Empirical research is usually so time consuming and expensive that before settling down to work, one should give the logic of the work a great deal of thought (not forgetting the possibility that a solution to the problem already could be found in existing literature). We no longer think like Descartes, who thought that there was only one "method of science", and by using this method, all the problems could be solved. Instead, there are usually several alternative approaches to any problem. The choice between them will depend primarily on the goal of the study.
The most common goal of empirical research is to gather information on the object of study. We call theory all the knowledge that has been accumulated through many research projects. The diagram on the right illustrates the idea of gathering data or information on a study object which resides in the empirical world. The data are collected, processed, and then the result is added to the collective structure of all theories; they are not used for making improvements to the object or its environment. Indeed, many scientists think that this descriptive approach is the only acceptable way of doing research: the scientist should by all means avoid disturbing the object, as this would of necessity corrupt the data so that they do not any more give a true picture of the object.
There is another, normative style of research where the prime objective is not to keep intact the object, but on the contrary to improve it by removing a practical problem around it or by developing a new product, for example. This philosophy of research is discussed on another page, Normative Point of View.
Each piece of knowledge that we have about any empirical object - in scientific research as well as in everyday life - can either pertain to one object, or to a class of more or less similar objects. It can be useful to differentiate between three ranges of validity of knowledge:
Each of these types of knowledge will be discussed below.
1. Facts that concern one specific case. Studies which aim at gathering this kind of specific knowledge are said to be idiographic (from Greek idios 'own' or 'peculiar', and 'graphein' describe). A typical method is a case study which makes possible to regard the object as an individual entity in its genuine environment with all its inborn attributes - in other words, taking a holistic view of it. It is often said that people, works of art and other products of human culture can be fully understood only this way. A study of a single object or a few objects is often called "intensive".
2. Knowledge about a class of cases, particularly about the population having a large number of cases as objects, consists of facts which are common to all or at least most of the cases in the population. Often these facts describe similarities between the cases, in other words patterns which do not vary from case to case and which we therefore can call invariances. There are two important types of invariance:
These two types of invariance are no opposites. The difference between them is only that dynamic invariances contain time as one of the variables. This difference, however, multiplies the amount of data to be analyzed and often necessitates particular research methods in these so called diachronic studies, as a contrast to the synchronic or "cross-sectional" study of static invariances.
Research which aims at finding invariances is sometimes called "extensive" or nomothetic (from Greek 'nomos' law, and 'tithemi' constate). Usually the great number of population prevents taking into account more than the most interesting attributes of the objects. For the analysis of this material a great number of sophisticated methods are available.
3. Seemingly universally valid knowledge, sometimes called a "law of nature", is in reality an assumption based on a great number of observations. Such assumptions can never be absolutely reliable, because it would never be possible to study all the pertinent empirical cases and thus reach a perfect reliability. Nevertheless, plausible universal assumptions are worth being mentioned as a special type of knowledge because they are very often used in practical life. It is often very convenient to apply a known "law of nature" without spending time for assessing its validity in your situation, without doing any modifications, let alone making an own empirical study on the matter. The convenience of application will usually compensate for the risk that the assumed "universal knowledge" is in fact not valid in your context. This risk is small especially in the fields of physical sciences and technology.
Because of the widespread practical demand for it, much research has been devoted for finding so universally valid knowledge as possible in various fields of science. The methods available are mostly the same as when searching invariances in the empirical cases of a definite population, mentioned above. Once an interesting structure has been found in a population, the researcher has the option of trying to assess its validity outside of the population, though there is no absolutely reliable method available for this speculative task. Nonetheless, the researcher can consider a few practical indicators which often hint towards an enlarged validity of the findings, outside of the originally studied population:
The criteria given above can be used by a researcher who wants to speculate on the scope of validity of his findings. Another situation where similar assessment is called for, is when a user of a published research report tries to assess whether he could use the findings. This type of evaluation is discussed in Assessing Validity of Information in Your Context and Utility Assessment.
Another useful angle of viewing research projects is the logical art of the resulting knowledge. In research reports, the most usual variety of knowledge is a description of the object. The purpose of description is usually to answer the question what, in so concise a way as possible and including only the relevant attributes of the object, whatever they may be in each project. Usually most interesting are the invariable attributes which are valid in a large number of cases if not universally.
Often the researcher wishes to continue the project to a deeper level than just to description: he wants to know why the object is such as it is. This knowledge helps summing up all that is known about the object, it helps to see it in its context and in a historical perspective, and it helps to forecast its future evolution when desired.
Finding the reasons, or explaining the phenomenon, can be done in a number of ways where the reasons are fetched either from the concurrent context of the phenomenon, or from the past or alternatively from the future. In the following are examples of usual types of explanation; the list is not exhaustive.
