ERIC Identifier: ED410316 Publication Date: 1996-12-00
Author: Bond, Linda A. Source: ERIC Clearinghouse on
Assessment and Evaluation Washington DC.
Norm- and Criterion-Referenced Testing. ERIC/AE Digest.
Tests can be categorized into two major groups: norm-referenced tests and
criterion-referenced tests. These two tests differ in their intended purposes,
the way in which content is selected, and the scoring process which defines how
the test results must be interpreted. This brief paper will describe the
differences between these two types of assessments and explain the most
appropriate uses of each.
The major reason for using a
norm-referenced tests (NRT) is to classify students. NRTs are designed to
highlight achievement differences between and among students to produce a
dependable rank order of students across a continuum of achievement from high
achievers to low achievers (Stiggins, 1994). School systems might want to
classify students in this way so that they can be properly placed in remedial or
gifted programs. These types of tests are also used to help teachers select
students for different ability level reading or mathematics instructional
With norm-referenced tests, a representative group of students is given the
test prior to its availability to the public. The scores of the students who
take the test after publication are then compared to those of the norm group.
Tests such as the California Achievement Test (CTB/McGraw-Hill), the Iowa Test
of Basic Skills (Riverside), and the Metropolitan Achievement Test
(Psychological Corporation) are normed using a national sample of students.
Because norming a test is such an elaborate and expensive process, the norms are
typically used by test publishers for 7 years. All students who take the test
during that seven year period have their scores compared to the original norm
While norm-referenced tests ascertains the rank of students,
criterion-referenced tests (CRTs) determine "...what test takers can do and what
they know, not how they compare to others (Anastasi, 1988, p. 102). CRTs report
how well students are doing relative to a pre-determined performance level on a
specified set of educational goals or outcomes included in the school, district,
or state curriculum.
Educators or policy makers may choose to use a CRT when they wish to see how
well students have learned the knowledge and skills which they are expected to
have mastered. This information may be used as one piece of information to
determine how well the student is learning the desired curriculum and how well
the school is teaching that curriculum.
Both NRTs and CRTs can be standardized. The U.S. Congress, Office of
Technology Assessment (1992) defines a standardized test as one that uses
uniform procedures for administration and scoring in order to assure that the
results from different people are comparable. Any kind of test--from multiple
choice to essays to oral examinations--can be standardized if uniform scoring
and administration are used (p. 165). This means that the comparison of student
scores is possible. Thus, it can be assumed that two students who receive the
identical scores on the same standardized test demonstrate corresponding levels
of performance. Most national, state and district tests are standardized so that
every score can be interpreted in a uniform manner for all students and schools.
SELECTION OF TEST CONTENT
Test content is an important
factor choosing between an NRT test and a CRT test. The content of an NRT test
is selected according to how well it ranks students from high achievers to low.
The content of a CRT test is determined by how well it matches the learning
outcomes deemed most important. Although no test can measure everything of
importance, the content selected for the CRT is selected on the basis of its
significance in the curriculum while that of the NRT is chosen by how well it
discriminates among students.
Any national, state or district test communicates to the public the skills
that students should have acquired as well as the levels of student performance
that are considered satisfactory. Therefore, education officials at any level
should carefully consider content of the test which is selected or developed.
Because of the importance placed upon high scores, the content of a standardized
test can be very influential in the development of a school's curriculum and
standards of excellence.
NRTs have come under attack recently because they traditionally have
purportedly focused on low level, basic skills. This emphasis is in direct
contrast to the recommendations made by the latest research on teaching and
learning which calls for educators to stress the acquisition of conceptual
understanding as well as the application of skills. The National Council of
Teachers of Mathematics (NCTM) has been particularly vocal about this concern.
In an NCTM publication (1991), Romberg (1989) cited that "a recent study of the
six most commonly used commercial achievement tests found that at grade 8, on
average, only 1 percent of the items were problem solving while 77 percent were
computation or estimation" (p. 8).
