimen of homework and memory drills, or a policy of holding kids back a grade, is ever likely to address.

Sometimes people cheat. The more emphasis a test is given, and the more widely the test results are publicized—a thing that happens invariably in this political climate—the more likely it is that people will cheat on them. And it isn’t the kids themselves who are doing the cheating; it’s the adults. Recent cases include a Virginia teacher who showed eighth-grade students copies of the test in advance, a Maryland principal who gave kids the answers to the tests, 18 Texas school officials who altered test results in their district, and 43 New York teachers from 32 schools who used a range of cheating methods.

Probably the most dispiriting case of adult cheating happened in neighboring Alabama last spring: Huge numbers of poor-performing students—10 percent of the entire student body at one Birmingham high school, for example—were simply withdrawn from school roles and told not to come back until the new school year began in August. As a result, according to a Birmingham News analysis, 13 schools in the Birmingham area “earned” achievement scores that showed more than twice the level of improvement that experts say indicates a likelihood of cheating. Meanwhile, the kids who most needed what school has to offer were banished from it for months.

Everyone I talked with for this story agreed there’s been absolutely no indication of cheating on the TCAP. But testing in individual classrooms works largely on the honor system here, with little or no external monitoring. And if the stakes go up—as whoever becomes our next president wants them to—there may well be some teachers who see no reason not to “help out” just a little: They might help a child interpret a test question, say, or define a word in the question that the child does not understand. That may not be as egregious as altering test answers after they’re turned in, but it’s cheating nonetheless. And even good people may be tempted to justify it.

The test doesn’t match the curriculum. Though the “T” in TCAP stands for Tennessee, it is not a test written specifically to match the Tennessee state curriculum; the name it’s called by CTB/McGraw Hill, and by other states that use it, is TerraNova. It is what educators call an “off the shelf” test. But that doesn’t mean, says Dr. Bob Crouch, director of research and evaluation for Metro public schools, “that it does not align, to a high degree, with [our] curriculum.”

Some teachers disagree. Particularly in some subjects, says one teacher, “it’s very different” from Metro’s core curriculum. “It matches a lot of what’s in our textbooks, [but our textbooks] don’t match up with the core curriculum.” For example, there are entire six-weeks grading periods when the core-curriculum guide for science teachers never refers to the science book at all: “It’s just like, ‘Good luck; you’re on your own,’ ” the teacher complains.

What all this means is that while teachers must scramble to make sure that they cover everything in the Metro core curriculum, they must also make sure students cover the textbook thoroughly enough to do well on the TCAP. Adjusting a curriculum to cover information and skills that test-makers, themselves educators, believe should be standard is not in and of itself a terrible state of affairs. But it’s one thing to introduce a concept in March, before the TCAP is administered, that won’t be covered in depth until May. It’s another thing entirely when the very methods by which a teacher covers a subject are circumscribed by the test itself.

It is undeniable that the best ways to improve test scores quickly are drills and practices of the skills required by the test—sometimes at the expense of the kind of teaching that leads to more thorough understanding and to higher-order thinking. “Around January and February, I start getting really nervous,” one Metro reading teacher explains, “because I feel like I’m doing things that are ridiculous, just because everyone wants the kids to do well on the TCAP. And to me that’s just not what it’s all about.”

This teacher, who like other teachers asked not to be identified for fear of reprisal, went on to describe a specific example of how the TCAP can limit innovation in the classroom: TCAP tests vocabulary as it appears in the standard reading textbook (which, not surprisingly, is sold by the same company that sells these tests). Last year, this teacher assigned full-book reading, rather than merely using lessons from the reader, and she created her own vocabulary lessons based on those books. Teaching vocabulary in the context of real literature is, as language-arts teachers fully know, the best possible way to increase both comprehension and vocabulary. And the students in this teacher’s class, it turns out, did earn extremely high scores in reading comprehension. But because the vocabulary words they learned were not the same words found in the basal reader, their TCAP vocabulary scores were very low.

When she received her students’ TCAP scores, “I got so upset about it at first,” the teacher says, “but then I said, ‘To hell with that test.’ Because at the end of the year I had every single parent come tell me, ‘My child loves to read because of you.’ And that’s the reason I teach. To make kids love reading.”

The tests don’t teach kids anything. Any teacher knows that multiple-choice tests are never the best way to measure learning, but sometimes an overwhelming class load can make giving such tests the only realistic way for a teacher to manage. The difference between a multiple-choice history test and the TCAP, however, is that the history test can be a learning tool. In a classroom, a child who fares poorly on the test is able to see which questions she got right and which she got wrong, and her teacher is able to explain to her why the correct answer is better than the answer she chose. Off-the-shelf standardized tests can do neither. So preparing for them takes up teaching time without teaching much—or, rather, without teaching the most important things.

