I want to leave the first comment!
It's not about whether these arguments are correct. They both are, to some extent. The question is: which side do you think has the best justification for its goals and imperfections?
It's kind of like thinking about American justice. We lean toward letting the guilty go free rather than allowing an innocent person to be imprisoned. The latter alternative is seen as more problematic and less in keeping with our values. It doesn't mean that the other alternative is incorrect; it's just in keeping with our values.
So the question these arguments really pose is: which side best suits your/our idea for getting the best quality education and outcomes for children?
This debate is a first in the EduSphere! What an outstanding concept! This new and positive aspect of EduBlogging only hints at all the possibilities for the free exchange of thoughts and ideas that are out there...
Both participants have represented their respective positions well; I'm looking forward to reading more!
Hmm...
A more focused question to explore in the future might be, "What would you change about ESEA in 2007?"
Both of these arguments have merit but neither addresses how could NCLB be improved. Both probably have something to say about that. It's clear that Mike thinks it should be changed (or abolished) and Kenneth's "it's better than nothing" perspective also implies NCLB could be improved. I'm curious as to what they both think should happen when NCLB is reauthorized.
Two points:
First, as KDeRosa says, "What they won’t tell you is how they plan to raise achievement (other than to continue to throw a lot more money at the problem) and why they haven't done it sooner. All we ever hear is lots of excuses why it can’t be done."
I've never heard a good response to this.
Second, Mike says "...the dropout rate of Texas minorities has actually increased significantly since the states testing based accountability went into effect."
And this proves what? Is it better to let kids who aren't able to do the work slide by via "social promotion"? Of course not! First you identify those who're failing, then you try to help them. Absent testing, these kids would graduate, walk out into the world, and get a sharp dose of reality as employer after employer told them they couldn't do the work.
At least with testing you can can catch the problem early, although KDeRosa is right in saying that the states have way too much leeway to dumb down the tests.
Chris,
We were not asked to address the issues of what we would do to replace or improve NCLB, we were asked to give our positions on the law.
Off the top of my head, here are a few suggested ways to improve NCLB:
1. Find a way to factor in SES into the achievement goals. Currently, affluent school districts are getting a pass on finding ways to improve achievement since they are loaded with high SES kids.
2. Concentrate effort on the K-5 grades. Until the elementery schools are fixed it will be difficult to fix the middle and high schools.
3. Along with Point 2, middle and high schools should be held repsonsible to maintain 100% of kids that they receieve who are at the proficient/advanced level at at least the proficient level today. By the same token, some allowance should be made for kids they recieve from the elementary schools who are already failing. From now until 2014, I'd set up a remedial category that will allow the middle/high schools some leeway to bring these kids up to speed.
4. Similiarly, we need to find a better way to deal with transfer students. Under the current system the school they transfer into is responsible for transferees who are failing. That isn't fair.
5. Provide more alternatives for kids in failing schools than is permitted uner the current law. In many school districts these children have nowhere to go.
Hi Mike,
I know that wasn't the question you were asked to respond to - It's just that after reading your positions I suspect both of you might have something interesting to say about this.
Perhaps that's the 'sequel' to this experiment... just a suggestion.
Ragnarok,
With all due respect, I think you're missing the point regarding Mike's mention of increasing dropout rates. It's not about social promotion. Many proponents of the testing program (including Paige) claimed the tests held administrators and teachers accountable and motivated them to provide better education for all students. In theory, schools would better prepare students so that fewer students would have to drop out. This was the 'Texas Miracle' that Governor Bush talked about so much in 2000.
In reality, that just isn't what happened. Instead, SES/ethnicity-based differences (in terms of attrition rates) has only widened since the introduction of the tests in Texas. The testing program isn't working the way some said it would.
Chris,
Understood, but I'm arguing "diagnosis first, then cure".
That is, I suspect that a lot of minority kids who used to get a free pass to the next grade are now going to be held back. The public schools will scramble to figure out how to teach them more effectively, and at some point they'll either figure it out or the kids will be given a chance to go somewhere else. Some will fail for sure, but some will do well - that's the nature of competition.
That's my take on NCLB's goals, but I could certainly be wrong.
Yes, that sounds about right.
However, the "scramble to teach them effectively" had allegedly already happened in Texas. That is, many claimed the tests motivated schools to do a better job and that all students were better served. The fact is, that hasn't happened (at least not in Texas).
