Though the testing and measuring of psychological traits is a relatively recent development, the movement has been in progress long enough to show very clearly that, if a test is to serve its purpose, it must meet certain definite requirements. These should be recognized at the outset as constituting the fundamental specifications of the present task. There are, of course, certain practical considerations to be met. Tests should not require too much time either for administration or for scoring; and, as a rule, they should not involve the use of elaborate apparatus, which may be unduly expensive, and which is often difficult to operate and repair. On the whole, the simpler the test in equipment, administration, and scoring, the wider will be the field of its usefulness. (Brewer & others, 1942)
These considerations, however important, have little direct bearing on the more fundamental question of what requirements a test must meet if it is to be adequate as an instrument of measurement. How does one know that a given test actually measures the trait it is intended to measure? And, assuming that it does measure the trait in question, how does one know that its quantitative results are dependable? These questions are usually discussed under the topics of validity and reliability, and since those terms have acquired a special and somewhat technical significance in the work on testing, it is perhaps desirable to define them at this point and to indicate their bearing on the problems of this research. (Hull, 1982)
A test is valid when it measures what it purports to measure, that is, when there is an actual empirical correspondence between test scores and proficiency in the activity chosen as a criterion.................
