10.4 Finding Out What is Not Known

The first step of planning is finding out how you can contribute new knowledge to your field. To do this, you must determine the boundaries of the current knowledge. As research papers will not be published if they do not contain new information, it is important to find out how to make a novel contribution. This may seem challenging to new Ph.D. students, since their whole field of research is new to them. As a rough rule of thumb, the first half of your time as a Ph.D. student will be used to reach the research frontier in your area. During the second half you are expected to make “an original contribution to knowledge”, which is what the Ph.D. is awarded for.

Getting familiar with the entire bulk of knowledge accumulated in your field over the years is a formidable or, quite likely, impossible task. You need to be selective. It is important that academic supervisors direct their students to research papers that have been influential. Get the “classics”, read them and reflect on them. You should regularly browse the scientific literature in your field to keep up to date with the latest developments. It is also important to be in touch with your part of the scientific community. The daily discussions with supervisors, other faculty members and more experienced Ph.D. students at your department are important, but it is just as important to listen to their discussions with other researchers. Conversations on a research topic will often give you valuable clues to “what is hot and what is not”. Every visit to a research conference is a golden opportunity to get an overview of research problems that are currently addressed and how they are studied. By taking in, processing and working with all this information, you gradually build a theoretical framework for your own research – a mental picture to orient yourself by. This picture will make it easier for you to see where and how you can make a contribution. Although the demand for novelty can be distressing for new Ph.D. students, their confidence tends to increase with time. As your mental picture gets clearer, you realize that making an original contribution is not an insurmountable obstacle.

Apart from knowing the boundaries of the current knowledge, it is also helpful to know something about what constitutes novelty in a piece of research. A list of the components of research is given below, followed by a discussion about the ways in which they can be new. Research results can be said to be the output from:

• Data
• Methods and techniques
• Research questions
• Research areas
• Analyses and syntheses.

Obtaining good and reliable data is central to all research. A unique experiment will provide unique data, but this is not sufficient to fulfill the need for novelty. New data that support an established theory, like Galileo's law of free fall, are not an especially valuable scientific contribution. Though such data are good news for the theory, confirmation of a thoroughly tested theory does not add much to our knowledge – unless it is confirmed under conditions where it is not expected to be valid. Data that show discrepancies with current theory, on the other hand, are interesting, since they may point to limitations in the current knowledge. But not all experiments are made to test theories. Their purpose may be to characterize a system or to investigate some aspect of its workings. The crucial point is that your data should bring new, useful information to your research community.

New methods and techniques can make it possible to study new aspects of a problem. They could be novel measurement techniques or new methods for analyzing data. Applying a known technique to a new area can be a sufficient criterion for novelty, as well as applying a new technique to an old area. The method does not even have to be applied since, in some fields, the development of new methods and techniques is the very subject of research. It is also worth mentioning models under this heading. Models are mathematical representations of systems, such as the earth's atmosphere, the combustion chamber of a diesel engine, an evolutionary process, or the traffic situation at a highway junction. Models are often complex, containing a large number of functions to account for various aspects of the system's properties. They are often based on a variety of simplifying assumptions. In contrast to theories models are not generally valid, but may be useful under well-defined conditions where they have been tested and calibrated. In most cases, models are used in computer simulations of a system's behavior, making it possible to study phenomena that are impossible or unpractical to study experimentally. You may, of course, make a novel contribution to research by developing a new model or carrying out a new modeling study.

Research questions are the very starting points of research. Finding a question, small or big, that has not yet been answered within your research area is one of the most promising ways of expanding the current knowledge. This is echoed in the quote at the beginning of Chapter 2, saying that it is less important for a researcher to know the right answers than to find out what the right questions might be. Finding new research questions may sound difficult but is often easier than one thinks. Looking at how research is done in your field, you will probably find that people tend to approach certain aspects of certain types of problems using certain methods. When you have identified such common characteristics, change your perspective and try to see what is not there. With a bit of creativity, you will probably find numerous questions that are short of attention.

Discoveries sometimes lead to new research areas. To name just one example, one of the central problems in physics at the turn of the last century was that there was no coherent theoretical treatment of heat radiation. Max Planck eventually discovered that this problem could be solved by treating light as discrete packages of energy; this was the starting point of quantum theory [5]. Very few can hope to make such grand contributions but, luckily, new research areas often occur on a much more modest scale. They can be the result of advances within an existing discipline. For example, a new method could make new types of studies possible. An interesting new research question or a new synthesis of results could also displace the interest towards an area that has not previously been studied. New research areas also emerge between existing disciplines. Recent decades have seen a boom of interdisciplinary research areas, combining methods and problems from several subjects.

Finally, novelty can consist of a new analysis or synthesis of existing evidence. As discussed in Chapter 5, questioning the validity of evidence and conclusions is an important way in which knowledge develops. Research papers are brief and often focus on a very specific question. Bringing together and comparing results from several previous studies that bear on a more general question can sometimes bring forth interesting new aspects of a problem, highlight limitations in a common approach or even point to misconceptions about a certain problem.

There are obviously many ways to be original without revolutionizing an entire discipline. Novelty is an imperative demand in all research but you will find that it is usually not a reason for worry. You could be original in testing someone else's idea or even by continuing a previously original work. In fact, previous work is one of the most common inspirations for new research tasks. Your own results will have a tendency to result in new research ideas, simply because your analysis involves a great deal of reflection. When you pursue these ideas, you are successively expanding the knowledge within your field.

Let us look at some methods for generating research ideas. The overall goal of science as it is defined in this book is to explain the world. It is therefore useful to make a habit of challenging your ability to explain phenomena within your field. One way to do this is to repeatedly ask yourself the question, “Why is that so?”, when you encounter statements about a phenomenon in a research paper or during a discussion. Anyone who has known a four-year old knows how frustrating such questions can be, because you very quickly tend to run out of answers. For example, if the question is, “Why is the sky blue?”, the answer is that sunlight is scattered from molecules in the air. “But why is it blue?” Well, it is because this process is more efficient at shorter wavelengths, so blue light is scattered more than red. “But why is that so?” At this point you need an explanatory theory to provide an answer, which has to do with how photons interact with molecules. Most parents probably resort at a much earlier stage to saying “because that's the way things are”, which, to a scientist, has a rather Aristotelian ring to it. The point of this technique is that the answers tend to move through a succession of descriptive statements before reaching a level where our understanding of a problem is finally challenged. Upon reaching that level you will know whether you, or anyone else come to that, can explain the phenomenon. That is why this technique is useful for finding out what is not known and, thereby, for generating new research ideas. For lack of a better term, we could call this technique why-analysis. When writing a scientific paper the same technique can be used as a quality control: act as the curious four-year old with every statement in your conclusions to see if you really can explain what you claim.

When you have explored the boundaries of the current knowledge and found a promising problem to work with, it is time to start generating hypotheses. In the hypothetico-deductive approach to research these tentative explanations of the observed phenomenon are the foundations of experiments, conducted to either support or refute the hypotheses. As described in Chapter 6, experiments may or may not be based on hypotheses. Non-hypothesis-based experiments are sufficient if you are only interested in connecting a cause with an effect. The hypothetico-deductive approach is favored here, since hypotheses allow us to draw more complex and interesting conclusions from our data. Depending on the nature of the hypotheses they may even be used to develop explanatory theory. Before looking at tools for generating hypotheses, however, we will discuss the importance of limiting our ambitions.