Design ofResearch 101
NEED FOR METHODOLOGICALLY DESIGNED RE-
Design of Research
results and thus to adjust the results with a view to minimizing the effects of the shortcomings.
The use of idealized research model or the research standard forthe adjustment of actual data is common throughout the sciences.
For example, the ideal model for detennining the acceleration of freely falling bodies requires a perfect vacuum in which the bodies could fall with complete freedom. But in actual practice, the physicist can never create aperfèct vacuum. Still he can conducthis experiment in such a way that he can determine how a body would fall if it were in a nearly perfect vacuum. He determines how a bodywould fall if it were in a perfect vacuum. He determines how acceleration is affected by variations in atmospheric pressure and changes in acceleration. On this basis, he determines what would occur in a completevacuum and can thus infer the acceleration of the freely falling bodies. The idealized research design then, comprisesthe specifications of the most efficient conceivable conditions and procedures for conducting the research. But the proceduresand conditions specified in the idealized research model can seldom if ever be met in practice.
The next design-job for the researcher is to translate the idealized research model into a practical one. The practical research design has a reference to the translation of idealized designinto a realizable working procedure. The practical research design is necessary because certain factors do keep the researcher from meeting the idealized conditions. In a concrete research situation, practically may impose many restrictions on the researcher'sactivities.
The number of subjects or events he may ideally want to study may be much larger than his time, money and energy would allow. In such a case, he can only observe a portion of the whole. Once this restriction is imposed, the use of statistics and sampling becomes necessary. Hence, the translation of theideal model into a statistical model is a necessary step for the actual conduct of research.
Even where there is only one subject, event or property to be observed, the researcher is aware of the fact That his observations are always subject to error and he will thus require morethan one
104 R&cearch Methodology
observation for each set of variable-values. He would like to make an infinite number of observations of some single subject. l'his is obviously impossible and impracticable. Hence, he must deal with a sample of the possible observations. Thus, sampling possible observations requires a translation of the idealize model into
a
practical statistical model.
Even if situations existed m which the researcher could make an infinite number of observations on each subjecl, it might be wasteful to do so. He may not even require the degree of accuracy that such a large number of observations would yield. Therefore, if he wants to do just as much work as is necessary to get the amount ofaccuracy he needs, he will again wish to use only a sample of the possible observations, which means that he will render a statistical translation of the idealized research model.
In many social situations, manipulationofthetcjtajjty of variables involved is not possible, hence research must be conducted in situations which differ from the idealized one. Thus, we must determine how we can infer from the result obtained in some real concrete situations what we might observe if one had managed to produce the ideal situation. This requires that we make explicit the kind of real situation it wili look for, how one will characterize it and how one will adjust the results observed so that assertions about the idealized situation can be made. This too will require the statistical translation of the idealized research design and formulation of the research operations to be actually performed.
The practical research design may be conceived of as
comprising the following four phases:
(a) the sampling design, which deals with the method of selecting the subjects to be observed for the given study.
(b) the observational design, which relates to the conditions under which the observations are to be made or the data are to be secured.
(c) the statistical design, which deals with the question of how many subjects are to be observed and how the observations are to be organized with a view to securing answers to the research problem.
(d) the operational design, which deals with the specific
techniques by which the procedures specified in the sampling, statistical and observational designs can be earned out.
It must be remembered that non the these sub-designs and the resultant models are autonomous vis-a-vis the others. A decision in respect of any one phase of the design may influence or affect a decision subsumed under any other phase. Consequently, these phases generally overlap.
It should be clear by now, the practical research design
represents a compromise prompted by a number of practical
considerations that are related to the actual conduct of social research.
E. A. Schuman puts it, "Research design is not a highly specific plan to be followed without deviations but rather a series of guide- posts to keep one headed in the right direction".
Research designs differ depending on the research purpose just as the plan of a building would differ on the purpose for which it is intended to be used. The research purposes may be grouped broadly under the following four broad categories:
(a) To gain familiarity with the phenomenon or to achieve new insights into it, often in order to fomiulate more precise research problems orto develop hypotheses. Studies having this purpose are known generally as Exploratory or Formulailve studies.
(b) To portray accurately the characteristics of a particular situation or group or individual (with or without specific initial hypotheses about the nature of these characteristics). Studies eharacterised by such aims are known as Descriptive studies.
(c) To deteniune the frequency with which something occurs or with which it is associated with something else (usually but not necessarily, with a specific initial hypothesis). Studies having
this purpose are known as Diagnostic studies.
(d) To test a hypothesis suggesting a causal relationship between variables. Studies characterised by this purpose are called Experimental studies.
106 ResearchMethodology
It must be remembered that a fixed typology of the studies suggested above is inevitably arbitrary in as much as the different types of studies are not absolutely separable from one another and therefore, for purposes of classification, the 'major intent' of each becomes the basis for assigning them to different categories. In this connection it needs to be recognized that the development of knowledge rarely progresses in a direct step-wise manner. Each step forward in the resolution of a problematic situation is, at the same time, a step in the direction of posing new questions and of reformulating older ones. Max Weber has said: "Every scientific fulfilment raises new questions it asks to be surpassed and outdated".
