The following points highlight the four main characteristics of a good test. The characteristics are: 1. Validity 2. Reliability 3. Objectivity 4. Norms.
Characteristic # 1. Validity:
The first important characteristic of a good test is validity. The test must really measure what it has been designed to measure. Validity is often assessed by exploring how the test scores correspond to some criteria, that is same behaviour, personal accomplishment or characteristic that reflects the attribute that the test designed to gauge.
Assessing the validity of any test requires careful selection of appropriate criterion measure and that reasonable people may disagree as to which criterion measure is best. This is equally true of intelligence test. Reasonable people may disagree as to weather the best criterion measure of intelligence in school grades, teacher ratings or some other measures.
If we are to check on the validity of a test, we must settle on one or more criterion measures of the attribute that the test is designed to test. Once the criterion measures have been identified people scores on the measures can be compared to their scores on the test and the degree of correspondence can be examined for what it tells us about the validity of the test.
Only valid test can give useful information about people but the correction coefficients for validity are never as high as those for reliability. Though we try for reliabilities of 90 or 60, validities which have corrections between test scores and criterion measures are not higher than that of several tests with low but significant validity can sometimes be useful, if they are given together as a battery and their results are considered together. One reason that validity coefficients are lower than reliability coefficient is that the reliability of test sets limits on how valid the test can be.
Characteristic # 2. Reliability:
A good test should be highly reliable. This means that the test should give similar results even though different testers administrate it, different people scores in different forms of the test are given and the same person takes that test at two or more different times. Reliability is usually checked by comparing different sets of scores.
In actual practise, psychological tests are never perfectly reliable. One reason is that changes do occur in individuals over time; for example, a person who scores low in her group at an initial testing may develop new skills that rise her to a higher position in the group at the time of the second testing.
Despite such real changes, the best intelligence test usually yields reliability correlation coefficient of 90 or higher (where 1.00), indicates perfect correspondence and 0.00 indicates number correspondence whatever.
If tests with low reliability are used, their scores should be interpreted with caution. To improve reliability we should ensure that the test is administered and scored by a truly standard procedure. Making the test procedure uniform might make the test more reliable.
Characteristic # 3. Objectivity:
By objectivity of a measuring instrument is meant for the degree to which equally competent users get the same results. This presupposes subjective factor. A test is objective when it makes for the elimination of the scorer’s personal opinion bias judgment. The recognition of the quality objectivity in a test has been largely responsible for the development of an arised and objective type tests.
Objective-based tests measure or evaluate the entire human development in three domains that is cognitive, affective and psychomotor. As the name itself indicates they are based on particular objective of teaching and evaluating. They provide proper direction, and thus streamline the whole process of evaluation. These tests are all comprehensives.
Characteristic # 4. Norms:
In addition to reliability and validity good test needs norms. Norms are sets of score obtained by whom the test is intended. The scores obtained by these groups provide a basic for interpreting any individual score.
To understand why norms are important, let us imagine a test that does not have any, suppose a person takes a newly developed intellectual aptitude test and requires a score of 437 I.D this is a “Good Score”, should the person be happy or unhappy. Obviously a score without any basis for comparison is not very useful.
In fact one of the first things a person in the factious situation might do is seek out others who have taken the test to find out how his or her score compares to theirs. Psychologists do something similar to this when they develop norms. They seek out comparison groups whose performance on the test can serve as standard of comparison for each individual who takes the test later.