Hypothesis is the most important element of scientific research. A study that does not provide for the formulation of a preliminary hypothesis will be incomplete and unfocused. Therefore, when formulating the research task, the project manager must answer the question of whether the results obtained correspond to the initial hypotheses and how convincing and justified these results are.
The logical construction of the hypothesis is a conditionally categorical conclusion “If …, then …”. The first premise puts forward a condition, and the second asserts the consequence of the given condition. If the study does not confirm the consequence, the hypothesis is refuted, but the confirmation of the consequence does not provide logical grounds for the validity of the hypothesis. Confirmation makes the hypothesis plausible, probable. Hence one of the fundamental requirements for a good hypothesis: the more consequences it contains, the more likely it is to be confirmed.
It is important that the hypotheses are logically linked into a system of evidence for the explanation put forward. In this case, the confirmation of one hypothesis provides additional grounds for accepting the premise associated with it. Verification of the next corollary on a common premise presupposes new confirmations, and so on. It is clear that the refutation of the first working hypothesis requires the development of new hypotheses.
So, the initial hypotheses must be expanded into a whole chain of inferential hypotheses-consequences (the operation of deductive processing of hypotheses). In an empirical study, it is the hypotheses-consequences that are formulated in less general terms than the initial assumptions that are tested. Otherwise, the hypothesis is untestable in empirical data.
The orientation of sociological research towards the verification of precisely formulated hypotheses makes it possible, in its conduct, to confine itself to checking only the necessary connections, that is, to pay attention to the most essential variables. Modern sociological research is too expensive to make its fruitfulness dependent on the effectiveness of any one hypothesis put forward, therefore, in the study of a specific problem, they always focus on putting forward a number of hypotheses that exhaust this problem as much as possible.
In this regard, the degree of abstractness of the hypothesis becomes important. An overly specific hypothesis, subordinating the program and methodology of research, as a rule, leads to trivial results, at best reproducing the limited information contained in the hypothesis itself.
Hypothesis testing. From a logical point of view, the process of a comprehensive practical testing of a hypothesis acts as a process of confirmation by experience of the consequences arising from these hypotheses. Moreover, the verification must take place along each alternative path determined by the main hypotheses, which may be mutually exclusive, but necessarily represent a logical whole with their own hypotheses-consequences.
At the same time, the empirical confirmation of each individual consequence of the main hypothesis cannot serve as a proof of the hypothesis itself – this is an illegitimate conclusion from the truth of the consequence to the truth of the foundation.
It is possible that this consequence follows not only from this hypothesis, but also from some other one. But the greater the number of different consequences of a hypothesis is confirmed by experience, the less likely it is that all of them could be deduced from another hypothesis or hypotheses.
The most powerful way to empirically test a hypothesis is a social experiment. Being the most reliable method for testing hypotheses, at the same time it imposes the most stringent requirements both on the hypotheses themselves and on the tools of sociological research. As a result, it has so far found rather limited application within the framework of sociology. However, as the understanding of social processes deepens, the weight of experimental methods increases.
The usual method of substantiating the truth of a hypothesis is to calculate average tendencies, find coefficients of interdependence, etc. Verification of inferential hypotheses is possible only if all the terms in which they are formulated have been subjected to empirical interpretation and operationalization.
The hypothesis is verifiable by the selected empirical features. But where is the guarantee that these signs are justified? Thus, not only a hypothetical judgment, but also its empirical interpretation is subject to empirical testing for reliability.
To increase the confirmability of a hypothetical judgment, one should be guided by the following rules:
a) strive to put forward a larger number of interrelated hypotheses;
b) strive to indicate for each hypothesis a greater number of its empirical indicators (referents).
Nevertheless, the problems of the truth of hypotheses are not solved in this way, but only the probability of their justification increases.
Hypotheses differ in the degree of generality of assumptions as hypothesis-foundations and hypotheses-consequences. By themselves, the concepts in which the initial hypothesis is formulated may not have direct empirical features, but the concepts of inferential hypotheses must certainly be correlated with empirical indicators. Confirmation or refutation of hypotheses-consequences – the way to prove the validity or refutation of hypotheses-grounds.
