Reliability Reliability may be thought of as the repeatability of test results. If a test is reliable a student's scores should not differ markedly when he takes the same test again. Some measurements must be exactly the same each time to be considered reliable, while other measurements are allowed more leeway. For example, a scale reading would not be acceptable at all if it were not the same each time a specified weight was placed on the scale. On the other hand, a student could not be expected to obtain the exact same score on a golf test involving the hitting of balls into numbered circles. In the latter example, too many factors are involved that tend to reduce the probability of identical scores. As noted earlier, reliability and validity are interrelated. In fact, reliability is a necessary part of validity, because if consistent measurements cannot be achieved, the test cannot be considered valid. To illustrate this further, suppose that a physical education teacher wishes to measure leg strength. Using a dynamometer and belt arrangement, he is able to obtain an accurate measure of force exerted by the legs. However, the exact angle at which the leg extension test is performed is difficult to determine. Consequently, although the teacher obtains an accurate (and valid) measure of a particular student's leg lift at a specified angle, he may find considerable variation in his scores on subsequent tests because of the difficulty in establishing the correct angle—resulting in a low reliability coefficient. Furthermore, when he tests a group of students for leg strength, his inability to establish the same angle for everyone results in an invalid measure of the group's leg strength at the desired position. In short in order for a test to be valid it has to be reliable. However, a test can be highly reliable but not valid. You know why …right! Don’t make me slap you! Because for a test to be valid it has to be measuring what it is suppose to measure. Just because every time you give a test you get the same answer, meaning it is reliable, does not mean it is measuring what it is suppose to measure. Conversely, if the test is not reliable it can’t be valid because you are obviously getting different results every time you administer it.
Objectivity This is a simple one. For a test to have high objectivity no matter who administers the test you are going to get the same results. Objectivity of a test pertains primarily to the clarity of the directions for administering and scoring the test. As indicated, high objectivity is obtained when different teachers give a test to the same individuals and obtain approximately the same results. Naturally, the assumption is that the testers are equally competent. Objectivity is a part of reliability, or a form of it. Testing and scoring are two major sources of potential problems regarding the reliability of any test. As with any testing situation, the competency of the tester and the skill and care with which he or she administers the test are determining factors in obtaining reliable and thus valid results. Objectivity is dependent to a large extent on how complete and clear the test instructions are. For example, if a tester was measuring an individuals leg strength on a squat, and the directions are not clear as to the depth of the squat …parallel to the floor or an inch below parallel…there is increased likelihood for error. Moreover, if procedures for scoring are not standardized, some testers, for example in a sprint test, might record scores to the nearest half second, others to the nearest tenth of a second, and so on. Generally the more sophisticated the test the more objectivity presents a problem. For example to evaluate body fat Lange skin calibers are often used. This devise requires that a trained individual use them. At the world championships my body fat was evaluated by four different testers.