![](https://csdnimg.cn/release/download_crawler_static/19928758/bg1.jpg)
Ptwholopr
~n
(he
.S<hook
IYXO.
17,
IYh-?01
USING AND UPDATING LOCAL NORMS
STEPHEN N. ELLIOTT AND
BURKE
H.
BRETZING
’
Arizona
State
University
The most
useful
kind
of
norms are local norms collected by a test interpreter on a
population with which s/he is familiar. We present a rationale
for
the
use
of
local
norms and a procedure, illustrated with an example,
for
updating a norm group’s test
score mean and variance. It was concluded that test interpreters should build local
norms to
be
used in conjunction with norms based on a national sample, since local
norms can be more recent, relevant, and representative
for
local groups.
The most commonly used norms are based on national samples; however, local
norms are often more useful. Both Angoff (1971) and Cronbach (1970) stressed that
whenever possible test interpreters should calculate norms for the local groups with
which they deal. Although local norms have been advocated
in
several measurement
texts (e.g., Cronbach, 1970; Mehrens
&
Lehmann,
1973),
two current assessment texts
(Salvia
&
Ysseldyke, 1978; Sattler, 1974) do not mention their use.
In
addition, we ex-
amined all issues of
Journal
of
School Psychology
and
Psychology
in
the Schools
since
1970, and did not find an article discussing local norms. This surprised us,
for
it seems of
all people who use psychological tests, school psychologists have the most to gain from
using local norms.
School psychologists play a major role
in
programming and placement decisions for
children and are legally and morally responsible for their decisions; therefore, they are
in
need of reliable and meaningful documentation of children’s abilities.
We
believe local
norms for standard measures of intelligence, achievement, and adaptive behavior will
provide test interpreters with as much information as do national norms, and
in
cases
concerning minority children, probably more information. Thus, the purpose of this arti-
cle is to provide a rationale for using local norms and to describe a procedure for up-
dating the mean and variance of a norm group.
A
DeJinirion
of
Local
Norms
Norms are percentiles
or
standard score conversions derived from a distribution of
scores earned by an identified group. These score conversions enable one to make
statements about an individual’s performance
in
comparison to a particular group
in
which s/he is either a member
or
to a group
in
which s/he seeks membership.
“Local”
is
a relative term.
For
example, for educational purposes, one may
think
of
various ways
to
group
or
categorize people, such as by classroom
--*
grade
--.
building
+
school system
--.
county
--L
state
-+
region
-+
or nation.
A
unit occurring earlier
in
this se-
quence is “local”
in
comparison to units occurring later
in
the sequence.
In
this article,
the boundaries for local include classroom, grade, building,
or
school system, and subsets
(i.e., minority groups) within these units.
Requests
for
reprints should
be
sent to Stephen
N.
Elliott,
322L
Payne Hall, Dept.
of
Educational
We
would like to thank David
Krus
and Robert Ceurvorst
for
permission to use their notation and
for-
Psychology, Arizona State University, Tempe,
AZ
85281,
mulas
for
updating means and variances. We also thank Raymond Dean
for
collection
of
data.
196