Page 88 - DEDU504_EDUCATIONAL_MEASUREMENT_AND_EVALUATION_ENGLISH

Educational Measurement and Evaluation

Notes find out their reliability, they are administered to certain group of students or individuals,
and then the scores obtained on these two parallel tests are used to calculate correlation
coefficient (r). This correlation multiple indicates their reliability.
Under this method, individual differences of experience, practice and students do not influence
reliability, because the material in the two tests is different. But if there is much similarity
between the questions and materials of the two tests, then their reliability will be enhanced.
Generally, if two tests are administered with a sufficient time interval, then experience,
memory and other factors do not influence scores. It is suitable to have a time interval of
four weeks between the two tests.
Following points should be kept in view while preparing parallel formats :
(a) The items should be distributed equally from the standpoint of difficulty.
(b) The items should be homogeneous.
(c) The administration and scoring methods of the two tests should be equal.
(d) The number of items in the two formats should be equal.
(e) Type of items, content, difficulty level and samples should be equal in the two formats.
Limitations
(a) Tests have to be conducted twice.
(b) This testing produces the problem of standardization, which in itself is a very complex
and expensive process.
(c) If the time interval between the administration of two tests is longer, then the error
of testing-retesting method will be repeated in this too.
(d) Exercise has its influence in this method, because the form of items is almost equal or
similar.
(e) It is a difficult task to prepare two equal and similar formats for testing.
Merits
(a) This method consumes far less time as compared to testing-retesting method.
(b) This method is the amended form of testing-retesting method.
(c) This has least influence of exercise and memory.
(d) This method can be used for follow-up purpose too.
(e) This method can be used to ascertain reliability of speed test too.
2. Method of Rational Equivalence or K-R Formula : Kuder and Richardson presented a method
to ascertain reliability, which is called method of rational equivalence or K-R formula.
Under this method, the limitations of all other methods have been eliminated. Under it, the
test has to be administered only once, and correlation between question items is found out
in order to see similarity between them. So, the reliability coefficient obtained from this
method is also called coefficient of internal consistency. The chief characteristic of internal
consistence is that different items of the test have high correlation with each other. The chief
assumption of the application of this method is that all items included in this test should be
homogeneous, else reliability coefficient will be less than the split-half method. It entails
that a test has the same type of questions, because the aim of all these questions is to
measure the elements which influence ability and personality.
Under this method, two points are given special importance, these are inter-correlation of
the items and correlation of items with the while test.
By inter-correlation of questions is meant to measure the consistency of reactions in a
subject which occur for different items. That is, how consistent are one statement in relation

82 LOVELY PROFESSIONAL UNIVERSITY