ADVERTISEMENTS:
Read this article to learn about the use of research design for solving problems in research.
The research designer understandably cannot hold all his decisions in his head. Even if he could, he would have difficulty in understanding how these are interrelated. Therefore, he records his decisions on paper or record disc by using relevant symbols or concepts. Such a symbolic construction may be called the research design or model.
The model makes possible an overall evaluation of the total plan. It is on this basis that the researcher can appreciate the whole study structure as also the operations, the place and importance of the successive steps that he will be required to take in the total scheme. The research design results, from certain decisions taken and ordered in a certain sequence by the scientist.
ADVERTISEMENTS:
The major design decisions are in reference to the following aspects:
(a) What the study is about and what are the types of data needed?
(b) Why the study is being made?
(c) Where the data needed, can be found?
ADVERTISEMENTS:
(d) Where or in what area the study will be carried out?
(e) What periods of time the study will include?
(f) How much material or how many cases will be needed?
(g) What bases will be used for selection of cases?
(h) What techniques of gathering data will be adopted?
(i) How will the data be analysed?
(j) How best can these above questions be decided upon and decisions articulated in a manner that social research purpose will be achieved with minimum expenditure of money, time and energy?
As Selltiz, Jahoda, Destsch and Cook describe, “A research design is the arrangement of conditions for collection and analysis of data in a manner that aims to combine relevance to the research purpose with economy of procedure.”
The decisions in respect of the data to be collected, the sample to’ be selected, the manner in which the collected data are to be organized etc., which constitute the trunk of the research design, must be based on good grounds.
ADVERTISEMENTS:
In the interest of science, the design decisions must be based on an accepted methodology. The researcher must investigate or make investigable the method of making design – decisions. To the extent he does it, he is methodologically designing the researcher.
It must be remembered, however that no inquiry is completely methodological just as none can be completely un-methodological. Researchers vary between these two extremes. A completely methodological design is a scientific ideal which the researcher may never attain but which he is obliged to constantly try approximating.
Let us now turn to consider the need for methodologically designed research:
(1) In many a research inquiry, the researcher has no idea as to how accurate the results of his study ought to be in order to be useful. Where such is the case, the researcher has to determine how much inaccuracy may be tolerated. In quite a few cases he may be in a position to know how much inaccuracy his method of research will produce.
In either case he should design his research if he wants to assure himself of useful results. A researcher gets into trouble not only when he fails to obtain results which are accurate enough but also when he gets results that are much too accurate.
If the required degree of accuracy can be obtained with little trouble, and greater accuracy only with very great difficulty, then the researcher would only be wasting time, efforts and funds in working for greater accuracy.
(2) In many research projects, the time consumed in trying to ascertain what the data mean after they have been collected is much greater than the time taken to design a research which yields data whose meaning is known as they are collected. Modern research suffers a great deal from a ‘lust’ for fresh data. Researchers rush off to collect data without a concern for what they mean until they are collected.
At this stage it is often very late to improve upon them. It is true that in some cases the delay produced by research-planning may result in obtaining stale data. But it should be remembered that failure to plan a research may produce more inaccuracy than designed research project run later. For example, suppose the researcher rushes to make observations before he has developed adequate instruments.
The errors produced by his inferior instruments may be greater than the degree of inaccuracy he would have obtained had he waited to develop better instruments even at the obvious risk of his data getting a bit stale.
(3) A scientist owes certain obligations to the institution of science. His right to the title ‘scientist’ rests in part on his ability to develop better and still better ways of inquiring. This is true of all sciences but particularly so, of social sciences today.
The hope of any future for sciences of society may well depend on the extent to which the social scientists demonstrate how major social problems can be solved effectively in a scientific way. The scientist cannot afford to remain complacent with his methods.
As a scientist he is obliged to question each and every phase of his method, opening up in effect, the possibilities of continuous improvement. He cannot wait for lucky discovery to improve his methods. He must improve upon them and take what luck has to offer as a bonus.