The prevalent style of explanation is a little different in humanistic sciences and in the natural (or technological) ones. In the latter, causal explanation is the rule. In fact, the paradigms of humanistic and technological sciences are so different that you could speak about two cultures of research. The division is regrettable from the viewpoint of the researcher of professions and products who normally has to deal with both inanimate objects and people using the objects.
All three styles of explanation can be used in two situations:
In any case, a plausible explanation is expected to fulfil the following requirements:
We should not exaggerate the difference between description and explanation. They are no opposites but rather they are two slightly different views at the object of study. Some philosophers of science have wanted to see them as two subsequent phases in the process of understanding the object: you should thus first describe the object, and finally you can expect to arrive at the moment of "Eureka", at the revelation which explains all the initially perplexing details amassed during the phase of description.
Such a dramatic progression up till "Eureka" is easily feasible when writing the final report of the study where it certainly can heighten the readability of the work. However, in factual research such a progression from description to explanation cannot be taken as a general rule. On the contrary, it is usual that a researcher starts from a tentative explanation, a preliminary understanding of the relationships contained in the object, and on this basis he then defines the descriptive facts that he shall start to collect.
Neither descriptive nor explaining-type research aim at ameliorating its object directly, that is: during the research project. On the contrary, many researchers think that ideally a scientific study should incite no changes in the object; in this way one also would get most reliable findings. This is the principle of "strenge Wissenschaft", "rigorous science" which is, however, more an ideal than real practice in most fields of research, because it would often severely hamper observing the object.
Beside descriptive and explaining-type studies which avoid influencing their object, there is also another style of research, where the final target is to influence and guide factual development. This is the normative type of science. Earlier it was often called "applied science" because it applies the findings of descriptive research which provide the grounds and show alternative means to steer evolution. In this respect, causal explanations are of special value. If we know the dynamic invariances and causes of change, we can often manipulate these causes and bring about empirical changes at will.
Let us have a closer look at the nature of the information a research project strives for. A century ago, some researchers thought that science should only aim at explicit, definite, and exact knowledge. If a research project cannot report such positive knowledge, it is better to publish nothing, said these supporters of the "positivistic" school of thought, for example Wittgenstein (B): "Whereof one cannot speak, thereof one must be silent."
However, in closer inspection it turns out that the positivistic ideal is attainable only in some areas of research. Restricting our studies to just those things that we can exactly measure would mean denying ourselves many of the benefits of research. For example, forecasting weather can be quite useful even if uncertain.
Particularly when we are researching the activity of a profession, such as industrial design or production, we should note that these operate not only on the basis of explicit and exact knowledge, but also very much on tacit knowledge. This knowledge could also be called know how, competence, proficiency, or skill, and what is typical of it is that we use it without making it explicit.
It can thus be quite well thought-out to gather and clarify the tacit knowledge of the artisan or other competent professional, even when it perhaps would never be possible to put all that knowledge into words. The result of such a scientific compilation of the tacit knowledge used in an occupation is often called theory of practice. A few examples of its content are given on the page Theories of Production, and the methods for gathering it are discussed on the page Developing an Activity.
Knowledge and opinion are no opposites. Instead we could regard them as statements which can be placed on a continuum depending on the number of people who support them. Let us consider the following statements made by various people:
The first piece of information is an example of objective knowledge. The measurement does not depend on the person doing it and we can say that the statement is true (or false). Measurements may differ a little, but we call it random variation and it does not interest us. Normally we eliminate the variation by calculating the mean of measurements.
The second statement is clearly a subjective opinion; there are not many people who would subscribe it. Nevertheless, it is a fact that at least one person has such an opinion, and this state of things can be studied objectively.
The third statement is an intermediate case. It would be regarded as an appropriate opinion, perhaps even as a true statement, by almost all Europeans but not by all the people in the world. This type might be called an "intersubjective opinion".
Objective knowledge has been the primary target of scientists from antiquity until now. Subjective (and intersubjective) opinions were long regarded as too fluctuant objects for serious study. They started to interest researchers first when industry needed them as a basis for the design of products for people. Today we understand that we can study subjective opinions in a perfectly objective manner. In their study there are two main lines:
The researcher's own opinions. The discussion above concerns the input material of an empirical study. Another question then is what the researcher finds and asserts about empiria and particularly about the object of study. The general principle is that the researcher should report what he has found in empiria but he should not add evaluative opinions of his own. This concerns even normative research which should only sum up those opinions on future development that exist in the sources.
August 3, 2007.
Comments to the author:
Original location: http://www2.uiah.fi/projects/metodi