My only point is that some have claimed that accountability schemes can provide the 'cure', but there's little evidence for that. The 'diagnosis' is an important first step, of course...
I generally buy the pro-NCLB view. But Mike does a nice job of citing the work of Tom Kane and others on the variability problem.
The most helpful edit to NCLB would be to integrate a value-added measurement component, tracing the sum of all "student gains over baseline", from year to year.
Well, I'm very sceptical about the ability of any large bureaucratic, unionised entity to do anything effectively. Much easier for them to claim victory, fudge the numbers, and then claim victory. So I'm quite willing to believe that the Texas public schools fudged their results.
What I do hope is that as more children are given a choice, some of them will forsake the public schools for charters/religious/private schools, and they will do better. It won't be a gentle process, as kids who've learned to slide by are suddenly expected to do real work; but time will smooth the experience.
I suspect that eventually the public schools will become the Delphi's of the education world - bloated, unionised, ineffective, unsustainable and much too expensive. And then they'll disintegrate.
It's very important for a country to have a well-educated workforce. This country has evaded that requirement by attracting large numbers of skilled immigrants, but with globalisation that well's about to run dry. That's one reason why it's critical for us to have a crackerjack K-12 system.
Personally, I lean pro-NCLB, but Mike raised what I think is a big problem. Schools that don't get 100 percent of kids to succeed on tests will be labeled as failing schools. I don't think that's possible. I think we canget close, but not all the way.
So the question is: If a school gets 99 percent, is it a failure?
I appreciate the problem in setting the goal lower. You can't really do it and be fair to every child. You acn't decide in advance that some will fail. But I'm not entirely comfortable labeling a school a failure after it has managed to go from 30 percent passing to 98 percent passing.
I'm not sure what to do, but it does concern me.
Jenny,
Take a look at KDeRosa explaination of how the 100% is actually 94% in the "More Upset..." thread.
I find myself constantly amazed by those who suggest that testing of the kind mandated by NCLB (and pioneered to some degree by Texas) is a panacea for all educational problems. Testing solves nothing. Nothing. It merely provides a statistical snapshot of a given group of students and their performance on one test on one day.
Those who really know how students are doing and what they need are those who are utterly ignored, not only by debates like this, but by educrats pushing NCLB and similar mischief: the teachers.
Of course, we could really go out on a pedagogical limb and suggest that students (and their parents--assuming they aren't all orphans) bear the ultimate responsibility for their educations, but that would go against everything NCLB and its proponents stand for. In other words, that would make sense.
Sorry, KDeRosa's "explanation".
Rag,
You fail to grasp what KDeRosa was saying about absences and exemptions. Yes, 95% of the students must pass the test, but in a typical elementary school 2 - 3 % are absent on any given day. DeRosa states in his piece that 6% of schoolchildren are truly disabled ((it has remained at a constant 6%)), so why would absences and exemptions only be allowable to 5%.
In addition, the law, as currently written, only allows for 1% of the population to be exempt for special education by the year 2014.
What it does require is every student taking the test to pass it. Where I come from EVERY=100%
Excuse, I should know better than to write before I've had coffee. My second sentence should have said:
95% of the students must take the test. That is the bare minimum.
I need an explanation about the 94% I'm afraid. In my district, students absent from school must take the test. There is a three week window for test taking, and if you miss a day, there are two weeks afterward for makeup tests. From what I can tell, about 99 percent of those required to take the test actually do.
Re: the 1% cap and the 5% subgroup absentee cap take a look at:
Children with Disabilities Under No Child Left Behind (NCLB): Myths and Realities
Briefly, there are many ways a state can soften the impact of the tests: alternate assessments, test accomodations, alernate standards, and using the 1% cap.
Even though 6% of students have a medical handicap only a small percentage of them have a handicap that affects their cognitive capacity. This percentage is about 1-2%.
mike's variability argument is only valid if NCLB were trying to measure pewrformance gains for cohorts over successive years. It isn't. NCLB is measuring school performance by testing successive cohorts at a fixed period of time. Each has its advantages and disadvantages.
I suggest reading FInding Additional Value in New Accountability Systems for much more information than you'll need.
Many proponents of the testing program (including Paige) claimed the tests held administrators and teachers accountable and motivated them to provide better education for all students. In theory, schools would better prepare students so that fewer students would have to drop out.