Rising scores don’t necessarily mean that students are learning more. These tests vary slightly from year to year, but according to Howard Gardner, a professor in the Harvard School of Education, as long as the format and type of test questions remain constant, scores will invariably rise each year as students and teachers become accustomed to the test. When a new test is introduced—when a state shifts from TerraNova to the Stanford Achievement Test, for instance—scores across the board will drop precipitously before beginning a steady climb again.

Finally, even very high scores can mean problems for a school. The TCAP scores reflect not only how well a particular student population has performed on that particular test, but also how much the school as a whole has gained in scores since the previous year. This second set of data can be reassuring to officials in schools with a large population of impoverished children: Though they may not be performing as well as more affluent children, their net gains can sometimes be higher, which suggests that they are learning—if only that that they are learning how to take a test.

But this is a potential problem for teachers and students in the already high-performing schools. Goal No. 1 of the Metropolitan Nashville Public Schools Accountability Framework, amended less than a year ago, is to “Raise expectations and performance standards and provide for all students to demonstrate mastery of continually higher expectations and monitor student performance and make appropriate adjustments of methods to achieve established student outcomes.”

What this garbled statement means is that no matter how well a school is doing, as measured by the TCAP, it needs to do even better the next year. And the most efficient way to get from 90-percent mastery to 95-percent mastery is—you got it—drill, drill, drill. So even doing well on these tests doesn’t mean that teachers have the luxury of trying out creative teaching methods, of enlarging the curriculum with in-depth multidisciplinary explorations—not even in what are acknowledged to be the very best public schools.

Many of these objections to standardized testing come from classroom teachers, but it’s only fair to note that there are plenty of Metro teachers who don’t object to the TCAP at all, either because they share the state’s testing philosophy or because their own students score very well on the tests. Some teachers, one educator explains, “are test crazy. They love the tests, and they tell everyone their students’ scores.”

And there are teachers, too, who don’t feel any particular pressure, good or bad, from the tests. “I don’t think that preparing for the test makes that much of a difference in test scores,” admits one teacher, who has taught in both urban and suburban Metro schools. “When you look at test scores, you’re not really looking at an actual reflection of teaching. You’re looking at demographics: You’re looking at the socioeconomic level of the students, and you’re looking at the educational level of the parents, especially the mothers. As much as some teachers might like to think they are, they’re not having that much of an effect on student test scores.”

Metro school board member Dave Shearon disagrees, citing TCAP data which suggest that hard work by the “most effective teachers,” even of impoverished children, can raise test scores considerably. Shearon understands that demographics play a role in test results: “Of course Williamson County has better scores [than Metro],” he says. “You can go look at the cars parked in the driveways of the houses that feed kids into that school, and you can tell where on the scale of scores they’re going to be. That’s true. But that’s not the issue.”

The issue, says Shearon, is whether a school system, through 13 years of concerted instruction, can even up the scores a bit. And he says it can—if the instruction is excellent. Dramatically improved test scores, he insists, are “an effect that is tied to the teacher.” Which places the testing ball right back in the teacher’s court.

The good news for Nashville parents and kids is that, for all the flaws inherent in standardized testing, and for all the emphasis the TCAP tends to get in our schools, this test is less problematic than some of the other achievement tests used around the country.

Unlike the Texas test that Governor Bush is holding out as a fine national example of school reform, for instance, TCAP is not a criterion-referenced test—that is, it is not a test based exclusively on grade-level-specific information that kids must know before they can move on to the next grade. The TerraNova does require the ability to recall certain facts, but it is also what teachers call a “thinking test”: It tests how well a student is able to apply information that the test actually supplies.

In addition, as Shearon points out, while the TCAP shows the student’s performance in reference to a national norm, Tennessee, “unique among all states..., uses a value-added analysis on top of that, so what the test actually becomes is a system where each student is his or her own reference point.” Rather than looking merely at what kids know in reference to a particular curriculum or national norm, Shearon says, “We’re looking for gains every year.” If you don’t believe a standardized test can accurately measure how much a particular child has learned or improved over the course of a school year, the difference between a criterion-referenced test and the TerraNova may not seem all that great. But if there has to be a test at all, the TerraNova is definitely the lesser of two evils.

The other good news is that TCAP is not used as a high-stakes assessment tool. In Metro, the decision to recommend summer-school and, subsequently, whether to repeat a grade, is based on three factors: TCAP scores, exit-skills assessment by the individual teacher, and minimum attendance. To fail a grade in Nashville, a kid has to fail in two out of three of these categories. Poor TCAP scores alone won’t sink anyone.

That’s the good news. The bad news is that Metro is completely hamstrung in what it is able to do for the kids who are at risk of being failed by their education.

