2.1 Characteristics of the Scientific Approach

Imagine that you are driving a rental car in a foreign country. You have never driven the model of car before and, despite the car being brand new, you find that it does not seem to work properly. Sometimes when you turn the ignition key the engine just will not start. Although you are not a specialist you do have a basic understanding of how an engine works and, starting from there, you begin to investigate the problem.

Based on your limited knowledge of engines you make a list of potential causes of the problem. Comparing the symptoms you would expect from these causes with your experience of the problem, you find yourself forced to discard one point after another on the list until there are no potential causes left. The next time the engine fails to start you are faced with the fact that you are completely clueless about what to do. In an act of desperation you decide to walk around the car before turning the ignition key again and, to your immense surprise, the engine now starts without a problem. Encouraged, you begin to experiment with this new method and find that walking in a clockwise direction around the car does not work. After a walk in a counter-clockwise direction, however, the engine always starts perfectly. So, in the course of your systematic investigation of the problem you have made a discovery, and a highly unexpected discovery at that!

Later, when returning the car to the rental car office, you complain about the problem. The woman behind the desk asks you if you remembered to push down the brake pedal when turning the key. She explains that this is a safety feature in some cars with automatic gearboxes. To decrease the risk of the car moving during starting, the brake must be pushed down. Thinking back you realize that you must unconsciously have put your foot on the brake pedal when entering the car after the counter-clockwise walk but not after the clockwise walk. That could explain why your method worked.

This may sound like a contrived example but I assure you that it is a true story from life, once told to me by a friend. (Being an engineer he was not particularly proud of trying to walk counter-clockwise around the car, but he admitted to being desperate when doing it.) We will return to this example later. It is useful here because the two methods presented in it could be said to represent two types of knowledge of a problem. Surely, many of us would agree that walking counter-clockwise around the car seems like a less “scientific” method than pushing on the brake. A possible justification is that the latter method is based on a deeper understanding of how cars work. On the other hand, the former method was discovered through a more or less structured investigation of the problem. Isn't that how scientists work? At any rate, we have seen that the fact that a method seems unscientific does not necessarily mean that it does not work.

Scientific research is about obtaining new knowledge, but what kind of knowledge becomes science and how is it obtained? In school, many of us have completed tasks that were called research. They generally involved a visit to the school library to collect information on a topic and summarizing this information in writing. Being able to find information is an important skill for researchers, and scientists do spend considerable time studying their literature, but simply collecting information from books is not research. Since the aim is to acquire new knowledge, it requires something beyond moving facts from one place to another, structuring them neatly and referencing the sources.

Science is often connected with measurements. Are measurements the crucial difference between science and non-science? Measurements are made in different parts of society, from laboratories to the fruit market. In applied research fields, academic researchers and development engineers in industry often make the same kind of measurements. How can it be that a Ph.D. student gets a fancy academic title after using a measurement technique for a few years at a university, while a young engineer making the same type of measurements in industry would be lucky to get a pat on the back for a job well done? If we are not to blame this difference on old tradition, and thereby deprive the Ph.D. degree of its value as an academic merit, there must be a central difference between what the engineer and the scientist do. Something beyond the everyday activities (like calibrating instruments, meticulously following procedures to assure good data quality and taking measurements) in the laboratory. To begin to cast some light over this difference I am going to borrow the following example from Molander [1]:

The main character of the example, Mr. Green, is a keen gardener who one day decides to count the apples in his garden. He goes about the task systematically and methodically and finds out that there are 1493 apples. This is definitely new knowledge, previously unknown to humanity, but is it science? Most people that I ask agree that it is not, even though they may have difficulties explaining why they think so. Those who have published their results in scientific journals and know how results are scrutinized in peer review processes may say that Mr. Green's chances of getting his results published are very slim. But why? It is not because Mr. Green does not have the proper scientific training – even with a Ph.D. in plant physiology he would not get these results published. For his results to become scientifically interesting they need to be incorporated into a greater context, a theoretical framework that gives them generality and helps us better understand some aspect of the world. If Mr. Green had counted his apples every year while also recording information about temperature, precipitation and hours of sun per day he could have searched for relationships in the data. That would approach a scientific way of obtaining new knowledge. When we judge scientific quality it also involves appraising the value of the new knowledge. What is it worth to know something about the number of apples in Mr. Green's garden? Is the garden unique in some sense, for example regarding microclimate, soil, or the type of apples grown there?

Collecting data is an important aspect of research, but it is also important in technical development, politics and other activities that are not considered to be science. There is a wealth of things that we do not yet know and that we could find out, like the oxygen content of the water in a particular bay, or the number of flowers in a particular field. Finding these things out requires planning, meticulous data collection and possibly statistical analysis to provide an unbiased picture of reality. Still, when finding them out we are only describing what we see, we are answering “what” questions. Science goes beyond pure description. It aspires to explain what happens in the world and to predict what will happen under certain circumstances. In other words, it aspires to answer what Phillips and Pugh [2] call “why” questions. Why is the oxygen level low in the bay? Why are there fewer flowers in the field some years? Answering such questions requires something more than careful collection of information. It requires a scientific approach. If you are an engineer working towards a strict deadline in a product development project, or a politician dealing with a problem that suddenly attracts media attention, your need to act can often be more pressing than your need to understand. For the scientist it is always the other way around. In research, trying to find a solution before you understand the nature of the problem is a bit like tying your shoelaces before you put your shoes on.

This means that there are two central and intimately interconnected aspects of science: one that has to do with investigating the world and one that is about interpreting what we see. By investigation we hope to acquire hard facts about reality, and we hope to obtain understanding by interpreting these facts. Interpretation is done within a theoretical framework that allows us to explain the facts. Philosophy books about scientific method often use the words observation and theory instead of investigation and interpretation. Observation sounds more passive than investigation and certainly has a narrower meaning. This is perhaps significant of the fact that such books often focus on how theories are developed. Both aspects are, however, two sides of a coin.

The next few sections are about two basic approaches to research that are described in the philosophy of science. They are written from a philosophical point of view. Even when I criticize the ideas I speak with the voice of a philosopher. In the remaining parts of this chapter I will again take a scientist's point of view by being a little bit more practically oriented, recognizing the fundamental role of observation in science. I hope to show that the philosophical concepts are useful for understanding the nature of scientific knowledge, although they do not cover all aspects of practical research. We can learn important things from these ideas but should not be too worried if some research does not fit perfectly into their framework.