Testing of Children

This was written on September 2, 2003 in response to a discussion of testing an the "No Child Left Behind" act:

I agree with this, but have to add some observations. About ten years ago when I was teaching I had a book that came with a "test bank" of questions. As a teacher I did not need to write test, I could just use the test bank to generate the tests. Once I was in a hurry, and I decided to take them at their word, and generated a test. Of course, I never used an answer key on the grounds that if I couldn't answer the questions off the top of my head, then I should expect the students to be able to answer the questions. After the test I went through the questions to check the tests. Each question was multiple choice with 5 possible answers. As a rule I found that three of the possible answers could be easily justified. I didn't use it again.

The truth is that different people do better on different tests. I gave a test once in three parts, each with slightly different types of questions. The expectation was that one section was easy, and one medium, and one hard. The problem is that some students did well on the easy part, and washed out on the "hard" part. Others did well on the "hard" part, but washed out totally on the easy stuff. After I graded the test it was really obvious the results were really arbitrary.

I read the story of a test given in New Mexico, I believe. One question was, "When do the Monarch butterflies arrive?" The student answered "October," as he saw when they arrived. However, the text was written in New England, so the answer sheet said they arrived in March or April. The question was marked incorrect, and when the student complained the teacher stuck to her guns.

This last example shows the real problem with testing. People substitute memorizing for learning. The people who are rewarded are the least creative individuals, the people who cannot see more than one answer to a question. This makes for good automatons when the students grow up, but don't expect any unexpected problems solved.

(When I was in North Dakota I gave a test with a "true-false" question on it, and "justify your answer." Many people guessed correctly, but got almost no credit because they could not justify their answer. One student gave the "wrong" answer, but had a very good justification, pointing out points I hadn't thought of, so he got full credit. The students quit pushing me to have true-false questions on the tests.)

There is also the cultural and linguistic problems. For example, I was talking with my daughter last Sunday. She has been studying election methods in school, and she told me Instant Runoff voting isn't that good. After we discussed it a bit, it became obvious that what everyone else calls "Instant Runoff voting" her class called "Sequential Runoff." Her class used the term "Instant Runoff" to mean an election where the top two candidates are pitted against each other. This is a slight difference in terminology, but if a standardized test were given covering election methods, either she would get the wrong answers or I would, not because of differences in understanding, but because of a slight difference in language.

Other differences exist as well. For example, when do you eat dinner? Depending on where you are at, it can mean at noon, or at 6:00, or sometimes it floats depending on where you are at. Milk Shake is another such word. This assumes everyone speaks English. Suppose the person speaks Spanish. I lived in Texas where the only time the students heard English was when they were forced to use it in school. Most students couldn't really express any complex ideas as a result - that is, they didn't have the experience in English, and they had never really studied Spanish, so they couldn't speak either language adequately.

And I haven't even started on the real differences. When I was growing up my parents had no access to Shakespeare's plays. We had no TV, and could not afford to go to plays. Was I less acceptable than a rich person who could go to the plays, with biweekly trips to the library. Most city folks don't know how to milk a cow. (I think it would be interesting watching them try. I suspect most would get it over time, but the first half hour of trying would be fun.)

All this must be seen in the backdrop of the real reason for this "No Child Left Behind" initiative. That is, an attempt by the Republicans to do away with the concept of public schools, in favor of a voucher program. To accomplish this the test don't really have to be very good.