Page 124 - DEDU504_EDUCATIONAL_MEASUREMENT_AND_EVALUATION_ENGLISH

Educational Measurement and Evaluation

Notes 9.4 Establishment of Final Validity and Reliability

Once the test is ready and given to representative group of pupils and norms are derived on the
basis of scores, final step is to establish validity and reliability of the test. Though preliminary
steps taken during construction of test should make reasonably certain to satisfy these criteria, as
a final check it is essential to report to users of the test the validity and reliability of these test to
enable them to evaluate it.
9.4.1 Validity of the Test

Validity of the test depends on the efficiency with which it measures what it attempts to measure,
i.e., it serves the purpose the user has in mind. Tests cannot be considered as valid in general
terms, but only in connection with their intended use and at the intended ability level of the
pupils. Three types of validity are : (a) curricular validity, (b) statistical validity, and psychological
and logical validity.
9.4.2 Curricular Validity
Basis for determining curricular validity are :
(i) Adequate sampling of content or course of study
(ii) Reflection of instructional objectives or outcomes of learning
(iii) Recommendations of committees, subject and test specialists.

9.4.3 Statistical Validity
It is calculated by determining correlation between test scores with criterion measure, which is
considered as measurement standard. Commonly used methods are :
(i) Correlation with school marks

(ii) Correlation with ratings of expert judges
(iii) Correlation with other known concurrent measures, (teachers’ ratings etc.)
(iv) Correlation with measures of future outcomes (predictive).
9.4.4 Psychological and Logical Validity
In complex fields such as language and reading and study-skills, subjects are made up of many
interrelated abilities, in contrast to practical-skill areas where tested performance is either an
exact representation of, or a very similar substitute for, the instructional outcomes sought. Analysis
of desired learning outcomes (Intended outcomes) and of the proposed test using logical and
psychological methods can reveal sufficient commonality or similarity to justify that the test is a
valid measure of the intended outcomes. This method is followed quite frequently in complex
fields like language and reading skills or study skills.

9.4.5 Reliability of Test
A test is said to be reliable if it functions consistently or the efficiency with which the test
measures what it does measure. What it does test or what it attempts to measure is the ‘validity’
of the test. How efficiently does a test measure What the user attempts to test (whatever it
measures) is the reliability of the test. A test may be reliable without being valid, but it cannot
be valid unless it is reliable. Thus reliability is really an aspect or a phase of validity. Reliability
has two aspects : adequacy and objectivity. It is usually expressed in terms of coefficient of
correlation. It is internal consistency or self-consistency, which is evaluated for estimating the
reliability coefficient. The following methods or statistical procedures are used in obtaining
various coefficients.

118 LOVELY PROFESSIONAL UNIVERSITY