What are the Qualities of a Good Hypothesis?

The hypothesis has to be formulated in such a form to attain the following qualities:

(i) It should be based on known facts and must have roots in the existing theories of knowledge.

(ii) The hypothesis should be testable. (This is the most important requirement).

(iii)The explanation offered by the hypothesis should be quite simple, straightforward and brief, as this will be helpful in making the hypothesis ‘testable’. If the explanation is complex, it has to be broken up into smaller hypotheses and each has to be tested separately.

(iv)A hypothesis is of very great value if its acceptance or rejection is utilizable for many purposes or opens the doors for further researchers. For instance, the acceptance of the hypothesis related with the research problem mentioned earlier will be helpful in providing appropriate creative opportunities to children of the different sexes and intelligence levels in school and homes.

Validity:

There are different types of validity depending both on what the scores measure and what for they do it, that is on the matter and the purpose of measure meant. The various types of validity may be classified as given below: Content Validity

This type of validity is concerned with the measurement of attainment during a course of training according to a prescribed syllabus with the purposes of accessing the achievement of the candidate.

Thus this validity is meant for what are called ‘achievement or ‘attainment tests, that have a backward look. In order to increase this validity the items in this type of test should cover both the content and the objectives of the course adequately and in proper proportion of the constituent both content and objective-wise.

For this objective the content and the objective may be analysed by consulting the teachers, text-books, and the previous question papers, and a sort of ‘blue-print prepared.

By passing a judgement on the extent of its being in keeping with the analysis of the content and the objectives of the course, the ‘content validity’ of a test is assessed in qualitative terms.

Predictive Validity:

It assesses those attributes that are likely to predict success during a course or job. The measures of a test having predictive validity should predict the success for which they are used as predictors.

This type of validity is forward looking. It is estimated statistically by taking the measures on test which is supposed to be a predicator and correlating these measures after the course or performance on the job, for any specified period, with the criterion measures of success that may be scores

In some examination or rating by the superiors. It has to be understood that the criterion measures are not immediately available, and one has to wait for them.

Concurrent Validity:

In many situations new tools may be developed In place of some more cumbersome process of collecting information that is already in use. For example pathological tests may be there to determine whether a person is neurotic or not, after a number of clinical tests in medical labs and one may try to cut short all this by producing an objective type test of a number of questions to determine the same thing.

Then, the validity of this new test will be determined statistically by finding the assessments made by the former procedure almost simultaneously. In this case or need not wait for the criterion measures to mature as in case of ‘predictive’ validity.

Construct Validity:

If the difference between the concurrent and the predictive validity is related to time, the difference between content and construct validity is with respect to the thing to be measured. In case of the former it is a well set and definite course or syllabus, whereas in the letter it is an abstract psychological construct (a complex concept), like intelligence, creativity, anxiety, motivation, etc as they are so ill-defined and controversial, no. ‘blue-print’ can be prepared for the construction of their tests.

Every test constructor has his own mental image of the ‘construct’ and prepares the test according to that. Now, for measuring its validity there is no universally agreed description of the ‘construct’ for qualitative logical assessment, nor any single and reliable criterion for finding out the statistical validity, except another existing test in case we are out to produce a parallel form of it irrespective of the fact that the construct validity of that test may itself be in doubt the ‘construct validity is established on the basis of a quite number of evidences, gradually gathered over considerable length of time. These may be qualitative or statistical or both.