I think that only makes sense if the output of the evaluation is the end state and not the learning of the student. Ideally teaching and evaluating (not grading) is baked into one action. A teacher works closely with a student, assists them in their studies, judges the amount they've learned (evaluates) and where they're struggling, and helps them with their weak points. Done. The standardized tests are then given (if they must), and the teacher is allowed to look over the results to judge from an external evaluation how well the student did and learn from it.
More or less. I'm pointing out that incentives are not set as well as they could be, without specifying in great detail how they should be set.
Formal end-of-term type grades do not directly serve the students. They are used by others to rank the students and decide who among them are useful for a given purpose (jobs, college admissions). Students are still going to want these so they can signal to employers that they are valuable.
The grades are incentives that the teacher can use to control students. The intended use is of course to get them to learn the material. One point of the article is that they are also easily used for other types of control.
Because teachers are to some extent judged by how well their students do, there is often pressure for them to give better grades. Occasionally there is some inflexible rubric of X% of each class must get grade Y. There is no direct pressure that these grades should correspond at all to mastery of material.
Grades on e.g. homework assignments and quizzes can also serve to help the students (or their parents) gauge their progress and find weak areas so they can better learn the material.
Most schools in the U.S. make the final grades incorporate these earlier grades. This means there is pressure on these grades as well making them less useful for the purpose of helping the student learn better. This also means that they are useful for control, because students care about their final grades.
Suppose instead that the final grades were not at all set by teachers. Imagine something like the AP tests. Because the homework and quiz grades would not factor in, they are less useful as means of control, and theri only use is now monitoring and helping he student learn. Because the testers are not teachers, there is much less incentive to score the students highly. They can measure mastery with much less bias.
I think that only makes sense if the output of the evaluation is the end state and not the learning of the student. Ideally teaching and evaluating (not grading) is baked into one action. A teacher works closely with a student, assists them in their studies, judges the amount they've learned (evaluates) and where they're struggling, and helps them with their weak points. Done. The standardized tests are then given (if they must), and the teacher is allowed to look over the results to judge from an external evaluation how well the student did and learn from it.
Is this different from what you're saying?