Scientific Reasoning

Scientific reasoning is a special form of systems of reasoning.  Following is a brief description of the characteristics of scientific reasoning:

• Science has two major branches: descriptive-based science and hypothesis-based science.
• Descriptive-based science works to observe, explore, and discover. The descriptive-based scientist organizes information derived from the natural world, endeavors to discern regularities and patterns, and creates taxonomies. The questions that are at the focus of descriptive science tend to be along the lines of Who, What, When, Where, and How.
• Hypothesis-based science begins with a specific question or problem and a potential answer or solution that can be tested. The questions that are at the focus of hypothesis-based science tend to be along the lines of a deeper description of How, and then Why.  The hypothesis-based scientist endeavors to provide explanations for the phenomena provided by the descriptive-based scientist.
• Descriptive-based science is primarily inductive in its reasoning, using related observations to arrive at general conclusions.
• Hypothesis-based science is primarily deductive in its reasoning, using a general principle or law to forecast outcomes. From the general principles, the hypothesis-based scientist endeavors to predict the specific results that would be valid if the general principles are valid.
• It is common for the two branches to work in conjunction with each other, but it is not necessary. If performed carefully, the information provided by descriptive science is generally valid, whether explanatory hypotheses are available or not.
• Often the descriptive-based science comes first. For example, a new instrumentation system, such as the microscope or the radio telescope, allows observations to be made that weren’t possible previously.
• Good guidance for working in a new scientific territory is to start from careful observation of phenomena, prior to forming any conclusions. Be as objective as possible in perceiving and recording the observation, whether or not the phenomena conform to one’s expectations or make sense in terms of what is already regarded as known with confidence.  Strive to be complete in describing the observation.
  • Carefully preserve the observational data. It might be many years before they can be coupled with other data relevant for generating an explanation.  It is important that the information from trustworthy observations not be lost, particularly if they are anomalous and puzzling.
  • Be careful about the strength of support for any assumptions or conclusions from the observations one makes. In particular, remember that correlation does not imply causation.  The presence of two factors appearing together does not necessarily imply that they are causally related.
• Hypothesis-based science is particularly focused on identifying regularities in natural phenomena that can be explained in a coherent and consistent manner with a body of theory. It endeavors to build up a structure that explains the set of previously-observed phenomena and can be extended without excessive distortion to explain newly observed phenomena.
• Hypothesis-based science considers the new observations to evaluate how well existing theory matches with them.
  • When it appears that the existing body of theory does not adequately encompass the new observations, the scientist forms a hypothesis space with possible hypotheses that could be consistent with what was observed.
  • In organizing the set of candidate hypotheses, apply Occam’s Razor and prefer simpler hypotheses to more complex ones with more assumptions.
  • Always consider the null hypothesis—that there is in fact no relationship between two measured or observed phenomena (e.g. that a proposed treatment for a disease has a measurable and significant effect). It is important to accept, reject, or disprove the null hypothesis in any investigation.
  • Always evaluate possible measurement error and its effect on any conclusions in the chain of reasoning.
• The next step in hypothesis-based science is to consider what experiments could be performed to generate evidence to support or disprove the candidate hypothesis/ hypotheses.
  • A well-formed scientific hypothesis needs to be falsifiable. In other words, it must be inherently disprovable.  Some observation might be able to contradict the hypothesis and thus show that it is not valid, or that it must be modified to be valid.
  • The experiment space needs to be considered carefully to create strategies for generating evidence for the hypotheses being examined. The goal is to have experimental results that clearly discriminate the validity or invalidity of the possible hypotheses.
• The goal in extending scientific theory in hypothesis-based science is simple: that the new version has a high degree of predictive power for phenomena yet to be observed, in addition to credibly explaining previously-observed phenomena.
• Scientific reasoning has to tread carefully in evaluating anomalous evidence that significantly differs from a well-established body of prior evidence. The outlier may be due to some flaw in the observational process (e.g., an instrument out of calibration).  On the other hand, the outlier may be the indication of a previously unobserved phenomenon (a so-called Black Swan) that is highly significant.
• The plausibility of a chain of scientific reasoning is a critical aspect in its evaluation. Are all of the factors in the argumentation consistent and mutually supporting?  Are there any unsupported assertions?  Are there any logical leaps?  What assumptions are made without examination?  What beliefs are taken for granted?  Are any logical fallacies in evidence?  What cognitive biases appear to be present?
• Good scientific reasoning always considers the existing body of knowledge as working hypotheses, ready to be modified when sufficient evidence is accepted. Nothing should be regarded as settled fact, immune to possible change.  Revolutions occur regularly in scientific fields, where existing belief structures are overturned on the basis of new insights.