[X] HOME PAGE (Main Menu)


The two most common types of testing used by educators are norm-referenced tests and criteria-based tests. As the following discussion explains, each testing method has its preferred environment.

Norm-Referenced Testing
"Norm-referenced tests" are based on the performance of a sample of a population of students. For example, a random sample of the nation's 2nd grade students are given a spelling test of 100 words. The test includes very easy words, typical 2nd-grade words, and more difficult words. Naturally, some students will do better than others on such a test. The test results are then norm-referenced by finding the mid-point or 50th percentile point. Students spelling above the mid-point are considered to be "at or above grade level". The other half of the students are rated "below 2nd grade level". Nationally norm-referenced tests allows one to compare the academic skills of students in different states or different regions of the country.

Naturally, all teachers would like to have all of their students with spelling skills at or above grade level. An obvious way to improve spelling skills is to spend more time teaching the subject. If this were to start happening throughout the whole nation, the country's 2nd grade students would gradually become better spellers. That is, the number of students "at or above grade level" would increase to significantly more than 50 percent. When this happens, the nation's norm-referenced test is no longer valid and will have to be recalibrated. After recalibration, the number of correctly spelled words at the mid-point would be higher than before. But, the percent of students at or above the new "grade level" will again be set at 50 percent. If a norm-referenced test is valid, only 50 percent of the target student population will ever be performing above grade level.

Incidentally, the federal No Child Left Behind Act says that one of its goals is to have all children achieving at or above "grade level". However, in this instance, the tern "grade level" is to be defined by the individual states.

Norm-referenced test results from a single classroom will tell the teacher how far each student is above or below grade level. If students scoring below grade level are no further down than, for example, the 45th percentile, this would be a fairly good showing for an intercity classroom. Again, not all students in all schools can be above average.

While nationally norm-referenced tests provide a comparison of academic skills of students in different states or different areas of the country, norm-referenced tests are not suitable for determining how well students have learned Ohio's academic content standards. This is done by using criteria-based testing where test questions focus on only material defined in the academic content standards.

Criteria-Based Testing
"Criteria-based testing" involves testing students against a predefined academic criteria. For example, students are given a list of 100 spelling words to study. Later, when the students are tested, only words in the 100-word list will be included on the test. If all students have learned how to spell all 100 words, they would spell all of the words on the test correctly. And, when this happens, no observer would say the test is invalid. But, all of these same students would not be expected to be in the 100th percentile on a norm-referenced spelling test.

The state-wide achievement tests used in Ohio are almost criteria-based tests. They are based on Ohio's academic content standards. If all students in Ohio learn both (a) the concepts defined in the various academic content standards and (b) the test designers used only grade-level word meanings, one would expect all students to pass the various state-wide tests.

[X] HOME PAGE (Main Menu)

Bill Buckel
1641 Hess Blvd.
Columbus, OH 43212
...Ph. 614--488-8963

Notice: This web site does not have an e-mail return address.
URL = http://www.billbuckel.com
This page last updated: February 8, 2012