Psychology Of Scientific Reasoning Research Paper

1. Interviews And The Historical Record

Two frequently used and related approaches that have been used to investigate scientific thinking have been interviews with scientists and the analysis of historical records and documents such as notebooks. One of the earliest accounts of scientific thinking and reasoning was the interview of Albert Einstein conducted by the Gestalt Psychologist Max Wertheimer (1959). Wertheimer argued that a key strategy used by Einstein was to search for invariants. Wertheimer saw the velocity of light as a key invariant around which Einstein built his theory. Wertheimer incorporated his analysis into a Gestalt theory of thought. More recently, researchers have conducted interviews in the context of principles from cognitive science. For example, Paul Thagard (1999) has conducted many interviews with the scientists who proposed that ulcers are caused by bacteria. Thagard has pointed to the important roles of serendipity, observation, and analogy in this discovery. A related line of inquiry is the use of historical documents. Using the scientists’ lab books, biographical, and autobiographical materials, researchers attempt to piece together the reasoning strategies that the scientists used in making a discovery. For example, Nersessian (1992) has conducted extensive analyses of the physicist Faraday’s notebooks and has argued that the key to understanding his discoveries is in terms of his use of mental models. By mapping out the types of mental models that Faraday used and showing how these types of models shaped the discoveries that Faraday made, Nersessian offered a detailed account of the mental processes that led to a particular discovery. Another cognitive approach using the historical record is to take a real scientific discovery, such as Monod and Jacob’s Nobel Prize-winning discovery of a mechanism of genetic control and give people the same problem, using a simulated scientific laboratory, and determine whether people use the same discovery strategies that the scientists used to make the discovery, such as focusing on unexpected findings (Dunbar 1993).

2. Scientific Reasoning As Problem Solving And Concept Formation

Two common approaches to scientific thinking have been to see it as a way of discovering new concepts or as a form of problem solving. Beginning with Bruner et al.’s (1956) classic experiments in which college students were asked to induce the rule that determines whether an item is, or is not, a member of a category, these researchers attempted to discover the types of inductive reasoning strategies used to acquire new concepts. Bruner et al. argued that much of science consists of inducing new concepts from data and that the memory loads that different strategies require will make certain types of inductive reasoning strategies more common than others. More recently, Holland et al. (1986) provided an account of the different inductive learning procedures that could be used to acquire new concepts in science.

Herbert Simon (1977) argued that concept formation is a form of problem solving, thus the two approaches can be seen as complimentary (see Klahr et al. 2000). Simon argued that scientific thinking consists of a search in a problem space with the two main spaces being a hypothesis space and an experiment space. The hypothesis and experiment spaces refer to all possible hypotheses, experiments, and operators that can be used to get from one part of the space to the next part of the space, such as grouping common elements (the grouping operator) in sets of results to form a new hypothesis. Simon has taken specific scientific discoveries and mapped out the types of heuristics (or strategies), such as heuristics for designing experiments that a scientist used in searching the experiment space. Using the notion of searching in a problem space, other researchers have analyzed the types of search heuristics that are used in all aspects of scientific thinking and have conducted experiments on the problem-solving heuristics that people use in designing experiments, formulating hypotheses, and interpreting results (Dunbar 1993, Klahr et al. 2000). These approaches specify the types of knowledge that an individual must possess and the heuristics that are used to formulate hypotheses, design experiments, and interpret data.

3. Errors In Scientific Thinking

One of the most frequently investigated aspects of scientific thinking and reasoning has been the finding that both scientists and participants in psychology experiments attempt to confirm their hypothesis when they conduct an experiment, sometimes called ‘confirmation bias’ (see Tweeney et al. 1982). Following from the writings of the philosopher Karl Popper, many psychologists have assumed that attempting to confirm a hypothesis is a faulty reasoning strategy. Numerous studies have revealed that, when given an hypothesis to test, people will design experiments that will confirm their hypothesis and not conduct experiments that could falsify their own hypothesis. This is a pervasive phenomenon that is difficult to eradicate; even when given instructions to falsify hypotheses, people find it difficult to do. Thus, many researchers have concluded that both people in psychology experiments as well as scientists at large make this faulty reasoning error. However, Klayman and Ha (1987) argued that conducting experiments that confirm a hypothesis is not necessarily a scientific reasoning error. They argued that if the prior probability of confirming one’s hypothesis is low, then even if the scientist is attempting to confirm the hypothesis, it can still be disconfirmed. One other interpretation of the phenomenon of confirmation bias is that early in developing a theory or a hypothesis people will attempt to confirm the hypothesis; however once the hypothesis is fleshed out and confirmed, people will attempt to conduct disconfirming experiments (see Tweeney et al. 1980).

4. Science ‘In Vivo’: How Scientists Think In Naturalistic Contexts

One important issue in scientific reasoning and discovery is that most accounts have tended to use indirect evidence such as lab notebooks, biographies, and interviews with scientists to determine the thinking and reasoning strategies that scientists use. Another approach is to conduct experiments on isolated aspects of scientific thinking. See Dunbar (1995) for an analysis of these standard approaches that he has termed ‘in itro.’ Both approaches, while very informative, do not look at scientists directly. Thus, a complimentary approach has been to investigate real scientists’ thinking and reasoning strategies while they are conducting real research. Using this ‘in i o’ approach, Dunbar (1999) has identified the specific ways that scientists use analogies, deal with unexpected findings, and use collaborative reasoning strategies in their research. He found that scientists use analogies to similar entities (or ‘local analogies’) when fixing experimental problems, analogies to entities from the same class of items (or ‘regional analogies’) when formulating new hypotheses, and analogies to very dissimilar domains (‘long-distance analogies’) when explaining scientific issues to others. Furthermore, Dunbar found that over half the findings that scientists obtain are unexpected, and that the scientists have specific strategies for dealing with these unexpected findings: First, scientists provide methodological explanations using local analogies that suggest ways of changing their experiments to obtain the desired result. If the changes to experiments do not provide the desired results, then the scientists switch from blaming the method to formulating hypotheses; this involves the use of ‘regional analogies,’ as well as collaborative reasoning in which groups of scientists build models and theories together. Dunbar has further brought back these ‘in i o’ findings into the cognitive laboratory to conduct controlled experiments, which, taken together, have been used to build new accounts of the ways that analogy, collaborative reasoning, and causal reasoning are used in scientific thinking (Dunbar 1999).

