If we consider for example, a psychologist has made mechanical aptitude test and then it is discovered later that, persons getting higher scores on the test are nearly always found to be good mechanics. Whereas the ones who were getting lower scores on it are always found to be unsuccessful. In this case the psychologist might call the test valid. However it takes quite an amount of time to determine, the validity of a test.
The applicant at first must be tested, that is to say his test score must be determined; prior to his being hired or prior to being put to work, on a mechanical task. After some period of time his performance on the job will be measured, and then a comparative study will be made to determine whether the applicants who have scored higher on the test, can really do justice to their post or are clearly the ones who can do best on their job. Also the individuals who made poor scores on the test are also crossed checked, whether they are actually poor in their performance or not.
Validity of an online aptitude test is expressed in the form of coefficient of correlation in which, the test scores obtained are correlated with some criteria of performance. For example, the validity of intelligence test can be determined by coordinating the test score with the Ss’ marks in the examinations, all the failures in examination, their teachers’ ratings regarding their intelligence and various other such criteria.
Quite a number of statistical procedures for the determination of the correlation are available. However it should be known that this is a method of determination of the correspondence, between two sets of scores. For example, when we want to find out a validity coefficient, we actually try to correlate the test scores with some others scores.
Provided that the other set of scores should be held by the same group of subjects, on some other criteria of actual success the person have in the behaviours, which can be measured by their test scores. For instance, if all the higher scorers on a test of intelligence also score high in their examination averages, those scoring intermediate scores on the exam also score intermediate scoring averages in their examination averages. Furthermore, if the poor scorers on the test also scores poor marks in the examination averages, it can be concluded that there is a high correlation or correspondence between the test scores and the examination averages.
This was however a simple example of the many criteria which has been pointed out. Those many criteria can be legitimately used for validating a test. When correlation techniques are used, such a correlation is expressed in the numerical terms. As for example, if there is a very high correspondence, the coefficient of correlation will also be very high. The coefficient of correlation will be as high as +1, which corresponds to an ideal and perfect correlation between the test scores and the criteria scores. However such a case is quite rare.
