Twenty percent of New York students opt out of high stakes testing

[Left_Hand_of_Dorkness]

@Sam_Lowry

I think first before measuring what schools are doing a good job and which are not, I would try to get resources more equitably distributed amongst schools. I know that there are some schools with an excess of resources, and some that are hurting. A school with up to date computers, calculators, textbooks, and enough desks, pens, teachers, and other resources, is going to do better than schools where teachers have to pay for everything in their classroom and are stressed out about the school supplies being lost or stolen.

This is a big piece of it. But there are some other pieces:
-A serious campaign to reduce homelessness in the United States so that it approaches Scandinavian levels. Then you won’t have homelessness interfering with a kid’s education.
-A serious campaign to end food insecurity in the United States. Then you won’t have hunger interfering with a kid’s education.
-A serious campaign to fix housing disparities in the United States. Then you won’t have substandard housing interfering with a kid’s education.
-A serious campaign to reform the prison system in the United States. Then you won’t have as many kids with imprisoned parents, which interferes with their education.

I know these are huge things. But if we’re not willing to tackle these problems, we’re not serious about fixing education in our country. We KNOW that factors like these are major impediments to equal education. We need to address them.

[Bone]

@Left_Hand_of_Dorkness

“Fair” in the sense that it measures actual quality of teacher pedagogy.

I think this is part of the disagreement. I don’t think the tests are designed to measure the quality of teacher pedagogy. That’s not the point in my mind. The point is to gather data on student proficiency over certain specific skills or knowledge.

I’m unconvinced you can make a standardized test that accurately, precisely, and objectively measures a student’s ability to conduct a short research project, or to write a persuasive letter, or to design an experiment using valid scientific procedures. These are things much better assessed through observing students as they do them.

I think that’s probably true. Those skills are difficult to test in a standardized way. But I don’t think the inability to test for those things are a criticism of testing as a whole. Testing for mathematics, reading, comprehension, etc. are also important. Do you think those items are critical to graduate from high school, or at lower levels? I think getting the basics are a first step, and there are places where even that is failing. It’s not like you can develop skills to do a research project without first developing skills at the basics. If we’re failing at the basic level, then those other items wouldn’t be and shouldn’t be taught or focused on regardless.

Really, you would? School A has impoverished students who read novels, who have art and music and science and social studies, who do math in real-world contexts. School B has eliminated all extra activities, spends all day drilling students on multiple choice reading passages and multiple choice math tests. If School B gets more students to pass the end of grade test, that’s a success?

But you left out the part about the ability of schools to overcome their hurdles of poverty. How would you know if that is happening without measuring it?

In your example, can the enriched but impoverished students demonstrate mastery of the basics that the standardized tests would be focused on? Performing Hamlet and having a robust music program is great, but if you can’t read or do math I’d say the education is failing. But that’s not a great comparison because what I’d really like to see is School B underperforming on standardized tests, then focusing their efforts on them, then the students passing them. I would consider that at least a positive development. I’m not sure success is how I’d characterize it because there would be more distance to go.

[magellan01]

@Sam_Lowry

I’d keep standardized testing, but work out what would be a reasonable amount of it. I don’t know how much would be reasonable, but do something so that so much emphasis isn’t put towards teaching towards the test.

And there’s the rub. I think we need a way to evaluate schools. That mean some kind of student testing. But once you have the testing, and the school has something on the line, they will try to increase the likelihood that kids will do well. I don’t know the answer, but I do think we should constantly being evaluating the schools.

[magellan01]

@Left_Hand_of_Dorkness

This is a big piece of it. But there are some other pieces:
-A serious campaign to reduce homelessness in the United States so that it approaches Scandinavian levels. Then you won’t have homelessness interfering with a kid’s education.
-A serious campaign to end food insecurity in the United States. Then you won’t have hunger interfering with a kid’s education.
-A serious campaign to fix housing disparities in the United States. Then you won’t have substandard housing interfering with a kid’s education.
-A serious campaign to reform the prison system in the United States. Then you won’t have as many kids with imprisoned parents, which interferes with their education.

