Content validity outlines the straight correspondence of the contents of the test tasks to the trait or particularity that is measured. It is an “inspection” of the content of the test made in order to determinate this correspondence. To make the content validity high the test is required to “embrace” all the aspects of the phenomena under study in the right proportion. Predictive validity works on the possibility to “judge” the aspect of the behavior the psychologist is particularly interested in a client according to the results of the test. In order to define this validity in the process of test-execution it is necessary to correlate the contents of the test to the criterions it is supposed to predict theoretically. In different cases this validities may substitute each other. For instance, we can use it in case of working with a test that deals with the examination of a wide range of characteristics similar to MMPI.
As we need to define the test’s ability to measure every single trait it is supposed to measure, the correct choice of the target group and so on this type of validity becomes the best solution. In other words it may be used in order to determine the correspondence of the test to the name it carries and the trait it measures. This type of validity deals with observable constructs and is one of the widespread validities due to its effectiveness in eliminating non- eligible tests. Predictive validity comes into play when it is necessary to theorize. For instance, it possible to claim that a measure of creative ability should be able to predict how well a man will do in an art-based profession. The scores should correlate with the salaries these people get paid. Therefore a high correlation means that this test can correctly predict what it theoretically should have the possibility to predict. Or for example, a measure in math ability may predict high performance of a worker on a math-based job and so on. In these cases predictive validity will be more important than content validity due to the peculiarity of the situation or in other words the unobservability of the construct that it is intended to measure.
Therefore, it is impossible to say that either content validity or predictive validity is more or less important. The choice made is completely based on the measured construct and its nature.Nevertheless, both of them provide information about the validity of the inferences and the predictions made on the bases of the test results.