Clearly, they underestimated the resistance of the edublob which is taking a huge gamble by complaining instead of doing in the hopes of getting th law repealed before 2014.
I suspect this gambit will backfire. People are starting to get fed up. Take a look at:
this Fordham Foundation Poll Report.
Ohio's politicians should heed the survey's findings—especially the one indicating that the state's residents don't have much confidence in their leaders' efforts to improve the schools. Just four percent of survey respondents say that state elected officials and legislators are "doing a good job" when it comes to public education.
This would be okay if nothing needed to change, but Ohioans surely don't think so—and plenty of objective evidence says they are correct. Only a third of survey respondents—and fewer than one in five African Americans—believe their local public schools are "doing pretty well and need little change." Virtually all others want "major change" or "a whole new system." This is no surprise in a state where close to half of respondents also see the economy as a serious issue. Ohioans know that education and economic opportunity are connected, and they're worried about both..
Testing solves nothing. Nothing. It merely provides a statistical snapshot of a given group of students and their performance on one test on one day.
But testing is a necessary first step in any feedback loop system and provides a bseline upon which future error correction can be based and measured.
You con't actually think education can be improved without an accurate objective measure of student achievement?
By the way, it is quite likely that the 1% cap gets raised to about 2.5% when NCLB is reauthorized.
Those who really know how students are doing and what they need are those who are utterly ignored, not only by debates like this, but by educrats pushing NCLB and similar mischief: the teachers.
Of course, we could really go out on a pedagogical limb and suggest that students (and their parents--assuming they aren't all orphans) bear the ultimate responsibility for their educations.
Notice the subtle shift of responsibilty away from the teachers. They don't want to be responsible for education. Also, I'm wondering why, if teachers know so much, they've failed to act on it lo these many years. I am dubious to say the least.
Mike makes an important point about what is being measured, and this is the greatest flaw in the law. I generally like anything, law, regulation or custom, that is going to hold people accountable for their actions. The intent behind the law, I think cannot be argued as a bad thing. But, as it is said, the devil is in the details and Mike is right, the details of measurement are wrong.
You cannot measure one group against another group a year later to determine if progress is being made. What should be measured, obviously, is the success of each class of kids as they move through school. For example, Class A starts in grade 1 and they have 60% proficient rating in grade 1. As they progress through the school, that 60% should improve over time. Say by grade 3, this group of kids should have 65% of the kids rated as proficient, Grade 6 should be 75% and so on.
Of course, variability will come in through a mobile society, in which students come and go in a class, but that may affect proficiency only tangentially since ther percentage of change from year to year is relatively small.
From reading these statements, my guess is that Mike supports the ends of NCLB, just not the means.
JennyD-fabulous idea.
matt, NCLB is measuring school performance, not student performance. NCLB's intention is to look at certain grades and measure performance every year with the idea that improvemenr should be made toward the final goal. because of this the variability between successive cohorts is irrelevant.
KDeRosa,
Your links to the Myths are, as so much of the pro-NCLB beliefs, unsubstantiated opinions. The author of that website is NOT an expert on learning disorders or medical disabilities.
In addition, it should matter to you that the improvement of schools is being measured using a flawed process. You wrote, "NCLB is measuring school performance by testing successive cohorts at a fixed period of time" and clearly you fail to see that is the exact point the research showed, that this process is flawed.
Mike, I am citing the NCLB Myths as a summary of the issue. I've already citing the statutes and regulations in the other comments thread which show the same thing.
I understand the article contains opinion. I also understand that the "research" and "studies" you've cited are also opinion. Calling them research doesn't make it so.
As far as the medical disabilities issue goes, there is hard evidence to support the article's and my contention of a 6% disability rate. I 'm sure I or someone else can dig it up for you.
The NCLB scheme is not flawed in the way you are suggesting. AYP is a measure of the school's improvement not students' improvement. If it were the latter, then you might have a valid point. Under NCLB's AYP we aer measuring what each child knows on an absolute standard, not how much more he has learned from a baseline.
Let's use a simple example. Let's say a state has defined AYP for a fifth grade class as being 35% in 2001 and linearly progressing up to 100% in 2014. This merely means that every year we test a different fifth grade class (cohort) to see how much they know. If they meet that year's threshold, the school passes. It doesn't matter whether that cohort was duller or brighter than the last cohort, we only care that that particular cohort meets the standard.
