Abstract

PADILLA-SANTAMARIA, Fernando. The p-value: how to analyze it to separate from extreme positivism and naive inductivism?. Investigación educ. médica [online]. 2022, vol.11, n.44, pp.105-114.  Epub Jan 06, 2023. ISSN 2007-5057.  https://doi.org/10.22201/fm.20075057e.2022.44.22442.

Despite the new reflections and ideological currents in medical sciences, many researchers remain firmly attached to scientific positivism, naive inductivism and totally reductionist views, where the use of inferential statistics is considered almost indispensable to consider a “high quality” research, and within it, consider the famous p-value as the number that determines that a study is “good” or “bad”, that it is worthwhile or not, that it came out “good” or “bad”. Although inferential statistics is the most frequent in medical sciences, many researchers continue with epistemological problems for the p-value interpretation and make statistical decisions; Therefore, the main objective of this work is to provide a reflection and dynamic analysis of what is p-value, how it is obtained, how it is usually interpreted, and how it should be interpreted. It should be noted that this paper does not intend to teach statistics, but rather tries to change the way in which students and health professionals interpret inferential statistics, in order to encourage critical reading and thus provide weapons for self-taught learning. To arrive at the appropriate analysis of the p-value, throughout the work I carry out a general and graphic review about the construction of hypotheses, the normal distribution and the hypothesis tests. Although the simple fact that this work talks about inferential statistics already makes it (until a certain point) a positivist article, I hope that the new teaching in this area will allow the training of new professionals and researchers with broader visions of research, and thus, ending the promotion of reductionism and naive inductivism.

Keywords : Hypothesis tests; confidence intervals; normal distribution; quantitative methodologies.