Over on Pharyngula, PZ Myers posted a link to a very sad letter of resignation, written by a first year teacher. The struggles of teachers, both in the public and private schools, are very familiar to me — that letter could easily have been written by someone I know.
In the comments to the post, Alon Levy wrote a comment containing something that I have tremendous issues with.
Once in a while somebody comes and suggests to make graduation contingent on passing a standardized tests, and out come all the idiots who think either that testing is inherently evil or that standardized tests measure anything but how good you are at taking standardized tests.
Nobody likes to be called an idiot, of course, but the cavalier dismissal of the opinions of many educators is flat out wrong. At least he could provide some evidence that standardized testing has both necessary and sufficient value for the purpose he wants to use it.
In response to Alon's statement, I'm posting an essay by an educator, Joan Martin, who has many years of experience. That's practical, front-line experience. I've redacted some identifying information due to some privacy concerns. (Thank you, Joan for permission to reprint!)
In the essay, she's talking about the reasons why this particular school de-emphasizes the standardized tests that are mandated by accreditation association.
I'm sure that parents of third through fifth grade students got wind of the fact that ERB testing was taking place during the month of February. Yes, it was quietly going on in order to fulfill our requirement as an accredited school that, on a a biennial basis, must administer the Comprehensive Testing Program published by the Educational Records Bureau, a subsidiary of the well-known Educational Testing Service. Here, we refer to these tests as the ERBs, derived from the initials of the publishing house rather than the actual test name.
Why don't we advertise when we are administering the tests? We have learned from experience that when parents know about the testing dates, many believe that their children should be differently prepared for school — with a good night's sleep and a big breakfast. One parent actually asked me, "How will I know when to give my child a hearty breakfast and make sure he gets a good night's sleep if I don't know when the ERBs will be administered?" Obviously, children should ALWAYS come to school prepared with a good night's sleep and a healthy breakfast. Also, when parents make a special fuss about preparation for school on ERB days, the message children receive is that the ERBs constitute more important work than the work they do every single day. In fact, the opposite is true. Furthermore, when children sense that these tests somehow actually measure their academic worth, they become very anxious. I've seen test anxiety begin as early as kindergaten and I don't want to see that happen here. Finally, we like to administer the tests on days when the entire class is present so that we don't have to administer makeup tests. (Teachers have a window of time during which they can give the testing in order to conform to the constraints of the norming period.)
What do we say to the children about the ERBs?
We simply tell them that these tests represent a different kind of task that we do, because we are required to do so and that they will encounter this type of task in the future. We tell them that the teachers have to read the instructions exactly as they are printed and that we can't help them sort out the problems the way we usually do. We tell them to do their best work, take their time and check over their work when they are done. We provide them with practice tests prior to the real testing so that they can become acquainted with the directions.
Why don't we make a bigger deal out of these tests? Can't we learn something from them?
I have analyzed ERB tests for almost thirty years wanting to believe that there is meaningful data to be gleaned. My analyses have enabled me to look at every test item to see how our children do compared to their counterparts. I have tracked children's scores from year to year to see if they are gaining, maintaining or dropping their scores. What I have come to understand follows:
■ The tests measure analytical, logical, convergent thinking. They do NOT measure creativity, divergent thinking, in-depth analysis, comparative thinking, spatial intelligence and all the other types of intelligences that we value here.
■ Testing scores are a function of admissions criteria in independent schools. If you screen students for admissions based primarily on their cognitive development, you can be assured that our test scores will differ from a school that enrolls a diverse group of students based on many criteria other than cognitive development.
■ The tests are designed to make small differences among people appear large. Because item content differs from one test to another, even tests that claim to measure the same thing often produce very different results. Because of measurement error, two people with very different scores on one test administration might get the same scores on a second administration.