I know these are huge things. But if we’re not willing to tackle these problems, we’re not serious about fixing education in our country. We KNOW that factors like these are major impediments to equal education. We need to address them.

But that doesn’t address the problem of how we evaluate schools. Whether we fix all you list or don’t, there will be great schools, good schools, fair schools, bad schools, and horrible schools. How do we determine which is which? Testing does that. Is their another mechanism?

[Sam_Lowry]

@Left_Hand_of_Dorkness

This is a big piece of it. But there are some other pieces:
-A serious campaign to reduce homelessness in the United States so that it approaches Scandinavian levels. Then you won’t have homelessness interfering with a kid’s education.
-A serious campaign to end food insecurity in the United States. Then you won’t have hunger interfering with a kid’s education.
-A serious campaign to fix housing disparities in the United States. Then you won’t have substandard housing interfering with a kid’s education.
-A serious campaign to reform the prison system in the United States. Then you won’t have as many kids with imprisoned parents, which interferes with their education.

I know these are huge things. But if we’re not willing to tackle these problems, we’re not serious about fixing education in our country. We KNOW that factors like these are major impediments to equal education. We need to address them.

I totally agree, these all make sense. They are big problems, and they will take a whole lot of work to even partially fix, but without at least some work on them, we won’t make a lot of progress on education.

@magellan01

But that doesn’t address the problem of how we evaluate schools. Whether we fix all you list or don’t, there will be great schools, good schools, fair schools, bad schools, and horrible schools. How do we determine which is which? Testing does that. Is their another mechanism?

You keep a certain amount of testing to keep track of schools and to compare them, but only use the testing for that. Don’t base promotions or firing on test scores. Provide some incentive for the kids and teachers to do their best, but not so much incentive that all energy is placed on the tests and not on the rest of education. Use the grades from comparing schools, but have it just be one metric the school is ranked on. If I’m choosing on where to send my kids for school, and one school has slightly lower test scores, but fewer absences, more kids choosing to be in extracurriculars, and more kids getting into good colleges, I’ll maybe pick that school despite the lower test scores.

[magellan01]

@Sam_Lowry

You keep a certain amount of testing to keep track of schools and to compare them, but only use the testing for that. Don’t base promotions or firing on test scores. Provide some incentive for the kids and teachers to do their best, but not so much incentive that all energy is placed on the tests and not on the rest of education. Use the grades from comparing schools, but have it just be one metric the school is ranked on. If I’m choosing on where to send my kids for school, and one school has slightly lower test scores, but fewer absences, more kids choosing to be in extracurriculars, and more kids getting into good colleges, I’ll maybe pick that school despite the lower test scores.

That makes sense, but it still necessitates that 1) the test scores exist and 2) that they are accurate/useful. Correct?

[Frazzled]

If you have the time - check out John Oliver’s take on standardized tests:

Standardized Testing: Last Week Tonight with John Oliver (HBO)

[Miller]

@Bone

In some respects, it totally is. But eliminating it at the early levels while retaining it for college placement and other areas is worse. As long as that’s the system that is employed it should be taught.

I don’t really buy the idea that standardized testing is some special skill that kids need to learn in order to get into college. It’s a test. You learn the material for the test, then you take the test. I’m not sure why it makes a difference in a student’s performance if there’s another kid on the other side of the country taking the same test, versus the teacher creating his own, unique test for his class.

Anecdotally, when I was in school, the only standardized testing I did was the PSATs, and later, the SATs. Despite my total lack of prior experience with standardized testing, I scored well enough on my SATs that I was admitted to college despite my distinctly underwhelming GPA.

[CalMeacham]

@alphaboi867

Out of curiosity just what are those instructions?

Probably to subtract off 1/4 of the checked answers within the puked-on area, because by chance 1/4 of those answers will end up being correct.

[Left_Hand_of_Dorkness]

@magellan01

But that doesn’t address the problem of how we evaluate schools. Whether we fix all you list or don’t, there will be great schools, good schools, fair schools, bad schools, and horrible schools. How do we determine which is which? Testing does that. Is their another mechanism?

