Matt, my wife is a teacher in the Philadelphia school district.  And "teaching to the test" in Philly means a little more than it does when you bring it up here.  Teachers and faculty actively help the children cheat on the various evaluations the students are given.  My wife is regularly forced into situations where she feels obligated to report this abuse.  Problem is, so many people are in on it nothing is ever followed up.

 Admittedly, better outside enforcement would obviate this problem, but those kind of enforcement resources don't appear to be provided under NCLB.  Oy.

Since I was a bit harsh in my criticism of the earlier thread, I think I should thank Matt for responding more fully and not focusing just on the federal/state issue, as I accused him of doing before.

 Some of the responses Matt makes are at least moderately convincing.  On the other hand, I'm still not sure about the "teaching to the test" business.  I too took various AP (and IB) classes which, I suppose, could be said to be "teaching to the test."  But the basic fact is that the kind of basic proficiency tests that NCLB requires simply cannot be anything like AP tests.  Any test which is going to be made a standard like this is almost certain to be bad.

 I remember the Maryland Functional tests we had to take.  The teaching for the Functional Writing test was the most incredibly stupid thing I've ever come across.  We were taught that we had to write the most absurdly rigid five paragraph essays that you could possibly imagine.  I would guess that this is going to happen a lot.

1. Designing an appropriate test is not quite as simple as you imply when you dismiss as obvious that you shouldn't teach to a bad test. For instance, no achievement test that I'm aware of measures what it's supposed to measure with better than about 10-20% validity (r=.35 or so). No test perfectly measures achievement - not even close. Testing with a good test, let's say a test with a validity of 15%, is certainly better than a poor test, or nothing at all, but there are inherent and unavoidable limitations.

2. "Teaching to the test" sometimes means teaching the skills necessary to pass a good test, in which case it's a good thing. But sometimes it means teaching the specific items that will appear on a specific test. That is obviously not an appropriate way to learn, because the test is really just a sample of a broader set of items, all of which the student is presumed to know. The obvious solution to this is to make sure that teachers don't know the specific test items, but in practice that's not always the case.

3. Generally speaking, it's better to encourage a process orientation rather than an outcome orientation to learning. IMO, we have too much of an outcome orientation in our culture, and it results in an "anything goes" attitude, including rampant cheating.

For what it's worth, I think testing is a good idea, but we shouldn't underestimate how difficult it is to achieve and we shouldn't overestimate its benefits.

Testing is not bad in and of itself.  The problems are twofold.  One, you mentioned AP Physics, which, unless you're some kind of serious genius (which you may well be, for all I know), you took that class in your late high school years.  By then, you have (hopefully) had years of exposure to the critical thinking and problem solving that makes that kind of teaching to the test more than rote memorization.

I learned a lot while studying for my AP Chem exam too, but it's not the same for 5, 6, 7 year olds.

The second problem is how much of the time is spent teaching to the test.  When everything depends on every child passing these tests, that's all the teachers do.  There's little time for creativity or variation or free thought.  Nevermind time to stop and help the kids who aren't getting it as fast as the testing schedule demands.

Which brings me to another point, my mother's specialty: developmental age.  Not all 6-year olds are at the same level of development, but that's exactly what over-standardization forces on them.  Kids must be allowed some latitude to grow at their own pace, especially during these critical early years.

In looking over my posts, I notice that I refer to my family perhaps too often, so forgive me, but this is best example I can offer.  My sister and I learn very differently, always have.  I took to reading from a very early age and was reading full children's novels before I started Kindergarten.  My sister, four years older but raised by the same parents, struggled with reading and didn't really learn until one of her teachers finally noticed in the third grade.

One last thing has to do with retention.  I hear from almost all the teachers I know that many of the kids that are passing the test aren't retaining anything they learned for it a week or two weeks later.  I don't remember the molecular structure of sugar or any of the other hydrocarbons I had to memorize for my AP Chem exam, but if I picked up a Chemsitry book now or read an article in Science, I would understand it.  My sister comes back after winter break and often has to practically re-teach half the last semester because nothing stuck.