Critics of standardized testing tend to focus on one of two things: either the problematic high-stakes use of scores, or the limitations of these tests in evaluating basically competent students. But there is a way that standardized test scores can be of real value to a school system: They can point out students who are not even remotely able to do the academic work expected of children their age; they can identify the schools that need massive help. The objection to testing that Bostein makes on behalf of the child who incorrectly chooses World War I when the answer is World War II, for example, is utterly inapplicable to an eighth-grader who cannot read the question at all. In such cases, test results can sound a loud alarm that somebody—a teacher, a principal, someone—ought to heed.

As a major urban school district, Metro has more than a few children whose readiness to learn on entering kindergarten is basically nonexistent, and whose subsequent academic performance is dismally low. As a rule, abysmally poor scores are going to be clustered in schools that serve abysmally poor people. Which ought to tell anybody that fairly significant school-wide—even system-wide—adjustments need to happen as a result.

This doesn’t mean that teachers there should do even more drilling or more practice tests than other schools do. It means that the city of Nashville needs to ante up with enough money to pay for the kinds of programs that have been shown to help.

Telling the teachers in those schools to work harder, or the kids to study more, is useless. Test coaching can raise scores to a certain point, but as we’ve seen, there’s no necessary correlation between rising scores and increased learning in basically competent students. For the children who do stand to benefit by concerted intervention, though—for the kids who functionally can’t read, or who can’t perform basic mathematical computations—the best the schools can do with TCAP data, even if it comes on time, is to recommend summer school. When the scores come in late, the schools can’t do even that.

There are policies and programs that offer significant help to such children, but they cost a lot more money than standardized testing does. If Nashvillians are serious about raising standards in the Metro schools, they’re going to have to confront some serious funding needs. This doesn’t mean giving a bunch of unmarked money to the school system and saying, “Here’s some more money; spend it any way you like.” Metro’s needs are so great right now that it would be easy for all additional funds to go straight toward updating textbooks, improving crumbling buildings, and the like. These are legitimate, urgent needs, of course, but if what we want as a city is for all of our children to reach basic reading, writing, and mathematical competence—and for both humanitarian and very pragmatic reasons we absolutely should want it—we need some other things first.

We need to offer free preschool—at least at age 4, but preferably at age 3 too—to children zoned for the schools most at risk of failing. And we need to put those kids in very small classes. Last summer the Rand Corporation, a policy and research group out of Santa Monica, published a book-length, state-by-state study of standardized test results and the various methods of intervention states had used to raise scores. By far the greatest predictors of improved achievement, especially among impoverished children, was access to quality preschools and small class size. (Only the Gore education agenda, by the way, makes universal, federally funded preschools a goal—one of the few differences between Gore and Bush on education.)

We need to offer a transition grade to children who have completed kindergarten but who are not yet ready to learn to read. According to former Board of Education member Deirdre Macnab, Metro used to offer a program called a Transition 1 (or T-1) class that such kids took between kindergarten and first grade, but funding shortages killed it. “Developmental experts say that having a T-1 class is one of the most important components of a successful reading program,” Macnab says, “and we don’t have it anymore.”

We need high-quality after-school programs that keep grade-school kids at school until an adult comes home. I’m not suggesting that what poor kids need is more hours of test-preparation drilling. I’m saying that providing a safe place for them to play and to do their homework, ideally with a loving adult at hand to help and encourage, can go a long way toward instilling the desire to learn and do well in school.

We need to consider a year-round school calendar. During the summer, kids forget a lot of what they learned the previous year. A 1996 study conducted by the University of Missouri compared the findings of 13 other post-summer studies and concluded that students in all grades, and at all economic levels, lose an average of one month of skills during the summer. The study found that losses in math are pronounced for all children, but poor kids lose considerable ground in reading comprehension as well. A school calendar like one used in the W.A. Wright Elementary School in Mt. Juliet, which offers a nine-week-on/three-weeks-off program (with tutoring available during the break), has been shown to reduce this slide, giving teachers the chance to use more time teaching each year and less time reviewing.

In an election year it’s easy for politicians to say they want to raise standards and hold schools accountable for what kids learn. But if what they really mean by “high standards” and “accountability” is the inspiration and ability to succeed—for kids actually to learn, that is, and to develop the habits of mind that will keep them learning for the rest of their lives—then working to raise test scores isn’t going to do much. Like my newborn son’s perfect 10 on the APGAR test, improved TCAP scores will mean almost nothing about how smart our city’s kids are, or even how well their schools are serving them.

The other good news is that TCAP is not used as a high-stakes assessment tool. In Metro, the decision to recommend summer-school and, subsequently, whether to repeat a grade, is based on three factors: TCAP scores, exit-skills assessment by the individual teacher, and minimum attendance. To fail a grade in Nashville, a kid has to fail in two out of three of these categories. Poor TCAP scores alone won’t sink anyone.

