Page 44 - DEDU504_EDUCATIONAL_MEASUREMENT_AND_EVALUATION_ENGLISH
P. 44
Educational Measurement and Evaluation
Notes • Objectivity of a test can be increased by using more objective type test items and the answers
are scored according to model answers provided.
Objective Basedness
• The test should be based on pre-determined objectives.
• The test setter should have definite idea about the objective behind each item.
Comprehensiveness
• The test should cover the whole syllabus.
• Due importance should be given all the relevant learning materials.
• Test should be cover all the anticipated objectives.
Validity
• A said to be valid if it measures what it intends to measure.
There are different types of validity :
• Operational validity
• Predictive validity
• Content validity
• Construct validity
Operational Validity
• A test will have operational validity if the tasks required by the test are sufficient to evaluate
the definite activities or qualities.
Predictive Validity
• A test has predictive validity if scores on it predict future performance
Content Validity
• If the items in the test constitute a representative sample of the total course content to be tested,
the test can be said to have content validity.
Construct Validity
• Construct validity involves explaining the test scores psychologically. A test is interpreted in
terms of numerous research findings.
Reliability
• Reliability of a test refers to the degree of consistency with which it measures what it indented
to measure.
• A test may be reliable but need not be valid. This is because it may yield consistent scores, but
these scores need not be representing what exactly we want measurer.
• A test with high validity has to be reliable also. (the scores will be consistent in both cases)
• Valid test is also a reliable test, but a reliable test may not be a valid one
Different method for determining Reliability
• Test-retest method
• A test is administrated to the same group with short interval. The scores are tabulated and
correlation is calculated. The higher the correlation, the more the reliability.
• Split-half method
• The scores of the odd and even items are taken and the correlation between the two sets of
scores determined.
38 LOVELY PROFESSIONAL UNIVERSITY