Okay, I lied, but this is really the last thing.  I agree with you that testing can be a powerful tool to highlight deficiences and would then hopefully be used to bring extra support to those areas.  And accountability is certainly important - gym teachers shouldn't be teaching Algebra just for the extra money.  But NCLB as structured doesn't accomplish either of these goals and while it may be boosting test scores, I don't believe it's contributing to better minds in any sense.

We want children to be taught knowledge (both factual and operational) about a subject area. A test is at best a proxy for that knowledge. "Teaching to the test" makes the proxy the goal and appart from possibly confusing parents, kids and teachers as to the real goal, it can have a series of unfortunate consequences.

The incentive to cheat point has already been mentioned (see also Freakonomics by Levitt and Dubner for an example of how widespread the problem is) but I actually find the impact on education more troubling.

Tests are by their very nature constrained in a number of ways: they are a one-time sample (>fitness and motivation on that day, circadian rhythm) , wherein in a particular situation (group, seated) a limited number of response options (verbal, often selection from options) on a series of discrete problems is used to form an estimate. Presumably the average score of a school is free of these problems (provided we haven't forgotten to sacrifice several ducks to the deity of the uncorrelated error), but if you happen to be an indivual who performs best when applying and combining general knowledge from several areas to a new problem in the seclusion of your own home during the night through sketches and numbers, your test results will not reflect your knowledge particularly well. Much the same goes for people who have skills not reflected in traditional curicula (e.g. social ones) or pretty much anyone who isn't going to work behind a desk doing paperwork.

"Teaching to the test" invariably involves teaching test taking skills irrelevant to the actual knowledge, unless of course you believe "test taking" to be a valuable skill in itself.

Moreover, tests force you to break down the body of knowledge in particular ways, which usually ends up favouring short-term rote learning over long-term understanding. It also means focussing on matters which can be broken down into neat, independent components, to the detriment of the considerations of complex systems with multiple dependencies. Similarly, if favours broad and shallow knowledge over in-depth knowledge of a limited field.

Now, all these problems can be (and are) addressed, but if the main incentive is getting good test results, you can be fairly certain, that improvements which might make the test a better estimate of knowledge but yield worse overall results will not even be considered. Which is just another way of saying that replacing the real goal with a proxy goal is a bad idea. What kids should learn and checking on whether they actually did learn that should be done without limiting ourselves to the subset that is readily testable.

To the extent that test are a necessary evil and to the extent that they actually can measure a particular field accurately, we should of course use them, but just as we are not content to limit our effort to fight poverty to people below a certain income (the proxy), we shouldn't make tests the goal of education.

It's clear that it is standard pedagogical practice is to teach some material and then test students on what they learned. Matt defends this standard if imperfect practice, and claims he is defending "teaching to the test."

You know what? Maybe he's right. Nobody's ever fully explained to me what this phrase means. For all I know, it could just be what everyone does already.

Let me think back to grade school and high school.  What I recall is that the test covered a subset of the material taught. In theory, you were expected to learn everything, but in practice, you might be able to guess which parts were important. Barring that you raise your hand and whine "will this be on the exam?" although that was a little more prevalent in college than earlier.

Depending on the subject, you might learn a general principle to apply--say solving quadratic equations by "completing the square." or you might literally be tested on a subset of facts you were expected to know. In some classes, there could be essay questions, and there really weren't any stated guidelines; somehow the teacher would just figure out if what you wrote indicated that you had grasped the textbook and class discussions.

Now if that is "teaching to the test" I agree with Matt. But I'm pretty sure "teaching to the test" means something else. In conventional pedagogy, the teacher makes up a curriculum and then comes up with a test intended to determine how well the student learned it. I've done this myself as a TA.  You usually know you'll leave something out, but the student doesn't know what, so a student who does very well probably learned the untested parts as well. I would call this "testing what you taught."

My impression of teaching to the test is that you as a teacher see the test or a very typical sample test ahead of time. You then construct a curriculum in such a way that students who follow that curriculum will maximize scores on the test. In some cases--say a spelling test--the results might be identical to testing what you taught. In other cases, any resemblance is inadvertent.