In order to best prepare their students for the standardized achievement
tests, teachers usually devote much time to teaching the information which is
found on the standardized tests. This is particularly true if the standardized
tests are also used to measure an educator's teaching ability. The result of
this pressure placed upon teachers for their students to perform well on these
tests has resulted in an emphasis on low level skills in the classroom (Corbett
& Wilson, 1991). With curriculum specialists and educational policy makers
alike calling for more attention to higher level skills, these tests may be
driving classroom practice in the opposite direction of educational reform.
As mentioned earlier, a student's
performance on an NRT is interpreted in relation to the performance of a large
group of similar students who took the test when it was first normed. For
example, if a student receives a percentile rank score on the total test of 34,
this means that he or she performed as well or better than 34% of the students
in the norm group. This type of information can useful for deciding whether or
not students need remedial assistance or is a candidate for a gifted program.
However, the score gives little information about what the student actually
knows or can do. The validity of the score in these decision processes depends
on whether or not the content of the NRT matches the knowledge and skills
expected of the students in that particular school system.
It is easier to ensure the match to expected skills with a CRT. CRTs give
detailed information about how well a student has performed on each of the
educational goals or outcomes included on that test. For instance, "... a CRT
score might describe which arithmetic operations a student can perform or the
level of reading difficulty he or she can comprehend" (U.S. Congress, OTA, 1992,
p. 170). As long as the content of the test matches the content that is
considered important to learn, the CRT gives the student, the teacher, and the
parent more information about how much of the valued content has been learned
than an NRT.
Public demands for accountability, and consequently
for high standardized tests scores, are not going to disappear. In 1994,
thirty-one states administered NRTs, while thirty-three states administered
CRTs. Among these states, twenty-two administered both. Only two states rely on
NRTs exclusively, while one state relies exclusively on a CRT. Acknowledging the
recommendations for educational reform and the popularity of standardized tests,
some states are designing tests that "reflect, insofar as possible, what we
believe to be appropriate educational practice" (NCTM, 1991, p.9). In addition
to this, most states also administer other forms of assessment such as a writing
sample, some form of open-ended performance assessment or a portfolio
Before a state can choose what type of standardized test to use, the state
education officials will have to consider if that test meets three standards.
These criteria are whether the assessment strategy(ies) of a particular test
matches the state's educational goals, addresses the content the state wishes to
assess, and allows the kinds of interpretations state education officials wish
to make about student performance. Once they have determined these three things,
the task of choosing between the NRT and CRT will becomes easier.
Anastasi, A. (1988). Psychological Testing. New
York, New York: MacMillan Publishing Company.
Corbett, H.D. & Wilson, B.L. (1991). Testing, Reform and Rebellion.
Norwood, New Jersey: Ablex Publishing Company.
Romberg, T.A., Wilson, L. & Mamphono Khaketla (1991). "The Alignment of
Six Standardized Tests with NCTM Standards", an unpublished paper, University of
Wisconsin-Madison. In Jean Kerr Stenmark (ed; 1991). Mathematics Assessment:
Myths, Models, Good Questions, and Practical Suggestions. The National Council
of Teachers of Mathematics (NCTM)
Stenmark, J.K (ed; 1991). Mathematics Assessment: Myths, Models, Good
Questions, and Practical Suggestions. Edited by. Reston, Virginia: The National
Council of Teachers of Mathematics (NCTM)
Stiggins, R.J. (1994). Student-Centered Classroom Assessment. New York:
U.S. Congress, Office of Technology Assessment (1992). Testing in America's
Schools: Asking the Right Questions. OTA-SET-519 (Washington, D.C.: U.S.
Government Printing Office)
Please note that this site is privately owned and is in no way related
to any Federal agency or ERIC unit. Further, this site is using a
privately owned and located server. This is NOT a government sponsored
or government sanctioned site. ERIC is a Service Mark of the U.S. Government.
This site exists to provide the text of the public domain ERIC Documents
previously produced by ERIC. No new content will ever appear here
that would in any way challenge the ERIC Service Mark of the U.S. Government.