In the fonnulative or exploratory studies the premium is on discoveiyhe can borrow from other fields and from common language.
He needs to create his own guide-posts and schemes of classification.
He must decide what to look for, and what to ignore, what to record and what not to record, which clues to follow and which to abandon, what is of consequence and what is trivial. The explorer has great freedom but the same can so often be terrifying.
It has already been noted that more appropriate the exploratory study should be considered an initial step in a continuous research process rather than an exercise in isolation. The most difficult phase in an inquiry as one has had an occasion to see, is its initiation. The most careful methods during the later stages of enquiry are of little
worth if an incorrect or irrelevant start was made. Adequate
xplorationaisures iginst such an eventuaht) Seli Jahoda, Deutsch and Cook suggest that the following methods are likely to be very fruitful in exploratory research directed toward the search for meaningful hypothesis:(a) A review of related social science th other pertinent literature.
(b) A survey of people who have had practical experience of the broad problem are to be investigated.
Most exploratory researches utilize these methods. These methods to be used must be flexible. As the initial vaguely defined problem gets transfonned gradually into one with more precise meaning
and reference, frequent changes in the research procedures become necessary in order to provide for the gathering of data relevant to the evolving hypothesis.
LITERATURE SURVEY
Frequently, an exploratory study is concerned with an area of subject-matter in which explicit hypothesis have not yet been formulated. The researcher's task then is to review the available material with an eye on the possibilities of developing hypotheses from it. In some areas of the subject matter, hypotheses may have been formulated by previous research workers. The researcher has to take intostock of these various hypotheses with a view to evaluating their usefulness for further research and to consider whether they suggest any new hypothesis. A researcher working in the field of sociology will find that such publications as the Sociological Journals, Economic Reviews, the Bulletin of Abstracts of Current Social Science Research, Directory of doctoral dissertations accepted by Universities, c. afford a rich store of valuable clues. In addition to these general sources, some governmental agencies and voluntary organizations publish listings or summaries of research in their special fields of service. Professional organizations, research groups and voluntary organizations are a constant source of infonnation about unpublished works in their special fields. It could be too narrow an outlook, however, to restrict one's bibliographical survey to studies that are directly relevant to one's area of interest. The most fruitful means of developing hypothesis is an attempt to apply to the typhoid hypothesis, which is tenable. Ifthe post-treatment observations suggest un1vourable response, the typhoid hypothesis is falsified. Such a test of hypothesis does not belong to the realm of explorartory studies.
The above example illustrates the nature of an exploratory study and also how it differs from the problem-solving and hypothesis testing studies. In the initial stages when the doctor was asking the patient all sorts of questions and was examining him, using various instruments, scnitinizing various reports, the doctor was simply exploring the possibility of exploring for conducting some sort of an exploratory
108 Research Methodology
study. The end-result of this exploration was the questions that suggested itself to him as an alternative hypothesis.
After this phase of exploration, the doctor proceeded to test the proposed hypothesis by resorting to a more controlled or structured method of investigation. This second phase was the hypothesis testing phase of the inquiry. The exploratory study may thus be considered an earlier step consisting of problem finding or hypothesis formulation, to be followed by other steps aimed at problem solving or hypothesis-testing, on a continuum of research process. The flexible nature of the research design characteristic of exploratory studies should be clear from the above example. The doctors questions to the patient were not pre-determined nor was the use of certain instruments. The doctor was continually accommodating newer facts as they were becoming known to him, changing in effect, his tentative idea about the nature of disease from time to time till finally he could put forth his tentative diagnosis (hypothesis).
The relative youth of social science and researches in the realm of social science make it inevitable that much of social science research, for some time to come, will be of an exploratory nature.
Few well-trodden paths exist for the investigator of social life to follow. Most existing theories in social sciences are either too general or too specific to provide any clear guidance for empirical research.
Under the circumstances, exploratory research is necessary to obtain experience that will be helpful in fonnulating worthwhile hypothesis for more definitive investigations. For a general area of problems about which little knowledge is available and a general state of ignorance prevails, an exploratory study is most appropriate.
Quite often one sees a tendency to undermine the importance of exploratory research and to regard only experimental research as more scientific. But if experimental work is to have any theoretical or practical value it must be relevant to issues that are much more broader than those posed in the concrete confines of the experiment.
Such relevance can result only from adequate explorations of the dimensions of the problem with which the research is attempting to deal.
Design of Research
Path-breaking explorations or formulative researches are particularly complex affairs. One starts from the scratch, without guide-posts or yard sticks. Anyintellectualframework and categories within which to classify what one sees, are absent. The researcher's only resource is whatever studies must have enough flexibility to
pennit consideration of different aspects of a phenomenon.
In the descriptive and diagnostic studies, the major concern is with accuracy. Hence, the research design for such studies must be such that the bias will be minimized and the reliability of the evidence collected maximized. These two studies, namely, descriptive and diagnostic, though somewhat different in their aims, yet piesent similar requirements with respect to the research design.