Every year the school has ana dditional year to refine its teaching methodology and to see what is working, what is not, and to make adjustments. NCLB just ractchets up the goals every year with the expectation that progress is being made.
The variability between successive cohorts only matters for schools performing at the margin of success. If the schools were doing their jobs properly they would know well before fifth grade if they had a underperforming cohort on their hands and should have been taking measures to give them whatever help they needed to achieve. This is what NCLB is all about, making sure that schools are marshallingtheir resources as they see fit to assure that all children are performing at their potential.
Moreover, the system you allude to is equally as "flawed" but in a different way. First it requires twice as much data points to measure the delta of performance. Then you have the flawed notion that students are able to make tiny improvements in achievement every year so that in 2014 they'll all be at 100%. I child's fourth grade reading performance correlates highly with his 8th grade reading performace. If he's failing in 4th grade, most likely he'll be failing in 8th grade and beyond. If we have, say, a 35% passing rate in fifth grade, that cohort is not going to achieve any higher without serious remediation. You've effectively lost the 65% of the class who failed by fifth grade.
Here's an anecdote that may be instructive: my local high school has a senior class that tested extremely well. ACTs were high, extra-curricular activity is high, the football team is the most successful in the school's history, and so on. Next year's senior class (the one that caused my wife to quit teaching catechism when they were in fifth grade)has low ACTs, low grades, little extra-curricular involvement, the contempt of this year's senior class, and so on. The snapshot of the senior class, according to NCLB, would show a remarkably successful school; the same snapshot of the junior class would show a very different picture. The difference is the kids.
For those who think it's not fair to assess a school with cross-sectional approach:
Do you want twice as much testing to see the within-year changes for each class?
Rats, Mike in Texas, KDeRosa's answered you before I got to it. And then Jenny slipped in that bit about "...students absent from school must take the test. There is a three week window for test taking, and if you miss a day, there are two weeks afterward for makeup tests."
Not necessarily.
If all (and by less than 100%) students of both cohorts were achieving above the cut-off point, the variation between cohorts would not have shown up in NCLB even though the older cohort were higher achievers.
If the school were able to successfully bring the low performers up to the proficient level in the older cohort, they should have been also able to bring the low performers in the next cohort up to the same level even if there were more of them.
Also, NCLB isn't quite as simplistic as you've caricatured it. From the The Promise and Pitfalls of Using Imprecise School Accountability Measures article you cited:
The No Child Left Behind Act of 2001 provides states with the option of calculating three-year weighted averages of school performance on test scores. In Kane and Staiger (2001), we show that the optimal weights for averaging performance over time (which maximize the reliability of the resulting performance measure) incorporate information on the signal variance, sampling variation and the degree of persistence in the signal over time: One places more weight on past performance when a single year’s test score is unreliable and when the persistent component of performance is strongly correlated over time. The resulting estimate is a sophisticated moving average of each school’s prior test scores, and is equivalent to a Bayesian estimate of the posterior mean of the persistent component of each school’s test score. Our work suggests that this approach can greatly improve the accuracy of school ratings. For example, we found that the optimally weighted average of past test scores was much more successful in picking schools that were likely to perform well one or two years in the future, more than doubling the forecast R-squared compared to using a single year of data. But, in most cases, simple averages of past test scores achieved more than half of this gain in forecast performance. Thus, even simple averaging over several years leads to markedly improved performance measures.
...
Similarly, some accountability systems require schools to improve their student’s performance continuously. For example, the initial versions of the No Child Left Behind Act passed by the House of Representatives and the Senate in 2001 included provisions that required schools to show improvement in math and reading scores each year or be subject to a variety of sanctions. But even when a school is on the right track, the path to improved student performance is rarely a straight path. The natural fluctuation in noisy test score measures means that two steps forward are often followed by one step back. The No Child Left Behind Act had to be rewritten in conference committee after an initial analysis– carried out after passage by both houses -- suggested that virtually every school in the country would have failed to improve at least once in a five-year period because of the natural volatility in test score measures (Kane, Staiger and Geppert, 2002; Broder, 2001).
It would appear that your entire argument is a strawman based on the faulty premise that NCLB doesn't account for cohort variability and demands continual progress.
my 2nd paragraph above should read:
If all (and by "all" I mean less than 100%) students of both cohorts were achieving above the cut-off point, the variation between cohorts would not have shown up in NCLB even though the older cohort were higher achievers.