In theory, there might be tests that are so well-conceived that the process of teaching to these tests results in the construction of a very solid curriculum. I really don't know. But the process strikes me as ass-backwards. If we had good teachers and could count on them to devise a good curriculum first of all, we would have a direct rather than indirect way of getting students to learn a subject area.

Now some standardized tests (the SAT) test material needed for some many curricula that it is reasonable to assume that anyone who has been educated through 12th grade has been exposed to it. True, people take Kaplan courses to enhance their scores, but that was never the intent. Simply being well-read tended to be enough to get a good score on the old multiple choice SAT verbal for instance. I would call teaching to the test a failure of the testing process rather than a reasonable approach to teaching. 

I don't know any educational theory, so I can't reply to specifics. I think that it would certainly be reasonable to have a standardized set of subject goals--almost like a requirements document in software engineering--and then create both a teaching curriculum and a test from such a document. What bothers me is the idea of the test serving double duty as the learning goals, since tests are inevitably incomplete and not intended for that purpose.

It's not even that unreasonable to include material in the curriculum that you have not figured out how to test, since you have some probability of imparting the knowledge whether you can measure it or not.

To expand on the (maybe ill-suited) software engineering example: Backwards design sounds something like test driven development, in which you first design a test and then implement code intended to pass it. But both test and implementation are based on the requirements and design documents.

By contrast, "teaching to the test" sounds more like scamming the QA department. You find out what they're capable of testing in the software, and then you make it just good enough for them to let it through. If you have the world's best QA department, that might just be enough. But otherwise it seems like a crazy approach. 

As the parent of a child in a "failing" public school I have a little experience with this subject.

Firstly, the concept of teaching to the test is misguided when the focus is on test taking itself. Teaching a child how to do process of elimination and educated guessing is not particularly helpful when you are trying to teach long division.

Secondly, the whole concept of "high stakes testing" is completely insane. The Tacoma public Schools actually had to mandate elementary school recess this past year because principals across the city had cancelled it to create more class time for test prep. Here in Massachusetts, my child's school has after-school programs that focus purely on raising the math scores on the MCAS exams. You have 7 year olds that are being forced to do three hours of math homework in a detention-like setting every day. And guess what? The scores have not improved in two years. Oops. Wrong answer.

This leads me to my final point- the various college admissions tests. I have been studying for the LSAT which I will be taking on Monday. A while back, when I took my first first practice test I did fairly well with no preparation whatsoever. After a little tutoring on the games section, I have increased my practice scores dramatically into the highly competitive range. As an older, nontraditional law school applicant, my LSAT score will be weighted unusually heavily. Please tell me how in the world my new found ability to quickly diagram a twitty question about '6 students sharing 5 lockers' will make me a better lawyer. Is a high school student who spends $1000 bucks on a Kaplan SAT course actually better prepared for college than a kid who has to flip burgers to help his family pay rent?

I understand your point that *sometimes* teaching to the test is an efficient way to master certain subjects, but in this NCLB era, it is a lousy rule of thumb. 

The 100% passing requirement is "probably" too stringent? Good grief! In the law as written, all children in the eighth grade (for example) will pass the state achievement test by 2017 or the school fails. All means all; including students with less than 80 IQs. This provision virtually ensures that nearly every school will be failing by 2017.

"Probably" is a pretty weak protest for someone who purports to exist in the realty-based community.  The 100% requirement is outrageously unfair.
 

Testing to standards that are measured by tests aligned to those standards is fine. Rote memorization and drill will create many dropouts, such as those currently in Texas.

Even if we instituted merit pay, completely equalized funding passed out unlimited vouchers, and made it easier to fire teachers there would still be vast achievement gaps in our schools.

Many kids have parents that can't or won't support their education in anywhere near the fashion that their suburban peers recieve. There has been research demonstrating that children born in poverty are exposed to less than half the vocabulary in their homes than their middle class peers. Half!  Think of what an obstacle that is. Teachers in urban schools live with this daily. That is why there is a high attrition rate in the first five years of service; who wants to be beaten over the head for things they can't control and get paid $35000?

