Observer Forum

Psychology Needs to Take Scientific Method Seriously

In this brief commentary, I propose an important reform: that psychological scientists should take scientific method seriously. By this, I mean that they should be knowledgeable about, and be guided in their research by, theories of scientific method. I say “theories” because they are the primary vehicles for conveying knowledge about scientific method. However, psychologists show little interest in theories of method, and this constitutes a major defect in their methods teaching and research practice.

Theories of scientific method are important for two reasons: They are science’s centerpiece, and we cannot be properly informed about science without knowing about them. They also provide researchers with major sources of guidance in their quest to obtain knowledge about the world. We would do well to reflect on the fact that many of the greatest scientists (Newton, Darwin, and Einstein among them) wittingly subscribed to one or more theories of scientific method. For example, Charles Darwin’s 19th century research reported in the Origin of Species was guided in important ways by the rather sophisticated inductive methods of William Whewell and John Herschel. Similarly, faced with the complexity of psychology’s subject matter and our own cognitive limitations, we have much to gain by being guided in our research by contemporary theories of scientific method.

By common consent among philosophers of science, the “big four” theories of scientific method are inductive method, hypothetico-deductive method, Bayesian method, and inference to the best explanation (Nola & Sankey, 2007). The first two of these are often mentioned in psychology’s literature but are seldom given an informative treatment. Bayesian methods are making a comeback, but are still a minority practice. Inference to the best explanation is virtually unheard of in psychology, but it provides scientists with a means for evaluating the explanatory worth of their theories (Thagard, 1992).

Although each of the four theories has sometimes been proposed as the premier account of scientific method, they are all better thought of as restricted accounts of method that can be used to meet specific research goals, not global accounts of method that capture what is essential to all scientific inquiry. Inductive method is appropriate for detecting empirical phenomena, but not for theory construction. Inference to the best explanation is not an all-purpose form of inference, but is a method particularly suited for evaluating the worth of competing explanatory theories. The hypothetico-deductive method, appropriately modified, can usefully be used to test for the empirical adequacy of local hypotheses, whereas the Bayesian approach can be used to assign probabilities to hypotheses for which we have the relevant probabilistic information. All of these domain-specific methods are of vital importance to psychological science. Methodologists, researchers, and textbook writers should include them in their armamentarium.

My recent book, Investigating the Psychological World (2014), takes scientific method seriously by considering the topic in relation to psychological science. In the book, I say more about the four theories of scientific method noted above, and I develop a broad-ranging theory that offers a framework for the coherent treatment of a variety of behavioral research methods.

Brian D. Haig will present his Invited Talk, “Psychology’s Neglect of Scientific Method,” at the 2014 APS Annual Convention, May 22 –25 in San Francisco, California.

-Brian D. Haig
APS Fellow
University of Canterbury, New Zealand

References

Haig, B. D. (2014). Investigating the psychological world: Scientific method in the behavioral sciences. Cambridge, MA: MIT Press.

Nola, R., & Sankey, H. (2007). Theories of scientific method: An introduction. Montreal, Canada: McGill-Queen’s University Press.

Thagard, P. (1992). Conceptual revolutions. Princeton, NJ: Princeton University Press.

Observer Vol.27, No.5 May/June, 2014

Leave a comment below and continue the conversation.

Comments

An excellent piece – which should have been given a decent spread in the Observer instead of this cut-down summary.

While the APS trumpets itself as an association for ‘psychological science’, it’s senior executive members seem to actually have little clue what investigative science demands from those who would call themselves ‘scientists.

If this association was truly looking to ‘make a difference’ and provide real scientific leadership, it would convene a specialist symposium on ‘investigative science’ with Brian Haig, Joel Michell, Mike Maraun, Klaas Sitjsma, and James Grice as the keynote speakers and panel members. These are variously the thought-leaders, philosophers, and innovative empirical scientists in psychology today. The rest are just mildy-amusing ‘safe’ entertainment for the masses.

Science is about phenomena detection, constructing explanatory theories, constructing measurement where possible/sensible, and testing theory-based expectations.

To engage in these efforts productively, you need to understand methodology and theory construction, the constituent properties of quantitative and non-quantitative measurement, and how to test theory-statements in accordance with the properties of the phenomenal observations and desired knowledge claims you wish to make.

All that ‘New Statistics’ and Bayesian stuff is just tinkering with side-issues, like re-arranging the deck-chairs on the Titanic as it slowly sinks.

Those who can’t see this haven’t been reading the new facts and relevant articles and books published since 1997.

Instead of the ‘corporate’, anodyne, and ‘intellectually ‘safe’ symposia and keynotes the APS prides itself upon .. get these field-leaders on a stage somewhere and truly ‘make a difference’ for the future of psychological science.

Leave a comment.

Comments go live after a short delay. Thank you for contributing.

(required)

(required)