Friday, February 11, 2011

Scores from Reading Tests

“In this era of increased testing and expanding high stakes accountability systems, we need to remember the purpose for assessment. We want our schools to improve, and for this to happen, we have to do better at helping kids learn. Some of the tests teachers administer cannot help them much in this effort. Standardized measures (like those administered by states) and the outcome measures required under the No Child Left Behind law fall into this category. They are designed more to measure student achievement levels than to guide classroom instruction” (Santi, York, Foorman, & Francis, 2010, p. 1).

General Weaknesses of Standardized Reading Tests
But can individually administered standardized tests that compare children to one another—called norm-referenced tests—tell parents and teachers what children can and cannot easily read and what tasks and level of instruction will challenge but not frustrate them? Despite publishers’ claims, generally no. Here are two of many reasons:

  • Norm referenced tests usually have children complete tasks that differ greatly from what they’re asked to do in school or in real life situations. Rather than having children spell words, they’re asked to circle the correctly spelled words. Recognizing a word is easier than spelling one. And just because children can recognize the correct spelling of a word does not mean they can reproduce it from memory.
  • Norm-referenced tests usually have children complete tasks that are much shorter than those they have to do in school or in real life situations. On a test, children may be asked to answer questions after reading paragraphs that average only 30 or so words. In school, they have to read and comprehend much longer selections, selections that may average several hundred words. In line with this, here’s are some of the concerns and recommendations that Dr. Antonia D’Onofrio and I published about a widely used standardized test that we thought was pretty good: “Like all standardized tests, however, [The Woodcock Johnson-Diagnostic Reading Battery-III (WJ-III-DRB)] should not be used alone. It needs to be supplemented by other quality measures of reading ability, especially measures that more accurately reflect the reading demands placed on students (e.g., the need to comprehend lengthier, more complex reading materials)…. Together with tests that more closely approximate typical reading activities, a knowledgeable, informed examiner—who engages in diagnostic teaching, who observes the subjects in different instructional situations, who is familiar with the local reading curriculum and the skills and orientations of the instructors, who arranges to have the subject’s reading progress carefully monitored—should be able to use the WJ-III-DRB to gather the information needed to help many students improve their reading abilities. … But like a carpenter who needs far more than a hammer to build a house, examiners will need far more than the WJ-III-DRB to fully understand a student’s reading abilities” (Margolis & D’Onofrio, 2007, p. 871)

Instructional Utility of Standardized Reading Tests
One of the most important reasons for testing reading is to determine what the child needs to learn to advance. Here too, standardized tests are generally inadequate. As Carnine and his colleagues wrote:

  • An even greater disadvantage of norm-referenced tests is that they provide little information about how to instruct a student. That is, a percentile or grade equivalent score does not really indicate where to place a student in a commercial program or what a student’s skill deficits are. (Carnine, Silbert, & Kameenui, 1997, p. 43)

The Best Test
So, which test is best? Which are good?

It always depends on your purpose—what questions you want answered. (See chapters 4 and 5 of Reading Disabilities: Beating the Odds, www.reading2008.com). Moreover, the value of test scores always depends on the quality of the evaluators:

  • Too often examiners forget the dictum that ‘tests don’t diagnose, people do’ and base their diagnoses exclusively on test results, a hazardous enterprise at best. Test results are merely observations, not diagnoses. They specify a performance level at a given time under a particular situation, but they do not tell the examiner why a person performed as he or she did…. The questions concerning the why of test performance are the very essence of diagnosis, and they can be answered only by an insightful, competent test examiner. Test results make useful contributions to a diagnosis, but in the end, practical diagnoses rest on the clinical skills and experience of examiners. Test results are merely aids to clinical judgment. (Hammill & Newcomer, 1997, p. 40)

Further Information
For information on using reading evaluations to help your child, see chapter 5 of Reading Disabilities: Beating the Odds (www.reading2008.com).

No comments:

Post a Comment