Equalizing funding would help. Small class size can be effective with competent teachers.  A national curriculum is necessary so texts and effective teaching practices can be standardized.  Assessment should include a value-added component so that schools can be measured by how much progress is made.  In this scheme the poorest schools can be measured along side the wealthiest.  But there is no panacea and NCLB, in its current shape, will not achieve its stated goals.

I'm a sophomore at college currently, and I can tell you that throughout my years in elementary, middle, and high school, "teaching to the test" generally meant that teachers would spend an inordinate amount of time trying to make sure everyone learned the material to a minimum standard. Unfortunately, there were always a few kids in the classes who just didn't care and were truly determined to fail (since I went to the worst inner city public schools in the area, there were always people in the classes with that mindset regardless of what level the class was supposed to be at). So the end result was that even in advanced classes the teachers spent a lot of precious time trying to make everyone meet a minimum standard while leaving the competent kids to do whatever they please.

 The problem then is that the advanced students are getting woefully inadequate schooling so that they will go on to college less prepared than they should be, just because in the short term it's to the benefit of the teachers to focus on making sure everyone reaches a minimum standard than making sure that opportunities for learning are distributed uniformly across the classroom.

That situation is a common one as far as a I know, and is one that I think is necessarily induced by required standardized testing when the students can't be put into classes with their peers. That is precisely the reason that I feel obligated to oppose NCLB, because its methods for trying to help everyone are not guaranteed to ultimately help the lower level students really and has a good probability of just making it harder for good students to learn more.

The 79th Texas legislature closed on Monday. Figuring out how to fund public schools was supposed to be their top priority. They are under a drop-dead order to come up with a new system of funding to replace the old one, which has been deemed unconstitutional. If nothing is done by October of this year, all funding for public schools will cease. For the third year in a row, this Republican controlled legislature in this Republican controlled state failed to reach agreement. Every person in the state realizes that there will have to be some changes to the tax code, but none of the politicians is willing to stick his / her neck out to do anything about it. So, the schoolchildren of Texas are up a creek without a paddle unless Gov. Perry calls a special session or the State Supreme Court steps in. These legislators are the same ones who wasted weeks trying to pass unconstitutional bills outlawing gay foster parents and prohibiting suggestive high school cheerleading routines. And, of course, they are the legislators who are ultimately in charge of our public school accountability system.

As a parent and a teacher, I would describe this pitiful state of affairs in the language of school accountability testing. The 79th Do Nothing Texas Legislature is a low-performing legislature. They failed to meet minimally-acceptable standards. After three years of continued low-performance, they have triggered interventions and now outside management should be brought in.

Fiddling with the public schools is every politician's favorite hobby. They can do it with impunity. They are the only people in the system who, in practical terms, are accountable to no one. Most of these clowns are locked into office for life and run unopposed. If they are voted out of office, it usually isn't over their votes on public education.

The comparison to software development is not  bad one. In software design you have a robust set of requirements set by the appropriate stakeholders. At the same time you plan certain tests to see if the programs you develop meet these expectations. In the educational context these 'tests' may be traditional sit down exams, but would realistically include many types of assessment 'authentic' assessment as educators might say.

However, if you continue to follow the software design analogy you find that even in software design, a process which one would assume (because it involved computers) is rational and orderly is quite often anything but. I my experience software development is frequently far more fruitful when it is approached in a constructivist, 'feedback' loop model where specifications are continually reassessed as design moves forward. However, such a process is only possible when the people involved are quality participants that are actively involved in the process.

 The same is true for education. It is all well and good to talk about 'authentic' learning, and genuine understanding; and teaching each student to his or her own ability level rather than to a 'standard'... but schools must face important realities; 1. Not all teachers are as capable as others and 2. Not all students can be assumed to actively involved in the process of learning.

