I learned something new at school today

I wouldn't teach my kids mindless complicity for the sake of complying. I would teach them to make decisions based on their convictions and interests.

You don't learn about what he knows. You learn about what has been crammed into his short term memory without context.

I would teach them to make decisions based on their convictions and interests.

The objection is that testing is not a valid scientific measurement of education, or of anything meaningful that resembles education. The results are bogus.

In this way, you are mistaken to compare testing to a valid measurement like temperature. When you measure temperature, it has a well known and relevant scientific meaning. Something that is measured to have a high temperature is always "hot". A temperature reading can reliably indicate how well a given process heats a system.

Education testing is not anything like this. People can be well educated and do poorly on the test. People can be poorly educated and do well on the test. These tests don't indicate how well a process is serving a group of children.

Data for data's sake is worse than meaningless. It is misleading. The process that should be focused on what is best for the student's education instead is focusing on optimizing on a bogus metric.

The idea that educational testing represents some magic "thermometer" that tells us when students are cooked, ummm educated, isn't science.

A study of math scores on the 1992 National Assessment of Educational Progress found that the combination of four variables unrelated to instruction (number of parents living at home, parents’ educational background, type of community [e.g., “disadvantaged urban,” “extreme rural”], and state poverty rate) explained a whopping 89 percent of the differences in state scores. In fact, one of those variables, the number of students who had one parent living at home, accounted for 71 percent of the variance all by itself.

Sure it does. But when a concept is taught out of any context it leaves the short term memory before any real connections are made.

But the test was to study math scores of the students at the school in order to make funding decisions.

Should schools be punished for the demographics of the parent population?

Data like this is not misleading, it merely has other variables that have to be accounted for. Real life data is often like this. I am trying to measure how A impacts B, but C, D and E all also impact B, so I control for what I can and I randomize the rest.

The basic point is whether the tests are a good measure for education or a beneficial part of the educational process. I strongly argue that no, they are neither.

The task force’s recommendations focused on the above generic form as the best way to measure graduation, completion, and dropout rates in terms of cohort-based, cumulative methods; however, for each specific category, there were a number of alternative considerations NCES raised in order to keep “perverse incentives” from causing schools to manipulate graduation, completion, and transfer rates to artificially lower dropout rates. For example, the NCES realized that if schools were transferring students who were at a high risk for dropping out to adult or alternative educational programs, the methodology would not capture these transfers in the graduation rate.23 Therefore, the task force proposed a more “harsh” alternate methodology for measuring dropouts, but one that would seem to eliminate these perverse incentives to transfer failing students, with P being the size of the cohort and Y being a given academic year:



Another widely used methodology that came to popularity in 2003 amid the cries for statistical reform was the Cumulative Promotion Index (CPI), created by Christopher Swanson, then of the Urban Institute. The CPI “approximates the probability that a student entering the 9th grade will complete high school on time with a regular diploma.”24 Swanson takes the number of 10th graders in one year and compares that figure to the number of 9th graders the previous year, thereby calculating a percentage of those who were promoted to the next grade. This is repeated for grades 10th to 11th, from 11th to 12th, and from 12th to graduates in order to come up with a ration that approximates likelihood of graduation. According to this CPI methodology, the State of New York should have reported a graduation figure of 61% in 2003 instead of the 76% that was actually reported by the State.25

Quote:

Data like this is not misleading, it merely has other variables that have to be accounted for. Real life data is often like this. I am trying to measure how A impacts B, but C, D and E all also impact B, so I control for what I can and I randomize the rest.

Come on Engineer! You are arguing that higher test results mean higher test results. It is easy, yet quite bogus, to use a test to validate itself.

The question is whether these tests are a measure of meaningful education. Your obfuscations about As and Bs and Basketball players have nothing to do with this question. Basketball is a great example of one of my points, the object of basketball is clear-- the team with the highest score wins. Test scores are nothing like this. Test scores don't have any meaning outside of a meaningless test.

Basketball scores are the point of the game. Making test scores the point of education is exactly what I object so strongly to.

Let's talk about the issues at hand without obfuscation. Here are my hypotheses.

1. I assert (and open to discussion) that standardized test results do not accurately measure education in any meaningful way.

2. I assert that standardized tests are harmful to the process of good education. As evidence I submit that a clear majority of teachers will tell you as much.

I could imagine a study that compared how some aspect of performance in life, be it educational level, or salary or something else differed between tested students (or students from tested districts) with non-tested students. This would be a valid and interesting study.

Unfortunately, I don't think any such study has been done meaning that as far as the effectiveness of testing to do anything beyond raising test scores , your emotional arguments and anecdotal stories are no better then mine.

The positive impacts of these tests are circumstantial at best.

However, I can prove to you the negative effects of these tests. When I was teaching high school we had test preparation workshops that told us teachers how to raise test scores. We were instructed to focus on the mundane. This wasn't a mild distraction, this significantly took away from other educational goals, like teaching critical thinking skills or probing deeper for more understanding.

The fact that educators, who are the professionals who are actually responsible for education of our children, are widely opposed to testing should also tell you something.