Testing does not do that. Testing claims to do that. I’d rather have no system for evaluation than the current system, which punishes teachers who choose to work in a high-poverty school.

One alternative measure that’s being explored is the value-added model. It looks at students’ performance in previous years as a predictor of their growth in the current year, and compares their actual growth to predicted growth. Thus a student who learned 50% of what they should have learned in the past seven years, but who learned 75% of what they should have learned when I taught them, is a mark in my favor, whereas a student who learned 150% of what they should have learned in previous years, but only 125% of what they should have learned in my class, is a mark against me.

I’ve read articles talking about the problems with this model, but I honestly am not good enough at statistics to evaluate the claims about the problems, so I don’t have a very strong opinion on them.

[WhyNot]

The easiest non statistics problem with that is that it doesn’t account for the year the kid’s parents divorced and she’s had to spend weeks couch surfing while her mom works nights and is trying to find them a new apartment.

For example.

[DrCube]

@magellan01

But that doesn’t address the problem of how we evaluate schools. Whether we fix all you list or don’t, there will be great schools, good schools, fair schools, bad schools, and horrible schools. How do we determine which is which? Testing does that. Is their another mechanism?

Why is this an issue? Was there a problem figuring out which schools were good and which were bad before? (And in the US, people don’t even get to pick which public schools they send their kids to. It depends on where you live, period. So that’s not the issue.)

How do we currently manage to figure out which colleges are good and which aren’t without rampant standardized testing at the university level? How do we evaluate professors? Vocational schools? On the job training?

I think what we’re seeing is the result of massive lobbying by the companies that sell these tests. It’s good for their bottom line, but not so much for the rest of us.

[Left_Hand_of_Dorkness]

@WhyNot

The easiest non statistics problem with that is that it doesn’t account for the year the kid’s parents divorced and she’s had to spend weeks couch surfing while her mom works nights and is trying to find them a new apartment.

For example.

This is totally true; the counterargument is that, on average, these events don’t happen disproportionately in one year. If the data are used correctly, a teacher won’t be penalized for that divorce year, since it’ll average out over multiple years.

I’m not enough of a stats nerd to evaluate this claim, but it seems on surface reasonable.

[magellan01]

@DrCube

Why is this an issue? Was there a problem figuring out which schools were good and which were bad before? (And in the US, people don’t even get to pick which public schools they send their kids to. It depends on where you live, period. So that’s not the issue.)

It’s an issue because if we can identify the poorest performing schools and the best ones, we might be able to understand what makes schools poor or great.

I’s an issue because if you can assess schools regularly, you can see the ones that are sliding downhill and try to fix things before a generation of students go through the school getting a sub-par education.

(As far as having to attend the school where you live, that’s an argument for a massive voucher program. But that’s a subject that will hijack this thread, so let’s leave it aside.)

@DrCube

How do we currently manage to figure out which colleges are good and which aren’t without rampant standardized testing at the university level? How do we evaluate professors? Vocational schools? On the job training?

These are easily assessed through how the graduates do in the job market. It’s particularly obvious at elite schools, where one metric used is the number of students that make it into medical, law school, Wharton, etc.

[magellan01]

@Left_Hand_of_Dorkness

Testing does not do that. Testing claims to do that. I’d rather have no system for evaluation than the current system, which punishes teachers who choose to work in a high-poverty school.

One alternative measure that’s being explored is the value-added model. It looks at students’ performance in previous years as a predictor of their growth in the current year, and compares their actual growth to predicted growth. Thus a student who learned 50% of what they should have learned in the past seven years, but who learned 75% of what they should have learned when I taught them, is a mark in my favor, whereas a student who learned 150% of what they should have learned in previous years, but only 125% of what they should have learned in my class, is a mark against me.

I’ve read articles talking about the problems with this model, but I honestly am not good enough at statistics to evaluate the claims about the problems, so I don’t have a very strong opinion on them.

I think it makes perfect sense to take into account a baseline. Especially to ensure fairness to the teachers and administrators.