I guess that I did not sign in properly.&nbsp;<div><br /></div><div>This (&quot;<span class="Apple-style-span">Teaching to the test reduces schooling to that which is testable&quot;) was by me.</span></div>

There is a construct in educational psychology called Bloom's Taxonomy that defines six levels of cognition that people progress through as they learn information (rote knowledge, comprehension, application, analysis, synthesis, and evaluation).  "Teaching to the test" would not be problematic if the tests spanned all levels of Bloom's Taxonomy; however, studies have shown that 95% of standardized test questions involve only the first two rungs of the heirarchy: rote knowledge and comprehension.  There is a reason that AP and IB tests are frequently used as "good" examples of teaching to the tests -- they are the only tests in wide use that require a degree of analysis or synthesis. 

The reason that most tests measure only knowledge and comprehension is that tests become much more difficult and expensive to produce the higher the skill being measured is on Bloom's Taxonomy.  It is very difficult to control for the variables that make a test scientifically sound (reliability, validity, and accuracy) when you are measuring abstract skills.

With all of the money being thrown around by NCLB, one area that has been neglected is improving the science of testing.  We need to improve what can be accurately measured by tests so that "teaching to the test" will no longer be a euphamism for rote indoctrination.  Also, if these tests are going to be used for high-stakes rewards or penalties, we need to ensure that they are scientifically sound. 

there is a teaching shortage:

1) It is hard job. Try getting up at 5 or 6 every morning, and then imagine going into a lower-income underperforming school. Where you have to play policeman before you can even start trying to impart knowledge

2) This problem is compounded because teacher pay is not very good, maybe 30,000 for those just beginning, 70,000 for those at the highest end: at least thats what it is Washington state. Its not awful, but its not good enough for me - and I suspect many others - especially considering things like those mentioned in point 1.

3) Thirdly, I don't know if this is residual sexism or not, but teaching is generally considered as being as prestigious. This isn't help by a lot of the anti-public school/anti-teacher rhetoric that comes out of the kind TNR-style quasi-liberal circles. If you want people to support reform, at least first show some sympathy for people who are for the most part trying.

Ben P

As a 5th grade teacher in a low-income area, I have a couple of observations. First, I think it is important to make a distinction between standards-based teaching and high stakes testing. While I agree that high stakes testing has many problems, having a standardized list of learning goals for every grade is necessary. As a previous poster noted, without these standards, teachers frequently would spend a great deal of time on relatively minor points and often fail to address key areas of the curriculum. Furthermore, as a new teacher (only last year) having a set of state standards served as an invaluable map for where I wanted me students to be at the end of the year. If I knew where I wanted my students to be, I could design a plan to get them there. Unfortunately, that is not always the case. Even with state standards, testing based on these standards, and a pretty regimented administration, it is amazing how some teachers in my building spend their instructional time. I can only imagine how bad it must have been when teachers had complete curricular freedom.

On to the effects of state testing. I have a long, long list of problems that these tests cause, but I'll just post the "best" one for now. In California where I teach, schools get scored from 0-1000 based on their test results. My district, one of the worst in the state, decided that, instead of focusing on teaching and learning to improve our scores, they would take a short cut. Students can score in one of five performance levels (far below basic, below basic, basic, proficient, advanced) on their tests.  Schools earn bonus points when students move from one performance level to the next during a year. So if a student is basic in 4th grade and then moves to proficient in 5th grade, the schools gets extra points. Seems like a good idea to reward students making progress, but my district took it way too far. At the begining of the year, we identify the "bubble" students who are near the high end of a performance level and thus not far from making the jump to the next level. We are supposed to track these students carefully and monitor their progress. Every two weeks we meet with other teachers to compare notes on how well these "bubble" students are doing. At some schools in the district, teachers are told to give these students extra help, have them sit in the front of the room, and call on them first for answers. Obviously, this is a zero sum game as teachers don't have an infinite amount of time or energy, so students who are not on the bubble would recieve less attention. The most insidious part about this is that "bubble" does not mean that the student is most in need of help. There could be a student who is  so far below basic that they would not be deemed worthy of the attention, while a student who is nearly advanced and one of the highest students in the class would get extra help. Not the way to design effective instruction. The program, at least last year, was unsuccsesful. Shockingly, while the bubble students' scores went up, many other students' went down. I know, it's hard to imagine that students who recieve less attention and help will not do well, but that is sadly the state of many district officials who are trying to navigate NCLB.   

