posted by:
|
Mack McCary
on June 17, 1999
at 6:01PM
|
subject:
|
state tests vs. diagnostic measures
|
Brain observed that the measures teachers use to find out what students know often don't look like state tests, and that teachers report the state tests aren't very helpful. In our district, guess that would depend on the purpose of testing. On the plus side, our state tests in NC, at least in grades 3-8, are criterion referenced, and as surveys of whether the knowledge and skills in the curriculum have been taught, they aren't too bad, especially as measures of the growth of groups of students. However, with younger students, teachers report that the distractors in the multiple choice measures seemed design to "trick" novice students, and they are not always convinced the tests are accurately measuring what they claim to assess. However, as measures of individual student achievement, these state measures are terrible! There is a great deal of error for individuals. We have been working for several years in our district and with others to develop more individual diagnostic measures, as well as increase our use of authentic tasks which call for and demonstrate a greater depth of understanding than multiple choice measures. There are some rather standard, individualized measures we've found useful in reading, such as the John's inventory and Accelerated Reader reports which show a range of literal comprehension. However, math has been more problematic. We have had some luck in using teacher-designed and computer-designed multiple choice assessments, along with released state items, to assess knowledge and skill. When we have used more observational judgments in math, teachers are more comfortable when they are used to inform instructional decisions, less comfortable when those judgments are used in grading decisions. So I believe, Brian, our teachers would tend to agree with those you observed. The state type of measures are useful for making judgments of progress for groups of students, but more individual and observational assessments are needed for judging individual student progress. And those state measures usually tap a level of understanding, namely knowledge and skills, which are necessary but not sufficient to assess whether students understand the math well enough to explain it, make connections, and especially apply it flexibly in realistic contexts. We are very much concerned that the limitations of multiple choice testing are driving teachers towards short term gains at the expense of focusing on developing student understanding of math concepts.
=== Dr. C. E. McCary III Assistant Superintendent for Instructional Services Elizabeth City-Pasquotank Public Schools
|
|