Critical Thinking

Critical thinking is central to scientific endeavor. While there exist many definitions of critical thinking, some more esoteric than others, perhaps the one that resonates best with scientists and employers who value this skill in individuals with scientific training is that by Michael Scriven and Richard Paul for the National Council for Excellence in Critical Thinking. Critical thinking, according to them, “is the intellectually disciplined process of actively and skillfully conceptualizing, applying, analyzing, synthesizing, and/or evaluating information gathered from, or generated by, observation, experience, reflection, reasoning, or communication, as a guide to belief and action.” They note that critical thinking is based on “clarity, accuracy, precision, consistency, relevance, sound evidence, good reasons, depth, breadth, and fairness.” Critical thinking so defined essentially frames what scientific investigation is all about at an intellectual level. The following represent key stages of critical thinking within a scientific context and are in accord with commonly accepted facets of experimental design:

Key Stages of Critical Thinking

Acquisition of information.
It is important in beginning to frame a question or hypothesis to acquire all the information that is potentially relevant. This is a task that must be accomplished efficiently, thoroughly, and without bias. The skill is acquired through predoctoral training, and like anything of a sophisticated technical nature relies on intuition as well. The most important source of information are studies published in scientific journals. Tools used in exploring journals for specific types of information are search engines such as PubMed and Medline and links within journal articles of interest to later articles citing them, to articles which they themselves cite, or to related articles. Being made aware of emerging articles of interest by topics or authors can be achieved by services such as CiteTrack or sometimes the journals of interest themselves. We all have a shortlist of go-to journals as well, which send links to tables of contents as issues are published. Other sources of information include, of course, data from our own experiments, unpublished data of others to which we’re privy, conversations, posters, and formal presentations. The volume of published data alone is enormous and growing exponentially. Acquiring the right information in an efficient, thorough fashion is a considerable skill.

Conceptualization.
Conceptualization is a process of discernment – what of the acquired information is important, and what isn’t – and of identifying key relationships. Its purpose is to provide a construct that is meaningful and coherent in its application to a problem. Conceptualization is in many ways an intuitive process, a modeling, and therefore one that invites and is propelled by creativity. It should be noted that conceptualization and acquisition of information are reiterative processes, one informing the other, as a conceptual construct is increasingly better honed. Conceptualization relies on weighting information based on the ‘strength’ of data, which incorporates a statistical dimension. So conceptualization, while in many ways intuitive, requires technical savvy as well. The conceptual construct is in some ways related to the premise in experimental design.

Reasoning.
The conceptualization above leads to a construct, a model of the way things might work or might relate to something of experimental interest. The question now is how to frame a question or hypothesis to test the construct or take it further. This leads us to the domain of deductive and inductive reasoning, and the closely allied processes of hypothesis-testing and model-building. Deduction, according to David Glass in Experimental Design for Biologists, is a process of reasoning from a general statement (a premise) to reach a specific and logically certain conclusion. The conclusion is bounded by and dependent on the truth of the initial premise. Induction is a process of reasoning that uses past experience, e.g. data obtained, to provide a platform for prediction, one that involves verification as opposed to falsification. Deductive reasoning draws close parallels with hypothesis testing, or more formally in an experimental context, hypothesis falsification. Inductive reasoning is used in forming questions that lead to experimental models. Glass refers to hypotheses as backward looking, and models as inductive, or future-predicting. Both are relevant to scientific thought and process.