that my problem - and is a position I have adopted by dating a public school teacher at a "problem" school, and being friends with or family to at least a dozen or so people who are either public school teachers or educational researchers, none of whom I know have a positive word to say about NCLB, and who represents a fairly wide swath of political opinion - is the way in which NCLB places standardized testing at the center of public school teaching by determining funding decisions based on the results. As I outlined here, I do think "teaching to the test" becomes a problem in poor schools. And I am also very dubious of whether studying for standardized tests relates in any significant way to learning the kind of skills that are going to help a student in college or in a career. I don't think the AP test example is a good one for reasons mentioned above - ie nature of the students taking the AP test/studying for the AP test, but also the kind of things the AP test expects you to learn, the kind of skills the AP test actually "tests." I think a better comparison would be studying for the GRE.

I think standardized testing is fine as a kind of baseline/skills proficiency exercise that can help you determine the kind of ballpark your students are in. This, however, is not the motivation or intellectual rational behind NCLB. As far as I understand it, NCLB - and people like Bush, who I think knows next to nothing about education - style tests become central to determining how much students have learned in a given year, suggesting that a standardized test - which I assume is a kind of multiple choice affair almost always - is the best way of measuring learning. Basically, people that still think rote memorization curriculum, particularly in poorer, less well-performing schools - which this kind of testing philosophy has the danger of causing - is seen as a fringe theory, analogous to say, Janice Rogers Browns' thinking is to contemporary constitutional scholarship. Ie in priveleges learning a series of facts - which will be surely forgotten - at the expense of a process of thinking through problems, solving problems, skills which will be applicable even when the subject isn't, say, geometry or modern American literature (which of course, for almost all students, it surely won't be)

Its surprising to me that so many people don't know, or cannot visualize, what is meant by 'teaching to the test', or how that works. As a parent of a middle schooler in Houston that has been involved in the Texas version of this program from the start I'll explain exactly what it looks like.

What happens is the teachers use last year's TAKs test to prepare the kids for this year's test. For example, a 4th grade math section of the test includes simple exercises in adding coins. The test has a picture of 2 quarters, 1 nickle, 2 dimes, and 3 pennies, and the student is required to determine what that adds up to in dollars and cents. There is no 'rote memorization' of answers, because the actual test will not have an identical question. That's a myth. Further, if children are doing this work and the homework, and they end up able to calculate the value of coinage...isn't that exactly what we want to accomplish? Of course it is. "Teaching to the test" merely prepares kids for the look and feel of the test. They aren't 'memorizing' the test so they can cheat it, they are learning the material.

At the end of the year we get a report from the school that shows exactly how our child scored in each subject and broken down into individual sections. From that it is fairly easy to see what your child's weak areas may be. The report also gives global scores for the school and the district, and breaks them down by race, etc. It is a very informative report and quite useful if you are interested in what's going on in your local schools.

The teacher' unions don't like it. Meh, tough. Nobody likes to have their performance measured like this. But they aren't the only ones reasponsible for the results by any stretch and my sense of it is if they'd use the results to lobby for the reasources they need it would be more productive for them. At a local level the teachers I've talked to don't hate the testing. Some like it, some don't like it that much. I've talked to quite a few about it though and nobody is marching in the streets over it.

NCLB is a good law. It sets a foundation of measurable achievement for the schools in the country. There are other things that can be done but this was a much needed reform.

I was one of those bored kids in grade school, and my parents fortunately had the resources to take me out of public school, where I wasn't being challenged, and put me in a private Montessori school after 4th grade.

The structure of what you're talking about was similar to what we had there: 4th-6th graders were all together so that it didn't matter what age or grade you technically were.  If you were particularly quick with math or science, then you learned those subjects in a different section.  But there was no stigma attached to being in one section or another.

The other truly beneficial thing was the non-traditional teaching methods they used.  Math for us was a combination of algebra, geometry, even some basic trig, and it was taught with hands-on materials, not just out of books.  Science was a little bio, a little chem, a little physics.

