In his article “Mindless Statistics,” Gerd Gigerenzer focuses on discussing the problem of the ‘null ritual’ and provides arguments against the use of this procedure in social sciences because it is based on misinterpreting original statistical methods. Thus, the author argues that the ‘null ritual’ as a procedure of testing a null hypothesis with a focus on a p-value less than 0.05, 0.01, or 0.001 is not effective and rather ‘mindless’ because it ignores principles of statistical thinking.
In order to support his argument, Gigerenzer discusses the ‘null ritual’ in detail, concentrates on Fisher’s and Neyman-Pearson’s ideas and associated misunderstandings, refers to the researchers’ feeling of guilt, focuses on collective illusions, discusses the role of editors ‘with guts,’ refers to the superego, ego, and id analyzes Meehl’s and Feynman’s conjectures and predicts the progress of statistical thinking.
The author notes that the ‘null ritual’ “became institutionalized in curricula” and the practice of researchers and the problem is in misinterpreting Fisher’s and Neyman-Pearson’s concepts. Gigerenzer pays attention to the fact that many other statistical tools are often ignored, and they remain unknown to psychologists. The focus of researchers is on the standardized procedure that cannot be used in all cases, but it is helpful to address p-value and probability, as well as to eliminate Type-II errors. Thus, the ‘null ritual’ is a hybrid of misinterpreted methods offered by Fisher (null hypothesis testing) and Neyman and Pearson (the decision theory).
Gigerenzer also notes that when researchers do not know what p-value to refer to, they can feel guilty because of the necessity to adjust the overall procedure to their needs and determine the p-value after collecting the results. This issue is also related to collective illusions that are developed by researchers when they focus on rituals. Thus, researchers often misinterpret the level of significance while trying to focus on the p-value and “strike a balance between α, β, and sample size n”. According to the author, these researchers do not understand that cheating, dilemmas, and illusions are caused by confusion about methods.
The author also argues that the problem is often in the impact of editors ‘with guts’ and the associated superego, ego, and id. On the one hand, psychologists often use the ‘null ritual’ because it allows them to organize studies according to the rules approved by editors and publish them quickly, without any barriers. On the other hand, researchers follow these rules because the Neyman-Pearson theory affects them as superego.
Fisher’s theory associated with the necessity of null hypothesis testing is the ego in this case, and the id is based on the Bayesian view. Thus, rituals play important roles in the researcher’s practice because of the developed misunderstandings.
Finally, Gigerenzer discusses Meehl’s and Feynman’s conjectures that are important to understand that a null hypothesis can be rejected or accepted in 50% of cases and that the formulation of a null hypothesis depends on the formulation of an alternative hypothesis. While concluding that these conjectures are important to disapprove the ‘null ritual,’ the author predicts the development of statistical thinking among those psychologists who choose to reject the ritual. Still, according to the author, “to stop the ritual, we also need more guts and nerves”.
It is possible to state that Gigerenzer primarily uses logos in his article while providing numerous reasons why the ‘null ritual’ is inappropriate for statistical thinking. However, he also uses pathos while proposing a variety of examples, anecdotes, and ironic stories that make the author’s argument more vivid and appealing. Ethos is applied in the work when Gigerenzer refers to prominent figures in the field of psychology and their arguments.
Gigerenzer’s article is interesting and important because it draws the reader’s attention to the controversial topic of using conventional procedures in statistics. The vividness of the author’s argumentation and examples make the reader pay attention to his ideas. As a result, some of these ideas seem to be convincing and well-grounded while provoking further discussion of the problem.
References
- G. Gigerenzer, “Mindless statistics,” J Socio Econ, vol. 33, pp. 587-606, 2004.
- R. M. Warner, Applied Statistics: From Bivariate Through Multivariate Techniques. New York, NY: Sage, 2012.