Hacker News Reader

Comment by WarOnPrivacy | Hacker News Reader

Twenty years ago I was sharing a conclusion w/ my kids' principal - that preparing for state assessment tests was cheating.

The point of the tests is to evaluate the schools overall effectiveness. As much as test-centric cramming effects those results, it creates a false picture.

Some notes and observations:

Because the state both accommodates and encourages this scenario, the fault can't reasonably be laid on the schools.

A more impeccable (but problematic) way to handle these tests would be for the state to show up with little/no notice and administer it.

I'm personally inclined to scrap state assessment testing altogether. The zero-sum nature of pitting the schools against each other for limited funding consumes a great deal of precious teaching time for little clear benefit.

This conversation happened at our county's first charter school. A few years earlier, this principal happened to be my son's PS teacher.

View on HN

"point of the tests"

The point of the tests should be to determine what topics a student has successfully mastered and which they need to spend more time on. If each school is allowed to come up with their own standard, it will be difficult to tell which schools are lowering standard. Using it as a measure of school effectiveness is a secondary measure that is really not possible without longitudinal data.

They're also useful for evaluating schools and teachers.

Schools on poor areas in general go bad on tests not just because the teachers are bad, there are other socioeconomic variables that influence much more on student grades, like poverty, nutrition, violence, and so on.

Even if you do believe evaluation works the way it is I still think its execution doesn't make any sense. You're taking schools that have bad grades and granting them less money on the next term, what positive effect is this supposed to have?

> Schools on poor areas in general go bad on tests not just because the teachers are bad...

True, but if you see scores consistently jump in one subject year-over-year, that's probably because of a good teacher. For evaluating teachers, relative test scores are more interesting.

As for schools, yeah, it's heard when you're bumping into a lot of external limitations. You almost need to model expected performance and benchmark against that because it's incredibly hard to get mediocre performance out of bad circumstances.

Usually the concept is to reward improvement. If your school helps the disadvantaged students more than other schools do, it gets more funding.

If the test is good, there is nothing wrong with teaching to the test.

If test scores are rewarded, teachers will teach to the test. The obvious solution is to accept the fact that people respond to incentives, and make the test better.

How about this, then: Each school gets a share of the economic benefit it generates. That is, every year after leaving the school, every student is graded by what they paid in taxes, earned in income, contributed to social security, or whatever scheme you can make up. This grade then goes back to the school. This should be arguably the best and most relevant test one can conceive. Do you see anything wrong with that?

Determining how well a school did at teaching people subjects like literature, foreign languages, science, mathematics and history now depends on how many of their graduates became stockbrokers and business executives, not on whether the school actually did a good job at educating their students.

Assessing the quality of education by economic output means the engineer who declines a job at Google or Facebook to take a job at a human rights non-profit is a "worse" engineer, and their education must be considered substandard compared to one who goes and takes a Big Tech job. Same with doctors who volunteer to run free HIV clinics in sub-Saharan Africa when they could get a job doing buttlifts for movie stars in Los Angeles, or lawyers who become public defenders or represent asylum seekers and other public interest clients rather than doing extremely well-compensated corporate work.

In as much as it reduces all of human life and values to numbers on a spreadsheet, yes, there is almost certainly something wrong with it.

yes. High income is more to do with where you are born, not how well you did in school. though there may be a gigantic gray area, this holds true in general. You will be throwing money where it's needed the least and will help the least.

Yes, it's quite problematic.

First, it's obviously not a fair measure; tax contributed is not a direct reflection of education quality. Parents, background, accidents, moving abroad... You can't possibly clean the statistics enough to get rid of all those factors.

Second, it's not the business of the school to know my tax information and employment status. You can't measure something based on a number that you can't know.

Reducing humans to hedonic consumption machines where life satisfaction is measured by consumption and value is measured by production is easy and terrible.

What is a better test?

To me this implies more material will be tested, to the point where the test dictates the course material.

Allow me to explain one reason to have standardised tests. My last teaching job before retiring was in a school where the administration expected 75% of the students to pass Yr 9 Maths. The Middle School Head had come to my class a number of times to help the students, so she certainly knew what their ability was. It would have been no surprise to her that, rather than 75% passing the test, 75% failed badly. Yet, she was not prepared to accept those results. She pointed out that one particularly disruptive student had passed the previous year. Fortunately, I could point to the standardised test that showed that that particular student was in the bottom 10% in the state. My implication was that his pass in the previous year was fabricated. I was instructed not to hand back any of the students' test results, at that time. I left the school shortly after and have no idea how they would get 75% of the students to pass. But I did know that the standardised results would be a truer indication of that class's Maths ability, than the school's results would be.

What is the expected benefit of punishing schools by underfunding them if they get beaten by other schools’ test scores?

Have school districts not yet discovered that competition applied to learning is counterproductive?

Personally, I feel we have reached the limits of what schools can do with the current structures, that there is very little improvement that can be mined.

The things we do to maximize fiscal efficiency (grouping by age, school start times during optimal sleep periods) aren't in the students' best interests. The OP's article seems to bear witness of that.

I'm not sure what it would look like, but I'd like to see kids a few years older teaching younger kids. You get a much better understanding of the subject when you have to teach it.

In the early 70s I was part of an experiment that did just that. We were one very large class of grades 1-6. Older kids were assigned younger ones to mentor.

I was in it for two years; it was a really good experience. As for the outcome, all I know it wasn't broadly adopted.

Not sure about the UK, but in the US, it's the opposite. Failing schools get more funding, at least temporarily, with the expectation they turn the scores around in a few years. They risk losing accreditation if they don't improve, but that's usually years away.

In Maine for example, when the boards are looking into renewing a charter school's license, their test results on the state exam is heavily weighed. Schools that are not performing well on the tests may get their charter revoked. So the benefit is saving children from incompetent management.

My town's school department had 5 elementary schools. During 4th grade, my elementary school which I lived closest to was underperforming in state evaluation tests. This meant that they allowed anybody attending it to transfer over to any other elementary school. I opted for this option and was able to re-attend my original elementary school before I had moved.

My (charter) high school consistently performed in the higher ranks for SAT & state tests. They didn't have any classes to prepare us for them, although that probably would have been useful from a college admissions perspective. If a school is performing really badly it's probably that their curriculum is flawed or that management isn't doing a good job. Competition applied to learning helps light a fire under people doing a bad job running a school.

I believe it is district funding that suffers for low grades and that it takes a few years of consistently low performance before the state applies any action. I should note that our schools are primarily funded from county taxes but state funding is enough to matter.

However, the testing dynamic tends to nurture a perverse incentive for counties to showcase their high performing schools and increase funding to them. Some school boards resist that temptation better than others.

Reply on news.ycombinator.com