This is not an ideal learning environment for every child - some kids do need more structure in order to focus, but wouldn't be nice if this was an option for kids who would benefit from it?

I also couldn't agree more with your point about stigma.  Study after study shows that it's not about biological age at all.  The fact that some kids can master trig as 13-14 year olds doesn't mean that the ones who don't get it until they're 17 are inferior.  It merely means that the kids have different brains that are good at different things.  Wouldn't be great if our educational system recognized that?

Matthew -- for those of us who live in the real world...  I have three kids in high school -- their school is one of four in the district.  Each high school must muster 95% compliance in taking the NCLB test.  The system does not adjust for drop outs; one high school has so many kids in jail that they don't reach 95% compliance -- so the whole district is on probation.  For a third of the students at my kids' school, English is a second language.  Those who don't achieve `proficient' or higher on this year's English test, must take a remedial course in addition to mainstream English next year -- yet there are no additional funds to add teachers or classes -- so teachers will be shifted and classes overcrowded.  Their principal invests an extraordinary amount of resources and energy into bringing marginal students up to speed, under enormous financial constraints.  This in a state (California) where the public education system in general & teachers in particular are under attack by the Governor.   Worrying about whether the NCLB test gets the formula just right, misses the point.  

My kids do fine on these exams, but they have two PhDs at home to edit essays and help them with french, calculus, and chemistry questions.  That's not true for most of their classmates.   And no, this is not a `failing' public high school.  My kids are getting a rigorous education from dedicated professionals.  Teachers who claim that '1984' is the most important book ever written, teachers who protest military recruiters on campus by passing out leaflets as students arrive in the morning, teachers who tutor physics on the weekend (gratis), teachers who teach latin off the books during the summer...  

I support the concept of national baseline standards in education, and national standardized tests are an obvious tool to measure progress.  But NCLB is quick to punish and slow to repair.  When a week every year is given over to taking these exams, they can't help but impoverish the gifts that qualified teachers impart.  The challenges I see won't be cured by another Bush policy of swatting flies, or is it mosquitos?  Now I've gone & mixed my metaphors.

Complaints with US elem/sec. education (and NCLB):

- Our system produces an inferior outcome compared to other countries (Japan, India, etc.).  Why is that?  Why do Japanese students outpace US students so much at the end of secondary education (including the large amount of homwork expected, and much tighter school discipline)?

- It has been recognized for some time that our students get their best world-class education in college, not in pre-college. What does this do to help non-college kids prepare for their life?

- Yes reading is critical to everything.  So is some facility in writing?  Math?  No so much beyond basic arithmetic, unless certain professions are the object.  Why teach algebra and trig to kids not going to college in those specific professions?

- Second language facility is increasingly important, but grossly neglected.  The time to being this is pre-school and early in school, but where does this happen.  Starting in high school is just silly.

I have to agree with this idea.  Standardized testing emphasizes only the most basic levels of knowlege, stressing memorization and limited synthesis.  The problem for most standardized tests is logistical:  How can you assess how well a student has truly integrated a concept into his or her world-view using a Scantron?  There has been a paradigm shift in education over the past decade with an increasing emphasis on critical thinking skills and analysis, and less emphasis on memorizing facts and dates.  Despite claims from the testing industry that they can accurately assess these skills, I don't buy their arguments.  If a student can respond to a critical thinking problem in the exact same way the test designer does, he gets it right.  If he approaches the problem from a different angle, he gets it wrong.  Ask a roomfull of adults what 53 + 39 is.  They will (hopefully) all get it right.  Then ask them how they arrived at the answer.  I think you would be surprised at the variety of methods people use for a simple problem like adding two numbers.

Our schools are struggling to meet the demands of our economy and the world of the 21st century.  Now is a time to allow educators the freedom to "think outside the box" and try a variety of strategies.  Unfortunately, the restrictive burden of high stakes testing is stiffling the creativity that may ultimately be our salvation.

that most teachers opposition exists because of having the quality of there work measured.

That might be the union's rationale, but not specific teachers.