That’s the good news. The bad news is that Metro is completely hamstrung in what it is able to do for the kids who are at risk of being failed by their education.

Critics of standardized testing tend to focus on one of two things: either the problematic high-stakes use of scores, or the limitations of these tests in evaluating basically competent students. But there is a way that standardized test scores can be of real value to a school system: They can point out students who are not even remotely able to do the academic work expected of children their age; they can identify the schools that need massive help. The objection to testing that Bostein makes on behalf of the child who incorrectly chooses World War I when the answer is World War II, for example, is utterly inapplicable to an eighth-grader who cannot read the question at all. In such cases, test results can sound a loud alarm that somebody—a teacher, a principal, someone—ought to heed.

As a major urban school district, Metro has more than a few children whose readiness to learn on entering kindergarten is basically nonexistent, and whose subsequent academic performance is dismally low. As a rule, abysmally poor scores are going to be clustered in schools that serve abysmally poor people. Which ought to tell anybody that fairly significant school-wide—even system-wide—adjustments need to happen as a result.

This doesn’t mean that teachers there should do even more drilling or more practice tests than other schools do. It means that the city of Nashville needs to ante up with enough money to pay for the kinds of programs that have been shown to help.

Telling the teachers in those schools to work harder, or the kids to study more, is useless. Test coaching can raise scores to a certain point, but as we’ve seen, there’s no necessary correlation between rising scores and increased learning in basically competent students. For the children who do stand to benefit by concerted intervention, though—for the kids who functionally can’t read, or who can’t perform basic mathematical computations—the best the schools can do with TCAP data, even if it comes on time, is to recommend summer school. When the scores come in late, the schools can’t do even that.

There are policies and programs that offer significant help to such children, but they cost a lot more money than standardized testing does. If Nashvillians are serious about raising standards in the Metro schools, they’re going to have to confront some serious funding needs. This doesn’t mean giving a bunch of unmarked money to the school system and saying, “Here’s some more money; spend it any way you like.” Metro’s needs are so great right now that it would be easy for all additional funds to go straight toward updating textbooks, improving crumbling buildings, and the like. These are legitimate, urgent needs, of course, but if what we want as a city is for all of our children to reach basic reading, writing, and mathematical competence—and for both humanitarian and very pragmatic reasons we absolutely should want it—we need some other things first.

We need to offer free preschool—at least at age 4, but preferably at age 3 too—to children zoned for the schools most at risk of failing. And we need to put those kids in very small classes. Last summer the Rand Corporation, a policy and research group out of Santa Monica, published a book-length, state-by-state study of standardized test results and the various methods of intervention states had used to raise scores. By far the greatest predictors of improved achievement, especially among impoverished children, was access to quality preschools and small class size. (Only the Gore education agenda, by the way, makes universal, federally funded preschools a goal—one of the few differences between Gore and Bush on education.)

We need to offer a transition grade to children who have completed kindergarten but who are not yet ready to learn to read. According to former Board of Education member Deirdre Macnab, Metro used to offer a program called a Transition 1 (or T-1) class that such kids took between kindergarten and first grade, but funding shortages killed it. “Developmental experts say that having a T-1 class is one of the most important components of a successful reading program,” Macnab says, “and we don’t have it anymore.”

We need high-quality after-school programs that keep grade-school kids at school until an adult comes home. I’m not suggesting that what poor kids need is more hours of test-preparation drilling. I’m saying that providing a safe place for them to play and to do their homework, ideally with a loving adult at hand to help and encourage, can go a long way toward instilling the desire to learn and do well in school.

We need to consider a year-round school calendar. During the summer, kids forget a lot of what they learned the previous year. A 1996 study conducted by the University of Missouri compared the findings of 13 other post-summer studies and concluded that students in all grades, and at all economic levels, lose an average of one month of skills during the summer. The study found that losses in math are pronounced for all children, but poor kids lose considerable ground in reading comprehension as well. A school calendar like one used in the W.A. Wright Elementary School in Mt. Juliet, which offers a nine-week-on/three-weeks-off program (with tutoring available during the break), has been shown to reduce this slide, giving teachers the chance to use more time teaching each year and less time reviewing.

In an election year it’s easy for politicians to say they want to raise standards and hold schools accountable for what kids learn. But if what they really mean by “high standards” and “accountability” is the inspiration and ability to succeed—for kids actually to learn, that is, and to develop the habits of mind that will keep them learning for the rest of their lives—then working to raise test scores isn’t going to do much. Like my newborn son’s perfect 10 on the APGAR test, improved TCAP scores will mean almost nothing about how smart our city’s kids are, or even how well their schools are serving them.