Home About Us Reading Education Writing Education Publicly Reported Tests of Reading and Writing Achievement
 
Definitions
Criterion (and standards) referenced tests:
Criterion referenced tests, such as a driver's license test, measure how well a person has learned a specific body of knowledge. Everyone correctly answering a pre-determined number of questions passes.

In education, the percent of students passing criterion referenced tests is determined by the types of questions included in the test, the cut-off criteria (the number of questions the test makers or policy makers decide students need to pass, or to be classified as "basic", "proficient" or "advanced"), and which students take the test (e.g., whether non-native speakers of English still learning English are included or excluded in pass rates in a test administered in English).

Well known criterion-referenced tests are the
  • National Assessment of Educational Progress (NAEP)
  • Texas Assessment of Academic Skills (TAAS).
For more information on criterion (and standards) referenced tests in education see http://www.fairtest.org/facts/csrtests.html
Norm-referenced tests:
Norm-referenced tests are designed to compare and rank test-takers. By definition, they are constructed so that 50 percent of the test takers will be in the top 50 percent and 50 percent will be in the bottom 50 percent, so that 25 percent will be in the top 25 percent and 25 percent will be in the bottom 25 percent. If a norm-referenced test does not do that, the test makers change the questions until it does.

In education, when test makers construct norm-referenced tests, they select which questions to include in the test or exclude from the test according to how many students can answer them correctly, not necessarily according to whether the questions are over content that has been taught.

Results are reported in terms of how test takers do in comparison to the first group of students who took the test. This first group is called the norming group. To be valid, group comparisons should be made between similar students (e.g., the percent of children learning English should be the same in the norming and comparison groups).

Well known norm-referenced achievement tests include the
  • California Achievement Test (CAT)
  • Comprehensive Test of Basic Skills (CTBS), which includes the "Terra Nova"
  • Iowa Test of Basic Skills (ITBS)
  • Massachusetts Comprehensive Assessment System

  • (See http://www.edweek.org/ew/newstory.cfm?slug=42haney.h21 and at http://epaa.asu.edu/epaa/v10n24)
  • Metropolitan Achievement Test (MAT)
  • Stanford Achievement Test (commonly called the Stanford 9 or the SAT 9, pronounced "sat nine"). This test is not the same as the college admissions SAT (pronounced by saying the names of the letters: "S-A-T").
For more information on publicly reported norm-referenced tests see http://www.fairtest.org/facts/nratests.html.
Reading at grade level:
Reading at grade level is reading at an average level for a given grade. For example, a third grade child who is an average third grade reader is reading at grade level.

It is mathematically impossible for all children to read at grade level. When below-average readers read better, the average changes.

It is also mathematically impossible, long-term, for all children to read at or above the 50th percentile or higher on norm-referenced tests (e.g., the SAT 9). Norm-referenced tests are, by definition, constructed so that 50 percent of the test-takers will always be in the bottom 50 percent. When this changes, the test questions are changed so that once again 50 percent of the test-takers will be in the bottom 50 percent.

 
Back to Top | Home | About Us | Reading Education | Writing Education | Publicly Reported Tests of Reading and Writing Achievement