Page 121 - DEDU504_EDUCATIONAL_MEASUREMENT_AND_EVALUATION_ENGLISH

Unit 9 : Test Standardization

2. Develop two or more test items on the same concept, principle, law or generalisation. In Notes
fact when a blueprint is developed on the basis of design of the test, it provides good basis
to prepare two or more items of the same form, using the same concept, testing the same
objective or learning outcomes, as reflected in such a table of specification. This ensures
much better equivalence of items.
3. Another procedure is to use derived scores for establishing comparable forms of tests;
though the complexity of statistical techniques makes it impracticable at this point. There
are widely used derived scores that have constant meaning, whether or not they are obtained
on the same form of the test of from the same pupil group.

9.3 Derivation of Test Norms

Norms are tables of information necessary for interpretation of test scores and are obtained by
giving the particular test to a large and representative sample of pupils in the same grades with
which teachers will use the test. Establishment of norms that furnish reliable and useful basis for
interpretation depends on the extent to which sample used in obtaining the norms is distributed
over a large population in typical school situations and the conditions under which tests are to be
administered are rigidly followed by teachers using the tests. Norms provide the users of a
standardised test a basis for practical interpretation and application of the results. Existence of
norm is the most distinctive feature of standardised tests, though not the only characteristic
feature.
9.3.1 Types of Norms

The form in which norms for a test are provided depends largely on the level in the school
system where the test is used. It is also conditioned by the nature of the test itself. Tests designed
for elementary school grades are usually accompanied by age norms and grade norms and also
sometimes by percentile norms based on grade placement. Tests for use at secondary stage are
more frequently provided by percentile and grade norms only, because the age norms are not
considered useful since growth curve at 16th and 17th years appears to flatten out rapidly.
9.3.2 Grade Norms

These are based on median scores obtained by giving the tests to a large groups of pupils within
each grade. It is a common but not a universal practice to express these norms in terms of end of
the year’s achievement. These norms clearly indicate the period they are designed to cover. They
help in expressing the progress of pupils through grades by converting their raw scores or
standard scores into grade-equivalent scores. If seventh grade end of the year norm of a test was
120 points and the eighth grade end of the year norm is 140 points, then a score of 130 points will
be treated as representing achievement half way through 80 grade or 8.5 grade equivalent. In
most of the tests composed of several parts, raw scores are frequently changed into standard
scores before establishing grade norms (Iowa Language Abilities Test). Raw scores on each subtest
are changed into standard scores. Total score on all the parts of the test is represented by the
median of the several standard scores.
9.3.3 Age Norms
Age norms appear to provide more adequate basis for the interpretation of individual pupil
achievement at elementary school level than is possible with grade norms or percentile grade
norms alone. It involves re-grouping of all pupils used in grade tabulation into chronological
age groups regardless of the grade location or school progress. Test scores of these chronological
age groups are then tabulated and the means or medians computed, which becomes the basis for
setting up tables of scores corresponding to several age groups. Factors like overageness, retardation
and acceleration do influence the average achievement of pupils grouped in grades. For example,

LOVELY PROFESSIONAL UNIVERSITY 115