Also, the Promise and Pitfalls of Using Imprecise School Accountability Measures article I cited above is from the same authors, Kane and Stagier, as the article Mike cited, but it is a different article.
Perhaps the difficulty isn't in testing students for what they know (and can do), but in the method of assessment. A body of knowledge can be tested using multiple choice questions, but skills cannot. In my field, English, highly quantifiable areas such as grammar can be tested using multiple choice questions. The skill of writing does not fit into an easily measurable test structure.
Sadly, the Virginia state standardized SOL tests try to test writing skills with a multiple choice format. While the structure of the questions is clever in that it contorts writing skills into this format, what results is too contorted for kids to readily understand. They have difficulty figuring out what is asked; they are asked to work with written pieces that are presented in a way that is found nowhere else -- The assessment winds up being disconnected from the real world of actual writing.
Yes, there is a written component to the SOL tests. They are given in March in order to allow scoring time. Since reading and writing and grammar are all closely intertwined, all the year's curriculum must be completed before the test is taken.
Surely there is a way to adequately test kids to give a picture of a school's performance on a large enough scale (portfolio assessment isn't practical for the data that we want) without relying on a multiple choice format to reflect knowledge and skills.
95% of the students must pass the test
Many states do take advantage of this option, or have done so in their earlier AYP goals. However, after 3 years the AYP goal must increase at a rate enough to cover. For example, a school can achieve AYP for 3 years in a row with the same score, but in the 4th year the required score must go up to achieve the same incremental increases that would have occured over the 3 years.
Two additional points to make, both from other articles by Kane and Steiger (and Geppert).
a disproportionate number of the schools that enroll disadvantaged minority subgroups are likely to fail. The minimum proficiency rate that schools are required to meet will be raised gradually to 100 percent over the next 12 years.
There's that pesky 100% figure Ragnarok wants to deny, and there is another opinion that minority enrollment effects tests scores, or in this case the lucky and the unlucky.
Another quote from the article:
Multiple years of data are required to measure improvements in performance reliably.
Unfortunately for schools, sanctions begin to occur after only 1 year of failure to meet AYP.
This is getting to be pretty funny, Mike. Now you're arguing with yourself.
You said that 95% of the students have to take the test. You also said that 100% of those who take the test have to pass it. Right? There's no contradiction here, Mike, 100% of the 95% who're tested must score proficient or better.
We won't even get into the fact that an additional 1% can take a state-designed test in place of the test. Or the fact that a state can ask for additional exemptions.
"There's that pesky 100% figure Ragnarok wants to deny, and there is another opinion that minority enrollment effects [sic] tests scores, or in this case the lucky and the unlucky."
Huh? NP-complete has nothing on parsing this here so-called sentence.
Actually Rag, I had some trouble with HalScan and it would not let me post a correction. What I meant to paste there was:
The No Child Left Behind Act of 2001 provides states with the option of calculating three-year weighted averages of school performance on test scores.
HaloScan would not let me post another time last night for some reason.
What does proficent mean? It varies per state. But, in Kentucky it means scoring 100 on a 140 point test or 71.4%.
Then if you add in confidence intervals, which Kentucky has like 26 other states. Then the precentage can drop even lower.
Confidence intervals vary according to the size of the sample, and there is some advantage to smaller schools. But for agruement sake lets use +-5%. Then this brings the lowest acceptable score down to 66.4%.
So, what the you are saying is we can't get 100% (or 95%) of our students to score a D on a test if we work on it over the next 9 years?
In my district we have an elementary school that's Title 1, 60% Free and Reduced Lunch, Urban that's scored a 98.1 on our CATS Test. The free and reduced lunch kids scored higher than the kids who didn't get free and reduced lunch.
I no longer accept poverty as an excuse, it leadership and commitment to excell that make the difference.
Bob, it's even worse than that. Typically, the state tests are much simpler than the NAEP exam which tests far below grade level. So by 11th grade, the NAEP is mostly testing 8th grade material and lower. The State tests are even easier and, as you point out, students only need to know 60% - 70% of the material for compliance with NCLB.
It just goes to show you how poor a job our educators are doing by and large. I supposed I'd be complaining about NCLB too if I were in their shoes.
Commenting by HaloScan
