Ideally, however, students should be subject to a wide variety of tests to paint a clearer picture of their proficiency. Norm-based tests provide an idea of how well a student has performed in relation to peers with a similar educational background; subjective tests can reveal creativity or talents not scored on a standardized test.
Q3. The validity coefficient is calculated in terms of whether the content of what is being tested (such as intelligence) is comprehensive enough; if the face or surface validity meets a commonsensical test; if the construct of the test (such as how intelligence is defined) is valid; and if the outcomes of the test are supported by other, existing measures of validity (criterion-based validity).
A reliable test produces the same results on a consistent basis, when it is administered to the same test population. If the same person, or people with similar backgrounds and demographic information, gets the same score on the test, it is judged to be reliable.
A reliable test that aims to test certain concepts should also determine if test-takers score similarly on these individual units: for example, if a test subject gets the same types of answers on the same types of questions that measure a specific content domain, the test has also demonstrated its internal reliability.
A valid test measures what it is supposed to be measuring. Even if an IQ test may be reliable, it may not be valid. For example, some psychologists deny the existence of something known as general intelligence, and believe that intelligence should be understood as multiple intelligences or different types of abilities. For the latter type of psychologist, a test of general intelligence may produce reliable results, but is not seen as valid. Validity is of greater concern in test development, as reliable tests that do not provide a meaningful portrait of the phenomenon they attempt to study are of little use..