[Bone]

@Miller

I don’t really buy the idea that standardized testing is some special skill that kids need to learn in order to get into college. It’s a test. You learn the material for the test, then you take the test. I’m not sure why it makes a difference in a student’s performance if there’s another kid on the other side of the country taking the same test, versus the teacher creating his own, unique test for his class.

Because how is an entity supposed to evaluate the ability of a student or groups of students en masse without testing? I don’t think standardized testing is a necessary special skill to get into college, but like you say, you did well enough on the SAT - a standardized test, to gain entrance into college. Any criticism of the SAT would apply to other standardized tests at grade levels, right?

@Left_Hand_of_Dorkness

One alternative measure that’s being explored is the value-added model. It looks at students’ performance in previous years as a predictor of their growth in the current year, and compares their actual growth to predicted growth. Thus a student who learned 50% of what they should have learned in the past seven years, but who learned 75% of what they should have learned when I taught them, is a mark in my favor, whereas a student who learned 150% of what they should have learned in previous years, but only 125% of what they should have learned in my class, is a mark against me.

Again, I don’t think this is about individual teachers. This is about macro level information. But even in your example, you are creating some quantitative measure - “what they should have learned” and various percentages. How would you measure that? Couldn’t that be applied at a more macro level? What level of students pass a certain test, what the level of improvement should be, vs. what it actually is? A good way to measure that at a macro level is by testing. It sounds like you wouldn’t be opposed if the testing were about gathering data, and using it to guide future analysis. If the data were never revealed about individual teachers, and only schools or districts, would that be better?

@DrCube

Why is this an issue? Was there a problem figuring out which schools were good and which were bad before? (And in the US, people don’t even get to pick which public schools they send their kids to. It depends on where you live, period. So that’s not the issue.)

It’s not like people pick where the live at random. I would expect most parents to assess the schools where they choose to live - I certainly did. How would you go about that assessment if you were moving to a new neighborhood?

[Left_Hand_of_Dorkness]

@Bone

Again, I don’t think this is about individual teachers. This is about macro level information. But even in your example, you are creating some quantitative measure - “what they should have learned” and various percentages. How would you measure that? Couldn’t that be applied at a more macro level? What level of students pass a certain test, what the level of improvement should be, vs. what it actually is? A good way to measure that at a macro level is by testing. It sounds like you wouldn’t be opposed if the testing were about gathering data, and using it to guide future analysis. If the data were never revealed about individual teachers, and only schools or districts, would that be better?

The problem creeps in when the data is used to put pressure on any level of the hierarchy to improve scores on standarized tests, and where this is the only or primary measure of a school’s or district’s success. When such pressure occurs, there are proven methods that can be used to increase test scores at the expense of a useful education (e.g., drilling students in test-taking strategies, eliminating art and music and science and social studies, making all other assessments of students mimic standardized tests in order to familiarize them with the format, etc.).

If scores are not used to apply such pressure, I’d be a lot more okay with them. I’m deeply suspicious of their efficacy, and I’m not convinced it’s humane to subject eight-year-olds to a three-hour test with no water or snack breaks, but at least the lack of such pressure would be a real positive.

The other possibility is to put multiple conflicting measures in place. Standardized tests would stand next to parent and student ratings of their happiness in school, their own ownership of their learning, etc. Administrators, unable to maximize one measure, would need to balance measures.

It’s not like people pick where the live at random. I would expect most parents to assess the schools where they choose to live - I certainly did. How would you go about that assessment if you were moving to a new neighborhood?

I’d ask parents. I’d ask teachers. I’d ask for a tour of the school.

As someone with a decade’s experience in the classroom (if I count student teaching), the last thing I’d do is look at standardized test scores. I genuinely believe they don’t reflect a school’s quality.

[Bone]

@Left_Hand_of_Dorkness

The problem creeps in when the data is used to put pressure on any level of the hierarchy to improve scores on standarized tests, and where this is the only or primary measure of a school’s or district’s success. When such pressure occurs, there are proven methods that can be used to increase test scores at the expense of a useful education (e.g., drilling students in test-taking strategies, eliminating art and music and science and social studies, making all other assessments of students mimic standardized tests in order to familiarize them with the format, etc.).