5. The Development Of Scientific Thinking Skills

Beginning with the work of Piaget, many researchers have noted that children are similar to scientists. This ‘child-as-scientist’ metaphor has two main strands. First, children’s acquisition of new concepts and theories is said to be similar to the large conceptual changes that occur in scientific fields. Researchers investigating this view have pointed to parallels between changes in children’s concepts, such as their concepts of heat and temperature, and changes in the concepts of heat and temperature in the history of physics (see Chi 1992, Carey 1992). The second strand of the child-as-scientist metaphor is that children reason in identical ways to scientists, ranging from deduction and induction to experimental design. Some researchers have argued that there is little difference between a scientist and a three-year-old; while scientists clearly have considerably more knowledge of specific domains than children, their underlying competencies are viewed as the same. Other researchers have taken this ‘child-as-scientist’ view even further and argued that infants are basically scientists. Yet other researchers have argued that there are fundamental differences between children and scientists and that scientific thinking skills follow a developmental progression (see Klahr et al. 2000 for an overview of this debate).

6. Cognitively Driven Computational Discovery: Twenty-first Century Scientific Discovery

One development that has occurred in many sciences is the placing of vast amounts of information in computer databases. In the year 2000 the entire human genome was mapped and now exists in databases. Similar developments have occurred in physics, where the entire universe has been mapped and put on a database. In the case of the human genome, data consists of long sequences of nucleotides, each of which is represented by a letter such as A for Adenine. Databases consist of strings of letters such as ATGTC with each letter representing a particular nucleotide. These strings or sequences can extend for hundreds of millions of nucleotides without interruption. Buried in these sequences are genes, and most of the genes are of unknown function. One goal of researchers is to search the database, find genes, and determine the function of the genes and how the genes interact with each other. This new wave of scientific investigation incorporates many of the principles of analogical reasoning, inductive reasoning and problem-solving strategies dis-cussed here, as well as many of the algorithms (neural nets, Markov models, and production systems) discovered by cognitive psychologists in the 1980s and 1990s to discover the functions of genetic sequences and the properties of certain types of matter in the universe. One interesting aspect of this development has been that rather than computer programs being expected to make an entire discovery from beginning to end, they are a tool that can be used by scientists to help make a discovery. The cognitive psychology of scientific reasoning has moved from being a description of the scientific mind to an active participant in scientific practice.

Bibliography:

1. Bruner J S, Goodnow J J, Austin G A 1956 A Study of Thinking. Wiley, New York
2. Carey S 1992 The origin and evolution of everyday concepts. In: Giere R N (ed.) Minnesota Studies in the Philosophy of Science. Vol. XV: Cognitive Models of Science. University of Minnesota Press, Minneapolis, MN, pp. 89–128
3. Chi M 1992 Conceptual change within and across ontological categories. Examples from learning and discovery in science. In: Giere R N (ed.) Minnesota Studies in the Philosophy of Science. Vol. XV: Cognitive Models of Science. University of Minnesota Press, Minneapolis, MN, pp. 129–86
4. Dunbar K 1993 Concept discovery in a scientific domain. Cognitive Science 17: 397–434
5. Dunbar K 1995 How scientists really reason: Scientific reasoning in real-world laboratories. In: Sternberg R J, Davidson J E (eds.) Mechanisms of Insight. MIT Press, Cambridge, MA, pp. 365–95
6. Dunbar K 1999 The scientist in vivo: How scientists think and reason in the laboratory. In: Magnani L, Nersessian N, Thagard P (eds.) Model-based Reasoning in Scientific Discovery. Kluwer Academic Plenum Publishers, New York, pp. 89–98
7. Giere R N (ed.) 1992 Minnesota Studies in the Philosophy of Science. Vol. XV: Cognitive Models of Science. University of Minnesota Press, Minneapolis, MN
8. Holland J H, Holyoak K J, Nisbett R E, Thagard P R 1986 Induction: Processes of Inference, Learning, and Discovery. MIT Press, Cambridge, MA
9. Klahr D, Dunbar K, Fay A, Penner D, Schunn C 2000 Exploring Science: The Cognition and Development of Discovery Processes. MIT Press, Cambridge, MA
10. Klayman J, Ha Y W 1987 Confirmation, disconfirmation, and information in hypothesis testing. Psychological Review 94: 211–28
11. Nersessian N J 1992 How do scientists think? Capturing the dynamics of conceptual change in science. In: Giere R N (ed.) Minnesota Studies in the Philosophy of Science. Vol. XV: Cognitive Models of Science. University of Minnesota Press, Minneapolis, MN
12. Simon H A 1977 Models of Discovery: and Other Topics in the Methods of Science. D Reidel Publishing, Dordrecht, The Netherlands
13. Thagard P 1999 How Scientists Explain Disease. Princeton University Press, Princeton, NJ
14. Tweney R D, Doherty M E, Mynatt C R (eds.) 1981 On Scientific Thinking. Columbia University Press, New York
15. Wertheimer M 1959 Productive Thinking. Harper and Row, New York