From the point of view of the objectives of the study, the hypotheses are divided into basic and non-basic. Unlike the hypotheses of foundations and consequences, which are logically interconnected, these hypotheses relate to different tasks and, as it were, coexist with each other. Naturally, the main attention in hypotheses is given to the main assumptions related to the central question of the study.
According to the degree of development and validity, primary and secondary hypotheses are distinguished. Secondary ones are put forward to replace the first ones if they are refuted by empirical data. Sometimes primary hypotheses are called “working” hypotheses in the sense that they are used as scaffolding for the construction of more valid hypotheses. Good research usually relies on a series of alternative hypotheses. Then the test allows one to obtain higher grounds for accepting those assumptions that remain after discarding other alternatives.
According to the content of the assumptions about the subject area of the problem, descriptive and explanatory hypotheses can be distinguished. Descriptive – these are assumptions about the essential properties of objects (classification), about the nature of the relationships between the individual elements of the object under study (structural). Explanatory hypotheses refer to the assumptions about the degree of closeness of interaction links (functional) and cause-and-effect dependencies in the studied social processes and phenomena. These are the strongest hypotheses requiring experimental verification.
We can formulate some general requirements that a successful hypothesis must satisfy in order to be subject to direct empirical testing:
– the hypothesis should not contain concepts that have not received an empirical interpretation, otherwise it is unverifiable;
– it should not contradict previously established scientific facts. In other words, the hypothesis explains all known facts without exceptions to the general assumption. For example, the hypothesis “the more diverse the work, the greater the job satisfaction” should be rejected, because it contradicts the data available in psychology. It is known that with a certain psycho-physiological type of personality, it is monotonous and monotonous work that gives satisfaction, while diverse work does not. Another hypothesis – “the functional content of work (that is, including the degree of monotony and diversity of work) determines job satisfaction” – does not contradict this information and can be accepted for verification;
– the requirement of simplicity of the hypothesis follows from the previous rule. It should not be overgrown with a whole list of possible assumptions and restrictions, it is better to proceed from the simplest and most general basis. Our assumption about the influence of the functional content of labor on the attitude to work had a single limitation: “under the given socio-economic conditions.” Meanwhile, the hypothesis was tested only on workers. Does this mean that the assumption should be limited by additional conditions? No. Because no facts are known that contradict the stated assumption, if we apply it to engineering labor or to rural labor;
– this is all the more important to keep in mind if another requirement is taken into account. A good hypothesis is applicable to a wider range of phenomena than the area directly observed in the study;
– the hypothesis should be fundamentally verifiable at a given level of theoretical knowledge, methodological equipment and practical possibilities of research. Although this requirement is also obvious, it is often violated;
– finally, the working hypothesis should be specified in the sense that the formulation itself should indicate the method of testing it in this study. This requirement sums up all the previous ones. It assumes that there are no unclear terms in the formulation of the hypothesis, the expected connection of events is clearly indicated, and verification of the assumption does not cause difficulties on the part of methods and organizational capabilities. Inferential hypotheses are specific, i.e., those particular consequences that are verified by direct comparison with facts.
The formal requirements listed above make a hypothesis “good” only on the condition that its content is not trivial and is not reduced to judgments of common sense. In a theoretically oriented study, it is sometimes useful to sacrifice the rigor of empirical testing of some particular consequences of a categorically rich interesting hypothesis that opens up the prospect of incrementing scientific knowledge. In applied research, the nontriviality of hypotheses, as a rule, is expressed in the formulation of alternative solutions to a practical problem. No matter how reasonable a solution may seem, it will always have weaknesses. An empirical test of the strengths and weaknesses of several options for possible solutions to a social problem will certainly be more valuable, since, after weighing all the pros and cons, the most effective course of action will be found.
The entire research process consists of, one might say, continuous formulation and verification of various assumptions: the central hypothesis of the entire study, the consequences of it, secondary hypotheses formulated in case of rejection of erroneous ones, the formulation of particular tasks of a methodological nature (also hypotheses, but already fulfilling “instrumental and auxiliary “role”). Within the framework of the research program, the sociologist’s attention should be focused, first of all, on the development of a central hypothesis orienting the whole work and the verifiable consequences arising from its content.