Once the research problem is formulated in clear-cut terms, the researcher is in a position to consider how he will try to solve it. The first step toward obtaining a solution should advisedly be in the nature of designing an ideal research procedure; that is, the procedure the researcher would have liked to adopt for solving a problem if he was completely unrestricted by practical exigencies and limitations.
This is the idealized research design. Ackoff defines the idealized research design thus:
“The idealized research design is concerned with specifying the optimum research procedure that could be followed were there no practical restrictions.”
At first glance, such a step, i.e., of designing an idealized plan, might seem very impracticable and even an unnecessary one. The researcher may be inclined to ask why he should bother himself with procedures that cannot be carried out.
Why dream idly of realizing an ideal almost impossible of attainment? The answer to this can be that concern with ideal or optimum research conditions is neither idle dreaming nor wishful thinking. It is really important, if we want to know how good the results are that we would eventually obtain.
The ideal conditions and procedures act as a standard by reference to which we can evaluate the practical research conditions and determine their shortcomings. If these shortcomings are made explicit, it is possible in many a case to determine their effects on the observed 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 for adjustment of actual data is common throughout the sciences. For example, the ideal model for determining 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 a perfect vacuum. Still he can conduct his experiment in such a way that he can determine how a body would fall if it were in a perfect vacuum. He determines how acceleration is affected by variations in atmospheric pressure. He calculates the relation between the changes in the atmospheric pressure and changes in acceleration.
On this basis, he determines what would occur in a complete vacuum and can thus infer the acceleration of freely falling bodies. The idealized research design then, comprises the specifications of the most efficient conceivable conditions and procedures for conducting the research. But the procedures and 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 design into 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, practicality may impose many restrictions on the researcher’s activities. 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 population or converse. Once this restriction is imposed, the use of statistics of sampling becomes necessary. Hence, the translation of the ideal 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 need more than one observation for each set of variable-values.
He would like to make an infinite number of observations of some single subject. This is obviously impossible. Hence, he must deal with a sample of the possible observations. Thus, sampling possible observations requires a translation of the idealized model into a practical statistical model.
Even if situations existed in which the researcher could make an infinite number of observations on each subject, it might be wasteful to do so. He may not even need 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 degree of accuracy he requires, he will again want 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, manipulation of the totality 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 situation what we might observe if we had managed to produce the ideal situation.
This requires that we make explicit the kind of real situation we will look for, how we will characterize it and how we 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 or eve-tents 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 the secured.
(c) The statistical design, 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 carried out smoothly. In other words, the operational design deals with the administration research project.
It must be remembered that none of these sub-designs and the resultant models is autonomous vis-a-vis the others. A decision in respect of any phase of the design may influence or affect a decision subsumed under any other phase. Consequently, these phases generally overlap.
As should be clear now, that the practical research design represents a compromise prompted by a number of practical considerations that are related to the actual conduct of social research. As E. A. Schuman puts it thus, “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 depended upon the purpose for which it is intended to be used.
The research purpose 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 formulate more precise research problem or to develop hypotheses. Studies having this type of a purpose are known generally as Exploratory or Formulative 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 characterised by such aims are known as Descriptive studies.
(c) To determine 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 type of 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.
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 purpose 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. As Max Weber has said: “Every scientific fulfillment raises new questions it asks to be surpassed and outdated.”
In the formulative or exploratory studies the premium is on discovery of ideas and insights. Therefore, the research design corresponding to such studies must have enough flexibility to permit 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, present similar requirements with respect to the research design.
Studies which aim at testing causal hypotheses, i.e., the experimental studies, require procedures that will not only minimize bias and increase reliability but also permit inference about causality. Experiments are particularly suited to this end.
In practice, however, these different types of studies or researches are not always sharply separable. Any given research may have in it elements of two or more types of the functions. In a single study, however, the primary emphasis is usually on only one of these functions and the study can be thought of as falling into the category corresponding to the major function aimed at.