■ Test taking habits of young children are VERY dicey. We try to give the children test-taking tips such as checking that the number of the question that you are answering is the same as the number you are "filling in" on the answer sheet. (Just last week, despite numerous reminders during the course of each test to check their test booklet number against their answer sheet number, several fifth graders were distraught to see that they had completed all the bubbles on their answer sheet, but still had one item left in the test booklet!) Such errors are more common than not.
■ Testing format is esoteric. Despite training in these strange formats, students are befulddled by the directions. Here is an example. To assess spelling, in the "writing mechanics" section, children are asked to identify the correctly spelled word amongst three words. I fnone of them are spelled correctly, then the child is to fill in the "d" bubble. Here's what a child might see:
- no error
For a child in third grade, the words are all neatly typed, and they ALL say "work," so the child will be apt to fill in the "d" bubble on the answer sheet. An analysis of third grade spelling responses shows strings of "d's" in the class profile. Of course, the real assessment of a child's spelling occurs in his or her daily writing.
Another example in the writing mechanics section follows:
The girl run down the street.
○a. ○b. ○c. ○d.
If there is an error above any of the letters, the child will mark that latter on the answer sheet. If the sentence is correct, the child will mark the bubble next to "d" for "no error." Most third grade students will mark "d" because they see correct beginning and ending punctuation and the word "run" is spelled correctly. In test preparation lessons, they have been encouraged to "say the sentence to themselves" to catch the syntactical errors, but they don't usually do this. Instead they just race on silently to the next test item.
■ If we spent hours on test taking skills and drilled and practiced with the kids to equip them to pay closer attention to test "tricks" such as those shown above, we might be able to elevate children's scores a bit, but we would rob our children of invaluable time from their real work — their reading, writing, mathematical problem solving — and we would be training them for a task that has little life-long value.
Finally, I wish to share a story with you. Many years ago, a mother and father walked into my office waving their fifth grade child's ERB scores in the air saying, "What can we do? Our child did abominably on this test!" They were absolutely furious. I asked them if they would describe their child as compassionate, kind, conscientious, creative, spirited, good-humoured, thoughtful, reflective. They answered "yes" to every adjective, and indeed every adjective was applicable to this child. I asked them if seeing their child's ERB scores made a difference in their understanding of their child. They sheepishly nodded and admitted that the testing scores made them take their evaluation of their child down a peg. I said that their child is still the same child they knew BEFORE he took the test and he would continue to be that very same child as long as they didn't make an issue of his test scores. I promised them that he was going to be an enormously successful student and human being. They said, "Ok, then what should we do with these test scores?" I said, "Rip them up into tiny, teeny pieces and pretend you never saw them." They deposited the test scores, in tiny pieces, in my trash basked and walked out of the office. That child, now in the upper school, is doing extraordinary scholastic work and continues to be the child described above. He is bound for college where he will continue to become more of the remarkable human being he was in fifth grade.
(Typos are mine.) I have a few comments on the essay:
The second bullet refers to independent (i.e. private) schools, with selective admissions criteria. The point can still be generalized to public schools, where neighborhoods tend to self-select and you end up with very different student bodies in the same district.
The second point relates to the spelling issues above. I'm certain that the standardized test fans in the audience will say something like this: "Well, the child should be able to spell 'work' correctly." I won't dispute that I'd be happier if they did than if they didn't. The point that Joan is making, though, is that this is a single data point among thousands (a preschool child has a vocabulary of between 10,000 and 14,000 words) Assessing one, or a dozen, words doesn't give a true picture of the child's development.
There is a fundamental issue that must be addressed before we can decide whether standardized testing as a graduation requirement is appropriate. That is, what do we want children to be like when they graduate? What are the minimum skills and abilities and what is our ideal? Is our goal to equip someone to work behind the counter at McDonalds, or is it to nurture the next Nobel Prize winner? Not everyone can reach the loftiest of goals, but I think that aiming lower hurts everyone, even the worker at McDonalds. Only once we have decided on our goals, can we decide how to measure them. Personally, no matter what the goals are, I don't think that standardized tests are a very good measurement tool.
More on this topic in another post.