From a 'Card-Carrying Psychometrician'
I wish that I were handsome. I also wish that I could afford to purchase three-quarter-page advertisements in issues of Education Week to give educators my perspectives on current, controversial topics involving education.
But I'm not necessarily photogenic like Keith Geiger, the president of the National Education Association, nor do I have the financial backing of a large organization like the N.E.A. However, I do pay attention to the key educational issues of the day, and I enjoy reading Mr. Geiger's analyses. Nevertheless, I am sometimes overcome by a troubling feeling when I read his display advertisement/analyses that appear in Education Week and other publications; it has only been recently that I have been able to pinpoint the reason for those feelings.
You see, I am a psychometrician. O.K., I'll state it plainly--I help to make tests. So I am understandably a little sensitive when criticisms are leveled at my vocation. Sure, some of the criticisms are valid (such as, poorly constructed tests measure only lower-order skills), but other criticisms are downright dopey (like, multiple-choice tests caused the banality of the 1988 Presidential campaign). On balance, I think that some educators appreciate the information that tests can provide, some are virulently opposed to all testing, and a lot are simply ambivalent or want more information.
But I can't figure out Mr. Geiger, who appears to both embrace and abhor testing. For example, in a June 6, 1990, advertisement/essay in Education Week, he appears to side with the "anti-testers.'' In that article, he reports favorably on one teacher's proposal to eliminate "the mandatory, standardized tests I give to my students.'' He also comments on the work of the National Commission on Testing and Public Policy, which emphasized the social injustice of current tests, their cultural bias, their failure to measure educationally important outcomes, and their assessment of trivial, proxy skills. Pretty strong stuff.
About four months after that article appeared, Mr. Geiger made another declaration, this time concerning class size (Education Week, Oct. 17, 1990). He reported on a "landmark study'' costing $12 million that "produced compelling evidence that in the critical early years reducing class size produces significantly higher achievement.'' He called the achievement gains of the Tennessee schoolchildren "impressive.''
Now here's what I can't figure out. How were the gains in achievement measured in the study that Mr. Geiger wholeheartedly endorsed? Well, it appears that the Stanford Achievement Test was used to assess the impressive gains in both math and reading.
What! How could the Stanford--that loathsome, standardized, ne'er-do-well, multiple-choice, lower-order, mother-of-all culturally biased tests--be deemed suitable for assessing the achievement of the Tennessee elementary-school children? If we believe Mr. Geiger's earlier article in which tests--including the Stanford--are flogged for being culturally biased, lower-order, etc., then it seems to me that reasonable people should have hoped for the Tennessee kids to perform worse on them when class size was reduced and better instruction was provided. We surely would not have wished for them to improve their performance on ostensibly trivial, biased, low-level tasks.
Of course, maybe I'm being overly critical. It is certainly possible to have a change of heart about the value of tests like the Stanford. I am certainly not one to cast stones, having hated standardized tests as an adolescent, and now finding myself to be a card-carrying psychometrician. But now it has happened again. Just when we might have thought it was safe to take another swim in the item pool, Mr. Geiger has pronounced that tests are, again, dangerous to our intellectual health. In the Dec. 11, 1991, issue of Education Week, we learn of his opinion that tests like the Stanford are not to be trusted. They "have little or no educational purpose ... [and] ... they "do nothing to improve teaching or learning,'' he avers. Tests like the Stanford "merely inform schools how well their students can regurgitate a given set of knowledge ... [and] ... waste precious instructional time,'' Mr. Geiger opines. He concludes that, ultimately, the tests "cause enormous student anxiety--for no good purpose.''
So I guess that the reason I'm troubled is that I can't grasp the paradox. Are testing specialists just being used? Are standardized tests good for describing the fullness of math and reading achievement when they point to smaller class size, but bad for other applications? And who decides when the information they provide is "impressive'' and when the information reveals social injustice or educational irrelevance? One hypothesis to explain the contradictions is based on geographical criteria: It may simply be that it is all right to use the deficient Stanford in Tennessee, but not in Michigan. Or, maybe there is a grade-level interaction: O.K. for elementary-school students, but nix on the high-school use. Perhaps it is only a mild case of wanting to bash the testing pie and eat it, too.
Personally, I'm not persuaded by any of these hypotheses. Like Mr. Geiger, I think that the use of standardized tests is a perplexing issue. I don't for a minute believe Gerald Bracey's churlish report that Mr. Geiger recently said, "I don't know anything about testing. I'm really better off not knowing anything about testing.'' (See Education Week, Dec.11, 1991, Letters.) I think that we could all benefit from more information about the proper role, construction, use, and interpretation of tests, and we should welcome new approaches to sound assessment practice.
So I have one suggestion. Perhaps the N.E.A. could establish a task force to study the issue. The report of the findings might even make a good topic for a future advertisement in Education Week. Such an investigation would not benefit me personally; I probably won't get any more handsome or any wealthier. But I think I know a lot of students, teachers, administrators, and parents--consumers and users of tests--who would welcome a more consistent message on the appropriateness of standardized tests and the believability of the conclusions that are drawn from their results.
Gregory J. Cizek is assistant professor of measurement and
educational research at the University of Toledo.