Advisedly, the hypotheses should be related to a body of theory or some theoretical orientation. This requirement concerns the theoretic rationale of a hypothesis, i.e., what will be the theoretical gains of testing the hypothesis?
If the hypothesis is related to some theory, research will help to qualify, support, correct or refute the theory. A science can become cumulative only through interchange between the existing body of fact and theory.
Will not deriving hypotheses as a rule from some theoretical base throttle ventures into new fields in which no articulate theoretical system has developed? Will not such hypotheses lead to unnecessary repetitions? Doubts of this order may be raised by some.
Theses objections do not have much substance since such hypotheses formulate imaginatively, besides serving the function of elaborating, extending and improving the theory, they may also suggest important links between it and certain other theories.
Thus, the exercise of deriving hypotheses from a body of theory may also be an occasion of a scientific leap into newer areas of knowledge. As Parsons put it, “Theory not only formulates what we know but also tells us what we want to know.”
If hypotheses were derived from a body of theory, to that extent it would be possible to formulate them as statements about what will happen, that is, the roots of hypotheses in theory would invest these hypotheses with the power of prediction.
One of the valuable attributes of a good hypothesis is its power of prediction. The potency of hypotheses in regard to predictive purposes constitutes a great advancement in scientific knowledge.
To quote Cohen and Nagel, “… the hypothesis to be preferred is one which can predict what will happen, and from which we can infer what has already happened, even if we did not know (it had happened) when the hypothesis was formulated.”
In the example cited earlier, the hypothesis that lower suicide rates should be expected among the Catholics than among the Protestants besides having a predictive potential would also afford by virtue of its theoretical moorings, the basis for saying that married persons or a minority community or a tribal community by virtue of high social cohesion would have lower suicide rates.
It is in this sense that a ‘good’ hypothesis helps us make statements about what is already there or what has already happened although we were not aware of it.