If scores are not used to apply such pressure, I’d be a lot more okay with them. I’m deeply suspicious of their efficacy, and I’m not convinced it’s humane to subject eight-year-olds to a three-hour test with no water or snack breaks, but at least the lack of such pressure would be a real positive.

The other possibility is to put multiple conflicting measures in place. Standardized tests would stand next to parent and student ratings of their happiness in school, their own ownership of their learning, etc. Administrators, unable to maximize one measure, would need to balance measures.

I don’t have a lot to disagree with here. Except that I’m as concerned with crowding out of other aspects of education in lieu of teaching to the test requirements. If that happens, then those areas needed improvement. If the scores are already at high levels, then the crowding out wouldn’t happen. In either case, students are brought up to standard in their ability to pass the tests. If it’s a choice between proficiency at math vs. learning art and music, I’m picking math every time.

I agree that scores shouldn’t be used to evaluate teachers. I think they should be used in conjunction with other measures to direct analysis regarding schools and districts.

I’d ask parents. I’d ask teachers. I’d ask for a tour of the school.

How does a remote monitoring entity assess the school without tests? I stated up front I think the federal dept of ed should be abolished but if it’s there, then they do need a way to gather data. This would apply generally to any entity charged with monitoring performance at macro levels.

[WhyNot]

Is there a school left in the US without online reviews? That’s where I started when choosing my daughter’s school. Then I got my butt over there to look at the place and get a feel for it and the staff. My state, like most states, also has a “report card” for every school with lots of information, also available online.

How did parents chose a school in 1970?

[Left_Hand_of_Dorkness]

@Bone

I don’t have a lot to disagree with here. Except that I’m as concerned with crowding out of other aspects of education in lieu of teaching to the test requirements. If that happens, then those areas needed improvement.

At the risk of using a bad analogy, imagine a kid is malnourished: poor levels of vitamin A, vitamin C, iron, and vitamin B12. You’ve got a test for vitamin C, but not for any other vitamin. The kid’s diet current contains low levels of all these nutrients. Does it make sense to eliminate the kid’s current sources for vitamins A and B12 and iron so that you can ensure adequate levels of vitamin C?

Because that’s what happens when you test only reading and math. Kids need science and social studies and PE and music and art. Eliminating those because you can’t test them is going to harm kids.

If it’s a choice between proficiency at math vs. learning art and music, I’m picking math every time.

I’m not. Beyond the ideal of a well-rounded person, art and music may keep kids engaged in school who would otherwise check out, and we have a real tendency to undervalue intangibles like engagement. Furthermore, art and music have math embedded in them, often in ways that are pracitcal (learning patterns and spatial relationships, e.g.). Finally, remember that what often happens at schools with poor test scores is faux teaching: you don’t teach “math” so much as you teach “math testing.” Some of what you teach is actual math, but a lot of it is teaching kids what to do when they encounter a multiple choice test.

How does a remote monitoring entity assess the school without tests? I stated up front I think the federal dept of ed should be abolished but if it’s there, then they do need a way to gather data. This would apply generally to any entity charged with monitoring performance at macro levels.

They do, but they need to be humble in their use of data. Right now data is practically worshiped.

My students once took a standardized, nationally normed test that gave me, for each student, five scores: an overall math score, and a score in each of four subcategories (algebraic operations, geometry, patterns, and something else, IIRC). I got a chart that showed me what they thought I should work on next with each kid. Kids at one score level should, they told me, work on concepts like “tens” and “hundreds.” Kids at the next score level should work on concepts like “quintillion.”

It was gibberish. I tried really, really hard to make sense out of it, asked a lot of questions of administration trying to figure out how I could use this data. Finally I started getting glares from administration and being told to make the best of it and to think positively.

Bad data is worse than no data. If my kids, instead of taking that computerized test, had sat down with me to explain their mathematical thinking, I would have learned so much more about them than I learned from having them take this test. The fact that it was nationally normed, completely standardized, is irrelevant if I’m trying to find useful data.