When elementary teachers walk children through a series of contrived lessons only so the children can go home and study for the test, what are they really demonstrating in passing or not passing? In some cases they are demonstrating how little sleep they've had, or how hungry they were when taking the test. In other cases, they're showing how much trivial knowledge they failed to remember so they could regurgitate it for the test.
Pencil and paper assessments are contrived by their very nature, in that they are more concerned with obtaining a grade for the grade book than anything else. Think about the number of times you as a teacher have made the number of questions on the test divisible by 4, 5, 10, 20....or any other factor of 100. I can recall instances where I've added a dumb question on a test simply to make a factor of 100 in the number of questions that I present. Then what of the spelling tests that children take where they try to memorize words. To what end? Oh, and there are usually ten or twenty questions on these tests as well. The number 100 is quite arbitrary in and of itself. Why not 200? Isn't that better than 100?
My point is, in making and presenting "tests" we don't often concern ourselves with truly identifying a child's progress. Does a pencil and paper test a student's progress, and if so at what? In my mind most pencil and paper tests are to test how well a child has memorized a bunch of facts. In some cases, facts have been proven false in many history textbooks. And if they remember all this trivial knowledge does this mean they have made progress?
Children's progress should be assessed and evaluated using real activities. What about a friendly letter they are writing to a child in another school? A teacher can make a nice assessment of student writing skills by reviewing a first draft friendly letter they have written. What about observing a science experiment or scrutinizing a progression of journal observations? What about engaging in a purposeful discussion with a child around a story the teacher and student both have read? Won't this discussion tell the teacher a lot about how the child understood and interpreted the text?
Assessment should be grounded in the everyday activities that the children in the class are performing, provided we're talking about a nontraditional classroom program. Otherwise, forget it. Pencil and paper assessment are designed for traditional classrooms. What else could the children possibly do in this setting other than take a ten question test? I tell my children every day is a test, but not my test. They are testing themselves. What they can do, what they want to be able to do, what they challenge themselves to do, and what they can do that they weren't able to do before. Under the right conditions children put themselves to the test, and teachers are easily able to ascertain progress without at all pulling teeth.