Again, standardized can be good to be to establish a kind of remeidal baseline. But multiple choice testing should not be central to determining whether or not schools get funding, because the kind of things it tries to evaluate are not central to what is valuable about education.

Your coin example shows this point. Sure its great that kids could learn that. But conveying specific facts, for the most part, is not the purpose of education. The purpose of education is to develop thought processes, critical thinking skills, problem solving ability in a more generalized sense. Learning thinks like what you mention above are only useful as a baseline: like memorizing your times tables. If you want to get the kind of jobs that college level education requires, the kind of learning encouraged by standardized testing should not be central to the education process.

And this ultimately is the problem. Wealthier schools will not have to make standardized test material central to their curriculum, whereas poor schools will be much more likely to, because, a) these children start from a higher baseline and thus need less preparation; and b) they won't depend on federal funding.

Most of the comments here focus on "teaching to the test", and whether tests are good measures of learning, etc. Personally, my belief is that well-designed tests can provide a reasonably good proxy measure of students' learning level. But a lot of NCLB is based on the supposition that school quality can be measured by looking at what percentage of students at the school reach a specified learning level, as measured by the test. Even if the test is well-designed to measure learning levels, the question remains as to whether this is a good way to measure the quality of a school. Because students enter different schools with dramatically different learning levels on average, and also learn dramatically different amounts during the summer, it does not seem sensible to measure school quality by simply looking at the percentage of students at the school above a certain learning level. Rather, we need to measure how much they have learned while at the school, the schools' "value added", which may be measured by the change in test scores, not the test score level. This may seem an academic debate, but under NCLB, schools that are deemed "failing schools" may be subject to severe sanctions. If in fact these schools are not really poor quality schools, we have a serious problem.

ERIC (Education Resources Information Center) has a collecton of articles about school testing and standards, and one of them discusses why teaching the test is bad.  (Appropriately named: "Why Teaching the Test is a Bad Thing")

First, what does teaching the test mean? It means that teachers only teach what is on the test, often in the specific format that the test takes.  Therefore, quality of the education depends heavily on the quality of the test.  And even a good test can only test a certain subset of skills in a certain subset of formats.

So, the danger is that if the teacher only teaches a certain test format, students will only learn how to answer problems posed in that manner.  In fact, if students are only taught one way to answer questions they might not actually have learned the main skill that the test was supposed to be assessing.  To cite an example from the article, one study showed that when students were taught how to translate from Roman to Arabic numerals, they couldn’t necessarily do the reverse.  In fact, their test scores went down 35% to 50% from the former to the latter.

Another problem is that the skills and subject areas that are not tested are often pushed slowly out the curriculum.  This is certainly true of classes in the arts and outdoor recreation.  However, when only reading and math skills are tested, areas such social studies and the sciences are often neglected, as well. 

Another problem is that standards of learning tests rarely test for critical thinking skills or skills involving the synthesis of information (not to mention creativity).  That means that these skills are often not taught when teachers are teaching the test.  For example, the article mentions that in one study, “elementary teachers had given up reading real books, writing and long term projects and were focusing on word recognition, recognition of spelling answers, language usage, punctuation, and arithmetic operations” .

The truth is, no test can adequately all the skills that are valuable to students.  That’s why teaching the test is so bad.  You leave out important information and skill areas. 

The link to the articles is here:

http://www.eric.ed.gov:80/ERICWebPortal/Home.portal?_nfpb=true&am p;ERICExtSearch_SearchValue_0=%22standards+and+testing+and+teache rs%22&ERICExtSearch_Operator_2=and&ERICExtSearch_SearchTy pe_2=eric_metadata&ERICExtSearch_SearchType_0=eric_metadata&a mp;ERICExtSearch_SearchCount=2&ERICExtSearch_PubDate_From=199 8&ERICExtSearch_FullText=true&ERICExtSearch_SearchValue_1 =%22teaching+the+test%22&ERICExtSearch_Operator_1=and&ERI CExtSearch_SearchType_1=eric_metadata&ERICExtSearch_PubDate_T o=2004&_pageLabel=RecordDetails&objectId=0900000b8013af14

(The comment editor isn't recognizing my html for some reason)